System

Abstract

To provide a system capable of concentrating on works by reducing labor burden of on-site managerial workers.SOLUTION: A system includes means for receiving a report about works of a user prepared by the user, means for deriving work problems, causes of the problems and solutions to the problems by using a prompt sentence to instruct to derive the work problems, the causes of the problems and the solutions to the problems from contents of the received report, and a generative AI model, means for preparing a report representing the work problems, the causes of the problems and the solutions of the problems by using the prompt sentence to instruct to prepare the report representing the derived work problems, causes of the problems and solutions to the problems and the generative AI model, and means for presenting the prepared report to the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] On-site managers spend a lot of time reviewing reports from their subordinates and customers, identifying problems, analyzing their causes, and coming up with solutions, which reduces the time they can spend concentrating on other important tasks and increases their workload. [Means for solving the problem]

[0005] This invention provides a means for reviewing reports submitted by subordinates and customers through AI, identifying problems and proposing causes and solutions, and a means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. Furthermore, it provides a means for AI to review meeting minutes, identifying problems and proposing causes and solutions, and a means for AI to make decisions on behalf of section managers, just as they would if they were making decisions based on past experience and future predictions. This reduces the workload of on-site managers, allowing them to focus on other tasks. [Brief explanation of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment 2. [Figure 15] FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, the processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0011] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0012] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0018] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0022] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0023] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0024] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] One aspect of the present invention is a system that uses AI to review reports submitted by subordinates and customers. Specifically, the AI ​​analyzes the text data of the report and extracts problems. The AI ​​then analyzes the causes of the problems and derives solutions. This AI uses natural language processing (NLP) technology to understand the content of the report and extract problems. The AI ​​also uses machine learning algorithms to learn from past reports and their results, and derives causes and solutions.

[0029] "Example 2"

[0030] Furthermore, one aspect of the present invention is a system that delegates managerial authority to AI. Specifically, AI takes over on-site decision-making that would normally be performed by managers, allowing them to focus on other tasks. For example, AI can automatically approve or reject vacation requests from subordinates. This decision is based on the subordinate's past vacation history, current workload, and other relevant information.

[0031] "Example 3"

[0032] Another aspect of the present invention is a system in which AI reviews meeting minutes, extracts problems, and derives their causes and solutions. Specifically, the AI ​​analyzes the meeting minutes and extracts problems from the content of the discussion. The AI ​​then analyzes the causes of the problems and derives solutions. This AI uses natural language processing (NLP) technology to understand the content of the minutes and extract the problems. The AI ​​also uses machine learning algorithms to learn from past minutes and their results, and derives causes and solutions.

[0033] The processing flow of each embodiment will be described below.

[0034] "Example 1"

[0035] Step 1: Reports from subordinates and customers are sent to the system.

[0036] Step 2: AI analyzes the text data of the report and extracts problem areas. This analysis is performed using natural language processing (NLP) technology.

[0037] Step 3: AI analyzes the cause of the problem and derives a solution. This analysis is done using machine learning algorithms.

[0038] "Example 2"

[0039] Step 1: Subordinate requests for vacation time and other matters requiring on-site decisions typically made by managers are sent to the system.

[0040] Step 2: The AI ​​evaluates the matter and makes a decision, based on the employee's past leave history, current workload, and other relevant information.

[0041] Step 3: The AI's decision is notified to management, allowing them to focus on other tasks.

[0042] "Example 3"

[0043] Step 1: The meeting minutes are sent to the system.

[0044] Step 2: AI analyzes the text data of the minutes and extracts issues from the content of the discussion. This analysis is carried out using natural language processing (NLP) technology.

[0045] Step 3: AI analyzes the cause of the problem and derives a solution. This analysis is done using machine learning algorithms.

[0046] Example 1

[0047] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0048] On-site managers are required to perform a wide range of tasks, including reviewing reports submitted by subordinates and customers, identifying problems, analyzing causes, and deriving solutions, which places a heavy workload on them and makes it difficult for them to concentrate on other important tasks. Accurately analyzing the contents of reports and deriving appropriate solutions also requires a high level of specialized knowledge and experience, which further increases the burden on on-site managers. Furthermore, reviewing meeting minutes and making decisions based on past experience and future predictions are equally burdensome.

[0049] The identification process by the identification processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means. In this invention, the server includes a means for reviewing reports submitted by subordinates and customers using artificial intelligence, extracting problems and deriving causes and solutions, a means for analyzing the contents of the reports using natural language processing technology, and a means for learning from past reports and their results using a machine learning algorithm to derive causes and solutions. This reduces the workload of on-site managers, allowing them to focus on other important tasks.

[0050] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and learn, reason, and self-correct.

[0051] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[0052] A "machine learning algorithm" is an algorithm that learns patterns from data and makes predictions and classifications.

[0053] A "report" is a document about work or a situation that is sent by a subordinate or a customer.

[0054] "Review" is the process of evaluating the contents of a report and identifying problems and areas for improvement.

[0055] "Problems" are issues or obstacles that need to be resolved and are identified in the report.

[0056] "Cause" refers to the reason or background behind the occurrence of a problem.

[0057] A "solution" is a specific method or means for solving a problem.

[0058] "Managerial authority" refers to the authority that on-site managers have regarding the execution of work and making decisions.

[0059] "On-site level judgment" is a judgment made based on the specific situation at the site.

[0060] "Labor burden" refers to the burden and stress of work that accompanies the performance of work.

[0061] "Text data" is textual information extracted from a report.

[0062] "Analysis results" are the results of analysis obtained using natural language processing technology and machine learning algorithms.

[0063] "Minutes" are documents that record the contents of a meeting.

[0064] "Past experience" refers to knowledge and data about tasks and situations that have been accumulated up to now.

[0065] "Future forecasting" is the prediction of future conditions based on past data and current conditions.

[0066] This invention is a system that uses artificial intelligence to review reports submitted by subordinates and customers, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system will be described below.

[0067] Program Generation

[0068] The server generates a program to review reports submitted by subordinates and customers. This program is built by combining natural language processing (NLP) technology and machine learning algorithms. Specifically, it uses SpaCy and NLTK as NLP libraries, and Scikit-learn and TENSORFLOW (registered trademark) as machine learning libraries.

[0069] Program processing

[0070] The server performs the following processing using the generated program.

[0071] 1. Receiving reports and obtaining text data

[0072] The terminal transmits the report created by the user to the server.

[0073] The server receives the report sent from the terminal and extracts the text data of the report. Specifically, if the report is in PDF format, it extracts the text using a PDF parser (e.g., PyPDF2).

[0074] 2. Analysis using natural language processing (NLP)

[0075] The server analyzes the extracted text data using NLP techniques, specifically using SpaCy and NLTK to tokenize the text, tag parts of speech, and analyze dependencies.

[0076] 3. Identifying the problem

[0077] The server extracts problem areas from the report based on the results of NLP analysis, specifically detecting predefined keywords and phrases (e.g., "declining sales" or "customer complaints").

[0078] 4. Cause analysis using machine learning algorithms

[0079] The server analyzes the causes of the extracted problems using a machine learning model trained on past report data. Specifically, it uses a model built using Scikit-learn and TensorFlow.

[0080] 5. Deriving a solution

[0081] The server then derives appropriate solutions based on the results of the root cause analysis, specifically by referencing past success stories and best practices.

[0082] 6. Return of results

[0083] The server compiles the identified problem, cause, and solution and returns it to the terminal.

[0084] The terminal displays the results received from the server to the user.

[0085] Specific examples

[0086] The user sends a "sales report" from the terminal to the server, which includes a description of the reasons for the decline in sales.

[0087] 1. The terminal sends the "sales report" created by the user to the server.

[0088] 2. The server receives the report sent from the device and extracts the text data using a PDF parser.

[0089] 3. The server tokenizes the extracted text data using SpaCy, and performs part-of-speech tagging and dependency analysis.

[0090] 4. The server extracts the problem of "declining sales" from the analysis results.

[0091] 5. The server uses a model trained on past sales reports using Scikit-learn to identify that the cause is a "reduced advertising budget."

[0092] 6. The server comes up with the solution: "Increase advertising spending."

[0093] 7. The server summarizes the problem, cause, and solution and returns it to the terminal.

[0094] 8. The terminal displays the results received from the server to the user.

[0095] Prompt Sentence Examples

[0096] "Analyze the following sales report and identify the problem, its cause, and a solution. Report content: 'Sales this month have decreased by 20% compared to last month. This is likely due to a reduction in advertising expenses.'"

[0097] In this way, the server performs specific processing at each step, and ultimately the system automatically performs everything from identifying problems to proposing solutions to the user.

[0098] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0099] Step 1: Receive the report

[0100] Input: Reports sent by users from their terminals

[0101] Specific operation: The terminal sends the report created by the user to the server in PDF or text format.

[0102] Output: Report data received by the server

[0103] Step 2: Extract text data

[0104] Input: Report data received by the server

[0105] What happens: The server extracts text data from the received report. If the report is in PDF format, it uses a PDF parser (e.g. PyPDF2) to extract the text.

[0106] Output: Extracted text data

[0107] Step 3: Natural Language Processing (NLP) Analysis

[0108] Input: Extracted text data

[0109] What it does: The server analyzes text data using NLP techniques, such as SpaCy and NLTK, to tokenize text, tag parts of speech, and analyze dependencies.

[0110] Output: Parsing results (tokenized text, part-of-speech tags, dependencies)

[0111] Step 4: Identify the problem

[0112] Input: NLP analysis results

[0113] What it does: The server extracts problems in the report based on the analysis results, detecting predefined keywords and phrases (e.g., "declining sales," "customer complaints").

[0114] Output: Extracted issues

[0115] Step 5: Root cause analysis using machine learning algorithms

[0116] Input: Extracted issues

[0117] How it works: The server analyzes the cause using a machine learning model trained on past report data. Specifically, it uses a model built using Scikit-learn and TensorFlow.

[0118] Output: Analysis results (cause of the problem)

[0119] Step 6: Derive a solution

[0120] Input: Analysis results (cause of problem)

[0121] Specific operation: The server derives appropriate solutions based on the results of the cause analysis, and proposes solutions by referring to past success stories and best practices.

[0122] Output: Proposed solution

[0123] Step 7: Returning the results

[0124] Input: Proposed Solution

[0125] Specific operation: The server compiles the identified problem, cause, and solution and returns it to the terminal.

[0126] Output: Results returned to the device (problem, cause, solution)

[0127] Step 8: Viewing the results

[0128] Input: Results returned to the device (problem, cause, solution)

[0129] Specific operation: The terminal displays the results received from the server to the user.

[0130] Output: The results displayed to the user (problem, cause, solution)

[0131] In this way, the server performs specific processing at each step, and ultimately the system automatically performs everything from identifying problems to proposing solutions to the user.

[0132] (Application example 1)

[0133] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0134] Conventional report review systems required humans to manually analyze the contents of reports submitted by subordinates and customers, extracting problems, analyzing causes, and proposing solutions, which was time-consuming, labor-intensive, and inefficient. This also placed a heavy burden on factory workers and managers, increasing the workload of on-site managers and preventing them from concentrating on other tasks. Furthermore, the system lacked the ability to learn from past reports, making it difficult to improve the accuracy of analysis.

[0135] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0136] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers through AI, identifying problems and proposing causes and solutions; a means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means installed on a robot used in a factory, which automatically reviews reports submitted by workers and managers, identifying problems, analyzing causes, and proposing solutions; and a means for learning from past reports and their results to improve analysis accuracy. This enables automatic analysis of reports, identifying problems, analyzing causes, and proposing solutions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. Furthermore, learning from past reports can improve analysis accuracy.

[0137] "Subordinates and customers" refers to employees in a company or organization who receive instructions, and external users of products and services.

[0138] A "report" is a document that summarizes information or results related to a particular task or project.

[0139] "AI" is an abbreviation for artificial intelligence, and refers to the technology that allows computers to imitate human intelligence to learn and reason.

[0140] "Review" refers to the process of evaluating and analyzing submitted documents and data to identify problems and areas for improvement.

[0141] A "problem" refers to a factor or issue that hinders normal progress in a particular task or project.

[0142] "Cause" refers to the fundamental reason or factor that causes a problem.

[0143] A "solution" refers to a specific method or means for solving a problem.

[0144] "Managerial authority" refers to the authority of a person in a position of authority who has the authority to give instructions and make decisions regarding work in a company or organization.

[0145] "On-site level judgment" refers to making the necessary judgments and decisions immediately at the actual work site.

[0146] "Labor burden" refers to the burden of labor and work necessary to carry out business.

[0147] "Robots used in factories" refers to machinery and devices used in manufacturing and production industries for the purpose of automating and streamlining work.

[0148] "Workers" refers to employees who actually perform work in factories or on-site.

[0149] A "manager" refers to a person in a position of authority who supervises workers and manages operations at a factory or on-site.

[0150] "Automated review" refers to artificial intelligence or machines analyzing and evaluating documents and data without human intervention.

[0151] "Learning" refers to the process by which artificial intelligence acquires new knowledge and skills from past data and experiences.

[0152] "Analysis accuracy" refers to the accuracy and reliability of analyzing data or documents.

[0153] The following system configuration will be described as an embodiment of the present invention.

[0154] The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. Specifically, the server uses natural language processing (NLP) technology to analyze the content of the reports and extract problems. Furthermore, it uses machine learning algorithms to learn from past reports and their results and derive causes and solutions.

[0155] The server also includes a means to delegate managerial authority to AI and have it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. This frees on-site managers from routine decision-making tasks and allows them to focus on more strategic tasks.

[0156] Furthermore, it includes a means to be installed on robots used in factories and automatically review reports submitted by workers and managers, extract problems, analyze their causes, and propose solutions.The robot analyzes reports submitted by workers and managers in real time and provides immediate feedback.

[0157] The server includes a means for learning from past reports and their results to improve the accuracy of the analysis, which improves the accuracy of the analysis over time, enabling more accurate problem detection and solution proposals.

[0158] In terms of specific hardware, the server uses a computer system equipped with a high-performance processor and large memory capacity. For software, spaCy is used for natural language processing, and scikit-learn is used for machine learning. These software programs provide the functions necessary to analyze the text data in the reports, extract problems, analyze their causes, and propose solutions.

[0159] As a concrete example, consider the following report submitted:

[0160] "Machine A's operation is unstable."

[0161] "Product B's quality varies."

[0162] "Production speed on Line C is slow."

[0163] In response to these reports, the server makes the following suggestions:

[0164] "Machine A is operating unstable. Please perform maintenance on the machine."

[0165] "The quality of Product B is inconsistent. -> Please review your quality control process."

[0166] "The production speed on line C is slow. Please adjust the speed of the production line."

[0167] Example prompts to input to a generative AI model:

[0168] "Machine A's operation is unstable." "Product B's quality varies." "Production speed on line C is slow." Please suggest causes and solutions for these problems.

[0169] By using this prompt, the generative AI model is expected to propose an appropriate cause analysis and solution.

[0170] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0171] Step 1:

[0172] The server receives reports submitted by subordinates and customers. The input is report data in text format. The server analyzes this data using natural language processing (NLP) technology and divides it into sentences. Specifically, it uses the spaCy library to analyze the text and divide it into sentences.

[0173] Step 2:

[0174] The server extracts problems from the divided sentences. The input is the sentence data divided in step 1. The server analyzes each sentence and identifies sentences that contain problems. Specifically, it uses spaCy's dependency analysis function to extract keywords and phrases that indicate problems.

[0175] Step 3:

[0176] The server analyzes the causes of the extracted problems. The input is the problem sentence data extracted in step 2. The server uses a machine learning algorithm to identify the cause using a model trained from past report data and its results. Specifically, the server uses the scikit-learn library to convert the problem sentence into a feature vector, which is then input into the trained model to predict the cause.

[0177] Step 4:

[0178] The server proposes a solution based on the identified cause. The input is the cause data identified in step 3. The server searches a historical database for similar problems and their solutions and proposes the optimal solution. Specifically, it executes a database query to retrieve similar problems and their solutions, and selects the optimal solution.

[0179] Step 5:

[0180] The server generates a report with the proposed solution and provides feedback to the user. The input is the solution data proposed in step 4. The server generates a report including the solution and sends it to the user. Specifically, the server embeds the solution in a report template and notifies the user via email or a dashboard.

[0181] Step 6:

[0182] The server learns from past reports and their results to improve analysis accuracy. The input is user feedback and past report data. The server analyzes the feedback and updates the machine learning model. Specifically, it uses the scikit-learn library to retrain the model with new data and improve analysis accuracy.

[0183] Example 2

[0184] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0185] On-site managers are required to handle a wide range of tasks, such as reviewing reports from subordinates and customers and approving vacation requests, which increases their workload and prevents them from concentrating on other important tasks. In addition, these judgments must be made quickly and accurately, further increasing the burden on managers.

[0186] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0187] In this invention, the server includes a means for reviewing reports submitted by subordinates and customers using artificial intelligence to identify problems and derive their causes and solutions, a means for delegating managerial authority to the artificial intelligence and having the AI ​​make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, a means for accepting vacation requests, collecting and preprocessing related information, and inputting prompts into the generative AI model to obtain a judgment, and a means for obtaining the judgment result and notifying the user. This reduces the workload of managers, allowing them to focus on other important tasks.

[0188] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and perform functions such as learning, reasoning, and problem-solving.

[0189] "Managerial authority" refers to the authority that managers have within an organization to make decisions and give instructions.

[0190] "On-site level decisions" are decisions made to respond quickly and appropriately to specific problems and situations that arise on-site.

[0191] "Labor burden" refers to the physical and mental burden incurred when performing work.

[0192] A "leave application" is the application documents and procedures submitted by an employee to obtain leave.

[0193] A "generative AI model" is an artificial intelligence model that uses natural language processing and machine learning techniques to generate appropriate responses and decisions based on input data.

[0194] A "prompt" is text containing instructions or questions to be input into a generative AI model.

[0195] "Data preprocessing" is the process of formatting collected data into a form suitable for analysis or model input.

[0196] A "judgment result" is the response or decision that the generative AI model outputs based on the prompt.

[0197] "User" means an individual or member of an organization who uses the system.

[0198] This invention is a system that delegates the authority of managers to artificial intelligence and allows them to make on-site decisions, thereby reducing the workload of managers and allowing them to focus on other tasks. A specific embodiment of this system will be described below.

[0199] Hardware and software used

[0200] Hardware: Servers, terminals

[0201] Software: Generative AI models (e.g., GPT-4®), database management systems (e.g., MySQL®), business management software (e.g., Jira)

[0202] System Overview

[0203] The system delegates tasks like reviewing reports from subordinates and clients, approving vacation requests, and other on-the-ground decisions to artificial intelligence, allowing managers to focus on other important tasks.

[0204] Data processing and calculation flow

[0205] 1. Acceptance of leave requests

[0206] User: A subordinate uses a terminal to apply for leave. Specifically, the subordinate enters the necessary information into the leave application form in the business management software and clicks the submit button.

[0207] Terminal: Sends the information entered in the vacation request form to the server.

[0208] 2. Data Collection

[0209] Server: After receiving the leave request, access the database management system to collect relevant information such as the subordinate's past leave history, current workload, and workload of team members.

[0210] What it does: Executes an SQL query to retrieve the required data.

[0211] 3. Data Preprocessing

[0212] Server: Preprocesses the collected data, specifically formatting the data and filling in missing values.

[0213] Specific operation: Standardize the data format and impute missing values ​​with the mean or median.

[0214] 4. Prompt generation

[0215] Server: Based on the preprocessed data, it generates prompt sentences to be input to the generative AI model.

[0216] Specific behavior: Generates a prompt like the following:

[0217] Please use the following information to help you decide whether your subordinates should request time off.

[0218] Past vacation history: 10 days of vacation taken in the past year

[0219] Current workload: Project A is 70% complete

[0220] Team member workload: High

[0221] 5. Input and judgment for AI model

[0222] Server: The generated prompt sentence is input into the generative AI model to determine whether to approve or reject the vacation request.

[0223] Specific operation: Call the API of the generative AI model and send a prompt.

[0224] 6. Obtaining the judgment result

[0225] Server: Obtains the judgment results from the generative AI model.

[0226] Specific operation: Receives the API response and analyzes the judgment result.

[0227] Approve the vacation request, but be careful not to delay progress on Project A.

[0228] 7. Notification of Results

[0229] Server: Sends the decision result to the device.

[0230] Terminal: Notifies the user (subordinate) of the received judgment results. Specifically, the notification function of the work management software is used to display the results to the subordinate.

[0231] Specific behavior: Display "Vacation request approved" on the vacation request form along with detailed comments.

[0232] In this way, artificial intelligence can make on-site decisions on behalf of managers, reducing the workload on managers and allowing them to focus on other important tasks.

[0233] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0234] Step 1: Accepting the leave request

[0235] User: A subordinate uses a terminal to apply for leave. Specifically, the subordinate enters the necessary information into the leave application form in the business management software and clicks the submit button.

[0236] Input: Information entered into the leave request form (e.g. applicant name, leave period, reason, etc.).

[0237] Output: The vacation request data is sent to the server.

[0238] Specific operation: The terminal sends the information entered in the vacation application form to the server.

[0239] Step 2: Data collection

[0240] Server: After receiving the leave request, access the database management system to collect relevant information such as the subordinate's past leave history, current workload, and workload of team members.

[0241] Input: Leave application data, access information to database management system.

[0242] Output: Relevant information collected (e.g., past leave history, current workload, workload of team members).

[0243] What happens: The server executes an SQL query to retrieve the required data from the database.

[0244] Step 3: Data Preprocessing

[0245] Server: Preprocesses the collected data, specifically formatting the data and filling in missing values.

[0246] Input: Relevant information collected.

[0247] Output: Preprocessed data.

[0248] Specific operation: The server unifies the data format and imputes missing values ​​with the mean or median.

[0249] Step 4: Generate a prompt statement

[0250] Server: Based on the preprocessed data, it generates prompt sentences to be input to the generative AI model.

[0251] Input: Preprocessed data.

[0252] Output: The generated prompt statement.

[0253] Specific behavior: The server generates a prompt like this:

[0254] Please use the following information to help you decide whether your subordinates should request time off.

[0255] Past vacation history: 10 days of vacation taken in the past year

[0256] Current workload: Project A is 70% complete

[0257] Team member workload: High

[0258] Step 5: Input and judgment of the AI ​​model

[0259] Server: The generated prompt sentence is input into the generative AI model to determine whether to approve or reject the vacation request.

[0260] Input: The generated prompt statement.

[0261] Output: The decision result from the generative AI model.

[0262] Specific operation: The server calls the API of the generative AI model and sends a prompt.

[0263] Step 6: Obtaining the decision

[0264] Server: Obtains the judgment results from the generative AI model.

[0265] Input: The API response from the generative AI model.

[0266] Output: The analyzed decision result.

[0267] Specific operation: The server receives the API response and analyzes the decision result.

[0268] Approve the vacation request, but be careful not to delay progress on Project A.

[0269] Step 7: Notification of results

[0270] Server: Sends the decision result to the device.

[0271] Terminal: Notifies the user (subordinate) of the received judgment results. Specifically, the notification function of the work management software is used to display the results to the subordinate.

[0272] Input: The parsed decision result.

[0273] Output: The decision result communicated to the user.

[0274] Specific behavior: The device will display "Vacation request approved" on the vacation request form, along with any detailed comments.

[0275] (Application example 2)

[0276] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0277] On-site managers have to handle a wide range of tasks, including reviewing reports from subordinates and customers, reviewing meeting minutes, and managing worker schedules, placing a heavy workload on them. Furthermore, approving or rejecting workers' vacation requests requires consideration of their past vacation history and current workload, making the decision difficult. This creates a problem where managers are unable to focus on other important tasks.

[0278] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0279] In this invention, the server includes a means for using AI to review reports submitted by subordinates and customers, extracting problems and identifying their causes and solutions, a means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, and a means for managing worker schedules and automatically approving or rejecting vacation requests based on past vacation history and current workload. This reduces the workload of managers, allowing them to focus on other important tasks.

[0280] A "report" is a document that describes the work content, progress, problems, etc. submitted by subordinates or customers.

[0281] "Review" is the process of carefully examining documents such as reports and minutes, confirming their contents, and extracting problems.

[0282] A "problem" is an obstacle or challenge that arises in the progress of a task or project.

[0283] A "cause" is the underlying reason or factor that causes a problem.

[0284] A "solution" is a specific method or means for solving a problem.

[0285] "Managerial authority" refers to the authority that managers have to give instructions and make decisions regarding work.

[0286] "On-site level judgment" refers to judgments made in response to specific problems or situations that arise on-site.

[0287] "Labor burden" refers to the effort and time required to carry out work.

[0288] "Schedule management" refers to managing schedules such as workers' working hours and holidays.

[0289] "Leave Request" means a request submitted by a worker to take leave.

[0290] "Past vacation history" refers to a record of vacations taken by a worker in the past.

[0291] The "current workload" refers to the workload or workload that a worker is currently responsible for.

[0292] "Approval" means to acknowledge an application or proposal.

[0293] "Rejection" means not approving an application or proposal.

[0294] A system for carrying out this invention has the following configuration: The server includes means for using AI to review reports submitted by subordinates and customers, extracting problems and deriving their causes and solutions, means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, and means for managing worker schedules and automatically approving or rejecting vacation requests based on past vacation history and current workload.

[0295] Program processing explanation

[0296] The server uses an AI system implemented in Python to process and calculate data as follows: The hardware used is a robot operating in the factory, and the software is an AI system implemented in Python.

[0297] 1. Report Review:

[0298] The server receives reports sent by subordinates and customers and analyzes the content using an AI model. The AI ​​model uses natural language processing technology to understand the content of the report and extract problems. After the problems are extracted, the model identifies their causes and proposes appropriate solutions.

[0299] 2. Delegation of Management Authority:

[0300] The server delegates on-site decision-making that would normally be performed by managers to AI. This allows the AI ​​to monitor the situation on-site in real time and make necessary decisions. For example, by making decisions regarding work progress and quality control, it reduces the workload of managers.

[0301] 3. Schedule management and leave request processing:

[0302] The server manages the schedules of workers. When a vacation request is received from a worker, the AI ​​automatically approves or rejects the request, taking into account past vacation history and current workload. This saves managers the trouble of processing vacation requests individually.

[0303] Specific examples

[0304] For example, consider a scenario in which a robot operating in a factory accepts vacation requests from workers. When a worker requests vacation, the server receives the request, and the AI ​​makes a decision based on past vacation history and current workload. If the worker has taken too many vacations in the past or the current workload is high, the AI ​​will reject the vacation request, but if not, it will approve it.

[0305] Example prompt sentence:

[0306] "Evaluate workers' time off requests and approve or deny them based on their past leave history and current workload."

[0307] In this way, the server can use AI to streamline managerial tasks and reduce their workload.

[0308] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0309] Step 1:

[0310] The server receives reports from subordinates and customers. The input is the report data sent by the subordinates and customers, and the output is the received report data. The server prepares this data to be passed to the AI ​​model.

[0311] Step 2:

[0312] The server inputs the received report data into the AI ​​model and analyzes the content. The input is the received report data, and the output is the analysis results. The AI ​​model uses natural language processing technology to understand the report content and extract problems.

[0313] Step 3:

[0314] The server identifies the cause of the problem and proposes a solution based on the analysis results obtained from the AI ​​model. The input is the analysis results from the AI ​​model, and the output is the cause of the problem and a solution. The server provides this information to management.

[0315] Step 4:

[0316] The server delegates managerial authority to the AI, allowing it to make on-site decisions. The input is on-site situation data, and the output is the AI's judgment results. The server monitors the on-site situation in real time and makes the necessary decisions.

[0317] Step 5:

[0318] The server manages the schedules of workers. The input is the schedule data of the workers, and the output is the updated schedule data. The server manages the schedules of workers, such as their working hours and holidays.

[0319] Step 6:

[0320] The server receives vacation requests from workers. The input is the vacation request data from the workers, and the output is the received vacation request data. The server prepares this data to be passed to the AI ​​model.

[0321] Step 7:

[0322] The server inputs vacation request data into the AI ​​model, which then makes a decision based on past vacation history and current workload. The inputs are vacation request data, past vacation history data, and current workload data, and the output is a decision on whether to approve or reject the vacation request. The AI ​​model analyzes this data and makes the appropriate decision.

[0323] Step 8:

[0324] The server notifies the worker of the decision result obtained from the AI ​​model. The input is the decision result from the AI ​​model, and the output is a notification message to the worker. The server informs the worker whether the vacation request has been approved or rejected.

[0325] In this way, the server can use AI to streamline managerial tasks and reduce their workload.

[0326] Example 3

[0327] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0328] On-site managers are required to review reports and meeting minutes submitted by their subordinates and customers, identify problems, analyze their causes, and derive solutions, which increases their workload and makes it difficult for them to concentrate on other tasks. They are also required to make quick and accurate decisions at the on-site level, which also places a burden on managers. Furthermore, it is important to use past experience and data to predict the future and make appropriate decisions, which requires advanced analytical skills. To solve these issues, an efficient system utilizing artificial intelligence is needed.

[0329] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0330] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using artificial intelligence to identify problems and derive their causes and solutions; a means for delegating managerial authority to the artificial intelligence and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing meeting minutes using natural language processing technology and extracting problems from the content of discussions; a means for learning from past minutes and their results using a machine learning algorithm to derive causes and solutions; and a means for inputting prompts into a generative AI model to obtain analysis results. This reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, the use of artificial intelligence enables rapid and accurate problem identification, cause analysis, and solution deriving, enabling future predictions and appropriate decision-making using past data.

[0331] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and automatically perform tasks such as learning, reasoning, and problem-solving.

[0332] A "report" is a document sent by subordinates or customers that describes the progress of work, problems, proposals, etc.

[0333] "Review" is the process of checking the contents of submitted reports and minutes, extracting problems, analyzing their causes, and deriving solutions.

[0334] "Problems" refer to issues or obstacles that need to be resolved in the content of reports or minutes.

[0335] "Cause" refers to the reason or background behind the occurrence of a problem.

[0336] A "solution" refers to a specific method or means for solving a problem.

[0337] "Managerial authority" refers to the authority that on-site managers have to give instructions and make decisions regarding work.

[0338] "On-site level judgment" refers to the process of making quick and appropriate decisions based on the situation on-site.

[0339] "Labor burden" refers to the time and effort required to carry out work.

[0340] "Natural language processing technology" is a technology that allows computers to understand, analyze, and generate human language.

[0341] A "machine learning algorithm" is a mathematical technique for learning patterns from data and making predictions or classifications.

[0342] A "generative AI model" is an artificial intelligence model that is trained to produce optimal results for a specific task.

[0343] A "prompt" is an instruction given to a generative AI model, containing the text to be analyzed and generated.

[0344] This invention is a system that uses artificial intelligence to review meeting minutes and reports, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system is described below.

[0345] System configuration

[0346] Hardware:

[0347] Server: Use a server equipped with a high-performance processor and large memory capacity, which allows for rapid processing of large amounts of data.

[0348] Device: A device such as a computer or tablet used by a user to upload minutes and reports and check results.

[0349] software:

[0350] Natural language processing libraries: Use Python's spaCy and NLTK for text tokenization, part-of-speech tagging, and dependency analysis.

[0351] Machine learning algorithms, such as random forests and support vector machines (SVMs), learn from past data to identify causes and solutions.

[0352] Generative AI models: Use artificial intelligence models that are trained to generate optimal results for specific tasks.

[0353] Program processing

[0354] Upload the minutes:

[0355] The user uploads the minutes of the meeting from the terminal to the server. The minutes are in text file format. The user selects the minutes file using the file selection dialog on the browser and clicks the "Upload" button.

[0356] Natural Language Processing (NLP) text analysis:

[0357] The server receives the uploaded minutes file and analyzes the text using NLP techniques, specifically using the Python libraries spaCy and NLTK. The server tokenizes the minutes text, tags it with parts of speech, and performs dependency analysis.

[0358] Identifying the problem:

[0359] The server uses the analysis results to understand the content of the discussion and extract issues. Issues are identified by detecting specific keywords and phrases (such as "issue," "problem," or "concern"). The server searches for keywords such as "issue" or "problem" in the text and extracts sentences related to them.

[0360] Analysis of the cause:

[0361] The server uses machine learning algorithms to analyze the causes of the extracted problems. Using past minutes and their results as learning data, it applies algorithms such as random forests and support vector machines (SVMs) to identify the causes. The server references a database of past minutes and applies a model to identify the causes of similar problems.

[0362] Solution derivation:

[0363] The server then derives a solution based on the cause, again using a machine learning model based on past data to suggest the optimal solution. The server references past success stories and applies the model to suggest solutions to similar problems.

[0364] The resulting output:

[0365] The server outputs the extracted problems, causes, and solutions to the terminal. The user can check this information through the terminal. The server generates the analysis results in HTML format and displays them in the user's browser.

[0366] Specific examples

[0367] Let's say a user uploads "Minutes of the project meeting held on October 1, 2023." The minutes contain the following content:

[0368] Agenda Item 1: Project Status

[0369] Challenge: Schedule Delays

[0370] Cause: Insufficient resources

[0371] Solution: Add a new member

[0372] The server parses the minutes and outputs the following results:

[0373] Problem: Schedule Delays

[0374] Cause: Insufficient resources

[0375] Solution: Add a new member

[0376] Prompt Sentence Examples

[0377] An example of a prompt sentence to input to the generative AI model is as follows:

[0378] Analyze the minutes of the following meeting and extract the problems, causes, and solutions.

[0379] minutes:

[0380] Agenda Item 1: Project Status

[0381] Challenge: Schedule Delays

[0382] Cause: Insufficient resources

[0383] Solution: Add new members

[0384] By inputting this prompt sentence into the generative AI model, the model analyzes the contents of the minutes and extracts the problem, its cause, and a solution. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0385] Step 1:

[0386] Uploading minutes

[0387] The user uploads the minutes of the meeting from the terminal to the server. The minutes are in text file format. The user selects the minutes file using the file selection dialog on the browser and clicks the "Upload" button.

[0388] Input: A minutes file selected by the user

[0389] Output: Meeting minutes file uploaded to the server

[0390] Step 2:

[0391] Text analysis using natural language processing (NLP)

[0392] The server receives the uploaded minutes file and analyzes the text using NLP techniques, specifically using the Python libraries spaCy and NLTK. The server tokenizes the minutes text, tags it with parts of speech, and performs dependency analysis.

[0393] Input: Uploaded minutes file

[0394] Output: Tokenized, part-of-speech tagged, dependency parsed text data

[0395] Step 3:

[0396] Identifying the problem

[0397] The server uses the analysis results to understand the content of the discussion and extract issues. Issues are identified by detecting specific keywords and phrases (such as "issue," "problem," or "concern"). The server searches for keywords such as "issue" or "problem" in the text and extracts sentences related to them.

[0398] Input: Parsed text data

[0399] Output: Extracted problem statement

[0400] Step 4:

[0401] Analysis of the cause

[0402] The server uses machine learning algorithms to analyze the causes of the extracted problems. Using past minutes and their results as learning data, it applies algorithms such as random forests and support vector machines (SVMs) to identify the causes. The server references a database of past minutes and applies a model to identify the causes of similar problems.

[0403] Input: Extracted problem statements, database of past meeting minutes

[0404] Output: Identified cause

[0405] Step 5:

[0406] Derivation of a solution

[0407] The server then derives a solution based on the cause, again using a machine learning model based on past data to suggest the optimal solution. The server references past success stories and applies the model to suggest solutions to similar problems.

[0408] Input: Identified causes, database of past success stories

[0409] Output: Proposed solution

[0410] Step 6:

[0411] Output of results

[0412] The server outputs the extracted problems, causes, and solutions to the terminal. The user can check this information through the terminal. The server generates the analysis results in HTML format and displays them in the user's browser.

[0413] Input: Extracted problem, identified cause, proposed solution

[0414] Output: Analysis results generated in HTML format

[0415] (Application example 3)

[0416] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0417] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. There is also the issue of it being difficult to efficiently analyze the contents of minutes from production meetings and quality control meetings, identify problems, analyze their causes, and propose solutions. Furthermore, there is a lack of systems that can learn from past minutes data and support efficient problem-solving.

[0418] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0419] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers through AI, identifying problems and proposing causes and solutions; a means for delegating managerial authority to AI and allowing it to make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing minutes of production meetings and quality control meetings within the factory, identifying problems, analyzing causes, and proposing solutions; and a means for learning from past minutes data to support efficient problem solving. This reduces the workload of on-site managers, allowing them to focus on other tasks. It also makes it possible to efficiently analyze the contents of minutes from production meetings and quality control meetings, identifying problems, analyzing causes, and proposing solutions. Furthermore, it is possible to learn from past minutes data to support efficient problem solving.

[0420] A "report" is a report or record document submitted by a subordinate or customer.

[0421] "AI" is an abbreviation for artificial intelligence, a system that mimics human intelligence using technologies such as machine learning and natural language processing.

[0422] "Review" is the process of analyzing the contents of reports and minutes and evaluating and examining them.

[0423] "Problems" are issues or obstacles that need to be resolved in the content of a report or minutes.

[0424] A "cause" is the fundamental reason or factor that causes a problem.

[0425] A "solution" is a specific method or means for solving a problem.

[0426] "Managerial authority" refers to the authority that managers have to make decisions and give instructions.

[0427] "On-site level judgment" refers to specific judgments and decisions made based on the situation on-site.

[0428] "Labor burden" refers to the effort and time required to carry out work.

[0429] A "production meeting" is a meeting held within a factory to discuss production-related topics.

[0430] A "quality control meeting" is a meeting within the factory to discuss quality control topics.

[0431] "Minutes" are documents that record the contents of a meeting and the decisions made.

[0432] "Analysis" is the process of analyzing and understanding data and information in detail.

[0433] "Learning" is the process of acquiring knowledge from past data and experiences and applying it to future decisions and actions.

[0434] "Efficiency" means achieving a goal effectively and without waste.

[0435] "Support" refers to helping with the activities or work of others.

[0436] As an embodiment of the present invention, the following system is constructed.

[0437] System Configuration

[0438] The server includes the following means:

[0439] 1. A method of using AI to review reports submitted by subordinates and customers, extracting problems and identifying their causes and solutions.

[0440] 2. Delegating managerial authority to AI and having it make on-site decisions reduces the workload of on-site managers, allowing them to focus on other tasks.

[0441] 3. A means of analyzing the minutes of production meetings and quality control meetings within the factory to identify problems, analyze their causes, and propose solutions.

[0442] 4. A means of learning from past meeting minutes data to support efficient problem solving.

[0443] Hardware and software used

[0444] Hardware: Servers, terminals (PCs, tablets, smartphones), robots in factories

[0445] Software: Natural language processing libraries (spaCy, NLTK), machine learning libraries (scikit-learn), databases (MySQL, PostgreSQL)

[0446] Data processing and calculation

[0447] The server does the following:

[0448] 1. Data collection: Reports submitted by subordinates and customers, as well as minutes of production meetings and quality control meetings within the factory, are stored in a database.

[0449] 2. Natural language processing: Using a natural language processing library (e.g., spaCy), the contents of reports and minutes are analyzed and problems are extracted.

[0450] 3. Machine learning: Using a machine learning library (e.g., scikit-learn), learn from past minutes data, analyze the causes of problems, and propose solutions.

[0451] 4. Displaying the results: The analysis results and proposed solutions are displayed on the device for the user to review.

[0452] Specific examples

[0453] For example, if the minutes of a production meeting include issues such as "delays on the production line" or "a high number of defective products found during quality inspections," the server will extract these issues and, based on the results of learning from past data, propose solutions such as "automating the production line" or "strengthening quality inspections."

[0454] Prompt Sentence Examples

[0455] "Analyze the minutes of the production meeting, extract the problems, analyze the causes, and propose solutions. The minutes contain the following information: 'There are delays on the production line' and 'There are many defective products during quality inspections'."

[0456] In this way, the workload of on-site managers can be reduced, allowing them to concentrate on other tasks. It also makes it possible to efficiently analyze the contents of minutes from production meetings and quality control meetings, extract problems, analyze their causes, and propose solutions. Furthermore, it is possible to learn from past minutes data and support efficient problem solving.

[0457] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0458] Step 1:

[0459] Data collection

[0460] The server stores reports submitted by subordinates and customers, as well as minutes of production and quality control meetings held within the factory, in a database. The input includes the text data of the reports and minutes. The server receives this data and stores it in a database (for example, MySQL or PostgreSQL). The output is the text data stored in the database.

[0461] Step 2:

[0462] Natural Language Processing

[0463] The server uses a natural language processing library (e.g., spaCy) to analyze the contents of reports and minutes and extract problems. The input includes text data obtained from a database. The server analyzes the text data and extracts sentences containing problems. The output is a list of extracted problems.

[0464] Step 3:

[0465] Machine Learning

[0466] The server uses a machine learning library (e.g., scikit-learn) to learn from past meeting minutes data, analyze the causes of problems, and propose solutions. The input includes a list of extracted problems and past meeting minutes data. The server uses this data to train a machine learning model, analyze the causes of problems, and generate solutions. The output is a list of causes and solutions.

[0467] Step 4:

[0468] Displaying the results

[0469] The server displays the analysis results and proposed solutions on the terminal for the user to review. The input includes a list of causes and solutions. The server sends this data to the terminal and displays it for the user to review. The output is a screen display that the user can review.

[0470] In this way, the server performs a series of processes from data collection to analysis, learning, and display of results, reducing the workload of on-site managers and creating a system that supports efficient problem-solving.

[0471] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0472] "Example 1"

[0473] In one embodiment of the present invention, AI reviews reports from subordinates and customers and extracts problems. It then uses an emotion engine to recognize the user's emotions. For example, it can sense the user's anger or frustration from the report text. This emotional information is then used to analyze the causes of the problems and derive solutions.

[0474] "Example 2"

[0475] In another implementation, the AI ​​can act on behalf of managers, making decisions on the ground, such as when employees request time off. It uses an emotion engine to recognize the employee's emotions and adjust its decisions accordingly. For example, if it recognizes that an employee is stressed, it can prioritize approving the vacation request.

[0476] "Example 3"

[0477] In yet another embodiment, the AI ​​reviews meeting minutes and extracts issues. It uses an emotion engine to recognize participants' emotions from the text in the minutes. For example, if a discussion gets heated or participants seem confused, the AI ​​can understand the flow of the discussion from the emotional information and identify the cause and solution of the problem.

[0478] The processing flow of each embodiment will be described below.

[0479] "Example 1"

[0480] Step 1: The AI ​​receives reports from subordinates and customers.

[0481] Step 2: The AI ​​reviews the report and identifies problems.

[0482] Step 3: Using an emotion engine, we recognize the user's emotions from the report text. Step 4: Based on the emotion information, we analyze the cause of the problem and derive solutions.

[0483] "Example 2"

[0484] Step 1: The AI ​​receives information such as vacation requests from subordinates.

[0485] Step 2: Use the emotion engine to recognize your subordinates' emotions.

[0486] Step 3: Decide whether to approve or deny the vacation request based on the employee's feelings.

[0487] "Example 3"

[0488] Step 1: The AI ​​receives the meeting minutes.

[0489] Step 2: Use the emotion engine to recognize participants' emotions from the text of the minutes.

[0490] Step 3: Understand the flow of the discussion from emotional information and identify the cause of the problem and solutions.

[0491] Example 1

[0492] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0493] On-site managers spend a great deal of time and effort reviewing reports from their subordinates and customers, identifying problems, analyzing their causes, and deriving solutions. This makes it difficult for them to concentrate on other important tasks, increasing their workload. Accurately understanding the contents of reports and deriving appropriate solutions requires advanced expertise and experience, further increasing the burden on on-site managers. Furthermore, it is difficult to properly recognize user sentiment contained in reports and utilize it to solve problems. An effective system to resolve these issues is needed.

[0494] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0495] In this invention, the server includes: means for receiving a report on the user's work created by the user; a prompt statement instructing the system to derive problems in the work, causes of the problems, and solutions to the problems from the content of the received report; means for deriving the problems in the work, causes of the problems, and solutions to the problems using a generative AI model; a prompt statement instructing the system to generate a report indicating the derived problems in the work, causes of the problems, and solutions to the problems using the generative AI model; and means for presenting the generated report to the user. The server further includes means for recognizing the user's emotion from the report using an emotion engine, and the report generating means generates the report using the prompt statement instructing the system to generate a report indicating the derived problems in the work, causes of the problems, and solutions to the problems based on the recognized user's emotion, and the generative AI model. Furthermore, the report generating means in the server generates the report using the prompt sentence instructing the generation of the report with the priority of the solutions adjusted based on the recognized user's emotions and the generation AI model. This reduces the workload of on-site managers, allowing them to focus on other important tasks. It also enables accurate understanding of the report content and promptly deriving appropriate solutions. Furthermore, it enables the user's emotions to be properly recognized and utilized in problem solving.

[0496] A "report" is a document sent by subordinates or clients that describes the progress of a project, problems, proposals, etc.

[0497] "Artificial intelligence" refers to the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence, including, among other things, natural language processing and machine learning.

[0498] A "review" is a process in which the contents of a report are examined in detail, problems are identified, and suggestions for improvement are made.

[0499] "Problems" refer to elements or issues described in the report that hinder the progress of the project or the execution of work.

[0500] "Cause" refers to the fundamental reason or background behind the occurrence of a problem.

[0501] A "solution" refers to a specific method or means for solving the identified problem.

[0502] "Natural language processing technology" refers to technology that enables computers to understand, analyze, and generate human language, and includes text tokenization and grammatical analysis.

[0503] "Machine learning algorithms" refer to mathematical models and methods for learning patterns from data and making predictions or classifications.

[0504] An "emotion engine" is a technology for recognizing and analyzing user emotions from text data, and includes an emotion analysis library.

[0505] "Managerial authority" refers to the authority to make decisions regarding business operations and problem solving.

[0506] "On-site decision-making" refers to the process of making decisions quickly based on the specific situation at the site.

[0507] "Labor burden" refers to the amount of work and stress that managers bear in carrying out their duties.

[0508] This invention is a system that uses artificial intelligence to review reports submitted by subordinates and customers, extracting problems, analyzing their causes, and deriving solutions. This system works in cooperation with a server, a terminal, and a user.

[0509] Hardware and software used

[0510] The server uses the following hardware and software:

[0511] Hardware: A server machine with a powerful processor, sufficient memory, and ample storage.

[0512] software:

[0513] NLP libraries such as "spaCy" and "NLTK" for implementing natural language processing techniques.

[0514] Machine learning libraries such as "Scikit-learn" and "TensorFlow" for implementing machine learning algorithms.

[0515] Sentiment analysis libraries such as "TextBlob" and "VADER" for implementing sentiment engines.

[0516] Program processing

[0517] The server receives reports sent from users who are subordinates or customers, and performs the following processes.

[0518] 1. Receiving the report

[0519] The server receives the text data of the report via an HTTP request.

[0520] 2. Analyzing the report

[0521] The server uses "spaCy" or "NLTK" to tokenize the text data of the received report and analyze its grammatical structure.

[0522] 3. Identifying the problem

[0523] The server extracts issues from the report based on the analysis results, using rule-based filtering to match the issues against a predefined list of issues.

[0524] 4. Analysis of the cause of the problem

[0525] The server then uses machine learning algorithms to analyze the causes of the extracted problems. Specifically, it uses Scikit-learn's random forest and TensorFlow's neural network to identify the causes using past report data.

[0526] 5. Deriving a solution

[0527] The server then generates solutions based on the results of the root cause analysis, generating specific proposals such as "increasing the resources of the design department" or "introducing a project management tool." In this embodiment, "2. Report Analysis" to "5. Deriving a Solution" are performed using a generative AI model (GPT-4). Specifically, the server derives the business problem, the cause of the problem, and the solution to the problem from the content of the received report using a prompt statement instructing the server to derive the business problem, the cause of the problem, and the solution to the problem, and using the generative AI model. At this time, by providing past report data to the generative AI model, the past report data can be taken into consideration when deriving the solution.

[0528] 6. Emotion recognition

[0529] The server uses an emotion engine to recognize user emotions from the report text. Specifically, it uses "TextBlob" and "VADER" to detect user emotions such as "anger" and "frustration." This information is used to adjust the priority of solutions.

[0530] 7. Outputting the results

[0531] The server compiles the analysis results, cause analysis, solutions, and emotion recognition results and provides feedback to the user. Specifically, the server generates the report using a generative AI model and a prompt statement instructing the server to generate a report that shows the derived business problems, causes of the problems, and solutions to the problems based on the recognized user emotions. At this time, the prompt statement instructs the server to generate a report in which the priority of the solutions is adjusted based on the recognized user emotions. For example, the server generates a report in HTML format and sends it to the user's device.

[0532] Specific examples

[0533] Suppose a user submits the following report:

[0534] "Recently, projects have been delayed, and we've been receiving more complaints from clients. This seems to be due in particular to the slow response of the design department."

[0535] Process flow:

[0536] 1. Receiving the report

[0537] The server receives the reports sent by the user via HTTP requests.

[0538] 2. Analyzing the report

[0539] The server uses "spaCy" to tokenize the report text and parse the grammatical structure.

[0540] 3. Identifying the problem

[0541] The server extracts problems such as "delays in project progress," "client complaints," and "delays in response from the design department" from the analysis results.

[0542] 4. Analysis of the cause of the problem

[0543] The server uses Scikit-learn's random forest to analyze past data and determine that the cause is a "lack of resources in the design department."

[0544] 5. Deriving a solution

[0545] Server proposes solutions such as "increasing the resources of the design department" and "introducing project management tools."

[0546] 6. Emotion recognition

[0547] The server uses "TextBlob" to sense the user's "dissatisfaction" from the report and adjust the priority of the solution.

[0548] 7. Outputting the results

[0549] The server compiles the analysis results, cause analysis, solutions, and emotion recognition results into an HTML report and sends it to the user's device.

[0550] Example prompt sentence:

[0551] Analyze the following report and identify the problem, cause, and solution. Also, recognize the user's feelings.

[0552] Report: Project progress has been slow recently, and we have received an increasing number of complaints from clients. This is likely due in particular to the slow response of the design department.

[0553] In this way, the server provides a system that efficiently reviews reports from subordinates and customers, extracts problems, analyzes their causes, and proposes solutions.

[0554] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0555] Step 1: Receive the report

[0556] The server receives the report sent by the user via an HTTP request. As input, it receives the text data sent by the user and passes the text data to the next processing step as output. Specifically, the server listens for HTTP requests and temporarily stores the received data.

[0557] Step 2: Analyze the report

[0558] The server uses "spaCy" or "NLTK" to tokenize the received report text data and analyze its grammatical structure. It uses the text data received in step 1 as input and generates tokenized text data and grammatical analysis results as output. Specifically, the server passes the text data to an NLP library to perform tokenization and grammatical analysis.

[0559] Step 3: Identify the problem

[0560] The server extracts issues from the report based on the analysis results. It uses the tokenized text data generated in step 2 and the grammar analysis results as input and generates a list of extracted issues as output. Specifically, the server performs rule-based filtering to match the list of extracted issues against a predefined list of issues.

[0561] Step 4: Analyze the cause of the problem

[0562] The server uses a machine learning algorithm to analyze the causes of the extracted issues. It uses the list of issues generated in step 3 as input and generates a list of causes as output. Specifically, the server uses past report data to identify causes using Scikit-learn's random forest or TensorFlow's neural network.

[0563] Step 5: Derive a solution

[0564] The server derives solutions based on the results of the cause analysis. It uses the list of causes generated in step 4 as input and generates a list of solutions as output. Specifically, the server uses a model trained on past successful cases to propose optimal solutions.

[0565] Step 6: Emotion Recognition

[0566] The server uses an emotion engine to recognize the user's emotions from the report text. It uses the tokenized text data generated in step 2 as input and generates emotion recognition results as output. Specifically, the server uses "TextBlob" and "VADER" to detect the user's emotions, such as "anger" and "dissatisfaction."

[0567] Step 7: Output the results

[0568] The server compiles the analysis results, cause analysis, solutions, and emotion recognition results and provides feedback to the user. It uses the various data generated in steps 3 to 6 as input and generates an HTML report as output. Specifically, the server integrates this data and generates an HTML report to send to the user's device.

[0569] (Application example 1)

[0570] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0571] With conventional report review systems, it was difficult to accurately analyze the content of reports submitted by subordinates and customers, extract problems, analyze the causes, and derive appropriate solutions. This also created a heavy workload for on-site managers, making it difficult for them to concentrate on other tasks. Furthermore, because the system was unable to recognize the user's emotions from the report or assess the seriousness of the problem, appropriate responses were delayed.

[0572] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means. In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using AI to identify problems and derive their causes and solutions; a means for delegating managerial authority to AI and having AI make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for recognizing user emotions from reports using an emotion engine and assessing the severity of problems; a means for analyzing report content and extracting problems using natural language processing technology; and a means for learning from past reports and their results using a machine learning algorithm to derive causes and solutions. This enables accurate analysis of report content, extraction of problems, analysis of causes, and deriving appropriate solutions. It also reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, recognizing user emotions from reports and assessing the severity of problems enables prompt and appropriate responses.

[0573] "Subordinates and customers" refers to employees who carry out work in a company or organization, and customers who use the services or products of a company or organization.

[0574] A "report" refers to a document that describes the progress of work, problems, proposals, etc.

[0575] "AI" is an abbreviation for artificial intelligence, and refers to the technology that allows computers to learn and reason by imitating human intelligence.

[0576] "Review" refers to the process of checking the contents of a submitted report and evaluating and analyzing it.

[0577] "Identifying problems" refers to finding business problems and issues from the contents of the report.

[0578] "Identifying causes and solutions" refers to analyzing the causes of the problems that have been identified and proposing appropriate solutions to them.

[0579] "Delegating managerial authority to AI" refers to transferring the decision-making authority held by on-site managers to artificial intelligence.

[0580] "On-site level judgment" refers to judgment on specific problems and situations that arise on-site.

[0581] "Reduction of labor burden" refers to reducing the labor burden associated with performing work.

[0582] "Emotion engine" refers to technology for analyzing and recognizing user emotions from text data.

[0583] "Natural language processing technology" refers to technology that enables computers to understand and analyze human language.

[0584] A "machine learning algorithm" refers to a computational method for learning patterns from data and making predictions or classifications.

[0585] "Past reports and their results" refers to previously submitted reports and the actions and results taken based on those reports.

[0586] As an embodiment of this invention, we will build a system in which AI automatically reviews reports submitted by factory workers and managers, extracts problems, analyzes their causes, and proposes solutions. Furthermore, we will use an emotion engine to recognize the emotions of workers and evaluate the severity of the problems.

[0587] The server implements the system using the following hardware and software.

[0588] Hardware and software used:

[0589] Hardware: Robots and servers in factories

[0590] Software: Python, spaCy (NLP library), TextBlob (sentiment analysis library), scikit-learn (machine learning library)

[0591] Data processing and calculation:

[0592] 1. Report analysis: The server uses spaCy to analyze the text data of reports submitted by workers and managers and extracts problems.

[0593] 2. Emotion Recognition: The server analyzes the emotion of the report using TextBlob and calculates a positive / negative score.

[0594] 3. Cause analysis: The server uses scikit-learn's RandomForestClassifier to predict the cause of the problem using a model learned from past data.

[0595] 4. Solution Proposal: The server proposes an appropriate solution based on the cause.

[0596] Examples:

[0597] The report read: "The machine was down due to a maintenance issue. The operator was very frustrated."

[0598] Results: Sentiment score: -0.5, Problem: "Maintenance issues", Cause: "Lack of maintenance", Solution: "Schedule maintenance"

[0599] Example prompt for a generative AI model:

[0600] The report read: "The machine was down due to a maintenance issue. The operator was very frustrated."

[0601] Analyze this report and suggest the problem, cause, and solution.

[0602] This system makes it possible to accurately analyze the contents of reports, extract problems, analyze the causes, and derive appropriate solutions. It also reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, by recognizing the user's emotions from the reports and assessing the seriousness of the problem, it becomes possible to take appropriate action quickly.

[0603] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0604] Step 1:

[0605] The server receives the text data of reports submitted by workers and managers. The input is the text data of the report, and the output is the text data to be analyzed. This text data is used in subsequent processing steps.

[0606] Step 2:

[0607] The server analyzes the report content using natural language processing technology. Specifically, it uses the spaCy library to analyze the text data and extract issues. The input is the text data to be analyzed, and the output is a list of extracted issues.

[0608] Step 3:

[0609] The server uses an emotion engine to recognize user emotions from reports. Specifically, it uses the TextBlob library to calculate emotion scores for text data. The input is the text data to be analyzed, and the output is the emotion score.

[0610] Step 4:

[0611] The server uses a machine learning algorithm to learn from past reports and their results and predict the cause of the problem. Specifically, it uses scikit-learn's RandomForestClassifier to predict the cause of the problem. The input is a list of extracted problems, and the output is a list of predicted causes.

[0612] Step 5:

[0613] The server proposes appropriate solutions based on the predicted causes. Specifically, it retrieves solutions corresponding to the causes from a database and proposes them. The input is a list of predicted causes, and the output is a list of proposed solutions.

[0614] Step 6:

[0615] The server compiles the analysis results, sentiment scores, predicted causes, and proposed solutions and outputs them as a report. The inputs are the analysis results, sentiment scores, predicted causes, and proposed solutions, and the output is the final report.

[0616] Through these processing steps, the server can accurately analyze the report content, extract the problem, analyze the cause, and derive an appropriate solution. It also recognizes the user's emotions and evaluates the seriousness of the problem, allowing for a prompt and appropriate response.

[0617] Example 2

[0618] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0619] On-site managers are required to perform a wide range of tasks, such as reviewing reports from subordinates and customers and approving vacation requests, which increases their workload and prevents them from concentrating on other important tasks. Furthermore, it is sometimes difficult for them to properly understand and judge the emotions and work situation of their subordinates, which can lead to poor decisions. This leads to a decline in work efficiency and delays in improving the work environment.

[0620] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0621] In this invention, the server includes a means for retrieving subordinates' past vacation history, current workload, and other related information from a database; a means for inputting the retrieved data into a generative AI model to determine whether to approve or reject the vacation request; and a means for recognizing subordinates' emotions using an emotion engine and adjusting decisions based on those emotions. This reduces the workload of on-site managers, allowing them to focus on other important tasks. Furthermore, by properly understanding subordinates' emotions and work situations and making more appropriate decisions, it is expected to improve work efficiency and the work environment.

[0622] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and automatically perform tasks such as learning, reasoning, and problem-solving.

[0623] A "generative AI model" is a type of artificial intelligence that generates new data and decisions based on input data, and is a model trained to provide optimal results for a specific task.

[0624] An "emotion engine" is a technology that analyzes human emotions from data such as text, audio, and images, and recognizes their emotional state.

[0625] A "database" is a system for efficiently storing, managing, and searching data, and is capable of handling large amounts of data while maintaining the consistency and integrity of the information.

[0626] A "terminal" is a device that a user directly operates, and includes a computer, a smartphone, a tablet, and the like.

[0627] "Vacation history" is a record of vacations taken by subordinates in the past, and includes information such as the type of vacation, duration, and date taken.

[0628] "Workload" is an indicator that shows the amount and content of work that a subordinate is responsible for within a specific period of time, and includes the number of tasks, difficulty, and time required.

[0629] A "field manager" is a position that is responsible for managing operations at the field level, instructing and evaluating subordinates, etc., and typically corresponds to a section chief or chief.

[0630] "Labor burden" refers to the effort and time required to carry out work, and excessive labor burden can lead to reduced work efficiency and stress.

[0631] A "judgment result" is a final conclusion or decision reached based on data analyzed by a generative AI model or emotion engine.

[0632] MODE FOR CARRYING OUT THE INVENTION

[0633] This invention is a system that delegates managerial authority to artificial intelligence and automates on-site decision-making. Specifically, the server retrieves subordinates' past vacation history, current workload, and other relevant information from a database, inputs it into a generative AI model, and determines whether to approve or reject vacation requests. It also uses an emotion engine to recognize the subordinates' emotions and adjusts the decision based on those emotions. The final decision is notified to the terminal.

[0634] Hardware and software used

[0635] Hardware: Servers, terminals

[0636] Software: Generative AI model, emotion engine, database management system (DBMS)

[0637] Data collection

[0638] The server retrieves the subordinate's past vacation history, current workload, and other relevant information from the database. For example, the server retrieves data from the DBMS using an SQL query. Specifically, it executes the query "SELECT FROM Vacation History WHERE Subordinate ID = 'A'".

[0639] Data analysis

[0640] The server inputs the acquired data into the generative AI model and determines whether to approve or reject the vacation request. Specifically, it inputs the prompt statement, "Considering subordinate A's past vacation history and current workload, should you approve or reject the vacation request?" into the generative AI model.

[0641] emotion recognition

[0642] The server recognizes the emotions of subordinates using the emotion engine. Specifically, it inputs a prompt sentence to the emotion engine: "Analyze the current emotional state of subordinate A and determine whether he is feeling stressed."

[0643] Adjusting judgment

[0644] The server adjusts the generative AI model's judgment based on the results of the emotion engine. For example, if subordinate A is recognized as feeling stressed, the server prioritizes approving his vacation request.

[0645] Notification of results

[0646] The server notifies the terminal of the final decision result. Specifically, the server generates a message saying "Subordinate A's vacation request has been approved" and sends it to the terminal of subordinate A. The terminal displays the received message.

[0647] Specific examples

[0648] Example prompt sentence:

[0649] "Should I approve or reject subordinate A's vacation request, taking into account his or her past vacation history and current workload?"

[0650] "Analyze subordinate A's current emotional state and determine whether he is feeling stressed."

[0651] This system allows managers to delegate on-site decision-making to AI, allowing them to focus on other tasks. For example, when a subordinate requests time off, the AI ​​automatically approves or rejects the request, reducing the burden on managers.

[0652] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0653] Step 1:

[0654] The server retrieves the subordinate's past vacation history, current workload, and other relevant information from a database.

[0655] Input: Subordinate ID

[0656] Data processing: The server uses SQL queries to retrieve subordinates' past vacation history and current workload from the DBMS.

[0657] Output: Acquired data (leave history, workload)

[0658] Specific operation: The server establishes a database connection and retrieves data by executing the query "SELECT FROM Vacation History WHERE Subordinate ID = 'A'". The retrieved data is temporarily stored in memory.

[0659] Step 2:

[0660] The server inputs the acquired data into a generative AI model to determine whether to approve or reject the vacation request.

[0661] Input: Acquired data (leave history, workload)

[0662] Data calculation: The server inputs the prompt statement to the generative AI model: "Considering subordinate A's past vacation history and current workload, should you approve or reject his vacation request?"

[0663] Output: The generative AI model's decision (accept or reject)

[0664] Specific operation: The server calls the API of the generative AI model and sends the prompt and data. The generative AI model analyzes the input data and returns a decision of approval or rejection.

[0665] Step 3:

[0666] The server recognizes the emotions of the subordinates using an emotion engine.

[0667] Input: Subordinates' emotional data

[0668] Data calculation: The server inputs a prompt statement to the emotion engine: "Analyze the current emotional state of subordinate A and determine whether he is feeling stressed."

[0669] Output: Emotion engine result (stress level)

[0670] Specific operation: The server calls the emotion engine's API and sends the emotion data of subordinate A. The emotion engine analyzes the emotion data and returns the stress level.

[0671] Step 4:

[0672] The server adjusts the generative AI model's decisions based on the results of the emotion engine.

[0673] Input: Generative AI model's judgment results, emotion engine results

[0674] Data calculation: The server receives the results of the emotion engine and reevaluates the judgment results of the generative AI model.

[0675] Output: Adjusted decision result (approve or reject)

[0676] Specific operation: The server obtains the results of the emotion engine and compares them with the judgment results of the generative AI model. If necessary, it corrects the judgment results.

[0677] Step 5:

[0678] The server notifies the terminal of the final decision result.

[0679] Input: Adjusted decision result

[0680] Data processing: The server generates a notification message.

[0681] Output: Information message

[0682] Specific operation: The server generates a message saying "Subordinate A's vacation request has been approved" and sends it to the terminal of subordinate A. The terminal displays the received message.

[0683] (Application example 2)

[0684] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0685] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. It also makes it difficult to monitor the operating status of machinery in factories in real time and perform appropriate maintenance and troubleshoot any abnormalities. This can lead to machinery breakdowns and reduced production efficiency.

[0686] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0687] In this invention, the server includes means for using AI to review reports submitted by subordinates and customers, extracting problems and identifying their causes and solutions, means for delegating managerial authority to AI and having it make on-site-level decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, means for monitoring the operating status of machinery within the factory, automatically detecting machinery requiring maintenance, and adjusting the maintenance schedule, and means for detecting machinery in an abnormal state and handling the abnormality. This reduces the workload of on-site managers, allowing them to focus on other tasks, and makes it possible to monitor the operating status of machinery within the factory in real time and perform appropriate maintenance and handle abnormalities.

[0688] "Reports from subordinates and customers" refers to reports and feedback submitted by employees within the organization or external customers.

[0689] "AI" is an abbreviation for artificial intelligence, a technology that enables computer systems to mimic human intelligence by learning, reasoning, and self-correcting.

[0690] A "review" is a process in which submitted information and data are examined in detail and evaluated and analyzed.

[0691] "Problem extraction" means finding issues or obstacles from data or information.

[0692] "Identifying causes and solutions" means identifying the root cause of the problem that has occurred and proposing appropriate measures to address it.

[0693] "Managerial authority" refers to the rights and responsibilities that managers have within an organization to make decisions and give instructions.

[0694] "On-site level decisions" refer to decisions made based on the specific circumstances and conditions at the site.

[0695] "Reducing labor burden" means reducing the workload and stress of employees.

[0696] "Focusing on other tasks" means redirecting the time and resources freed up from a specific task to other important tasks.

[0697] "Machinery and equipment in a factory" refers to production equipment and machinery used in a factory.

[0698] "Monitoring operational status" means checking the operating status and performance of machinery and equipment in real time.

[0699] "Machinery requiring maintenance" refers to machinery that is in a state that requires regular maintenance or repair.

[0700] "Adjusting the maintenance schedule" means planning and determining the timing of maintenance and repairs of machinery and equipment.

[0701] "Machinery in an abnormal state" refers to machinery that is deviating from normal operation and experiencing a breakdown or malfunction.

[0702] "Abnormality processing" refers to detecting abnormalities in mechanical equipment and taking appropriate measures to return it to a normal state.

[0703] The system for implementing this invention consists of three main elements: a server, a terminal, and a user. The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. It also delegates managerial authority to AI, allowing it to make on-site decisions, reducing the workload of on-site managers and allowing them to focus on other tasks. It also monitors the operating status of machinery within a factory, automatically detecting machinery requiring maintenance and adjusting maintenance schedules. It is also possible to detect machinery in an abnormal state and handle the abnormality.

[0704] The server runs a program implemented using Python. Specifically, it periodically checks reports stored in a database (e.g., MySQL) and machine operation data, and uses AI models to identify problems and analyze their causes. It uses a real-time monitoring system (e.g., Prometheus) to monitor the operation status of the machine and respond immediately if an abnormality occurs.

[0705] Terminals include sensors installed on each machine in the factory, as well as computers and smartphones used by managers. These terminals communicate with the server to send and receive necessary data. For example, they send operational data and abnormal status information on machines to the server and receive instructions from the server.

[0706] The users are factory managers and engineers who take appropriate action based on the information provided by the server. For example, if the server detects that a machine needs maintenance, the users follow the instructions to carry out the maintenance. Also, if an abnormality occurs, the users respond by following the instructions from the server to handle the abnormality.

[0707] For example, if a robot in a factory has not undergone maintenance for 40 days, the server will detect the robot and adjust its maintenance schedule. Also, if a robot is in an error state, the server will detect the error and take appropriate action.

[0708] An example of a prompt to input to a generative AI model is as follows:

[0709] Create a Python program that monitors the operating status of robots in a factory, automatically detects robots that require maintenance, and adjusts the maintenance schedule. Also, add a function to detect robots that are in an error state and handle the error.

[0710] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0711] Step 1:

[0712] The server retrieves reports submitted by subordinates and customers from the database. It uses the report data stored in the database as input and generates a dataset to input into the AI ​​model as output. Specifically, it executes SQL queries to extract the report data and convert it into the required format.

[0713] Step 2:

[0714] The server uses an AI model to review the report, identify issues, their causes, and solutions. It uses the dataset generated in step 1 as input and generates a list of issues, causes, and solutions as output. Specifically, it uses natural language processing technology to analyze the report and extract key information.

[0715] Step 3:

[0716] The server delegates managerial authority to the AI ​​and makes on-site decisions. It uses leave requests and work status data from subordinates as input, and generates a decision on whether to approve or reject the leave request as output. Specifically, it takes into account past leave history and current workload to make the decision.

[0717] Step 4:

[0718] The server monitors the operation status of machinery in the factory in real time. It uses sensor data sent from the terminal as input and generates a report on the operation status of the machinery as output. Specifically, it collects data from the sensors and checks for any abnormalities.

[0719] Step 5:

[0720] The server automatically detects machines that require maintenance and adjusts the maintenance schedule. It uses the operation status report generated in step 4 as input and generates a maintenance schedule as output. Specifically, it determines the need for maintenance based on the number of operating days and abnormality data.

[0721] Step 6:

[0722] The server detects abnormal machines and handles the abnormalities. It uses the operational status report generated in step 4 as input and generates instructions for handling the abnormalities as output. Specific operations include analyzing the abnormal data and proposing appropriate countermeasures.

[0723] Step 7:

[0724] The user takes appropriate action based on the information provided by the server. The input is the maintenance schedule and abnormality handling instructions from the server, and the output is the actual maintenance work and abnormality handling. Specific actions include maintenance and repair of machinery.

[0725] Example 3

[0726] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0727] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. Furthermore, it is difficult to solve problems efficiently because it takes time and effort to extract problems from meeting minutes and derive causes and solutions. Furthermore, it is difficult to understand the flow of discussions and the emotions of participants, which hinders problem solving.

[0728] The identification process by the identification processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using artificial intelligence to extract problems and derive causes and solutions; a means for delegating managerial authority to artificial intelligence and having it make on-site-level decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing the contents of minutes using natural language processing technology and extracting problems; a means for learning from past data using a machine learning algorithm and deriving causes and solutions; and a means for recognizing participants' emotions from text using an emotion engine and understanding the flow of the discussion. This reduces the workload of on-site managers and enables efficient problem solving.

[0729] "Artificial intelligence" is a technology that enables computer systems to mimic human intelligence and automatically perform tasks such as learning, reasoning, and problem-solving.

[0730] "Managerial authority" refers to the authority that managers have within an organization to make decisions and give instructions.

[0731] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[0732] A "machine learning algorithm" is an algorithm that learns patterns from data and makes predictions and classifications based on the results of that learning.

[0733] An "emotion engine" is a technology that recognizes and analyzes human emotions from data such as text and voice.

[0734] Minutes are documents that record the contents of meetings and discussions.

[0735] A "problem" is a challenge or obstacle that needs to be resolved in a particular situation or system.

[0736] A "cause" is the factor or reason why a problem occurs.

[0737] A "solution" is a specific method or means for solving a problem.

[0738] A "report" is a document that compiles information about a particular topic or situation.

[0739] "On-site level judgment" refers to decisions and responses made based on the situation on-site.

[0740] This invention is a system that uses artificial intelligence to analyze minutes of a meeting, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system will be described below.

[0741] First, the user uploads the meeting minutes to the system. Minutes are often entered as text files or PDF format. The user uploads a file called "Minutes.txt," and the system receives the file and sends it to the server.

[0742] The server uses natural language processing technology to analyze the received minutes. Specifically, it uses the Python libraries NLTK and spaCy. The server analyzes the text using import spacy and extracts sentence structure and keywords. For example, it analyzes the text using nlp = spacy.load("en_core_web_sm") and understands the sentence structure using doc = nlp(text).

[0743] Next, the server uses the analysis results to understand the content of the discussion and extract issues. This is done using text classification algorithms and topic modeling. Specifically, from sklearn.feature_extraction.text import CountVectorizer is used to extract frequently occurring keywords and phrases from the text data. For example, vectorize the text using vectorizer = CountVectorizer(), and identify frequently occurring issues using X = vectorizer.fit_transform(doc).

[0744] The server then analyzes the causes of the extracted issues using a machine learning model trained on past meeting minutes data and their results. Specifically, it predicts the causes of issues using from sklearn.ensemble import RandomForestClassifier. For example, train the model using clf = RandomForestClassifier() and predict the cause using clf.predict(X).

[0745] The server then derives a solution based on the cause. It uses a recommendation system to refer to past success stories. Specifically, it uses from surprise import SVD to propose the optimal solution. For example, build a recommendation system using algo = SVD() and derive a solution using algo.fit(trainset).

[0746] In another embodiment, the server uses an emotion engine to recognize participants' emotions from the text of the minutes. This allows the flow of the discussion to be understood and the causes and solutions of problems to be derived. Specifically, emotions are extracted from the text using from transformers import pipeline. For example, emotions are analyzed using emotion_pipeline = pipeline("sentiment-analysis"), and emotion information is obtained using results = emotion_pipeline(text).

[0747] As a concrete example, the following prompt sentence is input into the generative AI model:

[0748] Examples:

[0749] Meeting minutes: "Project X is progressing slowly due to a lack of resources and communication issues. The proposed solutions are to provide additional resources and set up regular meetings."

[0750] Prompt statement:

[0751] "Analyze the minutes of the following meeting and extract the problem, cause, and solution. Minutes: 'Project X is lagging behind. The cause is a lack of resources and communication issues. The proposed solutions are to allocate additional resources and set up regular meetings.'"

[0752] In this way, the server realizes a system that analyzes the minutes of the meeting, extracts problems, analyzes the causes, and derives solutions. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0753] Step 1: Enter the minutes

[0754] Users upload meeting minutes to the system, often as text files or PDFs.

[0755] Input: The minutes file uploaded by the user (e.g. "minutes.txt")

[0756] Output: Minutes data sent to the server

[0757] Specific operation: The user uploads "Minutes.txt" through the system interface, and the system receives the file and sends it to the server.

[0758] Step 2: Analyze the minutes

[0759] The server uses natural language processing techniques, specifically the Python libraries NLTK and spaCy, to analyze the received minutes.

[0760] Input: Minutes data sent to the server

[0761] Output: Analyzed text data (sentence structure and keywords)

[0762] Specific operation: The server analyzes the text using import spacy to extract sentence structure and keywords. For example, analyze the text using nlp = spacy.load("en_core_web_sm") and understand the sentence structure using doc = nlp(text).

[0763] Step 3: Identify the problem

[0764] The server uses the analysis results to understand the content of the discussion and extract issues, using text classification algorithms and topic modeling.

[0765] Input: Parsed text data

[0766] Output: Extracted issues

[0767] What it does: The server extracts frequently occurring keywords and phrases from the text data using from sklearn.feature_extraction.text import CountVectorizer. For example, it vectorizes the text using vectorizer = CountVectorizer() and identifies frequently occurring issues using X = vectorizer.fit_transform(doc).

[0768] Step 4: Analyze the cause

[0769] The server analyzes the causes of the extracted problems using a machine learning model that has learned from past meeting minutes data and their results.

[0770] Input: Extracted issues

[0771] Output: Analyzed causes

[0772] Specific operation: The server predicts the cause of the problem using from sklearn.ensemble import RandomForestClassifier. For example, train the model using clf = RandomForestClassifier() and predict the cause with clf.predict(X).

[0773] Step 5: Derive a solution

[0774] The server derives solutions based on the cause, using a recommendation system to reference past success stories.

[0775] Input: Analyzed cause

[0776] Output: Proposed solution

[0777] What it does: The server uses from surprise import SVD to propose the optimal solution. For example, build a recommendation system using algo = SVD() and derive a solution using algo.fit(trainset).

[0778] Step 6: Emotion Recognition (Alternative Embodiment)

[0779] The server uses an emotion engine to recognize participants' emotions from the text of the minutes, thereby understanding the flow of the discussion and identifying the causes of problems and solutions.

[0780] Input: Text data of meeting minutes

[0781] Output: Recognized emotion information

[0782] Specific operation: The server extracts emotions from text using from transformers import pipeline. For example, it analyzes emotions using emotion_pipeline = pipeline("sentiment-analysis") and obtains emotion information using results = emotion_pipeline(text).

[0783] (Application example 3)

[0784] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0785] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. Analyzing meeting minutes, identifying problems, analyzing causes, and proposing solutions takes time and effort, making it difficult to carry out these tasks efficiently. Furthermore, it is difficult to recognize the emotions of participants during meetings and understand the flow of discussions.

[0786] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0787] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using AI to identify problems and derive their causes and solutions; a means for delegating managerial authority to AI and allowing it to make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing meeting minutes in real time, identifying problems, analyzing causes, and proposing solutions; and a means for using an emotion engine to recognize participants' emotions from the text in the minutes and understand the flow of the discussion. This reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, the system can efficiently analyze meeting minutes, identify problems, analyze causes, and propose solutions, while recognizing participants' emotions and understanding the flow of the discussion.

[0788] "Reports submitted by subordinates or customers" refers to reports or documents created and submitted by subordinates or customers in a company or organization.

[0789] "AI" stands for artificial intelligence, a system that mimics human intellectual tasks using technologies such as machine learning and natural language processing.

[0790] A "review" is the process of examining specific documents or data in detail and evaluating or analyzing them.

[0791] "Problem extraction" refers to the process of finding problematic parts or elements in specific documents or data.

[0792] "Cause analysis" is the process of investigating and understanding the reasons and background behind a particular problem.

[0793] "Solution proposal" is the act of proposing effective measures or methods for a specific problem.

[0794] "Managerial authority" refers to the decision-making and command powers held by managers in a company or organization.

[0795] "Field-level decisions" are decisions made based on the situation and conditions at the field.

[0796] "Reducing labor burden" means reducing the workload and stress of employees and managers.

[0797] A "meeting minutes" is a document that records the contents and statements made at a meeting.

[0798] "Real-time analysis" refers to the process of processing data and information instantly and obtaining results.

[0799] An "emotion engine" is a technology or system for recognizing and analyzing emotions from text and speech.

[0800] "Participant emotions" refers to the emotional state and feelings of people attending a meeting or event.

[0801] "Discussion flow" refers to the progression and development of topics in a meeting or discussion.

[0802] To implement the present invention, the following system configuration and program are required.

[0803] System Configuration

[0804] The server uses the following hardware and software:

[0805] Hardware: Robots (e.g. Pepper, NAO, etc.) installed in conference rooms within the factory

[0806] Software: Python, Spacy, Transformers, TextBlob

[0807] Program processing explanation

[0808] The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. Specifically, it uses natural language processing (NLP) technology to analyze the content of the reports and extract problems. It then uses machine learning algorithms to learn from past reports and their results and derive causes and solutions.

[0809] The server also analyzes meeting minutes in real time, extracting problems, analyzing their causes, and proposing solutions. It uses an emotion engine to recognize participants' emotions from the text in the minutes and understand the flow of the discussion. This makes it possible to instantly identify problems during a meeting and propose appropriate solutions.

[0810] Specific examples

[0811] For example, if the minutes of a factory meeting state that "delays on the production line are a problem," the server analyzes the minutes in real time and identifies "delays on the production line" as the problem. It then references past minutes data to analyze the cause of similar problems. For example, if it identifies "machine failure" as the cause, it suggests "performing regular maintenance" as a solution.

[0812] Furthermore, by using an emotion engine to recognize the emotions of participants, it is possible to understand the flow of the discussion from emotional information, such as when the discussion is heated or when participants are confused, and to present solutions at the appropriate time.

[0813] Prompt Sentence Examples

[0814] “Analyze meeting minutes, extract issues, identify sentiment, and propose solutions.

[0815] Minutes: 'Meeting minutes. The problem is delays on the production line. Participants are unhappy.'

[0816] In this way, the server can reduce the workload of on-site managers, allowing them to focus on other tasks. It can also efficiently analyze meeting minutes, identify problems, analyze causes, and propose solutions, and it can recognize participants' emotions and understand the flow of discussions.

[0817] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0818] Step 1:

[0819] The server receives reports submitted by subordinates and customers. The input is the text data of the report, and the output is the storage of that text data. Specifically, when a report is sent to the server, the server stores its contents in a database.

[0820] Step 2:

[0821] The server analyzes the text data of the saved reports using natural language processing (NLP) technology. The input is the text data of the saved reports, and the output is a list of issues as a result of the analysis. Specifically, Spacy is used to tokenize the text and extract important keywords and phrases.

[0822] Step 3:

[0823] The server analyzes the causes of the extracted problems by referencing past report data. The input is a list of problems and past report data, and the output is a list of causes. Specifically, a machine learning algorithm is used to learn patterns from past data and identify the causes of current problems.

[0824] Step 4:

[0825] The server proposes solutions based on the identified causes. The input is a list of causes and the output is a list of solutions. Specifically, it uses TextBlob to parse the text of the causes and generate positive solutions.

[0826] Step 5:

[0827] The server receives meeting minutes in real time and stores them as text data. The input is the audio data of the meeting, and the output is the minutes converted into text. Specifically, the audio data is converted into text using speech recognition technology and stored in a database.

[0828] Step 6:

[0829] The server analyzes the text data of the saved minutes in real time and extracts issues. The input is the text of the minutes, and the output is a list of issues. Specifically, it uses Spacy to tokenize the text and extract important keywords and phrases.

[0830] Step 7:

[0831] The server recognizes participants' emotions from the text data of the minutes. The input is the transcribed minutes, and the output is a list of emotions. Specifically, emotions are extracted from the text using the Transformers sentiment analysis pipeline.

[0832] Step 8:

[0833] The server understands the flow of the discussion based on the extracted issues and emotion information and proposes appropriate solutions. The input is a list of issues and a list of emotions, and the output is a list of solutions. Specifically, it uses TextBlob to analyze the issues and emotion information and generate appropriate solutions.

[0834] In this way, the server can reduce the workload of on-site managers, allowing them to focus on other tasks. It can also efficiently analyze meeting minutes, identify problems, analyze causes, and propose solutions, and it can recognize participants' emotions and understand the flow of discussions.

[0835] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0836] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0837] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0838] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0839] [Second embodiment]

[0840] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0841] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0842] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0843] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0844] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0845] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0846] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0847] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0848] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0849] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0850] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0851] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0852] "Example 1"

[0853] One aspect of the present invention is a system that uses AI to review reports submitted by subordinates and customers. Specifically, the AI ​​analyzes the text data of the report and extracts problems. The AI ​​then analyzes the causes of the problems and derives solutions. This AI uses natural language processing (NLP) technology to understand the content of the report and extract problems. The AI ​​also uses machine learning algorithms to learn from past reports and their results, and derives causes and solutions.

[0854] "Example 2"

[0855] Furthermore, one aspect of the present invention is a system that delegates managerial authority to AI. Specifically, AI takes over on-site decision-making that would normally be performed by managers, allowing them to focus on other tasks. For example, AI can automatically approve or reject vacation requests from subordinates. This decision is based on the subordinate's past vacation history, current workload, and other relevant information.

[0856] "Example 3"

[0857] Another aspect of the present invention is a system in which AI reviews meeting minutes, extracts problems, and derives their causes and solutions. Specifically, the AI ​​analyzes the meeting minutes and extracts problems from the content of the discussion. The AI ​​then analyzes the causes of the problems and derives solutions. This AI uses natural language processing (NLP) technology to understand the content of the minutes and extract the problems. The AI ​​also uses machine learning algorithms to learn from past minutes and their results, and derives causes and solutions.

[0858] The processing flow of each embodiment will be described below.

[0859] "Example 1"

[0860] Step 1: Reports from subordinates and customers are sent to the system.

[0861] Step 2: AI analyzes the text data of the report and extracts problem areas. This analysis is performed using natural language processing (NLP) technology.

[0862] Step 3: AI analyzes the cause of the problem and derives a solution. This analysis is done using machine learning algorithms.

[0863] "Example 2"

[0864] Step 1: Subordinate requests for vacation time and other matters requiring on-site decisions typically made by managers are sent to the system.

[0865] Step 2: The AI ​​evaluates the matter and makes a decision, based on the employee's past leave history, current workload, and other relevant information.

[0866] Step 3: The AI's decision is notified to management, allowing them to focus on other tasks.

[0867] "Example 3"

[0868] Step 1: The meeting minutes are sent to the system.

[0869] Step 2: AI analyzes the text data of the minutes and extracts issues from the content of the discussion. This analysis is carried out using natural language processing (NLP) technology.

[0870] Step 3: AI analyzes the cause of the problem and derives a solution. This analysis is done using machine learning algorithms.

[0871] Example 1

[0872] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0873] On-site managers are required to perform a wide range of tasks, including reviewing reports submitted by subordinates and customers, identifying problems, analyzing causes, and deriving solutions, which places a heavy workload on them and makes it difficult for them to concentrate on other important tasks. Accurately analyzing the contents of reports and deriving appropriate solutions also requires a high level of specialized knowledge and experience, which further increases the burden on on-site managers. Furthermore, reviewing meeting minutes and making decisions based on past experience and future predictions are equally burdensome.

[0874] The identification process by the identification processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means. In this invention, the server includes a means for reviewing reports submitted by subordinates and customers using artificial intelligence, extracting problems and deriving causes and solutions, a means for analyzing the contents of the reports using natural language processing technology, and a means for learning from past reports and their results using a machine learning algorithm to derive causes and solutions. This reduces the workload of on-site managers, allowing them to focus on other important tasks.

[0875] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and learn, reason, and self-correct.

[0876] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[0877] A "machine learning algorithm" is an algorithm that learns patterns from data and makes predictions and classifications.

[0878] A "report" is a document about work or a situation that is sent by a subordinate or a customer.

[0879] "Review" is the process of evaluating the contents of a report and identifying problems and areas for improvement.

[0880] "Problems" are issues or obstacles that need to be resolved and are identified in the report.

[0881] "Cause" refers to the reason or background behind the occurrence of a problem.

[0882] A "solution" is a specific method or means for solving a problem.

[0883] "Managerial authority" refers to the authority that on-site managers have regarding the execution of work and making decisions.

[0884] "On-site level judgment" is a judgment made based on the specific situation at the site.

[0885] "Labor burden" refers to the burden and stress of work that accompanies the performance of work.

[0886] "Text data" is textual information extracted from a report.

[0887] "Analysis results" are the results of analysis obtained using natural language processing technology and machine learning algorithms.

[0888] "Minutes" are documents that record the contents of a meeting.

[0889] "Past experience" refers to knowledge and data about tasks and situations that have been accumulated up to now.

[0890] "Future forecasting" is the prediction of future conditions based on past data and current conditions.

[0891] This invention is a system that uses artificial intelligence to review reports submitted by subordinates and customers, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system will be described below.

[0892] Program Generation

[0893] The server generates a program to review reports submitted by subordinates and customers. This program is built by combining natural language processing (NLP) technology and machine learning algorithms. Specifically, it uses SpaCy and NLTK as NLP libraries and Scikit-learn and TensorFlow as machine learning libraries.

[0894] Program processing

[0895] The server performs the following processing using the generated program.

[0896] 1. Receiving reports and obtaining text data

[0897] The terminal transmits the report created by the user to the server.

[0898] The server receives the report sent from the terminal and extracts the text data of the report. Specifically, if the report is in PDF format, it extracts the text using a PDF parser (e.g., PyPDF2).

[0899] 2. Analysis using natural language processing (NLP)

[0900] The server analyzes the extracted text data using NLP techniques, specifically using SpaCy and NLTK to tokenize the text, tag parts of speech, and analyze dependencies.

[0901] 3. Identifying the problem

[0902] The server extracts problem areas from the report based on the results of NLP analysis, specifically detecting predefined keywords and phrases (e.g., "declining sales" or "customer complaints").

[0903] 4. Cause analysis using machine learning algorithms

[0904] The server analyzes the causes of the extracted problems using a machine learning model trained on past report data. Specifically, it uses a model built using Scikit-learn and TensorFlow.

[0905] 5. Deriving a solution

[0906] The server then derives appropriate solutions based on the results of the root cause analysis, specifically by referencing past success stories and best practices.

[0907] 6. Return of results

[0908] The server compiles the identified problem, cause, and solution and returns it to the terminal.

[0909] The terminal displays the results received from the server to the user.

[0910] Specific examples

[0911] The user sends a "sales report" from the terminal to the server, which includes a description of the reasons for the decline in sales.

[0912] 1. The terminal sends the "sales report" created by the user to the server.

[0913] 2. The server receives the report sent from the device and extracts the text data using a PDF parser.

[0914] 3. The server tokenizes the extracted text data using SpaCy, and performs part-of-speech tagging and dependency analysis.

[0915] 4. The server extracts the problem of "declining sales" from the analysis results.

[0916] 5. The server uses a model trained on past sales reports using Scikit-learn to identify that the cause is a "reduced advertising budget."

[0917] 6. The server comes up with the solution: "Increase advertising spending."

[0918] 7. The server summarizes the problem, cause, and solution and returns it to the terminal.

[0919] 8. The terminal displays the results received from the server to the user.

[0920] Prompt Sentence Examples

[0921] "Analyze the following sales report and identify the problem, its cause, and a solution. Report content: 'Sales this month have decreased by 20% compared to last month. This is likely due to a reduction in advertising expenses.'"

[0922] In this way, the server performs specific processing at each step, and ultimately the system automatically performs everything from identifying problems to proposing solutions to the user.

[0923] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0924] Step 1: Receive the report

[0925] Input: Reports sent by users from their terminals

[0926] Specific operation: The terminal sends the report created by the user to the server in PDF or text format.

[0927] Output: Report data received by the server

[0928] Step 2: Extract text data

[0929] Input: Report data received by the server

[0930] What happens: The server extracts text data from the received report. If the report is in PDF format, it uses a PDF parser (e.g. PyPDF2) to extract the text.

[0931] Output: Extracted text data

[0932] Step 3: Natural Language Processing (NLP) Analysis

[0933] Input: Extracted text data

[0934] What it does: The server analyzes text data using NLP techniques, such as SpaCy and NLTK, to tokenize text, tag parts of speech, and analyze dependencies.

[0935] Output: Parsing results (tokenized text, part-of-speech tags, dependencies)

[0936] Step 4: Identify the problem

[0937] Input: NLP analysis results

[0938] What it does: The server extracts problems in the report based on the analysis results, detecting predefined keywords and phrases (e.g., "declining sales," "customer complaints").

[0939] Output: Extracted issues

[0940] Step 5: Root cause analysis using machine learning algorithms

[0941] Input: Extracted issues

[0942] How it works: The server analyzes the cause using a machine learning model trained on past report data. Specifically, it uses a model built using Scikit-learn and TensorFlow.

[0943] Output: Analysis results (cause of the problem)

[0944] Step 6: Derive a solution

[0945] Input: Analysis results (cause of problem)

[0946] Specific operation: The server derives appropriate solutions based on the results of the cause analysis, and proposes solutions by referring to past success stories and best practices.

[0947] Output: Proposed solution

[0948] Step 7: Returning the results

[0949] Input: Proposed Solution

[0950] Specific operation: The server compiles the identified problem, cause, and solution and returns it to the terminal.

[0951] Output: Results returned to the device (problem, cause, solution)

[0952] Step 8: Viewing the results

[0953] Input: Results returned to the device (problem, cause, solution)

[0954] Specific operation: The terminal displays the results received from the server to the user.

[0955] Output: The results displayed to the user (problem, cause, solution)

[0956] In this way, the server performs specific processing at each step, and ultimately the system automatically performs everything from identifying problems to proposing solutions to the user.

[0957] (Application example 1)

[0958] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0959] Conventional report review systems required humans to manually analyze the contents of reports submitted by subordinates and customers, extracting problems, analyzing causes, and proposing solutions, which was time-consuming, labor-intensive, and inefficient. This also placed a heavy burden on factory workers and managers, increasing the workload of on-site managers and preventing them from concentrating on other tasks. Furthermore, the system lacked the ability to learn from past reports, making it difficult to improve the accuracy of analysis.

[0960] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0961] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers through AI, identifying problems and proposing causes and solutions; a means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means installed on a robot used in a factory, which automatically reviews reports submitted by workers and managers, identifying problems, analyzing causes, and proposing solutions; and a means for learning from past reports and their results to improve analysis accuracy. This enables automatic analysis of reports, identifying problems, analyzing causes, and proposing solutions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. Furthermore, learning from past reports can improve analysis accuracy.

[0962] "Subordinates and customers" refers to employees in a company or organization who receive instructions, and external users of products and services.

[0963] A "report" is a document that summarizes information or results related to a particular task or project.

[0964] "AI" is an abbreviation for artificial intelligence, and refers to the technology that allows computers to imitate human intelligence to learn and reason.

[0965] "Review" refers to the process of evaluating and analyzing submitted documents and data to identify problems and areas for improvement.

[0966] A "problem" refers to a factor or issue that hinders normal progress in a particular task or project.

[0967] "Cause" refers to the fundamental reason or factor that causes a problem.

[0968] A "solution" refers to a specific method or means for solving a problem.

[0969] "Managerial authority" refers to the authority of a person in a position of authority who has the authority to give instructions and make decisions regarding work in a company or organization.

[0970] "On-site level judgment" refers to making the necessary judgments and decisions immediately at the actual work site.

[0971] "Labor burden" refers to the burden of labor and work necessary to carry out business.

[0972] "Robots used in factories" refers to machinery and devices used in manufacturing and production industries for the purpose of automating and streamlining work.

[0973] "Workers" refers to employees who actually perform work in factories or on-site.

[0974] A "manager" refers to a person in a position of authority who supervises workers and manages operations at a factory or on-site.

[0975] "Automated review" refers to artificial intelligence or machines analyzing and evaluating documents and data without human intervention.

[0976] "Learning" refers to the process by which artificial intelligence acquires new knowledge and skills from past data and experiences.

[0977] "Analysis accuracy" refers to the accuracy and reliability of analyzing data or documents.

[0978] The following system configuration will be described as an embodiment of the present invention.

[0979] The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. Specifically, the server uses natural language processing (NLP) technology to analyze the content of the reports and extract problems. Furthermore, it uses machine learning algorithms to learn from past reports and their results and derive causes and solutions.

[0980] The server also includes a means to delegate managerial authority to AI and have it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. This frees on-site managers from routine decision-making tasks and allows them to focus on more strategic tasks.

[0981] Furthermore, it includes a means to be installed on robots used in factories and automatically review reports submitted by workers and managers, extract problems, analyze their causes, and propose solutions.The robot analyzes reports submitted by workers and managers in real time and provides immediate feedback.

[0982] The server includes a means for learning from past reports and their results to improve the accuracy of the analysis, which improves the accuracy of the analysis over time, enabling more accurate problem detection and solution proposals.

[0983] In terms of specific hardware, the server uses a computer system equipped with a high-performance processor and large memory capacity. For software, spaCy is used for natural language processing, and scikit-learn is used for machine learning. These software programs provide the functions necessary to analyze the text data in the reports, extract problems, analyze their causes, and propose solutions.

[0984] As a concrete example, consider the following report submitted:

[0985] "Machine A's operation is unstable."

[0986] "Product B's quality varies."

[0987] "Production speed on Line C is slow."

[0988] In response to these reports, the server makes the following suggestions:

[0989] "Machine A is operating unstable. Please perform maintenance on the machine."

[0990] "The quality of Product B is inconsistent. -> Please review your quality control process."

[0991] "The production speed on line C is slow. Please adjust the speed of the production line."

[0992] Example prompts to input to a generative AI model:

[0993] "Machine A's operation is unstable." "Product B's quality varies." "Production speed on line C is slow." Please suggest causes and solutions for these problems.

[0994] By using this prompt, the generative AI model is expected to propose an appropriate cause analysis and solution.

[0995] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0996] Step 1:

[0997] The server receives reports submitted by subordinates and customers. The input is report data in text format. The server analyzes this data using natural language processing (NLP) technology and divides it into sentences. Specifically, it uses the spaCy library to analyze the text and divide it into sentences.

[0998] Step 2:

[0999] The server extracts problems from the divided sentences. The input is the sentence data divided in step 1. The server analyzes each sentence and identifies sentences that contain problems. Specifically, it uses spaCy's dependency analysis function to extract keywords and phrases that indicate problems.

[1000] Step 3:

[1001] The server analyzes the causes of the extracted problems. The input is the problem sentence data extracted in step 2. The server uses a machine learning algorithm to identify the cause using a model trained from past report data and its results. Specifically, the server uses the scikit-learn library to convert the problem sentence into a feature vector, which is then input into the trained model to predict the cause.

[1002] Step 4:

[1003] The server proposes a solution based on the identified cause. The input is the cause data identified in step 3. The server searches a historical database for similar problems and their solutions and proposes the optimal solution. Specifically, it executes a database query to retrieve similar problems and their solutions, and selects the optimal solution.

[1004] Step 5:

[1005] The server generates a report with the proposed solution and provides feedback to the user. The input is the solution data proposed in step 4. The server generates a report including the solution and sends it to the user. Specifically, the server embeds the solution in a report template and notifies the user via email or a dashboard.

[1006] Step 6:

[1007] The server learns from past reports and their results to improve analysis accuracy. The input is user feedback and past report data. The server analyzes the feedback and updates the machine learning model. Specifically, it uses the scikit-learn library to retrain the model with new data and improve analysis accuracy.

[1008] Example 2

[1009] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1010] On-site managers are required to handle a wide range of tasks, such as reviewing reports from subordinates and customers and approving vacation requests, which increases their workload and prevents them from concentrating on other important tasks. In addition, these judgments must be made quickly and accurately, further increasing the burden on managers.

[1011] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1012] In this invention, the server includes a means for reviewing reports submitted by subordinates and customers using artificial intelligence to identify problems and derive their causes and solutions, a means for delegating managerial authority to the artificial intelligence and having the AI ​​make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, a means for accepting vacation requests, collecting and preprocessing related information, and inputting prompts into the generative AI model to obtain a judgment, and a means for obtaining the judgment result and notifying the user. This reduces the workload of managers, allowing them to focus on other important tasks.

[1013] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and perform functions such as learning, reasoning, and problem-solving.

[1014] "Managerial authority" refers to the authority that managers have within an organization to make decisions and give instructions.

[1015] "On-site level decisions" are decisions made to respond quickly and appropriately to specific problems and situations that arise on-site.

[1016] "Labor burden" refers to the physical and mental burden incurred when performing work.

[1017] A "leave application" is the application documents and procedures submitted by an employee to obtain leave.

[1018] A "generative AI model" is an artificial intelligence model that uses natural language processing and machine learning techniques to generate appropriate responses and decisions based on input data.

[1019] A "prompt" is text containing instructions or questions to be input into a generative AI model.

[1020] "Data preprocessing" is the process of formatting collected data into a form suitable for analysis or model input.

[1021] A "judgment result" is the response or decision that the generative AI model outputs based on the prompt.

[1022] "User" means an individual or member of an organization who uses the system.

[1023] This invention is a system that delegates the authority of managers to artificial intelligence and allows them to make on-site decisions, thereby reducing the workload of managers and allowing them to focus on other tasks. A specific embodiment of this system will be described below.

[1024] Hardware and software used

[1025] Hardware: Servers, terminals

[1026] Software: Generative AI models (e.g., GPT-4), database management systems (e.g., MySQL), business management software (e.g., Jira)

[1027] System Overview

[1028] The system delegates tasks like reviewing reports from subordinates and clients, approving vacation requests, and other on-the-ground decisions to artificial intelligence, allowing managers to focus on other important tasks.

[1029] Data processing and calculation flow

[1030] 1. Acceptance of leave requests

[1031] User: A subordinate uses a terminal to apply for leave. Specifically, the subordinate enters the necessary information into the leave application form in the business management software and clicks the submit button.

[1032] Terminal: Sends the information entered in the vacation request form to the server.

[1033] 2. Data Collection

[1034] Server: After receiving the leave request, access the database management system to collect relevant information such as the subordinate's past leave history, current workload, and workload of team members.

[1035] What it does: Executes an SQL query to retrieve the required data.

[1036] 3. Data Preprocessing

[1037] Server: Preprocesses the collected data, specifically formatting the data and filling in missing values.

[1038] Specific operation: Standardize the data format and impute missing values ​​with the mean or median.

[1039] 4. Prompt generation

[1040] Server: Based on the preprocessed data, it generates prompt sentences to be input to the generative AI model.

[1041] Specific behavior: Generates a prompt like the following:

[1042] Please use the following information to help you decide whether your subordinates should request time off.

[1043] Past vacation history: 10 days of vacation taken in the past year

[1044] Current workload: Project A is 70% complete

[1045] Team member workload: High

[1046] 5. Input and judgment for AI model

[1047] Server: The generated prompt sentence is input into the generative AI model to determine whether to approve or reject the vacation request.

[1048] Specific operation: Call the API of the generative AI model and send a prompt.

[1049] 6. Obtaining the judgment result

[1050] Server: Obtains the judgment results from the generative AI model.

[1051] Specific operation: Receives the API response and analyzes the judgment result.

[1052] Approve the vacation request, but be careful not to delay progress on Project A.

[1053] 7. Notification of Results

[1054] Server: Sends the decision result to the device.

[1055] Terminal: Notifies the user (subordinate) of the received judgment results. Specifically, the notification function of the work management software is used to display the results to the subordinate.

[1056] Specific behavior: Display "Vacation request approved" on the vacation request form along with detailed comments.

[1057] In this way, artificial intelligence can make on-site decisions on behalf of managers, reducing the workload on managers and allowing them to focus on other important tasks.

[1058] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1059] Step 1: Accepting the leave request

[1060] User: A subordinate uses a terminal to apply for leave. Specifically, the subordinate enters the necessary information into the leave application form in the business management software and clicks the submit button.

[1061] Input: Information entered into the leave request form (e.g. applicant name, leave period, reason, etc.).

[1062] Output: The vacation request data is sent to the server.

[1063] Specific operation: The terminal sends the information entered in the vacation application form to the server.

[1064] Step 2: Data collection

[1065] Server: After receiving the leave request, access the database management system to collect relevant information such as the subordinate's past leave history, current workload, and workload of team members.

[1066] Input: Leave application data, access information to database management system.

[1067] Output: Relevant information collected (e.g., past leave history, current workload, workload of team members).

[1068] What happens: The server executes an SQL query to retrieve the required data from the database.

[1069] Step 3: Data Preprocessing

[1070] Server: Preprocesses the collected data, specifically formatting the data and filling in missing values.

[1071] Input: Relevant information collected.

[1072] Output: Preprocessed data.

[1073] Specific operation: The server unifies the data format and imputes missing values ​​with the mean or median.

[1074] Step 4: Generate a prompt statement

[1075] Server: Based on the preprocessed data, it generates prompt sentences to be input to the generative AI model.

[1076] Input: Preprocessed data.

[1077] Output: The generated prompt statement.

[1078] Specific behavior: The server generates a prompt like this:

[1079] Please use the following information to help you decide whether your subordinates should request time off.

[1080] Past vacation history: 10 days of vacation taken in the past year

[1081] Current workload: Project A is 70% complete

[1082] Team member workload: High

[1083] Step 5: Input and judgment of the AI ​​model

[1084] Server: The generated prompt sentence is input into the generative AI model to determine whether to approve or reject the vacation request.

[1085] Input: The generated prompt statement.

[1086] Output: The decision result from the generative AI model.

[1087] Specific operation: The server calls the API of the generative AI model and sends a prompt.

[1088] Step 6: Obtaining the decision

[1089] Server: Obtains the judgment results from the generative AI model.

[1090] Input: The API response from the generative AI model.

[1091] Output: The analyzed decision result.

[1092] Specific operation: The server receives the API response and analyzes the decision result.

[1093] Approve the vacation request, but be careful not to delay progress on Project A.

[1094] Step 7: Notification of results

[1095] Server: Sends the decision result to the device.

[1096] Terminal: Notifies the user (subordinate) of the received judgment results. Specifically, the notification function of the work management software is used to display the results to the subordinate.

[1097] Input: The parsed decision result.

[1098] Output: The decision result communicated to the user.

[1099] Specific behavior: The device will display "Vacation request approved" on the vacation request form, along with any detailed comments.

[1100] (Application example 2)

[1101] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1102] On-site managers have to handle a wide range of tasks, including reviewing reports from subordinates and customers, reviewing meeting minutes, and managing worker schedules, placing a heavy workload on them. Furthermore, approving or rejecting workers' vacation requests requires consideration of their past vacation history and current workload, making the decision difficult. This creates a problem where managers are unable to focus on other important tasks.

[1103] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1104] In this invention, the server includes a means for using AI to review reports submitted by subordinates and customers, extracting problems and identifying their causes and solutions, a means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, and a means for managing worker schedules and automatically approving or rejecting vacation requests based on past vacation history and current workload. This reduces the workload of managers, allowing them to focus on other important tasks.

[1105] A "report" is a document that describes the work content, progress, problems, etc. submitted by subordinates or customers.

[1106] "Review" is the process of carefully examining documents such as reports and minutes, confirming their contents, and extracting problems.

[1107] A "problem" is an obstacle or challenge that arises in the progress of a task or project.

[1108] A "cause" is the underlying reason or factor that causes a problem.

[1109] A "solution" is a specific method or means for solving a problem.

[1110] "Managerial authority" refers to the authority that managers have to give instructions and make decisions regarding work.

[1111] "On-site level judgment" refers to judgments made in response to specific problems or situations that arise on-site.

[1112] "Labor burden" refers to the effort and time required to carry out work.

[1113] "Schedule management" refers to managing schedules such as workers' working hours and holidays.

[1114] "Leave Request" means a request submitted by a worker to take leave.

[1115] "Past vacation history" refers to a record of vacations taken by a worker in the past.

[1116] The "current workload" refers to the workload or workload that a worker is currently responsible for.

[1117] "Approval" means to acknowledge an application or proposal.

[1118] "Rejection" means not approving an application or proposal.

[1119] A system for carrying out this invention has the following configuration: The server includes means for using AI to review reports submitted by subordinates and customers, extracting problems and deriving their causes and solutions, means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, and means for managing worker schedules and automatically approving or rejecting vacation requests based on past vacation history and current workload.

[1120] Program processing explanation

[1121] The server uses an AI system implemented in Python to process and calculate data as follows: The hardware used is a robot operating in the factory, and the software is an AI system implemented in Python.

[1122] 1. Report Review:

[1123] The server receives reports sent by subordinates and customers and analyzes the content using an AI model. The AI ​​model uses natural language processing technology to understand the content of the report and extract problems. After the problems are extracted, the model identifies their causes and proposes appropriate solutions.

[1124] 2. Delegation of Management Authority:

[1125] The server delegates on-site decision-making that would normally be performed by managers to AI. This allows the AI ​​to monitor the situation on-site in real time and make necessary decisions. For example, by making decisions regarding work progress and quality control, it reduces the workload of managers.

[1126] 3. Schedule management and leave request processing:

[1127] The server manages the schedules of workers. When a vacation request is received from a worker, the AI ​​automatically approves or rejects the request, taking into account past vacation history and current workload. This saves managers the trouble of processing vacation requests individually.

[1128] Specific examples

[1129] For example, consider a scenario in which a robot operating in a factory accepts vacation requests from workers. When a worker requests vacation, the server receives the request, and the AI ​​makes a decision based on past vacation history and current workload. If the worker has taken too many vacations in the past or the current workload is high, the AI ​​will reject the vacation request, but if not, it will approve it.

[1130] Example prompt sentence:

[1131] "Evaluate workers' time off requests and approve or deny them based on their past leave history and current workload."

[1132] In this way, the server can use AI to streamline managerial tasks and reduce their workload.

[1133] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1134] Step 1:

[1135] The server receives reports from subordinates and customers. The input is the report data sent by the subordinates and customers, and the output is the received report data. The server prepares this data to be passed to the AI ​​model.

[1136] Step 2:

[1137] The server inputs the received report data into the AI ​​model and analyzes the content. The input is the received report data, and the output is the analysis results. The AI ​​model uses natural language processing technology to understand the report content and extract problems.

[1138] Step 3:

[1139] The server identifies the cause of the problem and proposes a solution based on the analysis results obtained from the AI ​​model. The input is the analysis results from the AI ​​model, and the output is the cause of the problem and a solution. The server provides this information to management.

[1140] Step 4:

[1141] The server delegates managerial authority to the AI, allowing it to make on-site decisions. The input is on-site situation data, and the output is the AI's judgment results. The server monitors the on-site situation in real time and makes the necessary decisions.

[1142] Step 5:

[1143] The server manages the schedules of workers. The input is the schedule data of the workers, and the output is the updated schedule data. The server manages the schedules of workers, such as their working hours and holidays.

[1144] Step 6:

[1145] The server receives vacation requests from workers. The input is the vacation request data from the workers, and the output is the received vacation request data. The server prepares this data to be passed to the AI ​​model.

[1146] Step 7:

[1147] The server inputs vacation request data into the AI ​​model, which then makes a decision based on past vacation history and current workload. The inputs are vacation request data, past vacation history data, and current workload data, and the output is a decision on whether to approve or reject the vacation request. The AI ​​model analyzes this data and makes the appropriate decision.

[1148] Step 8:

[1149] The server notifies the worker of the decision result obtained from the AI ​​model. The input is the decision result from the AI ​​model, and the output is a notification message to the worker. The server informs the worker whether the vacation request has been approved or rejected.

[1150] In this way, the server can use AI to streamline managerial tasks and reduce their workload.

[1151] Example 3

[1152] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1153] On-site managers are required to review reports and meeting minutes submitted by their subordinates and customers, identify problems, analyze their causes, and derive solutions, which increases their workload and makes it difficult for them to concentrate on other tasks. They are also required to make quick and accurate decisions at the on-site level, which also places a burden on managers. Furthermore, it is important to use past experience and data to predict the future and make appropriate decisions, which requires advanced analytical skills. To solve these issues, an efficient system utilizing artificial intelligence is needed.

[1154] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1155] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using artificial intelligence to identify problems and derive their causes and solutions; a means for delegating managerial authority to the artificial intelligence and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing meeting minutes using natural language processing technology and extracting problems from the content of discussions; a means for learning from past minutes and their results using a machine learning algorithm to derive causes and solutions; and a means for inputting prompts into a generative AI model to obtain analysis results. This reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, the use of artificial intelligence enables rapid and accurate problem identification, cause analysis, and solution deriving, enabling future predictions and appropriate decision-making using past data.

[1156] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and automatically perform tasks such as learning, reasoning, and problem-solving.

[1157] A "report" is a document sent by subordinates or customers that describes the progress of work, problems, proposals, etc.

[1158] "Review" is the process of checking the contents of submitted reports and minutes, extracting problems, analyzing their causes, and deriving solutions.

[1159] "Problems" refer to issues or obstacles that need to be resolved in the content of reports or minutes.

[1160] "Cause" refers to the reason or background behind the occurrence of a problem.

[1161] A "solution" refers to a specific method or means for solving a problem.

[1162] "Managerial authority" refers to the authority that on-site managers have to give instructions and make decisions regarding work.

[1163] "On-site level judgment" refers to the process of making quick and appropriate decisions based on the situation on-site.

[1164] "Labor burden" refers to the time and effort required to carry out work.

[1165] "Natural language processing technology" is a technology that allows computers to understand, analyze, and generate human language.

[1166] A "machine learning algorithm" is a mathematical technique for learning patterns from data and making predictions or classifications.

[1167] A "generative AI model" is an artificial intelligence model that is trained to produce optimal results for a specific task.

[1168] A "prompt" is an instruction given to a generative AI model, containing the text to be analyzed and generated.

[1169] This invention is a system that uses artificial intelligence to review meeting minutes and reports, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system is described below.

[1170] System configuration

[1171] Hardware:

[1172] Server: Use a server equipped with a high-performance processor and large memory capacity, which allows for rapid processing of large amounts of data.

[1173] Device: A device such as a computer or tablet used by a user to upload minutes and reports and check results.

[1174] software:

[1175] Natural language processing libraries: Use Python's spaCy and NLTK for text tokenization, part-of-speech tagging, and dependency analysis.

[1176] Machine learning algorithms, such as random forests and support vector machines (SVMs), learn from past data to identify causes and solutions.

[1177] Generative AI models: Use artificial intelligence models that are trained to generate optimal results for specific tasks.

[1178] Program processing

[1179] Upload the minutes:

[1180] The user uploads the minutes of the meeting from the terminal to the server. The minutes are in text file format. The user selects the minutes file using the file selection dialog on the browser and clicks the "Upload" button.

[1181] Natural Language Processing (NLP) text analysis:

[1182] The server receives the uploaded minutes file and analyzes the text using NLP techniques, specifically using the Python libraries spaCy and NLTK. The server tokenizes the minutes text, tags it with parts of speech, and performs dependency analysis.

[1183] Identifying the problem:

[1184] The server uses the analysis results to understand the content of the discussion and extract issues. Issues are identified by detecting specific keywords and phrases (such as "issue," "problem," or "concern"). The server searches for keywords such as "issue" or "problem" in the text and extracts sentences related to them.

[1185] Analysis of the cause:

[1186] The server uses machine learning algorithms to analyze the causes of the extracted problems. Using past minutes and their results as learning data, it applies algorithms such as random forests and support vector machines (SVMs) to identify the causes. The server references a database of past minutes and applies a model to identify the causes of similar problems.

[1187] Solution derivation:

[1188] The server then derives a solution based on the cause, again using a machine learning model based on past data to suggest the optimal solution. The server references past success stories and applies the model to suggest solutions to similar problems.

[1189] The resulting output:

[1190] The server outputs the extracted problems, causes, and solutions to the terminal. The user can check this information through the terminal. The server generates the analysis results in HTML format and displays them in the user's browser.

[1191] Specific examples

[1192] Let's say a user uploads "Minutes of the project meeting held on October 1, 2023." The minutes contain the following content:

[1193] Agenda Item 1: Project Status

[1194] Challenge: Schedule Delays

[1195] Cause: Insufficient resources

[1196] Solution: Add a new member

[1197] The server parses the minutes and outputs the following results:

[1198] Problem: Schedule Delays

[1199] Cause: Insufficient resources

[1200] Solution: Add a new member

[1201] Prompt Sentence Examples

[1202] An example of a prompt sentence to input to the generative AI model is as follows:

[1203] Analyze the minutes of the following meeting and extract the problems, causes, and solutions.

[1204] minutes:

[1205] Agenda Item 1: Project Status

[1206] Challenge: Schedule Delays

[1207] Cause: Insufficient resources

[1208] Solution: Add new members

[1209] By inputting this prompt sentence into the generative AI model, the model analyzes the contents of the minutes and extracts the problem, its cause, and a solution. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1210] Step 1:

[1211] Uploading minutes

[1212] The user uploads the minutes of the meeting from the terminal to the server. The minutes are in text file format. The user selects the minutes file using the file selection dialog on the browser and clicks the "Upload" button.

[1213] Input: A minutes file selected by the user

[1214] Output: Meeting minutes file uploaded to the server

[1215] Step 2:

[1216] Text analysis using natural language processing (NLP)

[1217] The server receives the uploaded minutes file and analyzes the text using NLP techniques, specifically using the Python libraries spaCy and NLTK. The server tokenizes the minutes text, tags it with parts of speech, and performs dependency analysis.

[1218] Input: Uploaded minutes file

[1219] Output: Tokenized, part-of-speech tagged, dependency parsed text data

[1220] Step 3:

[1221] Identifying the problem

[1222] The server uses the analysis results to understand the content of the discussion and extract issues. Issues are identified by detecting specific keywords and phrases (such as "issue," "problem," or "concern"). The server searches for keywords such as "issue" or "problem" in the text and extracts sentences related to them.

[1223] Input: Parsed text data

[1224] Output: Extracted problem statement

[1225] Step 4:

[1226] Analysis of the cause

[1227] The server uses machine learning algorithms to analyze the causes of the extracted problems. Using past minutes and their results as learning data, it applies algorithms such as random forests and support vector machines (SVMs) to identify the causes. The server references a database of past minutes and applies a model to identify the causes of similar problems.

[1228] Input: Extracted problem statements, database of past meeting minutes

[1229] Output: Identified cause

[1230] Step 5:

[1231] Derivation of a solution

[1232] The server then derives a solution based on the cause, again using a machine learning model based on past data to suggest the optimal solution. The server references past success stories and applies the model to suggest solutions to similar problems.

[1233] Input: Identified causes, database of past success stories

[1234] Output: Proposed solution

[1235] Step 6:

[1236] Output of results

[1237] The server outputs the extracted problems, causes, and solutions to the terminal. The user can check this information through the terminal. The server generates the analysis results in HTML format and displays them in the user's browser.

[1238] Input: Extracted problem, identified cause, proposed solution

[1239] Output: Analysis results generated in HTML format

[1240] (Application example 3)

[1241] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1242] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. There is also the issue of it being difficult to efficiently analyze the contents of minutes from production meetings and quality control meetings, identify problems, analyze their causes, and propose solutions. Furthermore, there is a lack of systems that can learn from past minutes data and support efficient problem-solving.

[1243] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1244] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers through AI, identifying problems and proposing causes and solutions; a means for delegating managerial authority to AI and allowing it to make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing minutes of production meetings and quality control meetings within the factory, identifying problems, analyzing causes, and proposing solutions; and a means for learning from past minutes data to support efficient problem solving. This reduces the workload of on-site managers, allowing them to focus on other tasks. It also makes it possible to efficiently analyze the contents of minutes from production meetings and quality control meetings, identifying problems, analyzing causes, and proposing solutions. Furthermore, it is possible to learn from past minutes data to support efficient problem solving.

[1245] A "report" is a report or record document submitted by a subordinate or customer.

[1246] "AI" is an abbreviation for artificial intelligence, a system that mimics human intelligence using technologies such as machine learning and natural language processing.

[1247] "Review" is the process of analyzing the contents of reports and minutes and evaluating and examining them.

[1248] "Problems" are issues or obstacles that need to be resolved in the content of a report or minutes.

[1249] A "cause" is the fundamental reason or factor that causes a problem.

[1250] A "solution" is a specific method or means for solving a problem.

[1251] "Managerial authority" refers to the authority that managers have to make decisions and give instructions.

[1252] "On-site level judgment" refers to specific judgments and decisions made based on the situation on-site.

[1253] "Labor burden" refers to the effort and time required to carry out work.

[1254] A "production meeting" is a meeting held within a factory to discuss production-related topics.

[1255] A "quality control meeting" is a meeting within the factory to discuss quality control topics.

[1256] "Minutes" are documents that record the contents of a meeting and the decisions made.

[1257] "Analysis" is the process of analyzing and understanding data and information in detail.

[1258] "Learning" is the process of acquiring knowledge from past data and experiences and applying it to future decisions and actions.

[1259] "Efficiency" means achieving a goal effectively and without waste.

[1260] "Support" refers to helping with the activities or work of others.

[1261] As an embodiment of the present invention, the following system is constructed.

[1262] System Configuration

[1263] The server includes the following means:

[1264] 1. A method of using AI to review reports submitted by subordinates and customers, extracting problems and identifying their causes and solutions.

[1265] 2. Delegating managerial authority to AI and having it make on-site decisions reduces the workload of on-site managers, allowing them to focus on other tasks.

[1266] 3. A means of analyzing the minutes of production meetings and quality control meetings within the factory to identify problems, analyze their causes, and propose solutions.

[1267] 4. A means of learning from past meeting minutes data to support efficient problem solving.

[1268] Hardware and software used

[1269] Hardware: Servers, terminals (PCs, tablets, smartphones), robots in factories

[1270] Software: Natural language processing libraries (spaCy, NLTK), machine learning libraries (scikit-learn), databases (MySQL, PostgreSQL)

[1271] Data processing and calculation

[1272] The server does the following:

[1273] 1. Data collection: Reports submitted by subordinates and customers, as well as minutes of production meetings and quality control meetings within the factory, are stored in a database.

[1274] 2. Natural language processing: Using a natural language processing library (e.g., spaCy), the contents of reports and minutes are analyzed and problems are extracted.

[1275] 3. Machine learning: Using a machine learning library (e.g., scikit-learn), learn from past minutes data, analyze the causes of problems, and propose solutions.

[1276] 4. Displaying the results: The analysis results and proposed solutions are displayed on the device for the user to review.

[1277] Specific examples

[1278] For example, if the minutes of a production meeting include issues such as "delays on the production line" or "a high number of defective products found during quality inspections," the server will extract these issues and, based on the results of learning from past data, propose solutions such as "automating the production line" or "strengthening quality inspections."

[1279] Prompt Sentence Examples

[1280] "Analyze the minutes of the production meeting, extract the problems, analyze the causes, and propose solutions. The minutes contain the following information: 'There are delays on the production line' and 'There are many defective products during quality inspections'."

[1281] In this way, the workload of on-site managers can be reduced, allowing them to concentrate on other tasks. It also makes it possible to efficiently analyze the contents of minutes from production meetings and quality control meetings, extract problems, analyze their causes, and propose solutions. Furthermore, it is possible to learn from past minutes data and support efficient problem solving.

[1282] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1283] Step 1:

[1284] Data collection

[1285] The server stores reports submitted by subordinates and customers, as well as minutes of production and quality control meetings held within the factory, in a database. The input includes the text data of the reports and minutes. The server receives this data and stores it in a database (for example, MySQL or PostgreSQL). The output is the text data stored in the database.

[1286] Step 2:

[1287] Natural Language Processing

[1288] The server uses a natural language processing library (e.g., spaCy) to analyze the contents of reports and minutes and extract problems. The input includes text data obtained from a database. The server analyzes the text data and extracts sentences containing problems. The output is a list of extracted problems.

[1289] Step 3:

[1290] Machine Learning

[1291] The server uses a machine learning library (e.g., scikit-learn) to learn from past meeting minutes data, analyze the causes of problems, and propose solutions. The input includes a list of extracted problems and past meeting minutes data. The server uses this data to train a machine learning model, analyze the causes of problems, and generate solutions. The output is a list of causes and solutions.

[1292] Step 4:

[1293] Displaying the results

[1294] The server displays the analysis results and proposed solutions on the terminal for the user to review. The input includes a list of causes and solutions. The server sends this data to the terminal and displays it for the user to review. The output is a screen display that the user can review.

[1295] In this way, the server performs a series of processes from data collection to analysis, learning, and display of results, reducing the workload of on-site managers and creating a system that supports efficient problem-solving.

[1296] Furthermore, an emotion engine that estimates the user's emotion may be combined.

[1297] The identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1298] "Example 1"

[1299] In one embodiment of the present invention, AI reviews reports from subordinates and customers and extracts problems. It then uses an emotion engine to recognize the user's emotions. For example, it can sense the user's anger or frustration from the report text. This emotional information is then used to analyze the causes of the problems and derive solutions.

[1300] "Example 2"

[1301] In another implementation, the AI ​​can act on behalf of managers, making decisions on the ground, such as when employees request time off. It uses an emotion engine to recognize the employee's emotions and adjust its decisions accordingly. For example, if it recognizes that an employee is stressed, it can prioritize approving the vacation request.

[1302] "Example 3"

[1303] In yet another embodiment, the AI ​​reviews meeting minutes and extracts issues. It uses an emotion engine to recognize participants' emotions from the text in the minutes. For example, if a discussion gets heated or participants seem confused, the AI ​​can understand the flow of the discussion from the emotional information and identify the cause and solution of the problem.

[1304] The processing flow of each embodiment will be described below.

[1305] "Example 1"

[1306] Step 1: The AI ​​receives reports from subordinates and customers.

[1307] Step 2: The AI ​​reviews the report and identifies problems.

[1308] Step 3: Using an emotion engine, we recognize the user's emotions from the report text. Step 4: Based on the emotion information, we analyze the cause of the problem and derive solutions.

[1309] "Example 2"

[1310] Step 1: The AI ​​receives information such as vacation requests from subordinates.

[1311] Step 2: Use the emotion engine to recognize your subordinates' emotions.

[1312] Step 3: Decide whether to approve or deny the vacation request based on the employee's feelings.

[1313] "Example 3"

[1314] Step 1: The AI ​​receives the meeting minutes.

[1315] Step 2: Use the emotion engine to recognize participants' emotions from the text of the minutes.

[1316] Step 3: Understand the flow of the discussion from emotional information and identify the cause of the problem and solutions.

[1317] Example 1

[1318] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1319] On-site managers spend a great deal of time and effort reviewing reports from their subordinates and customers, identifying problems, analyzing their causes, and deriving solutions. This makes it difficult for them to concentrate on other important tasks, increasing their workload. Accurately understanding the contents of reports and deriving appropriate solutions requires advanced expertise and experience, further increasing the burden on on-site managers. Furthermore, it is difficult to properly recognize user sentiment contained in reports and utilize it to solve problems. An effective system to resolve these issues is needed.

[1320] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1321] In this invention, the server includes: means for receiving a report on the user's work created by the user; a prompt statement instructing the system to derive problems in the work, causes of the problems, and solutions to the problems from the content of the received report; means for deriving the problems in the work, causes of the problems, and solutions to the problems using a generative AI model; a prompt statement instructing the system to generate a report indicating the derived problems in the work, causes of the problems, and solutions to the problems using the generative AI model; and means for presenting the generated report to the user. The server further includes means for recognizing the user's emotion from the report using an emotion engine, and the report generating means generates the report using the prompt statement instructing the system to generate a report indicating the derived problems in the work, causes of the problems, and solutions to the problems based on the recognized user's emotion, and the generative AI model. Furthermore, the report generating means in the server generates the report using the prompt sentence instructing the generation of the report with the priority of the solutions adjusted based on the recognized user's emotions and the generation AI model. This reduces the workload of on-site managers, allowing them to focus on other important tasks. It also enables accurate understanding of the report content and promptly deriving appropriate solutions. Furthermore, it enables the user's emotions to be properly recognized and utilized in problem solving.

[1322] A "report" is a document sent by subordinates or clients that describes the progress of a project, problems, proposals, etc.

[1323] "Artificial intelligence" refers to the technology that enables computer systems to learn, reason, and self-correct by imitating human intelligence, including, among other things, natural language processing and machine learning.

[1324] A "review" is a process in which the contents of a report are examined in detail, problems are identified, and suggestions for improvement are made.

[1325] "Problems" refer to elements or issues described in the report that hinder the progress of the project or the execution of work.

[1326] "Cause" refers to the fundamental reason or background behind the occurrence of a problem.

[1327] A "solution" refers to a specific method or means for solving the identified problem.

[1328] "Natural language processing technology" refers to technology that enables computers to understand, analyze, and generate human language, and includes text tokenization and grammatical analysis.

[1329] "Machine learning algorithms" refer to mathematical models and methods for learning patterns from data and making predictions or classifications.

[1330] An "emotion engine" is a technology for recognizing and analyzing user emotions from text data, and includes an emotion analysis library.

[1331] "Managerial authority" refers to the authority to make decisions regarding business operations and problem solving.

[1332] "On-site decision-making" refers to the process of making decisions quickly based on the specific situation at the site.

[1333] "Labor burden" refers to the amount of work and stress that managers bear in carrying out their duties.

[1334] This invention is a system that uses artificial intelligence to review reports submitted by subordinates and customers, extracting problems, analyzing their causes, and deriving solutions. This system works in cooperation with a server, a terminal, and a user.

[1335] Hardware and software used

[1336] The server uses the following hardware and software:

[1337] Hardware: A server machine with a powerful processor, sufficient memory, and ample storage.

[1338] software:

[1339] NLP libraries such as "spaCy" and "NLTK" for implementing natural language processing techniques.

[1340] Machine learning libraries such as "Scikit-learn" and "TensorFlow" for implementing machine learning algorithms.

[1341] Sentiment analysis libraries such as "TextBlob" and "VADER" for implementing sentiment engines.

[1342] Program processing

[1343] The server receives reports sent from users who are subordinates or customers, and performs the following processes.

[1344] 1. Receiving the report

[1345] The server receives the text data of the report via an HTTP request.

[1346] 2. Analyzing the report

[1347] The server uses "spaCy" or "NLTK" to tokenize the text data of the received report and analyze its grammatical structure.

[1348] 3. Identifying the problem

[1349] The server extracts issues from the report based on the analysis results, using rule-based filtering to match the issues against a predefined list of issues.

[1350] 4. Analysis of the cause of the problem

[1351] The server then uses machine learning algorithms to analyze the causes of the extracted problems. Specifically, it uses Scikit-learn's random forest and TensorFlow's neural network to identify the causes using past report data.

[1352] 5. Deriving a solution

[1353] The server then generates solutions based on the results of the root cause analysis, generating specific proposals such as "increasing the resources of the design department" or "introducing a project management tool." In this embodiment, "2. Report Analysis" to "5. Deriving a Solution" are performed using a generative AI model (GPT-4). Specifically, the server derives the business problem, the cause of the problem, and the solution to the problem from the content of the received report using a prompt statement instructing the server to derive the business problem, the cause of the problem, and the solution to the problem, and using the generative AI model. At this time, by providing past report data to the generative AI model, the past report data can be taken into consideration when deriving the solution.

[1354] 6. Emotion recognition

[1355] The server uses an emotion engine to recognize user emotions from the report text. Specifically, it uses "TextBlob" and "VADER" to detect user emotions such as "anger" and "frustration." This information is used to adjust the priority of solutions.

[1356] 7. Outputting the results

[1357] The server compiles the analysis results, cause analysis, solutions, and emotion recognition results and provides feedback to the user. Specifically, the server generates the report using a generative AI model and a prompt statement instructing the server to generate a report that shows the derived business problems, causes of the problems, and solutions to the problems based on the recognized user emotions. At this time, the prompt statement instructs the server to generate a report in which the priority of the solutions is adjusted based on the recognized user emotions. For example, the server generates a report in HTML format and sends it to the user's device.

[1358] Specific examples

[1359] Suppose a user submits the following report:

[1360] "Recently, projects have been delayed, and we've been receiving more complaints from clients. This seems to be due in particular to the slow response of the design department."

[1361] Process flow:

[1362] 1. Receiving the report

[1363] The server receives the reports sent by the user via HTTP requests.

[1364] 2. Analyzing the report

[1365] The server uses "spaCy" to tokenize the report text and parse the grammatical structure.

[1366] 3. Identifying the problem

[1367] The server extracts problems such as "delays in project progress," "client complaints," and "delays in response from the design department" from the analysis results.

[1368] 4. Analysis of the cause of the problem

[1369] The server uses Scikit-learn's random forest to analyze past data and determine that the cause is a "lack of resources in the design department."

[1370] 5. Deriving a solution

[1371] Server proposes solutions such as "increasing the resources of the design department" and "introducing project management tools."

[1372] 6. Emotion recognition

[1373] The server uses "TextBlob" to sense the user's "dissatisfaction" from the report and adjust the priority of the solution.

[1374] 7. Outputting the results

[1375] The server compiles the analysis results, cause analysis, solutions, and emotion recognition results into an HTML report and sends it to the user's device.

[1376] Example prompt sentence:

[1377] Analyze the following report and identify the problem, cause, and solution. Also, recognize the user's feelings.

[1378] Report: Project progress has been slow recently, and we have received an increasing number of complaints from clients. This is likely due in particular to the slow response of the design department.

[1379] In this way, the server provides a system that efficiently reviews reports from subordinates and customers, extracts problems, analyzes their causes, and proposes solutions.

[1380] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1381] Step 1: Receive the report

[1382] The server receives the report sent by the user via an HTTP request. As input, it receives the text data sent by the user and passes the text data to the next processing step as output. Specifically, the server listens for HTTP requests and temporarily stores the received data.

[1383] Step 2: Analyze the report

[1384] The server uses "spaCy" or "NLTK" to tokenize the received report text data and analyze its grammatical structure. It uses the text data received in step 1 as input and generates tokenized text data and grammatical analysis results as output. Specifically, the server passes the text data to an NLP library to perform tokenization and grammatical analysis.

[1385] Step 3: Identify the problem

[1386] The server extracts issues from the report based on the analysis results. It uses the tokenized text data generated in step 2 and the grammar analysis results as input and generates a list of extracted issues as output. Specifically, the server performs rule-based filtering to match the list of extracted issues against a predefined list of issues.

[1387] Step 4: Analyze the cause of the problem

[1388] The server uses a machine learning algorithm to analyze the causes of the extracted issues. It uses the list of issues generated in step 3 as input and generates a list of causes as output. Specifically, the server uses past report data to identify causes using Scikit-learn's random forest or TensorFlow's neural network.

[1389] Step 5: Derive a solution

[1390] The server derives solutions based on the results of the cause analysis. It uses the list of causes generated in step 4 as input and generates a list of solutions as output. Specifically, the server uses a model trained on past successful cases to propose optimal solutions.

[1391] Step 6: Emotion Recognition

[1392] The server uses an emotion engine to recognize the user's emotions from the report text. It uses the tokenized text data generated in step 2 as input and generates emotion recognition results as output. Specifically, the server uses "TextBlob" and "VADER" to detect the user's emotions, such as "anger" and "dissatisfaction."

[1393] Step 7: Output the results

[1394] The server compiles the analysis results, cause analysis, solutions, and emotion recognition results and provides feedback to the user. It uses the various data generated in steps 3 to 6 as input and generates an HTML report as output. Specifically, the server integrates this data and generates an HTML report to send to the user's device.

[1395] (Application example 1)

[1396] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1397] With conventional report review systems, it was difficult to accurately analyze the content of reports submitted by subordinates and customers, extract problems, analyze the causes, and derive appropriate solutions. This also created a heavy workload for on-site managers, making it difficult for them to concentrate on other tasks. Furthermore, because the system was unable to recognize the user's emotions from the report or assess the seriousness of the problem, appropriate responses were delayed.

[1398] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means. In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using AI to identify problems and derive their causes and solutions; a means for delegating managerial authority to AI and having AI make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for recognizing user emotions from reports using an emotion engine and assessing the severity of problems; a means for analyzing report content and extracting problems using natural language processing technology; and a means for learning from past reports and their results using a machine learning algorithm to derive causes and solutions. This enables accurate analysis of report content, extraction of problems, analysis of causes, and deriving appropriate solutions. It also reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, recognizing user emotions from reports and assessing the severity of problems enables prompt and appropriate responses.

[1399] "Subordinates and customers" refers to employees who carry out work in a company or organization, and customers who use the services or products of a company or organization.

[1400] A "report" refers to a document that describes the progress of work, problems, proposals, etc.

[1401] "AI" is an abbreviation for artificial intelligence, and refers to the technology that allows computers to learn and reason by imitating human intelligence.

[1402] "Review" refers to the process of checking the contents of a submitted report and evaluating and analyzing it.

[1403] "Identifying problems" refers to finding business problems and issues from the contents of the report.

[1404] "Identifying causes and solutions" refers to analyzing the causes of the problems that have been identified and proposing appropriate solutions to them.

[1405] "Delegating managerial authority to AI" refers to transferring the decision-making authority held by on-site managers to artificial intelligence.

[1406] "On-site level judgment" refers to judgment on specific problems and situations that arise on-site.

[1407] "Reduction of labor burden" refers to reducing the labor burden associated with performing work.

[1408] "Emotion engine" refers to technology for analyzing and recognizing user emotions from text data.

[1409] "Natural language processing technology" refers to technology that enables computers to understand and analyze human language.

[1410] A "machine learning algorithm" refers to a computational method for learning patterns from data and making predictions or classifications.

[1411] "Past reports and their results" refers to previously submitted reports and the actions and results taken based on those reports.

[1412] As an embodiment of this invention, we will build a system in which AI automatically reviews reports submitted by factory workers and managers, extracts problems, analyzes their causes, and proposes solutions. Furthermore, we will use an emotion engine to recognize the emotions of workers and evaluate the severity of the problems.

[1413] The server implements the system using the following hardware and software.

[1414] Hardware and software used:

[1415] Hardware: Robots and servers in factories

[1416] Software: Python, spaCy (NLP library), TextBlob (sentiment analysis library), scikit-learn (machine learning library)

[1417] Data processing and calculation:

[1418] 1. Report analysis: The server uses spaCy to analyze the text data of reports submitted by workers and managers and extracts problems.

[1419] 2. Emotion Recognition: The server analyzes the emotion of the report using TextBlob and calculates a positive / negative score.

[1420] 3. Cause analysis: The server uses scikit-learn's RandomForestClassifier to predict the cause of the problem using a model learned from past data.

[1421] 4. Solution Proposal: The server proposes an appropriate solution based on the cause.

[1422] Examples:

[1423] The report read: "The machine was down due to a maintenance issue. The operator was very frustrated."

[1424] Results: Sentiment score: -0.5, Problem: "Maintenance issues", Cause: "Lack of maintenance", Solution: "Schedule maintenance"

[1425] Example prompt for a generative AI model:

[1426] The report read: "The machine was down due to a maintenance issue. The operator was very frustrated."

[1427] Analyze this report and suggest the problem, cause, and solution.

[1428] This system makes it possible to accurately analyze the contents of reports, extract problems, analyze the causes, and derive appropriate solutions. It also reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, by recognizing the user's emotions from the reports and assessing the seriousness of the problem, it becomes possible to take appropriate action quickly.

[1429] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1430] Step 1:

[1431] The server receives the text data of reports submitted by workers and managers. The input is the text data of the report, and the output is the text data to be analyzed. This text data is used in subsequent processing steps.

[1432] Step 2:

[1433] The server analyzes the report content using natural language processing technology. Specifically, it uses the spaCy library to analyze the text data and extract issues. The input is the text data to be analyzed, and the output is a list of extracted issues.

[1434] Step 3:

[1435] The server uses an emotion engine to recognize user emotions from reports. Specifically, it uses the TextBlob library to calculate emotion scores for text data. The input is the text data to be analyzed, and the output is the emotion score.

[1436] Step 4:

[1437] The server uses a machine learning algorithm to learn from past reports and their results and predict the cause of the problem. Specifically, it uses scikit-learn's RandomForestClassifier to predict the cause of the problem. The input is a list of extracted problems, and the output is a list of predicted causes.

[1438] Step 5:

[1439] The server proposes appropriate solutions based on the predicted causes. Specifically, it retrieves solutions corresponding to the causes from a database and proposes them. The input is a list of predicted causes, and the output is a list of proposed solutions.

[1440] Step 6:

[1441] The server compiles the analysis results, sentiment scores, predicted causes, and proposed solutions and outputs them as a report. The inputs are the analysis results, sentiment scores, predicted causes, and proposed solutions, and the output is the final report.

[1442] Through these processing steps, the server can accurately analyze the report content, extract the problem, analyze the cause, and derive an appropriate solution. It also recognizes the user's emotions and evaluates the seriousness of the problem, allowing for a prompt and appropriate response.

[1443] Example 2

[1444] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1445] On-site managers are required to perform a wide range of tasks, such as reviewing reports from subordinates and customers and approving vacation requests, which increases their workload and prevents them from concentrating on other important tasks. Furthermore, it is sometimes difficult for them to properly understand and judge the emotions and work situation of their subordinates, which can lead to poor decisions. This leads to a decline in work efficiency and delays in improving the work environment.

[1446] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1447] In this invention, the server includes a means for retrieving subordinates' past vacation history, current workload, and other related information from a database; a means for inputting the retrieved data into a generative AI model to determine whether to approve or reject the vacation request; and a means for recognizing subordinates' emotions using an emotion engine and adjusting decisions based on those emotions. This reduces the workload of on-site managers, allowing them to focus on other important tasks. Furthermore, by properly understanding subordinates' emotions and work situations and making more appropriate decisions, it is expected to improve work efficiency and the work environment.

[1448] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and automatically perform tasks such as learning, reasoning, and problem-solving.

[1449] A "generative AI model" is a type of artificial intelligence that generates new data and decisions based on input data, and is a model trained to provide optimal results for a specific task.

[1450] An "emotion engine" is a technology that analyzes human emotions from data such as text, audio, and images, and recognizes their emotional state.

[1451] A "database" is a system for efficiently storing, managing, and searching data, and is capable of handling large amounts of data while maintaining the consistency and integrity of the information.

[1452] A "terminal" is a device that a user directly operates, and includes a computer, a smartphone, a tablet, and the like.

[1453] "Vacation history" is a record of vacations taken by subordinates in the past, and includes information such as the type of vacation, duration, and date taken.

[1454] "Workload" is an indicator that shows the amount and content of work that a subordinate is responsible for within a specific period of time, and includes the number of tasks, difficulty, and time required.

[1455] A "field manager" is a position that is responsible for managing operations at the field level, instructing and evaluating subordinates, etc., and typically corresponds to a section chief or chief.

[1456] "Labor burden" refers to the effort and time required to carry out work, and excessive labor burden can lead to reduced work efficiency and stress.

[1457] A "judgment result" is a final conclusion or decision reached based on data analyzed by a generative AI model or emotion engine.

[1458] MODE FOR CARRYING OUT THE INVENTION

[1459] This invention is a system that delegates managerial authority to artificial intelligence and automates on-site decision-making. Specifically, the server retrieves subordinates' past vacation history, current workload, and other relevant information from a database, inputs it into a generative AI model, and determines whether to approve or reject vacation requests. It also uses an emotion engine to recognize the subordinates' emotions and adjusts the decision based on those emotions. The final decision is notified to the terminal.

[1460] Hardware and software used

[1461] Hardware: Servers, terminals

[1462] Software: Generative AI model, emotion engine, database management system (DBMS)

[1463] Data collection

[1464] The server retrieves the subordinate's past vacation history, current workload, and other relevant information from the database. For example, the server retrieves data from the DBMS using an SQL query. Specifically, it executes the query "SELECT FROM Vacation History WHERE Subordinate ID = 'A'".

[1465] Data analysis

[1466] The server inputs the acquired data into the generative AI model and determines whether to approve or reject the vacation request. Specifically, it inputs the prompt statement, "Considering subordinate A's past vacation history and current workload, should you approve or reject the vacation request?" into the generative AI model.

[1467] emotion recognition

[1468] The server recognizes the emotions of subordinates using the emotion engine. Specifically, it inputs a prompt sentence to the emotion engine: "Analyze the current emotional state of subordinate A and determine whether he is feeling stressed."

[1469] Adjusting judgment

[1470] The server adjusts the generative AI model's judgment based on the results of the emotion engine. For example, if subordinate A is recognized as feeling stressed, the server prioritizes approving his vacation request.

[1471] Notification of results

[1472] The server notifies the terminal of the final decision result. Specifically, the server generates a message saying "Subordinate A's vacation request has been approved" and sends it to the terminal of subordinate A. The terminal displays the received message.

[1473] Specific examples

[1474] Example prompt sentence:

[1475] "Should I approve or reject subordinate A's vacation request, taking into account his or her past vacation history and current workload?"

[1476] "Analyze subordinate A's current emotional state and determine whether he is feeling stressed."

[1477] This system allows managers to delegate on-site decision-making to AI, allowing them to focus on other tasks. For example, when a subordinate requests time off, the AI ​​automatically approves or rejects the request, reducing the burden on managers.

[1478] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1479] Step 1:

[1480] The server retrieves the subordinate's past vacation history, current workload, and other relevant information from a database.

[1481] Input: Subordinate ID

[1482] Data processing: The server uses SQL queries to retrieve subordinates' past vacation history and current workload from the DBMS.

[1483] Output: Acquired data (leave history, workload)

[1484] Specific operation: The server establishes a database connection and retrieves data by executing the query "SELECT FROM Vacation History WHERE Subordinate ID = 'A'". The retrieved data is temporarily stored in memory.

[1485] Step 2:

[1486] The server inputs the acquired data into a generative AI model to determine whether to approve or reject the vacation request.

[1487] Input: Acquired data (leave history, workload)

[1488] Data calculation: The server inputs the prompt statement to the generative AI model: "Considering subordinate A's past vacation history and current workload, should you approve or reject his vacation request?"

[1489] Output: The generative AI model's decision (accept or reject)

[1490] Specific operation: The server calls the API of the generative AI model and sends the prompt and data. The generative AI model analyzes the input data and returns a decision of approval or rejection.

[1491] Step 3:

[1492] The server recognizes the emotions of the subordinates using an emotion engine.

[1493] Input: Subordinates' emotional data

[1494] Data calculation: The server inputs a prompt statement to the emotion engine: "Analyze the current emotional state of subordinate A and determine whether he is feeling stressed."

[1495] Output: Emotion engine result (stress level)

[1496] Specific operation: The server calls the emotion engine's API and sends the emotion data of subordinate A. The emotion engine analyzes the emotion data and returns the stress level.

[1497] Step 4:

[1498] The server adjusts the generative AI model's decisions based on the results of the emotion engine.

[1499] Input: Generative AI model's judgment results, emotion engine results

[1500] Data calculation: The server receives the results of the emotion engine and reevaluates the judgment results of the generative AI model.

[1501] Output: Adjusted decision result (approve or reject)

[1502] Specific operation: The server obtains the results of the emotion engine and compares them with the judgment results of the generative AI model. If necessary, it corrects the judgment results.

[1503] Step 5:

[1504] The server notifies the terminal of the final decision result.

[1505] Input: Adjusted decision result

[1506] Data processing: The server generates a notification message.

[1507] Output: Information message

[1508] Specific operation: The server generates a message saying "Subordinate A's vacation request has been approved" and sends it to the terminal of subordinate A. The terminal displays the received message.

[1509] (Application example 2)

[1510] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1511] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. It also makes it difficult to monitor the operating status of machinery in factories in real time and perform appropriate maintenance and troubleshoot any abnormalities. This can lead to machinery breakdowns and reduced production efficiency.

[1512] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1513] In this invention, the server includes means for using AI to review reports submitted by subordinates and customers, extracting problems and identifying their causes and solutions, means for delegating managerial authority to AI and having it make on-site-level decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, means for monitoring the operating status of machinery within the factory, automatically detecting machinery requiring maintenance, and adjusting the maintenance schedule, and means for detecting machinery in an abnormal state and handling the abnormality. This reduces the workload of on-site managers, allowing them to focus on other tasks, and makes it possible to monitor the operating status of machinery within the factory in real time and perform appropriate maintenance and handle abnormalities.

[1514] "Reports from subordinates and customers" refers to reports and feedback submitted by employees within the organization or external customers.

[1515] "AI" is an abbreviation for artificial intelligence, a technology that enables computer systems to mimic human intelligence by learning, reasoning, and self-correcting.

[1516] A "review" is a process in which submitted information and data are examined in detail and evaluated and analyzed.

[1517] "Problem extraction" means finding issues or obstacles from data or information.

[1518] "Identifying causes and solutions" means identifying the root cause of the problem that has occurred and proposing appropriate measures to address it.

[1519] "Managerial authority" refers to the rights and responsibilities that managers have within an organization to make decisions and give instructions.

[1520] "On-site level decisions" refer to decisions made based on the specific circumstances and conditions at the site.

[1521] "Reducing labor burden" means reducing the workload and stress of employees.

[1522] "Focusing on other tasks" means redirecting the time and resources freed up from a specific task to other important tasks.

[1523] "Machinery and equipment in a factory" refers to production equipment and machinery used in a factory.

[1524] "Monitoring operational status" means checking the operating status and performance of machinery and equipment in real time.

[1525] "Machinery requiring maintenance" refers to machinery that is in a state that requires regular maintenance or repair.

[1526] "Adjusting the maintenance schedule" means planning and determining the timing of maintenance and repairs of machinery and equipment.

[1527] "Machinery in an abnormal state" refers to machinery that is deviating from normal operation and experiencing a breakdown or malfunction.

[1528] "Abnormality processing" refers to detecting abnormalities in mechanical equipment and taking appropriate measures to return it to a normal state.

[1529] The system for implementing this invention consists of three main elements: a server, a terminal, and a user. The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. It also delegates managerial authority to AI, allowing it to make on-site decisions, reducing the workload of on-site managers and allowing them to focus on other tasks. It also monitors the operating status of machinery within a factory, automatically detecting machinery requiring maintenance and adjusting maintenance schedules. It is also possible to detect machinery in an abnormal state and handle the abnormality.

[1530] The server runs a program implemented using Python. Specifically, it periodically checks reports stored in a database (e.g., MySQL) and machine operation data, and uses AI models to identify problems and analyze their causes. It uses a real-time monitoring system (e.g., Prometheus) to monitor the operation status of the machine and respond immediately if an abnormality occurs.

[1531] Terminals include sensors installed on each machine in the factory, as well as computers and smartphones used by managers. These terminals communicate with the server to send and receive necessary data. For example, they send operational data and abnormal status information on machines to the server and receive instructions from the server.

[1532] The users are factory managers and engineers who take appropriate action based on the information provided by the server. For example, if the server detects that a machine needs maintenance, the users follow the instructions to carry out the maintenance. Also, if an abnormality occurs, the users respond by following the instructions from the server to handle the abnormality.

[1533] For example, if a robot in a factory has not undergone maintenance for 40 days, the server will detect the robot and adjust its maintenance schedule. Also, if a robot is in an error state, the server will detect the error and take appropriate action.

[1534] An example of a prompt to input to a generative AI model is as follows:

[1535] Create a Python program that monitors the operating status of robots in a factory, automatically detects robots that require maintenance, and adjusts the maintenance schedule. Also, add a function to detect robots that are in an error state and handle the error.

[1536] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1537] Step 1:

[1538] The server retrieves reports submitted by subordinates and customers from the database. It uses the report data stored in the database as input and generates a dataset to input into the AI ​​model as output. Specifically, it executes SQL queries to extract the report data and convert it into the required format.

[1539] Step 2:

[1540] The server uses an AI model to review the report, identify issues, their causes, and solutions. It uses the dataset generated in step 1 as input and generates a list of issues, causes, and solutions as output. Specifically, it uses natural language processing technology to analyze the report and extract key information.

[1541] Step 3:

[1542] The server delegates managerial authority to the AI ​​and makes on-site decisions. It uses leave requests and work status data from subordinates as input, and generates a decision on whether to approve or reject the leave request as output. Specifically, it takes into account past leave history and current workload to make the decision.

[1543] Step 4:

[1544] The server monitors the operation status of machinery in the factory in real time. It uses sensor data sent from the terminal as input and generates a report on the operation status of the machinery as output. Specifically, it collects data from the sensors and checks for any abnormalities.

[1545] Step 5:

[1546] The server automatically detects machines that require maintenance and adjusts the maintenance schedule. It uses the operation status report generated in step 4 as input and generates a maintenance schedule as output. Specifically, it determines the need for maintenance based on the number of operating days and abnormality data.

[1547] Step 6:

[1548] The server detects abnormal machines and handles the abnormalities. It uses the operational status report generated in step 4 as input and generates instructions for handling the abnormalities as output. Specific operations include analyzing the abnormal data and proposing appropriate countermeasures.

[1549] Step 7:

[1550] The user takes appropriate action based on the information provided by the server. The input is the maintenance schedule and abnormality handling instructions from the server, and the output is the actual maintenance work and abnormality handling. Specific actions include maintenance and repair of machinery.

[1551] Example 3

[1552] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1553] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. Furthermore, it is difficult to solve problems efficiently because it takes time and effort to extract problems from meeting minutes and derive causes and solutions. Furthermore, it is difficult to understand the flow of discussions and the emotions of participants, which hinders problem solving.

[1554] The identification process by the identification processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means. In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using artificial intelligence to extract problems and derive causes and solutions; a means for delegating managerial authority to artificial intelligence and having it make on-site-level decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing the contents of minutes using natural language processing technology and extracting problems; a means for learning from past data using a machine learning algorithm and deriving causes and solutions; and a means for recognizing participants' emotions from text using an emotion engine and understanding the flow of the discussion. This reduces the workload of on-site managers and enables efficient problem solving.

[1555] "Artificial intelligence" is a technology that enables computer systems to mimic human intelligence and automatically perform tasks such as learning, reasoning, and problem-solving.

[1556] "Managerial authority" refers to the authority that managers have within an organization to make decisions and give instructions.

[1557] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[1558] A "machine learning algorithm" is an algorithm that learns patterns from data and makes predictions and classifications based on the results of that learning.

[1559] An "emotion engine" is a technology that recognizes and analyzes human emotions from data such as text and voice.

[1560] Minutes are documents that record the contents of meetings and discussions.

[1561] A "problem" is a challenge or obstacle that needs to be resolved in a particular situation or system.

[1562] A "cause" is the factor or reason why a problem occurs.

[1563] A "solution" is a specific method or means for solving a problem.

[1564] A "report" is a document that compiles information about a particular topic or situation.

[1565] "On-site level judgment" refers to decisions and responses made based on the situation on-site.

[1566] This invention is a system that uses artificial intelligence to analyze minutes of a meeting, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system will be described below.

[1567] First, the user uploads the meeting minutes to the system. Minutes are often entered as text files or PDF format. The user uploads a file called "Minutes.txt," and the system receives the file and sends it to the server.

[1568] The server uses natural language processing technology to analyze the received minutes. Specifically, it uses the Python libraries NLTK and spaCy. The server analyzes the text using import spacy and extracts sentence structure and keywords. For example, it analyzes the text using nlp = spacy.load("en_core_web_sm") and understands the sentence structure using doc = nlp(text).

[1569] Next, the server uses the analysis results to understand the content of the discussion and extract issues. This is done using text classification algorithms and topic modeling. Specifically, from sklearn.feature_extraction.text import CountVectorizer is used to extract frequently occurring keywords and phrases from the text data. For example, vectorize the text using vectorizer = CountVectorizer(), and identify frequently occurring issues using X = vectorizer.fit_transform(doc).

[1570] The server then analyzes the causes of the extracted issues using a machine learning model trained on past meeting minutes data and their results. Specifically, it predicts the causes of issues using from sklearn.ensemble import RandomForestClassifier. For example, train the model using clf = RandomForestClassifier() and predict the cause using clf.predict(X).

[1571] The server then derives a solution based on the cause. It uses a recommendation system to refer to past success stories. Specifically, it uses from surprise import SVD to propose the optimal solution. For example, build a recommendation system using algo = SVD() and derive a solution using algo.fit(trainset).

[1572] In another embodiment, the server uses an emotion engine to recognize participants' emotions from the text of the minutes. This allows the flow of the discussion to be understood and the causes and solutions of problems to be derived. Specifically, emotions are extracted from the text using from transformers import pipeline. For example, emotions are analyzed using emotion_pipeline = pipeline("sentiment-analysis"), and emotion information is obtained using results = emotion_pipeline(text).

[1573] As a concrete example, the following prompt sentence is input into the generative AI model:

[1574] Examples:

[1575] Meeting minutes: "Project X is progressing slowly due to a lack of resources and communication issues. The proposed solutions are to provide additional resources and set up regular meetings."

[1576] Prompt statement:

[1577] "Analyze the minutes of the following meeting and extract the problem, cause, and solution. Minutes: 'Project X is lagging behind. The cause is a lack of resources and communication issues. The proposed solutions are to allocate additional resources and set up regular meetings.'"

[1578] In this way, the server realizes a system that analyzes the minutes of the meeting, extracts problems, analyzes the causes, and derives solutions. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1579] Step 1: Enter the minutes

[1580] Users upload meeting minutes to the system, often as text files or PDFs.

[1581] Input: The minutes file uploaded by the user (e.g. "minutes.txt")

[1582] Output: Minutes data sent to the server

[1583] Specific operation: The user uploads "Minutes.txt" through the system interface, and the system receives the file and sends it to the server.

[1584] Step 2: Analyze the minutes

[1585] The server uses natural language processing techniques, specifically the Python libraries NLTK and spaCy, to analyze the received minutes.

[1586] Input: Minutes data sent to the server

[1587] Output: Analyzed text data (sentence structure and keywords)

[1588] Specific operation: The server analyzes the text using import spacy to extract sentence structure and keywords. For example, analyze the text using nlp = spacy.load("en_core_web_sm") and understand the sentence structure using doc = nlp(text).

[1589] Step 3: Identify the problem

[1590] The server uses the analysis results to understand the content of the discussion and extract issues, using text classification algorithms and topic modeling.

[1591] Input: Parsed text data

[1592] Output: Extracted issues

[1593] What it does: The server extracts frequently occurring keywords and phrases from the text data using from sklearn.feature_extraction.text import CountVectorizer. For example, it vectorizes the text using vectorizer = CountVectorizer() and identifies frequently occurring issues using X = vectorizer.fit_transform(doc).

[1594] Step 4: Analyze the cause

[1595] The server analyzes the causes of the extracted problems using a machine learning model that has learned from past meeting minutes data and their results.

[1596] Input: Extracted issues

[1597] Output: Analyzed causes

[1598] Specific operation: The server predicts the cause of the problem using from sklearn.ensemble import RandomForestClassifier. For example, train the model using clf = RandomForestClassifier() and predict the cause with clf.predict(X).

[1599] Step 5: Derive a solution

[1600] The server derives solutions based on the cause, using a recommendation system to reference past success stories.

[1601] Input: Analyzed cause

[1602] Output: Proposed solution

[1603] What it does: The server uses from surprise import SVD to propose the optimal solution. For example, build a recommendation system using algo = SVD() and derive a solution using algo.fit(trainset).

[1604] Step 6: Emotion Recognition (Alternative Embodiment)

[1605] The server uses an emotion engine to recognize participants' emotions from the text of the minutes, thereby understanding the flow of the discussion and identifying the causes of problems and solutions.

[1606] Input: Text data of meeting minutes

[1607] Output: Recognized emotion information

[1608] Specific operation: The server extracts emotions from text using from transformers import pipeline. For example, it analyzes emotions using emotion_pipeline = pipeline("sentiment-analysis") and obtains emotion information using results = emotion_pipeline(text).

[1609] (Application example 3)

[1610] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1611] This increases the workload of on-site managers, making it difficult for them to concentrate on other tasks. Analyzing meeting minutes, identifying problems, analyzing causes, and proposing solutions takes time and effort, making it difficult to carry out these tasks efficiently. Furthermore, it is difficult to recognize the emotions of participants during meetings and understand the flow of discussions.

[1612] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1613] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers using AI to identify problems and derive their causes and solutions; a means for delegating managerial authority to AI and allowing it to make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means for analyzing meeting minutes in real time, identifying problems, analyzing causes, and proposing solutions; and a means for using an emotion engine to recognize participants' emotions from the text in the minutes and understand the flow of the discussion. This reduces the workload of on-site managers, allowing them to focus on other tasks. Furthermore, the system can efficiently analyze meeting minutes, identify problems, analyze causes, and propose solutions, while recognizing participants' emotions and understanding the flow of the discussion.

[1614] "Reports submitted by subordinates or customers" refers to reports or documents created and submitted by subordinates or customers in a company or organization.

[1615] "AI" stands for artificial intelligence, a system that mimics human intellectual tasks using technologies such as machine learning and natural language processing.

[1616] A "review" is the process of examining specific documents or data in detail and evaluating or analyzing them.

[1617] "Problem extraction" refers to the process of finding problematic parts or elements in specific documents or data.

[1618] "Cause analysis" is the process of investigating and understanding the reasons and background behind a particular problem.

[1619] "Solution proposal" is the act of proposing effective measures or methods for a specific problem.

[1620] "Managerial authority" refers to the decision-making and command powers held by managers in a company or organization.

[1621] "Field-level decisions" are decisions made based on the situation and conditions at the field.

[1622] "Reducing labor burden" means reducing the workload and stress of employees and managers.

[1623] A "meeting minutes" is a document that records the contents and statements made at a meeting.

[1624] "Real-time analysis" refers to the process of processing data and information instantly and obtaining results.

[1625] An "emotion engine" is a technology or system for recognizing and analyzing emotions from text and speech.

[1626] "Participant emotions" refers to the emotional state and feelings of people attending a meeting or event.

[1627] "Discussion flow" refers to the progression and development of topics in a meeting or discussion.

[1628] To implement the present invention, the following system configuration and program are required.

[1629] System Configuration

[1630] The server uses the following hardware and software:

[1631] Hardware: Robots (e.g. Pepper, NAO, etc.) installed in conference rooms within the factory

[1632] Software: Python, Spacy, Transformers, TextBlob

[1633] Program processing explanation

[1634] The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. Specifically, it uses natural language processing (NLP) technology to analyze the content of the reports and extract problems. It then uses machine learning algorithms to learn from past reports and their results and derive causes and solutions.

[1635] The server also analyzes meeting minutes in real time, extracting problems, analyzing their causes, and proposing solutions. It uses an emotion engine to recognize participants' emotions from the text in the minutes and understand the flow of the discussion. This makes it possible to instantly identify problems during a meeting and propose appropriate solutions.

[1636] Specific examples

[1637] For example, if the minutes of a factory meeting state that "delays on the production line are a problem," the server analyzes the minutes in real time and identifies "delays on the production line" as the problem. It then references past minutes data to analyze the cause of similar problems. For example, if it identifies "machine failure" as the cause, it suggests "performing regular maintenance" as a solution.

[1638] Furthermore, by using an emotion engine to recognize the emotions of participants, it is possible to understand the flow of the discussion from emotional information, such as when the discussion is heated or when participants are confused, and to present solutions at the appropriate time.

[1639] Prompt Sentence Examples

[1640] “Analyze meeting minutes, extract issues, identify sentiment, and propose solutions.

[1641] Minutes: 'Meeting minutes. The problem is delays on the production line. Participants are unhappy.'

[1642] In this way, the server can reduce the workload of on-site managers, allowing them to focus on other tasks. It can also efficiently analyze meeting minutes, identify problems, analyze causes, and propose solutions, and it can recognize participants' emotions and understand the flow of discussions.

[1643] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1644] Step 1:

[1645] The server receives reports submitted by subordinates and customers. The input is the text data of the report, and the output is the storage of that text data. Specifically, when a report is sent to the server, the server stores its contents in a database.

[1646] Step 2:

[1647] The server analyzes the text data of the saved reports using natural language processing (NLP) technology. The input is the text data of the saved reports, and the output is a list of issues as a result of the analysis. Specifically, Spacy is used to tokenize the text and extract important keywords and phrases.

[1648] Step 3:

[1649] The server analyzes the causes of the extracted problems by referencing past report data. The input is a list of problems and past report data, and the output is a list of causes. Specifically, a machine learning algorithm is used to learn patterns from past data and identify the causes of current problems.

[1650] Step 4:

[1651] The server proposes solutions based on the identified causes. The input is a list of causes and the output is a list of solutions. Specifically, it uses TextBlob to parse the text of the causes and generate positive solutions.

[1652] Step 5:

[1653] The server receives meeting minutes in real time and stores them as text data. The input is the audio data of the meeting, and the output is the minutes converted into text. Specifically, the audio data is converted into text using speech recognition technology and stored in a database.

[1654] Step 6:

[1655] The server analyzes the text data of the saved minutes in real time and extracts issues. The input is the text of the minutes, and the output is a list of issues. Specifically, it uses Spacy to tokenize the text and extract important keywords and phrases.

[1656] Step 7:

[1657] The server recognizes participants' emotions from the text data of the minutes. The input is the transcribed minutes, and the output is a list of emotions. Specifically, emotions are extracted from the text using the Transformers sentiment analysis pipeline.

[1658] Step 8:

[1659] The server understands the flow of the discussion based on the extracted issues and emotion information and proposes appropriate solutions. The input is a list of issues and a list of emotions, and the output is a list of solutions. Specifically, it uses TextBlob to analyze the issues and emotion information and generate appropriate solutions.

[1660] In this way, the server can reduce the workload of on-site managers, allowing them to focus on other tasks. It can also efficiently analyze meeting minutes, identify problems, analyze causes, and propose solutions, and it can recognize participants' emotions and understand the flow of discussions.

[1661] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1662] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1663] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1664] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1665] [Third embodiment]

[1666] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1667] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1668] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1669] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1670] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1671] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1672] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1673] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1674] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1675] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1676] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1677] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1678] "Example 1"

[1679] One aspect of the present invention is a system that uses AI to review reports submitted by subordinates and customers. Specifically, the AI ​​analyzes the text data of the report and extracts problems. The AI ​​then analyzes the causes of the problems and derives solutions. This AI uses natural language processing (NLP) technology to understand the content of the report and extract problems. The AI ​​also uses machine learning algorithms to learn from past reports and their results, and derives causes and solutions.

[1680] "Example 2"

[1681] Furthermore, one aspect of the present invention is a system that delegates managerial authority to AI. Specifically, AI takes over on-site decision-making that would normally be performed by managers, allowing them to focus on other tasks. For example, AI can automatically approve or reject vacation requests from subordinates. This decision is based on the subordinate's past vacation history, current workload, and other relevant information.

[1682] "Example 3"

[1683] Another aspect of the present invention is a system in which AI reviews meeting minutes, extracts problems, and derives their causes and solutions. Specifically, the AI ​​analyzes the meeting minutes and extracts problems from the content of the discussion. The AI ​​then analyzes the causes of the problems and derives solutions. This AI uses natural language processing (NLP) technology to understand the content of the minutes and extract the problems. The AI ​​also uses machine learning algorithms to learn from past minutes and their results, and derives causes and solutions.

[1684] The processing flow of each embodiment will be described below.

[1685] "Example 1"

[1686] Step 1: Reports from subordinates and customers are sent to the system.

[1687] Step 2: AI analyzes the text data of the report and extracts problem areas. This analysis is performed using natural language processing (NLP) technology.

[1688] Step 3: AI analyzes the cause of the problem and derives a solution. This analysis is done using machine learning algorithms.

[1689] "Example 2"

[1690] Step 1: Subordinate requests for vacation time and other matters requiring on-site decisions typically made by managers are sent to the system.

[1691] Step 2: The AI ​​evaluates the matter and makes a decision, based on the employee's past leave history, current workload, and other relevant information.

[1692] Step 3: The AI's decision is notified to management, allowing them to focus on other tasks.

[1693] "Example 3"

[1694] Step 1: The meeting minutes are sent to the system.

[1695] Step 2: AI analyzes the text data of the minutes and extracts issues from the content of the discussion. This analysis is carried out using natural language processing (NLP) technology.

[1696] Step 3: AI analyzes the cause of the problem and derives a solution. This analysis is done using machine learning algorithms.

[1697] Example 1

[1698] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1699] On-site managers are required to perform a wide range of tasks, including reviewing reports submitted by subordinates and customers, identifying problems, analyzing causes, and deriving solutions, which places a heavy workload on them and makes it difficult for them to concentrate on other important tasks. Accurately analyzing the contents of reports and deriving appropriate solutions also requires a high level of specialized knowledge and experience, which further increases the burden on on-site managers. Furthermore, reviewing meeting minutes and making decisions based on past experience and future predictions are equally burdensome.

[1700] The identification process by the identification processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means. In this invention, the server includes a means for reviewing reports submitted by subordinates and customers using artificial intelligence, extracting problems and deriving causes and solutions, a means for analyzing the contents of the reports using natural language processing technology, and a means for learning from past reports and their results using a machine learning algorithm to derive causes and solutions. This reduces the workload of on-site managers, allowing them to focus on other important tasks.

[1701] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and learn, reason, and self-correct.

[1702] "Natural language processing technology" is a technology that enables computers to understand, interpret, and generate human language.

[1703] A "machine learning algorithm" is an algorithm that learns patterns from data and makes predictions and classifications.

[1704] A "report" is a document about work or a situation that is sent by a subordinate or a customer.

[1705] "Review" is the process of evaluating the contents of a report and identifying problems and areas for improvement.

[1706] "Problems" are issues or obstacles that need to be resolved and are identified in the report.

[1707] "Cause" refers to the reason or background behind the occurrence of a problem.

[1708] A "solution" is a specific method or means for solving a problem.

[1709] "Managerial authority" refers to the authority that on-site managers have regarding the execution of work and making decisions.

[1710] "On-site level judgment" is a judgment made based on the specific situation at the site.

[1711] "Labor burden" refers to the burden and stress of work that accompanies the performance of work.

[1712] "Text data" is textual information extracted from a report.

[1713] "Analysis results" are the results of analysis obtained using natural language processing technology and machine learning algorithms.

[1714] "Minutes" are documents that record the contents of a meeting.

[1715] "Past experience" refers to knowledge and data about tasks and situations that have been accumulated up to now.

[1716] "Future forecasting" is the prediction of future conditions based on past data and current conditions.

[1717] This invention is a system that uses artificial intelligence to review reports submitted by subordinates and customers, extract problems, analyze their causes, and derive solutions. A specific embodiment of this system will be described below.

[1718] Program Generation

[1719] The server generates a program to review reports submitted by subordinates and customers. This program is built by combining natural language processing (NLP) technology and machine learning algorithms. Specifically, it uses SpaCy and NLTK as NLP libraries and Scikit-learn and TensorFlow as machine learning libraries.

[1720] Program processing

[1721] The server performs the following processing using the generated program.

[1722] 1. Receiving reports and obtaining text data

[1723] The terminal transmits the report created by the user to the server.

[1724] The server receives the report sent from the terminal and extracts the text data of the report. Specifically, if the report is in PDF format, it extracts the text using a PDF parser (e.g., PyPDF2).

[1725] 2. Analysis using natural language processing (NLP)

[1726] The server analyzes the extracted text data using NLP techniques, specifically using SpaCy and NLTK to tokenize the text, tag parts of speech, and analyze dependencies.

[1727] 3. Identifying the problem

[1728] The server extracts problem areas from the report based on the results of NLP analysis, specifically detecting predefined keywords and phrases (e.g., "declining sales" or "customer complaints").

[1729] 4. Cause analysis using machine learning algorithms

[1730] The server analyzes the causes of the extracted problems using a machine learning model trained on past report data. Specifically, it uses a model built using Scikit-learn and TensorFlow.

[1731] 5. Deriving a solution

[1732] The server then derives appropriate solutions based on the results of the root cause analysis, specifically by referencing past success stories and best practices.

[1733] 6. Return of results

[1734] The server compiles the identified problem, cause, and solution and returns it to the terminal.

[1735] The terminal displays the results received from the server to the user.

[1736] Specific examples

[1737] The user sends a "sales report" from the terminal to the server, which includes a description of the reasons for the decline in sales.

[1738] 1. The terminal sends the "sales report" created by the user to the server.

[1739] 2. The server receives the report sent from the device and extracts the text data using a PDF parser.

[1740] 3. The server tokenizes the extracted text data using SpaCy, and performs part-of-speech tagging and dependency analysis.

[1741] 4. The server extracts the problem of "declining sales" from the analysis results.

[1742] 5. The server uses a model trained on past sales reports using Scikit-learn to identify that the cause is a "reduced advertising budget."

[1743] 6. The server comes up with the solution: "Increase advertising spending."

[1744] 7. The server summarizes the problem, cause, and solution and returns it to the terminal.

[1745] 8. The terminal displays the results received from the server to the user.

[1746] Prompt Sentence Examples

[1747] "Analyze the following sales report and identify the problem, its cause, and a solution. Report content: 'Sales this month have decreased by 20% compared to last month. This is likely due to a reduction in advertising expenses.'"

[1748] In this way, the server performs specific processing at each step, and ultimately the system automatically performs everything from identifying problems to proposing solutions to the user.

[1749] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1750] Step 1: Receive the report

[1751] Input: Reports sent by users from their terminals

[1752] Specific operation: The terminal sends the report created by the user to the server in PDF or text format.

[1753] Output: Report data received by the server

[1754] Step 2: Extract text data

[1755] Input: Report data received by the server

[1756] What happens: The server extracts text data from the received report. If the report is in PDF format, it uses a PDF parser (e.g. PyPDF2) to extract the text.

[1757] Output: Extracted text data

[1758] Step 3: Natural Language Processing (NLP) Analysis

[1759] Input: Extracted text data

[1760] What it does: The server analyzes text data using NLP techniques, such as SpaCy and NLTK, to tokenize text, tag parts of speech, and analyze dependencies.

[1761] Output: Parsing results (tokenized text, part-of-speech tags, dependencies)

[1762] Step 4: Identify the problem

[1763] Input: NLP analysis results

[1764] What it does: The server extracts problems in the report based on the analysis results, detecting predefined keywords and phrases (e.g., "declining sales," "customer complaints").

[1765] Output: Extracted issues

[1766] Step 5: Root cause analysis using machine learning algorithms

[1767] Input: Extracted issues

[1768] How it works: The server analyzes the cause using a machine learning model trained on past report data. Specifically, it uses a model built using Scikit-learn and TensorFlow.

[1769] Output: Analysis results (cause of the problem)

[1770] Step 6: Derive a solution

[1771] Input: Analysis results (cause of problem)

[1772] Specific operation: The server derives appropriate solutions based on the results of the cause analysis, and proposes solutions by referring to past success stories and best practices.

[1773] Output: Proposed solution

[1774] Step 7: Returning the results

[1775] Input: Proposed Solution

[1776] Specific operation: The server compiles the identified problem, cause, and solution and returns it to the terminal.

[1777] Output: Results returned to the device (problem, cause, solution)

[1778] Step 8: Viewing the results

[1779] Input: Results returned to the device (problem, cause, solution)

[1780] Specific operation: The terminal displays the results received from the server to the user.

[1781] Output: The results displayed to the user (problem, cause, solution)

[1782] In this way, the server performs specific processing at each step, and ultimately the system automatically performs everything from identifying problems to proposing solutions to the user.

[1783] (Application example 1)

[1784] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1785] Conventional report review systems required humans to manually analyze the contents of reports submitted by subordinates and customers, extracting problems, analyzing causes, and proposing solutions, which was time-consuming, labor-intensive, and inefficient. This also placed a heavy burden on factory workers and managers, increasing the workload of on-site managers and preventing them from concentrating on other tasks. Furthermore, the system lacked the ability to learn from past reports, making it difficult to improve the accuracy of analysis.

[1786] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1787] In this invention, the server includes: a means for reviewing reports submitted by subordinates and customers through AI, identifying problems and proposing causes and solutions; a means for delegating managerial authority to AI and having it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks; a means installed on a robot used in a factory, which automatically reviews reports submitted by workers and managers, identifying problems, analyzing causes, and proposing solutions; and a means for learning from past reports and their results to improve analysis accuracy. This enables automatic analysis of reports, identifying problems, analyzing causes, and proposing solutions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. Furthermore, learning from past reports can improve analysis accuracy.

[1788] "Subordinates and customers" refers to employees in a company or organization who receive instructions, and external users of products and services.

[1789] A "report" is a document that summarizes information or results related to a particular task or project.

[1790] "AI" is an abbreviation for artificial intelligence, and refers to the technology that allows computers to imitate human intelligence to learn and reason.

[1791] "Review" refers to the process of evaluating and analyzing submitted documents and data to identify problems and areas for improvement.

[1792] A "problem" refers to a factor or issue that hinders normal progress in a particular task or project.

[1793] "Cause" refers to the fundamental reason or factor that causes a problem.

[1794] A "solution" refers to a specific method or means for solving a problem.

[1795] "Managerial authority" refers to the authority of a person in a position of authority who has the authority to give instructions and make decisions regarding work in a company or organization.

[1796] "On-site level judgment" refers to making the necessary judgments and decisions immediately at the actual work site.

[1797] "Labor burden" refers to the burden of labor and work necessary to carry out business.

[1798] "Robots used in factories" refers to machinery and devices used in manufacturing and production industries for the purpose of automating and streamlining work.

[1799] "Workers" refers to employees who actually perform work in factories or on-site.

[1800] A "manager" refers to a person in a position of authority who supervises workers and manages operations at a factory or on-site.

[1801] "Automated review" refers to artificial intelligence or machines analyzing and evaluating documents and data without human intervention.

[1802] "Learning" refers to the process by which artificial intelligence acquires new knowledge and skills from past data and experiences.

[1803] "Analysis accuracy" refers to the accuracy and reliability of analyzing data or documents.

[1804] The following system configuration will be described as an embodiment of the present invention.

[1805] The server uses AI to review reports submitted by subordinates and customers, extracting problems and proposing causes and solutions. Specifically, the server uses natural language processing (NLP) technology to analyze the content of the reports and extract problems. Furthermore, it uses machine learning algorithms to learn from past reports and their results and derive causes and solutions.

[1806] The server also includes a means to delegate managerial authority to AI and have it make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks. This frees on-site managers from routine decision-making tasks and allows them to focus on more strategic tasks.

[1807] Furthermore, it includes a means to be installed on robots used in factories and automatically review reports submitted by workers and managers, extract problems, analyze their causes, and propose solutions.The robot analyzes reports submitted by workers and managers in real time and provides immediate feedback.

[1808] The server includes a means for learning from past reports and their results to improve the accuracy of the analysis, which improves the accuracy of the analysis over time, enabling more accurate problem detection and solution proposals.

[1809] In terms of specific hardware, the server uses a computer system equipped with a high-performance processor and large memory capacity. For software, spaCy is used for natural language processing, and scikit-learn is used for machine learning. These software programs provide the functions necessary to analyze the text data in the reports, extract problems, analyze their causes, and propose solutions.

[1810] As a concrete example, consider the following report submitted:

[1811] "Machine A's operation is unstable."

[1812] "Product B's quality varies."

[1813] "Production speed on Line C is slow."

[1814] In response to these reports, the server makes the following suggestions:

[1815] "Machine A is operating unstable. Please perform maintenance on the machine."

[1816] "The quality of Product B is inconsistent. -> Please review your quality control process."

[1817] "The production speed on line C is slow. Please adjust the speed of the production line."

[1818] Example prompts to input to a generative AI model:

[1819] "Machine A's operation is unstable." "Product B's quality varies." "Production speed on line C is slow." Please suggest causes and solutions for these problems.

[1820] By using this prompt, the generative AI model is expected to propose an appropriate cause analysis and solution.

[1821] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1822] Step 1:

[1823] The server receives reports submitted by subordinates and customers. The input is report data in text format. The server analyzes this data using natural language processing (NLP) technology and divides it into sentences. Specifically, it uses the spaCy library to analyze the text and divide it into sentences.

[1824] Step 2:

[1825] The server extracts problems from the divided sentences. The input is the sentence data divided in step 1. The server analyzes each sentence and identifies sentences that contain problems. Specifically, it uses spaCy's dependency analysis function to extract keywords and phrases that indicate problems.

[1826] Step 3:

[1827] The server analyzes the causes of the extracted problems. The input is the problem sentence data extracted in step 2. The server uses a machine learning algorithm to identify the cause using a model trained from past report data and its results. Specifically, the server uses the scikit-learn library to convert the problem sentence into a feature vector, which is then input into the trained model to predict the cause.

[1828] Step 4:

[1829] The server proposes a solution based on the identified cause. The input is the cause data identified in step 3. The server searches a historical database for similar problems and their solutions and proposes the optimal solution. Specifically, it executes a database query to retrieve similar problems and their solutions, and selects the optimal solution.

[1830] Step 5:

[1831] The server generates a report with the proposed solution and provides feedback to the user. The input is the solution data proposed in step 4. The server generates a report including the solution and sends it to the user. Specifically, the server embeds the solution in a report template and notifies the user via email or a dashboard.

[1832] Step 6:

[1833] The server learns from past reports and their results to improve analysis accuracy. The input is user feedback and past report data. The server analyzes the feedback and updates the machine learning model. Specifically, it uses the scikit-learn library to retrain the model with new data and improve analysis accuracy.

[1834] Example 2

[1835] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1836] On-site managers are required to handle a wide range of tasks, such as reviewing reports from subordinates and customers and approving vacation requests, which increases their workload and prevents them from concentrating on other important tasks. In addition, these judgments must be made quickly and accurately, further increasing the burden on managers.

[1837] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1838] In this invention, the server includes a means for reviewing reports submitted by subordinates and customers using artificial intelligence to identify problems and derive their causes and solutions, a means for delegating managerial authority to the artificial intelligence and having the AI ​​make on-site decisions, thereby reducing the workload of on-site managers and allowing them to focus on other tasks, a means for accepting vacation requests, collecting and preprocessing related information, and inputting prompts into the generative AI model to obtain a judgment, and a means for obtaining the judgment result and notifying the user. This reduces the workload of managers, allowing them to focus on other important tasks.

[1839] "Artificial intelligence" is the technology that enables computer systems to mimic human intelligence and perform functions such as learning, reasoning, and problem-solving.

[1840] "Managerial authority" refers to the authority that managers have within an organization to make decisions and give instructions.

[1841] "On-site level decisions" are decisions made to respond quickly and appropriately to specific problems and situations that arise on-site.

[1842] "Labor burden" refers to the physical and mental burden incurred when performing work.

[1843] A "leave application" is the application documents and procedures submitted by an employee to obtain leave.

[1844] A "generative AI model" is an artificial intelligence model that uses natural language processing and machine learning techniques to generate appropriate responses and decisions based on input data.

[1845] A "prompt" is text containing instructions or questions to be input into a generative AI model.

[1846] "Data preprocessing" is the process of formatting collected data into a form suitable for analysis or model input.

[1847] A "judgment result" is the response or decision that the generative AI model outputs based on the prompt.

[1848] "User" means an individual or member of an organization who uses the system.

[1849] This invention is a system that delegates the authority of managers to artificial intelligence and allows them to make on-site decisions, thereby reducing the workload of managers and allowing them to focus on other tasks. A specific embodiment of this system will be described below.

[1850] Hardware and software used

[1851] Hardware: Servers, terminals

[1852] Software: Generative AI models (e.g., GPT-4), database management systems (e.g., MySQL), business management software (e.g., Jira)

[1853] System Overview

[1854] The system delegates tasks like reviewing reports from subordinates and clients, approving vacation requests, and other on-the-ground decisions to artificial intelligence, allowing managers to focus on other important tasks.

[1855] Data processing and calculation flow

[1856] 1. Acceptance of leave requests

[1857] User: A subordinate uses a terminal to apply for leave. Specifically, the subordinate enters the necessary information into the leave application form in the business management software and clicks the submit button.

[1858] Terminal: Sends the information entered in the vacation request form to the server.

[1859] 2. Data Collection

[1860] Server: After receiving the leave request, access the database management system to collect relevant information such as the subordinate's past leave history, current workload, and workload of team members.

[1861] What it does: Executes an SQL query to retrieve the required data.

[1862] 3. Data Preprocessing

[1863] Server: Preprocesses the collected data, specifically formatting the data and filling in missing values.

[1864] Specific operation: Standardize the data format and impute missing values ​​with the mean or median.

[1865] 4. Prompt generation

[1866] Server: Based on the preprocessed data, it generates prompt sentences to be input to the generative AI model.

[1867] Specific behavior: Generates a prompt like the following:

[1868] Please use the following information to help you decide whether your subordinates should request time off.

[1869] Past vacation history: 10 days of vacation taken in the past year

[1870] Current workload: Project A is 70% complete

[1871] Team member workload: High

[1872] 5. Input and judgment for AI model

[1873] Server: The generated prompt sentence is input into the generative AI model to determine whether to approve or reject the vacation request.

[1874] Specific operation: Call the API of the generative AI model and send a prompt.

[1875] 6. Obtaining the judgment result

[1876] Server: Obtains the judgment results from the generative AI model.

[1877] Specific operation: Receives the API response and analyzes the judgment result.

[1878] Approve the vacation request, but be careful not to delay progress on Project A.

[1879] 7. Notification of Results

[1880] Server: Sends the decision result to the device.

[1881] Terminal: Notifies the user (subordinate) of the received judgment results. Specifically, the notification function of the work management software is used to display the results to the subordinate.

[1882] Specific behavior: Display "Vacation request approved" on the vacation request form along with detailed comments.

[1883] In this way, artificial intelligence can make on-site decisions on behalf of managers, reducing the workload on managers and allowing them to focus on other important tasks.

[1884] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1885] Step 1: Accepting the leave request

[1886] User: A subordinate uses a terminal to apply for leave. Specifically, the subordinate enters the necessary information into the leave application form in the business management software and clicks the submit button.

[1887] Input: Information entered into the leave request form (e.g. applicant name, leave period, reason, etc.).

[1888] Output: The vacation request data is sent to the server.

[1889] Specific operation: The terminal sends the information entered in the vacation application form to the server.

[1890] Step 2: Data collection

[1891] Server: After receiving the leave request, access the database management system to collect relevant information such as the subordinate's past leave history, current workload, and workload of team members.

[1892] Input: Leave application data, access information to database management system.

[1893] Output: Relevant information collected (e.g., past leave history, current workload, workload of team members).

[1894] What happens: The server executes an SQL query to retrieve the required data from the database.

[1895] Step 3: Data Preprocessing

[1896] Server: Preprocesses the collected data, specifically formatting the data and filling in missing values.

[1897] Input: Relevant information collected.

[1898] Output: Preprocessed data.

[1899] Specific operation: The server unifies the data format and imputes missing values ​​with the mean or median.

[1900] Step 4: Generate a prompt statement

[1901] Server: Based on the preprocessed data, it generates prompt sentences to be input to the generative AI model.

[1902] Input: Preprocessed data.

[1903] Output: The generated prompt statement.

[1904] Specific behavior: The server generates a prompt like this:

[1905] Please use the following information to help you decide whether your subordinates should request time off.

[1906] Past vacation history: 10 days of vacation taken in the past year

[1907] Current workload: Project A is 70% complete

[1908] Team member workload: High

[1909] Step 5: Input and judgment of the AI ​​model

[1910] Server: The generated prompt sentence is input into the generative AI model to determine whether to approve ...

Claims

1. A means for receiving a report created by a user regarding the user's business; A means for deriving a problem in the business, a cause of the problem, and a solution to the problem from the content of the received report using a prompt sentence that instructs the deriving of the problem, the cause of the problem, and a solution to the problem, and a generative AI model; A means for generating a report representing the derived problems in the business, causes of the problems, and solutions to the problems, using a prompt sentence instructing to generate a report representing the derived problems in the business, causes of the problems, and solutions to the problems, and the generative AI model; means for presenting the generated report to the user; A system including:

2. means for recognizing an emotion of the user from the report using an emotion engine; 2. The system of claim 1, wherein the report generating means generates the report using a prompt sentence instructing the generation of a report representing the derived business problems, causes of the problems, and solutions to the problems based on the recognized user's emotions, and the generative AI model.

3. 3. The system of claim 2, wherein the means for generating a report generates the report using the prompt sentence instructing the user to generate the report with adjusted priorities of the solutions based on the recognized user sentiment, and the generative AI model.

Images