system

Abstract

Provide a system. 【Solution means】Means for collecting communication records and related documents of all employees, Means for analyzing the collected data and creating an index, Means for training an AI model based on the analyzed data, Means for receiving inquiries from users, Means for analyzing the received inquiry content, Means for generating an appropriate answer based on the analysis result, Means for presenting the generated answer to the user, Means for collecting feedback from users and improving the AI model, A system including the above.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the modern business environment, many companies are troubled by the ambiguity of the application methods, rules, guidelines, and manuals of in-house systems. Due to this problem, it often takes a long time to search for information, resulting in a decline in business efficiency. In addition, since the business process is not documented, it may be difficult to find the necessary information. Furthermore, since a lot of resources are allocated to the inquiry response work, the productivity of employees may be impaired. In order to solve these problems, a system that can respond 24 hours a day, 365 days a year is required.

Means for Solving the Problems

[0005] To solve the above problems, the present invention proposes the following means. First, it collects communication records and related documents of all employees and analyzes this data to create an index. Next, it trains an AI model based on the analyzed data and builds a system that receives user inquiries in real time. It analyzes the content of the received inquiries and generates the optimal answer. The generated answer is presented to the user and includes links to relevant internal documents and manuals. Furthermore, it collects feedback from users and uses this feedback to improve the accuracy of the AI ​​model. This system makes it possible to free up resources for inquiry handling and improve employee productivity.

[0006] "Communication records" refer to records of digital messages, such as emails and chat logs, that employees send and receive during work hours.

[0007] "Related documents" refers to business-related documents such as guidelines, manuals, and application procedures that employees use in their work.

[0008] An "index" is a data structure that organizes collected data based on keywords and context, making it easier to search.

[0009] An "AI model" is an artificial intelligence algorithm that has been trained to perform a specific task based on a large amount of data.

[0010] A "user" refers to an employee or user who can make inquiries to the system and use the system to obtain answers.

[0011] "Inquiry" refers to a question or request for information that a user enters into the system.

[0012] "Analysis" is the process of understanding collected data and inquiry content using natural language processing technology and extracting necessary information.

[0013] "Generation" refers to the process where a system creates the optimal answer based on the analysis results.

[0014] "Feedback" refers to the evaluations and comments that users make regarding the system's responses. [Brief explanation of the drawing]

[0015] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14] It is a sequence diagram showing the processing flow of a data processing system in Application Example 2 when a sentiment engine is combined.

Mode for Carrying Out the Invention

[0016] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0017] First, the terms used in the following description will be explained.

[0018] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units 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), and the like.

[0019] In the following embodiments, a numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0020] In the following embodiments, a numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0021] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0022] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0023] [First Embodiment]

[0024] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0025] As shown in Figure 1, the 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.

[0026] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0027] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0028] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0029] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0030] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0031] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0033] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0034] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0035] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0036] This invention is a system designed to solve a business problem where companies face the issue of unclear internal system application procedures, rules, guidelines, and manuals, which require significant time to research. This system utilizes AI to provide intelligent inquiry support available 24 / 7, aiming to improve operational efficiency and optimize resources.

[0037] System Overview

[0038] This system operates using the following procedure:

[0039] 1. Data Collection

[0040] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[0041] 2. Data analysis and index creation

[0042] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. This allows for the rapid extraction of appropriate information in response to queries.

[0043] 3. Training the AI ​​model

[0044] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[0045] 4. Real-time inquiry reception

[0046] The device (PC or smartphone) receives user inquiries in real time. Users input their questions in natural language.

[0047] 5. Analysis of the inquiry content

[0048] The server analyzes the received inquiry and extracts relevant keywords and contextual information.

[0049] 6. Generating the answer

[0050] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[0051] 7. Providing the answer

[0052] The device presents the generated answer to the user.

[0053] 8. Gathering feedback and improving the AI ​​model

[0054] The device collects user feedback and sends it to the server. The server uses this feedback to improve the accuracy of the AI ​​model.

[0055] Specific example

[0056] Example 1: Data Collection

[0057] The server collects emails and chat logs sent and received by employee A over the past six months.

[0058] Example 2: Data analysis and index creation

[0059] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[0060] Example 3: Training an AI model

[0061] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[0062] Example 4: Real-time inquiry reception

[0063] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[0064] Example 5: Analysis of inquiry content

[0065] The server extracts and analyzes the keywords "new project" and "application method".

[0066] Example 6: Generating an answer

[0067] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0068] Example 7: Providing an answer

[0069] The generated response is displayed on User B's terminal screen.

[0070] Example 8: Gathering feedback and improving the AI ​​model

[0071] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[0072] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[0073] The following describes the processing flow.

[0074] Step 1: Data Collection

[0075] The server connects to all employees' email and chat servers and downloads communication records for the past six months. It also accesses the document management system to retrieve relevant documents such as company guidelines, manuals, and application procedures.

[0076] Step 2: Data Analysis and Indexing

[0077] The server uses a natural language processing (NLP) engine to analyze the collected data. Specifically, it extracts keywords from each document and message and understands their contextual information. Then, based on the analysis results, it creates an index and stores it in a database.

[0078] Step 3: Training the AI ​​model

[0079] The server uses the analyzed data to train an AI model using a supervised learning algorithm. The model is designed to generate the best possible answers to user inquiries. The training process involves numerous question-and-answer pairs.

[0080] Step 4: Real-time inquiry reception

[0081] The terminal (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the terminal's interface.

[0082] Step 5: Analyzing the inquiry content

[0083] The server receives inquiries from users and uses a natural language processing engine to analyze the question content. The analysis extracts important keywords and contextual information from the question and compares them against an index database.

[0084] Step 6: Generating the answer

[0085] The server generates an appropriate response based on the query analysis results. If necessary, it also retrieves links to relevant internal documents and manuals from the index database and includes them in the response.

[0086] Step 7: Present your answer

[0087] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[0088] Step 8: Gathering feedback and improving the AI ​​model

[0089] The device receives feedback provided by the user and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[0090] (Example 1)

[0091] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0092] Traditional corporate systems often lacked clear internal rules, guidelines, and manuals, resulting in employees spending a significant amount of time searching for this information. This led to decreased employee efficiency and a substantial decline in productivity.

[0093] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0094] In this invention, the server includes means for collecting communication records and business-related information, means for analyzing the collected data using natural language processing technology and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users in real time, means for analyzing the content of the received inquiries using natural language processing technology, means for generating appropriate answers based on the analysis results, means for presenting the generated answers to users, and means for collecting feedback from users and improving the artificial intelligence model. As a result, users can obtain information quickly and appropriately, enabling increased efficiency in internal operations and optimization of resources.

[0095] "Communication records" refer to records of communication between employees within a company, such as emails and chat logs.

[0096] "Business-related information" refers to documentary information such as guidelines, manuals, and procedural documents necessary for carrying out business operations within a company.

[0097] "Natural language processing technology" refers to technologies that enable computers to recognize, interpret, and generate human language, and specifically includes text analysis and contextual understanding.

[0098] An "artificial intelligence model" is an algorithm that learns from large amounts of data, extracts patterns and knowledge, and automatically generates the optimal answer to an inquiry.

[0099] An "index" refers to a systematic organization of keywords and contextual information used to quickly retrieve specific information from a large amount of data.

[0100] "Users" refers to company employees who use this system to make internal inquiries.

[0101] "Real-time" refers to a timeframe in which processing and responses occur almost instantly, meaning that a response is provided immediately after a user submits a question.

[0102] "Feedback" refers to information collected from users who provide evaluations and opinions on the answers they receive, in order to help improve the system.

[0103] "Analysis results" refer to data containing important keywords and contextual information obtained after analyzing the content of a query using natural language processing technology.

[0104] "Answer" refers to the information and instructions provided in response to a user's inquiry, generated based on the analysis results.

[0105] A "link" is a means of reference, such as a URL, that contains direct reference information to internal documents or manuals related to the generated response.

[0106] The system according to this invention collects communication records and business-related information, analyzes and utilizes it to provide prompt and appropriate responses to user inquiries. This system significantly improves operational efficiency within a company and optimizes resources.

[0107] Hardware and software configuration

[0108] server

[0109] 1. Data collection:

[0110] The server collects communication records such as employee emails and chat logs, as well as business-related information such as guidelines and manuals. This utilizes the APIs of Microsoft® Exchange Server and Slack.

[0111] 2. Data Analysis:

[0112] The server analyzes the collected data using natural language processing (NLP) techniques. In this process, it creates indexes using Apache® Lucene and Elasticsearch®.

[0113] 3. AI model training:

[0114] Based on the analyzed and indexed data, AI models are trained using machine learning frameworks such as TENSORFLOW® and PyTorch.

[0115] 4. Analysis of the inquiry content:

[0116] The server analyzes user inquiries using advanced NLP models such as BERT and GPT-3 (registered trademark).

[0117] 5. Generating the answer:

[0118] Based on the analysis results, generate appropriate answers. Include links to relevant internal documents and manuals as needed.

[0119] 6. Gathering feedback and improving the AI ​​model:

[0120] We collect user feedback and use it to continuously improve our AI models.

[0121] terminal

[0122] 1. Real-time inquiry reception:

[0123] The terminal receives user inquiries in real time via PCs and smartphones. Users input their questions in natural language.

[0124] 2. Providing the answer:

[0125] The device displays the responses sent from the server to the user in real time.

[0126] user

[0127] 1. Enter your inquiry:

[0128] Users input questions in natural language via their devices.

[0129] 2. Providing feedback:

[0130] Feedback on the provided answers is entered into the terminal, and that feedback is sent to the server.

[0131] Specific example

[0132] Example 1: Data Collection

[0133] The server collects emails and chat logs sent and received by employee A over the past six months.

[0134] Example 2: Data Analysis

[0135] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[0136] Example 3: Training an AI model

[0137] The server trains an AI model based on the collected data, learning appropriate response patterns based on specific keywords and contexts.

[0138] Example 4: Real-time inquiry reception

[0139] User B uses their device to type "Please tell me how to apply for a new project" and sends it.

[0140] Example 5: Analysis of inquiry content

[0141] The server extracts and analyzes the keywords "new project" and "application method".

[0142] Example 6: Generating an answer

[0143] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0144] Example 7: Providing an answer

[0145] The generated response is displayed on User B's terminal screen.

[0146] Example 8: Gathering feedback and improving the AI ​​model

[0147] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it to improve the AI ​​model.

[0148] Example of a prompt

[0149] "Please tell me how to apply for a new project."

[0150] "What is the procedure for applying for leave?"

[0151] "Could you show me the minutes of the recent meeting?"

[0152] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[0153] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0154] Step 1:

[0155] Data collection

[0156] input:

[0157] All employee communication records (emails, chat logs), and business-related information (guidelines, manuals)

[0158] Specific actions:

[0159] The server uses the Microsoft Exchange Server API and the Slack API to collect employee emails and chat logs. It also retrieves business-related documents (guidelines, manuals) from the company's file server.

[0160] output:

[0161] Collected communication records and work-related information of all employees

[0162] Specific example:

[0163] The server collects employee A's emails and chat logs from the past six months, and simultaneously retrieves guidelines and manuals.

[0164] Step 2:

[0165] Data analysis and index creation

[0166] input:

[0167] Collected communication records and work-related information of all employees

[0168] Specific actions:

[0169] The server analyzes the collected data using Apache Lucene and Elasticsearch and creates an index. This involves using natural language processing techniques to extract and classify keywords.

[0170] output:

[0171] Indexed dataset

[0172] Specific example:

[0173] The server analyzes the data for keywords such as "how to apply for a new project" and "guidelines," and creates an index.

[0174] Step 3:

[0175] AI model training

[0176] input:

[0177] Indexed dataset

[0178] Specific actions:

[0179] The server uses TensorFlow or PyTorch to train an AI model based on indexed data. This process uses a large amount of text data to learn question and associated answer patterns.

[0180] output:

[0181] Trained AI model

[0182] Specific example:

[0183] The server trains an AI model based on indexed data, learning appropriate response patterns regarding "how to submit a new project application."

[0184] Step 4:

[0185] Real-time inquiry reception

[0186] input:

[0187] User inquiries (questions entered in natural language)

[0188] Specific actions:

[0189] The user sends a question to the system using a terminal. The terminal sends this inquiry to the server in real time.

[0190] output:

[0191] Inquiry content sent to the server

[0192] Specific example:

[0193] User B uses their device and types and sends the message, "Please tell me how to apply for a new project."

[0194] Step 5:

[0195] Analysis of inquiry content

[0196] input:

[0197] Inquiry content sent from the device

[0198] Specific actions:

[0199] The server analyzes queries using natural language processing techniques such as BERT and GPT-3, and extracts key keywords and contextual information.

[0200] output:

[0201] Analyzed query content (main keywords and contextual information)

[0202] Specific example:

[0203] The server extracts and analyzes the keywords "new project" and "application method" from user B's inquiry.

[0204] Step 6:

[0205] Answer generation

[0206] input:

[0207] Analyzed inquiry content

[0208] Specific actions:

[0209] The server uses a trained AI model to generate appropriate answers based on the analyzed query. Links to relevant internal documents and manuals are included as needed.

[0210] output:

[0211] Generated answer

[0212] Specific example:

[0213] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0214] Step 7:

[0215] Providing an answer

[0216] input:

[0217] Generated answer

[0218] Specific actions:

[0219] The device displays the responses sent from the server to the user in real time.

[0220] output:

[0221] User-displayed response

[0222] Specific example:

[0223] The generated response is displayed on User B's device.

[0224] Step 8:

[0225] Gathering feedback and improving the AI ​​model

[0226] input:

[0227] User Feedback

[0228] Specific actions:

[0229] The user enters feedback on the answer into the device. The device sends this feedback to the server, which adds it to the AI ​​model's training data to make improvements.

[0230] output:

[0231] Improved AI model

[0232] Specific example:

[0233] User B sends feedback saying "The answer was helpful," and the server uses that feedback to improve the accuracy of the AI ​​model.

[0234] (Application Example 1)

[0235] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0236] There is a problem in factories where information regarding business processes, operating procedures, and troubleshooting is ambiguous, making it difficult for employees to quickly obtain this information. This often leads to decreased operational efficiency and delays in troubleshooting. This invention aims to solve these problems and improve operational efficiency within factories by providing an AI-powered system.

[0237] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0238] In this invention, the server includes means for collecting communication records and related documents of all employees, means for analyzing the collected data and creating an index, and means for training an AI model based on the analyzed data. This makes it possible to respond quickly and accurately to user inquiries. Furthermore, by adding means for factory workers to make inquiries using tablet devices or robots, and means for supporting questions regarding factory operating procedures and troubleshooting, it becomes possible to speed up and streamline information acquisition within the factory.

[0239] "Communication records" refer to records of digital communications, such as emails and chat logs, conducted by all employees within the company.

[0240] "Related documents" refer to documents containing information such as guidelines, manuals, and operational procedures used within a company.

[0241] An "index" is a list or database structure that organizes analyzed data into a searchable format.

[0242] An "AI model" is an artificial intelligence model generated using machine learning algorithms based on collected and analyzed data.

[0243] An "inquiry" refers to a question or request that a user enters into a system.

[0244] "Answer" refers to the information generated by the AI ​​model and provided in response to a user's inquiry.

[0245] "Feedback" refers to evaluations and opinions given to responses provided by users.

[0246] A "factory worker" refers to an employee who performs tasks such as production, management, and maintenance within a factory.

[0247] A "tablet device" is a portable computer device equipped with a touchscreen.

[0248] A "robot" is a machine that operates automatically and performs tasks within a factory.

[0249] "Operating procedures" refer to step-by-step instructions or methods for performing a specific task or operation.

[0250] "Troubleshooting" refers to the procedures and methods used to resolve problems that occur in a system or machine.

[0251] "To support" means to provide the necessary information and means to perform a specific task or operation.

[0252] This invention is a system for efficiently acquiring information on operating procedures and troubleshooting within a factory. The system functions by collecting communication records and related documents from all employees, analyzing this data to create an index, and training an AI model. It also includes a process of receiving user inquiries in real time, analyzing them, generating and presenting answers, and collecting user feedback to improve the AI ​​model.

[0253] The server first collects communication records and related documents, such as emails and chat logs, from all employees. Next, it analyzes the collected data using natural language processing technology and creates an index in a searchable format. The analyzed data is used to train an AI model, which is built to generate appropriate responses.

[0254] Users (factory workers) can make inquiries using tablet devices or robots. For example, if a worker asks "How do I configure machine A?" or "How do I troubleshoot error code E123?", the device sends this inquiry to the server in real time. The server analyzes the received inquiry, extracts relevant keywords and contextual information, generates an appropriate answer based on the analysis results, and presents the answer to the device.

[0255] The generated responses may include links to relevant internal documents and manuals. This allows workers to quickly obtain the necessary information and improve work efficiency. User feedback is also important; by collecting feedback on whether the provided responses were helpful, the server can continuously improve the AI ​​model.

[0256] This system is implemented using hardware such as NVIDIA GPUs, cloud instances from AWS® and Azure®, and software such as Hugging Face's Transformers library. This enables high-performance data processing and analysis.

[0257] Specific example

[0258] For example, a factory worker might type "How do I troubleshoot error code E123 on machine B?" into a tablet. In this case, the system would generate the following response: "Error code E123 requires you to first turn off the power and restart the machine. Detailed troubleshooting information can be found at the following link."

[0259] Example of a prompt

[0260] "Please tell me how to troubleshoot error code E123 on machine B."

[0261] In this way, this system can streamline business processes within the factory and improve worker productivity.

[0262] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0263] Step 1:

[0264] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from all employees. The input is the communication records and related documents of all employees, and the output is a collection of the collected data. This data is used for subsequent analysis and indexing.

[0265] Step 2:

[0266] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. The input is the data collected in step 1, and the output is the analyzed data and its index. In this process, the meaning of the text is understood and converted into a searchable format.

[0267] Step 3:

[0268] The server trains an AI model based on the analyzed data. The input is the indexed data obtained in step 2, and the output is the trained AI model. This gives the AI ​​the ability to generate appropriate answers to queries.

[0269] Step 4:

[0270] The user enters the inquiry in natural language using a tablet or robot. The input is the user's inquiry statement, and the output is the inquiry request. This request is sent to the server in real time.

[0271] Step 5:

[0272] The server analyzes the received inquiry. The input is the inquiry request sent in step 4, and the output is the analyzed inquiry content. Keywords and contextual information are extracted and used as basic data to generate appropriate responses.

[0273] Step 6:

[0274] The server generates an appropriate response based on the analysis results. The input is the query analyzed in step 5 and the trained AI model, and the output is the generated response. Links to relevant documents and manuals are included as needed.

[0275] Step 7:

[0276] The terminal presents the generated answer to the user. The input is the answer generated in step 6, and the output is the answer displayed to the user. The user can then use this answer to solve the problem.

[0277] Step 8:

[0278] The user provides feedback on the provided answers. The input is the user's feedback, and the output is a feedback request. This request is sent to the server and used to improve the AI ​​model.

[0279] Step 9:

[0280] The server improves the AI ​​model based on user feedback. The input is the feedback request submitted in step 8, and the output is the improved AI model. This improves the accuracy of answers to future inquiries.

[0281] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0282] This invention is an intelligent inquiry response system combined with an emotion engine that recognizes users' emotions. This system is designed to improve the efficiency of in-house operations and realizes inquiry support that can be provided 24 hours a day, 365 days a year. Also, by analyzing the emotions of users, the quality of answers can be further enhanced.

[0283] Overview of the System

[0284] This system operates in the following procedure.

[0285] 1. Data Collection

[0286] The server collects communication records (e-mails, chat logs) and related documents (guidelines, manuals, etc.) of all employees.

[0287] 2. Data Analysis and Index Creation

[0288] The server analyzes the collected data using an NLP engine and creates an index. This enables the rapid extraction of appropriate information for inquiry content.

[0289] 3. Learning of the AI Model

[0290] The server trains an AI model based on the analyzed data. This AI model can generate an optimal answer to an inquiry from a user.

[0291] 4. Real-time Inquiry Reception

[0292] [[ID=4!]]The terminal (PC or smartphone) receives inquiries from users in real time. The user inputs and sends a question in natural language using the terminal interface.

[0293] 5. Emotion Recognition by the Emotion Engine

[0294] The server uses an emotion engine to recognize emotions from the user's inquiry. This allows it to understand the user's emotional state and use the data to generate more appropriate responses.

[0295] 6. Analysis of the inquiry content

[0296] The server re-analyzes the query content, taking sentiment information into account, and then extracts keywords and contextual information.

[0297] 7. Generating the answer

[0298] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[0299] 8. Providing the answer

[0300] The device displays the generated response to the user. The user can view the generated response and associated links on the device screen.

[0301] 9. Gathering feedback and improving the AI ​​model

[0302] The device collects user feedback and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[0303] Specific example

[0304] Example 1: Data Collection

[0305] The server collects emails and chat logs sent and received by employee A over the past six months.

[0306] Example 2: Data analysis and index creation

[0307] The server analyzes keywords such as "How to Apply for a New Project" and "Guidelines" and creates an index.

[0308] Example 3: Training of AI Model

[0309] The server trains an AI model using the collected data and learns appropriate response patterns based on keywords and context.

[0310] Example 4: Receiving Real-Time Inquiries

[0311] User B uses the terminal to input and send "Please tell me how to apply for a new project."

[0312] Example 5: Sentiment Recognition by Sentiment Engine

[0313] The server uses the sentiment engine to recognize that the user is confused from the inquiry sentence of the user who said "Please tell me how to apply for a new project."

[0314] Example 6: Analysis of Inquiry Content

[0315] The server extracts keywords such as "new project" and "application method" and conducts further analysis based on sentiment information.

[0316] Example 7: Generation of Answer

[0317] The server generates an answer "The method to apply for a new project is to do it from the application form on the portal site. Please refer to the following link for detailed procedures." and provides it in an easily understandable form to resolve the user's confusion.

[0318] Example 8: Presentation of Answer

[0319] The generated answer is displayed on the terminal screen of User B.

[0320] Example 9: Gathering feedback and improving the AI ​​model

[0321] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[0322] In this way, the system analyzes the content of inquiries and the user's emotions to generate more appropriate and effective responses. This leads to increased operational efficiency and improved employee productivity.

[0323] The following describes the processing flow.

[0324] Step 1: Data Collection

[0325] The server connects to all employees' mail and chat servers to download emails and chat logs from the past six months. It also accesses the document management system to collect relevant documents such as company guidelines, manuals, and application procedures.

[0326] Step 2: Data Analysis and Indexing

[0327] The server analyzes the collected data using an NLP engine. Specifically, it extracts important keywords and contextual information from each document and message. Then, it creates an index based on the analysis results and stores it in a database.

[0328] Step 3: Training the AI ​​model

[0329] The server trains an AI model using a supervised learning algorithm based on the analyzed data. The training data includes pairs of past queries and their appropriate answers. This gives the model the ability to generate optimal answers to a variety of queries.

[0330] Step 4: Real-time inquiry reception

[0331] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions using natural language.

[0332] Step 5: Emotion recognition by the emotion engine

[0333] The server receives the user's inquiry and uses an emotion engine to analyze the user's emotions. The emotion engine uses NLP techniques to identify emotions from the inquiry text.

[0334] Step 6: Analyzing the inquiry content

[0335] The server takes sentiment information into account and re-analyzes the query using an NLP engine. This extracts keywords and contextual information, and adds sentiment information.

[0336] Step 7: Generating the answer

[0337] Based on the analysis results, the server generates the most appropriate response for the user's emotional state and inquiry. Links to relevant internal documents and manuals are also added as needed.

[0338] Step 8: Present your answer

[0339] The device displays the generated answers to the user. The user then views the provided answers and associated links on the device screen.

[0340] Step 9: Gathering Feedback and Improving the AI ​​Model

[0341] The device collects user feedback and sends it to the server. The server analyzes this feedback and adds it to the AI ​​model as new training data. This improves the accuracy of the AI ​​model.

[0342] In this way, the system comprehensively analyzes the user's inquiry and emotions, enabling a process that generates the optimal response. This leads to increased operational efficiency and improved employee productivity.

[0343] (Example 2)

[0344] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0345] Traditional inquiry handling systems struggled to provide 24 / 7 support and often suffered from low-quality responses due to their failure to consider the user's emotional state. Furthermore, they were inefficient in handling inquiries because they couldn't effectively utilize all employee communication records and related documents. A system is needed to solve these problems and improve the efficiency of internal operations.

[0346] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0347] In this invention, the server includes means for collecting communication records and related documents of all employees; means for analyzing the collected data using a natural language processing engine and creating an index; means for training a machine learning model based on the analyzed data; means for receiving user inquiries in real time; means for recognizing the user's emotions from the received inquiry content using an emotion engine; means for analyzing the inquiry content taking emotional information into account; means for generating appropriate answers based on the analysis results; means for presenting the generated answers to the user; and means for collecting user feedback and improving the machine learning model. This enables 24 / 7 inquiry support and the provision of high-quality answers that take user emotions into consideration.

[0348] "All employee communication records" refers to digital communication data such as emails and chat logs used by company employees for work purposes.

[0349] "Related documents" refer to documents related to business operations, such as guidelines and manuals, that the company possesses.

[0350] A "natural language processing engine" is software that analyzes text data to extract keywords and perform sentiment analysis.

[0351] A "machine learning model" is a collection of algorithms that learn from data and perform pattern recognition and prediction.

[0352] A "means of receiving in real time" refers to a system or interface that can immediately receive inquiries from users.

[0353] An "emotion engine" is a tool or software that recognizes a user's emotional state from text data.

[0354] "Index creation" refers to techniques that generate lists of keywords and topics to enable rapid searching of large amounts of data.

[0355] "Means for analyzing inquiry content" refers to technologies for understanding the questions and requests entered by users and extracting the necessary information.

[0356] "Means for generating appropriate answers" refers to the process of creating the optimal response to a user's inquiry based on analyzed data.

[0357] "Means of presenting to the user" refers to a system for displaying generated answers and information to the user.

[0358] "Means of collecting feedback" refers to a system for gathering evaluations and opinions from users and using them to improve the system.

[0359] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[0360] System Overview

[0361] Data collection

[0362] The server collects communication records (emails, chat logs) of all employees using Microsoft Exchange Server and the Slack API. It also collects relevant documents such as guidelines and manuals from the company's file server and document management system.

[0363] Data analysis and index creation

[0364] The server uses natural language processing engines such as Google® NLP and IBM Watson® to analyze the collected data and extract key keywords and topics. This creates an efficient data index, enabling rapid information retrieval.

[0365] AI model training

[0366] The server uses libraries such as Hugging Face's Transformers library to train a machine learning model based on the analyzed data. During the training process, the data is split into training and validation sets, and the model parameters are adjusted. The trained AI model is then saved and prepared to generate the optimal response to queries.

[0367] Real-time inquiry reception

[0368] The device (PC or smartphone) receives user inquiries in real time through a web interface built with React or Vue.js. Users input and submit questions in natural language.

[0369] Emotion recognition by an emotion engine

[0370] The server uses emotion engines such as Affectiva or IBM Watson Tone Analyzer to recognize emotions from user inquiries. This data is then used to understand the user's emotional state and generate appropriate responses.

[0371] Analysis of inquiry content

[0372] The server re-analyzes the query content using the spaCy library, taking sentiment information into account. It extracts key keywords and contextual information and performs a detailed analysis based on the user's emotional state.

[0373] Answer generation

[0374] The server uses models such as OpenAI's GPT-3 to generate appropriate responses based on the analysis results. Links to relevant documents and manuals are included as needed. For example, it might generate a response like, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0375] Providing an answer

[0376] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[0377] Gathering feedback and improving the AI ​​model

[0378] The device collects user feedback and sends it to the server. The server analyzes the feedback, adds it as training data for the model, and improves the accuracy of the AI ​​model by retraining it.

[0379] Specific example

[0380] Data collection

[0381] The server uses the Microsoft Exchange Server API to retrieve employee A's email data and the Slack API to download employee A's chat logs.

[0382] Data analysis and index creation

[0383] The server uses the Google NLP API to analyze keywords such as "how to apply for a new project" and "project guidelines" and creates an index.

[0384] AI model training

[0385] The server uses the Hugging Face Transformers library to train an AI model based on the analyzed data, splitting it into training data and validation data for further training.

[0386] Real-time inquiry reception

[0387] The user enters "How do I submit a new project?" into a web interface built with React and submits it.

[0388] Emotion recognition by an emotion engine

[0389] The server uses IBM Watson Tone Analyzer to recognize the emotion of "confusion" from the user's inquiry.

[0390] Analysis of inquiry content

[0391] The server uses the spaCy library to extract keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[0392] Answer generation

[0393] The server uses GPT-3 to generate the response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0394] Providing an answer

[0395] The device displays the generated answer and related links on the user's screen.

[0396] Gathering feedback and improving the AI ​​model

[0397] The user enters feedback such as "The answer was helpful" into the React interface, and the device sends this to the server. The server analyzes the feedback data and uses it as training data for the AI ​​model.

[0398] Example of a prompt

[0399] "I'd like to know how to apply for a new project. Please provide a link to the application form on the portal site."

[0400] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0401] Step 1:

[0402] Data collection

[0403] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals) from all employees. Specifically, it uses Microsoft Exchange Server and the Slack API to retrieve data. Input data includes emails and chat messages, while output data includes all communication records and related documents. The server stores this data in a database for later analysis.

[0404] Step 2:

[0405] Data analysis and index creation

[0406] The server analyzes the collected data using a natural language processing (NLP) engine. Specifically, it uses Google NLP or IBM Watson to extract key keywords and topics. The input data consists of collected emails, chat logs, and related documents, while the output data is an index based on keywords and topics. The server creates this index to enable efficient data retrieval.

[0407] Step 3:

[0408] AI model training

[0409] The server trains a machine learning model using the analyzed data. Specifically, it uses the Hugging Face Transformers library to create the model. The input data consists of the analyzed data and extracted indices, while the output data is the trained machine learning model. The server stores this model so that it can generate the best possible answers to queries.

[0410] Step 4:

[0411] Real-time inquiry reception

[0412] The terminal receives user inquiries in real time. Users enter their inquiries through the terminal's web interface (built with React or Vue.js). The input data is the inquiry text entered by the user, and the output data is the transmission of that inquiry text to the server.

[0413] Step 5:

[0414] Emotion recognition by an emotion engine

[0415] The server recognizes emotions from the content of the received inquiry. Specifically, it uses Affectiva or IBM Watson Tone Analyzer. The input data is the user's inquiry, and the output data is the user's emotional state (e.g., confused, angry, happy). The server uses this emotional information for subsequent processing.

[0416] Step 6:

[0417] Analysis of inquiry content

[0418] The server re-analyzes the query content, taking sentiment information into account. Specifically, it uses the spaCy library to extract key keywords and contextual information. The input data consists of the user's query and sentiment information, while the output data consists of the analyzed keywords and contextual information.

[0419] Step 7:

[0420] Answer generation

[0421] The server generates appropriate answers based on the analysis results. Specifically, it uses models such as OpenAI's GPT-3. The input data consists of analyzed keywords and contextual information, and the output data is the generated answer. For example, an answer such as "To apply for a new project, please use the application form on the portal site" might be generated.

[0422] Step 8:

[0423] Providing an answer

[0424] The terminal displays the response sent from the server to the user. The input data is the generated response sent from the server, and the output data is the response displayed on the user's terminal. The user can view the response and related links on the terminal screen.

[0425] Step 9:

[0426] Gathering feedback and improving the AI ​​model

[0427] The device collects user feedback and sends it to the server. The input data is the user's feedback, and the output data is the transmission of that feedback to the server. The server analyzes the feedback data and adds it as training data for a machine learning model. This improves the model's accuracy and the quality of its responses.

[0428] (Application Example 2)

[0429] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0430] Traditional customer support systems often fail to provide adequate support because they generate answers without considering the user's emotions. This is particularly problematic for inquiries involving feelings of confusion or frustration, which can lead to decreased user satisfaction. Furthermore, there is a lack of mechanisms to effectively utilize user feedback to improve AI models, making it difficult to enhance system quality. Therefore, there is a need for an intelligent customer support system that provides high-quality answers and improves user satisfaction.

[0431] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0432] In this invention, the server includes means for collecting communication records and related documents from the entire organization, means for analyzing the collected data and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users, means for analyzing the content of the received inquiries, means for recognizing the emotional state of the user using an emotion recognition engine, means for generating an appropriate response based on the analysis results and emotional state, means for presenting the generated response to the user, and means for collecting feedback from the user and improving the artificial intelligence model. This enables the generation of high-quality responses that take into account the user's emotions in real time, improving the quality of inquiry handling and increasing user satisfaction.

[0433] "Communication records of the entire organization" refers to the history of digital communications, such as emails and chat logs, conducted within an organization such as a company or group.

[0434] "Related documents" refer to documents within an organization that are related to business operations, such as guidelines, manuals, memos, and reports.

[0435] "Means of collection" refers to systems that collect necessary information using databases, cloud storage, log management systems, etc.

[0436] "Means of analysis" refer to methods that use natural language processing techniques and machine learning algorithms to convert collected data into a format that can be understood and used.

[0437] "Creating an index" refers to a method of organizing data by adding keywords and metadata so that it can be quickly searched and referenced.

[0438] An "artificial intelligence model" is a computer program that uses machine learning and deep learning algorithms to learn knowledge from data and has the ability to perform specific tasks.

[0439] "Means of receiving inquiries from users" refers to interfaces or applications that receive questions and requests that users make to the system.

[0440] "Means for analyzing received inquiries" refers to methods for interpreting inquiries received from users and extracting information necessary to generate appropriate responses.

[0441] An "emotion recognition engine" is an artificial intelligence technology that identifies a user's emotional state from data such as text and audio.

[0442] "Emotional state" refers to the psychological state a user is in when making an inquiry, and includes, for example, confusion, frustration, or joy.

[0443] "Means for generating appropriate answers" refers to methods for creating specific and helpful answers to user inquiries based on collected and analyzed data.

[0444] "Means of presenting generated answers to users" refers to interfaces or applications that display system-generated answers in a user-friendly format.

[0445] "Means of collecting feedback" refers to methods of receiving opinions and evaluations provided by users and processing them within the system.

[0446] "Methods for improving artificial intelligence models" refer to methods of improving the accuracy and performance of existing artificial intelligence models using collected feedback data.

[0447] This invention provides an intelligent inquiry response system that collects communication records and related documents from the entire organization, trains an artificial intelligence model based on them, and provides appropriate answers to user inquiries. In particular, this system can improve user satisfaction by recognizing the user's emotional state using an emotion recognition engine and generating answers based on that state.

[0448] System Configuration

[0449] 1. Data Collection

[0450] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from the entire organization via cloud storage (e.g., Amazon S3). This data is used for analysis and indexing, as described later.

[0451] 2. Data analysis and index creation

[0452] The server uses a natural language processing engine (e.g., Amazon Comprehend) to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[0453] 3. Training the AI ​​model

[0454] The server trains an artificial intelligence model (e.g., TensorFlow, PyTorch) based on the analyzed data. This model has the ability to generate the best possible answers to user inquiries.

[0455] 4. Real-time inquiry reception

[0456] The device (e.g., a smartphone application) receives inquiries from users in real time. Users input and submit questions in natural language using the device's interface.

[0457] 5. Emotion recognition

[0458] The server analyzes the received inquiry using an emotion recognition engine (e.g., Amazon Rekognition, IBM Watson Tone Analyzer) to recognize the user's emotional state. This emotional state is a crucial factor in generating the response.

[0459] 6. Answer generation

[0460] The server generates appropriate responses based on the analysis results and recognized emotional states. Highly accurate responses are provided by using a generative AI model (e.g., OpenAI's GPT-4®). These responses may include links to relevant internal documents and instructions.

[0461] 7. Answer presentation

[0462] The device displays the generated response to the user. The user can check the response on the device screen.

[0463] 8. Gathering feedback and improving the AI ​​model

[0464] The device collects feedback from users and sends the feedback data to a server. The server uses this feedback to retrain the artificial intelligence model and improve the overall accuracy of the system.

[0465] Specific example

[0466] For example, consider a scenario where a user inquires, "My product hasn't arrived yet." This system can recognize the user's feelings of confusion or frustration and generate a response that includes a gentler tone and specific solutions. An example of an input prompt for the AI ​​model in this case is shown below.

[0467] Example of a prompt:

[0468] User inquiry: "My item hasn't arrived yet."

[0469] Emotional state: "Confused", "Irritated"

[0470] In this way, the present invention makes it possible to generate high-quality responses that take into account the user's emotions in real time, thereby improving the quality of inquiry handling and increasing user satisfaction.

[0471] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0472] Step 1:

[0473] The server collects communication records and related documents from the entire organization's cloud storage. The input to this process is the location where the communication records and related documents are stored (e.g., the cloud storage bucket ID), and the output is the collected data set. This data primarily includes emails, chat logs, guidelines, manuals, etc. The server stores this data in a database for use in the next analysis step.

[0474] Step 2:

[0475] The server analyzes the collected data using a natural language processing engine and creates an index. The input to this process is the collected dataset, and the output is the analyzed data and its index information. The server performs keyword extraction and contextual analysis to generate an index that enables rapid data retrieval.

[0476] Step 3:

[0477] The server trains an artificial intelligence model based on the analysis results. The input to this process is the analyzed data and index information, and the output is the trained AI model. Machine learning libraries (e.g., TensorFlow, PyTorch) are used to train the model. The server tunes the model to optimal performance, improving its ability to generate answers to queries.

[0478] Step 4:

[0479] The terminal receives inquiries from users in real time. The input in this process is a natural language question from the user, and the output is a notification that the inquiry has been received. The terminal sends the text entered by the user to the server for preparation for analysis.

[0480] Step 5:

[0481] The server analyzes the received inquiry using an emotion recognition engine to recognize the user's emotional state. The input for this process is the user's inquiry text, and the output is the recognized emotional state (e.g., confused, irritated). The server utilizes the emotion recognition engine to analyze the emotional tone.

[0482] Step 6:

[0483] The server generates an appropriate response based on the inquiry content and the recognized emotional state. The input to this process is the inquiry content and emotional state, and the output is the generated response. The server uses a generative AI model (e.g., OpenAI's GPT-4) to create specific and helpful responses.

[0484] Step 7:

[0485] The terminal presents the generated response to the user. The input in this process is the generated response sent from the server, and the output is the response displayed in a format visible to the user. The terminal provides the response to the user through a GUI (Graphical User Interface).

[0486] Step 8:

[0487] The terminal collects feedback from users and sends it to the server. The input in this process is the feedback provided by the user, and the output is the recording of the feedback and its transfer to the server. The server uses this feedback as retraining data for machine learning to continuously improve the artificial intelligence model.

[0488] Thus, the system of the present invention can provide highly accurate answers to user inquiries in real time using emotion recognition and an AI model.

[0489] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0490] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0491] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0492] [Second Embodiment]

[0493] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0494] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0495] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0496] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0497] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0498] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0499] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0500] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0501] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0502] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0503] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0504] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0505] This invention is a system designed to solve a business problem where companies face the issue of unclear internal system application procedures, rules, guidelines, and manuals, which require significant time to research. This system utilizes AI to provide intelligent inquiry support available 24 / 7, aiming to improve operational efficiency and optimize resources.

[0506] System Overview

[0507] This system operates using the following procedure:

[0508] 1. Data Collection

[0509] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[0510] 2. Data analysis and index creation

[0511] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. This allows for the rapid extraction of appropriate information in response to queries.

[0512] 3. Training the AI ​​model

[0513] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[0514] 4. Real-time inquiry reception

[0515] The device (PC or smartphone) receives user inquiries in real time. Users input their questions in natural language.

[0516] 5. Analysis of the inquiry content

[0517] The server analyzes the received inquiry and extracts relevant keywords and contextual information.

[0518] 6. Generating the answer

[0519] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[0520] 7. Providing the answer

[0521] The device presents the generated answer to the user.

[0522] 8. Gathering feedback and improving the AI ​​model

[0523] The device collects user feedback and sends it to the server. The server uses this feedback to improve the accuracy of the AI ​​model.

[0524] Specific example

[0525] Example 1: Data Collection

[0526] The server collects emails and chat logs sent and received by employee A over the past six months.

[0527] Example 2: Data analysis and index creation

[0528] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[0529] Example 3: Training an AI model

[0530] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[0531] Example 4: Real-time inquiry reception

[0532] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[0533] Example 5: Analysis of inquiry content

[0534] The server extracts and analyzes the keywords "new project" and "application method".

[0535] Example 6: Generating an answer

[0536] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0537] Example 7: Providing an answer

[0538] The generated response is displayed on User B's terminal screen.

[0539] Example 8: Gathering feedback and improving the AI ​​model

[0540] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[0541] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[0542] The following describes the processing flow.

[0543] Step 1: Data Collection

[0544] The server connects to all employees' email and chat servers and downloads communication records for the past six months. It also accesses the document management system to retrieve relevant documents such as company guidelines, manuals, and application procedures.

[0545] Step 2: Data Analysis and Indexing

[0546] The server uses a natural language processing (NLP) engine to analyze the collected data. Specifically, it extracts keywords from each document and message and understands their contextual information. Then, based on the analysis results, it creates an index and stores it in a database.

[0547] Step 3: Training the AI ​​model

[0548] The server uses the analyzed data to train an AI model using a supervised learning algorithm. The model is designed to generate the best possible answers to user inquiries. The training process involves numerous question-and-answer pairs.

[0549] Step 4: Real-time inquiry reception

[0550] The terminal (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the terminal's interface.

[0551] Step 5: Analyzing the inquiry content

[0552] The server receives inquiries from users and uses a natural language processing engine to analyze the question content. The analysis extracts important keywords and contextual information from the question and compares them against an index database.

[0553] Step 6: Generating the answer

[0554] The server generates an appropriate response based on the query analysis results. If necessary, it also retrieves links to relevant internal documents and manuals from the index database and includes them in the response.

[0555] Step 7: Present your answer

[0556] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[0557] Step 8: Gathering feedback and improving the AI ​​model

[0558] The device receives feedback provided by the user and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[0559] (Example 1)

[0560] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0561] Traditional corporate systems often lacked clear internal rules, guidelines, and manuals, resulting in employees spending a significant amount of time searching for this information. This led to decreased employee efficiency and a substantial decline in productivity.

[0562] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0563] In this invention, the server includes means for collecting communication records and business-related information, means for analyzing the collected data using natural language processing technology and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users in real time, means for analyzing the content of the received inquiries using natural language processing technology, means for generating appropriate answers based on the analysis results, means for presenting the generated answers to users, and means for collecting feedback from users and improving the artificial intelligence model. As a result, users can obtain information quickly and appropriately, enabling increased efficiency in internal operations and optimization of resources.

[0564] "Communication records" refer to records of communication between employees within a company, such as emails and chat logs.

[0565] "Business-related information" refers to documentary information such as guidelines, manuals, and procedural documents necessary for carrying out business operations within a company.

[0566] "Natural language processing technology" refers to technologies that enable computers to recognize, interpret, and generate human language, and specifically includes text analysis and contextual understanding.

[0567] An "artificial intelligence model" is an algorithm that learns from large amounts of data, extracts patterns and knowledge, and automatically generates the optimal answer to an inquiry.

[0568] An "index" refers to a systematic organization of keywords and contextual information used to quickly retrieve specific information from a large amount of data.

[0569] "Users" refers to company employees who use this system to make internal inquiries.

[0570] "Real-time" refers to a timeframe in which processing and responses occur almost instantly, meaning that a response is provided immediately after a user submits a question.

[0571] "Feedback" refers to information collected from users who provide evaluations and opinions on the answers they receive, in order to help improve the system.

[0572] "Analysis results" refer to data containing important keywords and contextual information obtained after analyzing the content of a query using natural language processing technology.

[0573] "Answer" refers to the information and instructions provided in response to a user's inquiry, generated based on the analysis results.

[0574] A "link" is a means of reference, such as a URL, that contains direct reference information to internal documents or manuals related to the generated response.

[0575] The system according to this invention collects communication records and business-related information, analyzes and utilizes it to provide prompt and appropriate responses to user inquiries. This system significantly improves operational efficiency within a company and optimizes resources.

[0576] Hardware and software configuration

[0577] server

[0578] 1. Data collection:

[0579] The server collects communication records such as employee emails and chat logs, as well as business-related information such as guidelines and manuals. This utilizes the Microsoft Exchange Server API and the Slack API.

[0580] 2. Data Analysis:

[0581] The server analyzes the collected data using natural language processing (NLP) techniques. During this process, it creates indexes using Apache Lucene or Elasticsearch.

[0582] 3. AI model training:

[0583] Based on the analyzed and indexed data, AI models are trained using machine learning frameworks such as TensorFlow and PyTorch.

[0584] 4. Analysis of the inquiry content:

[0585] The server analyzes user inquiries using advanced NLP models such as BERT and GPT-3.

[0586] 5. Generating the answer:

[0587] Based on the analysis results, generate appropriate answers. Include links to relevant internal documents and manuals as needed.

[0588] 6. Gathering feedback and improving the AI ​​model:

[0589] We collect user feedback and use it to continuously improve our AI models.

[0590] terminal

[0591] 1. Real-time inquiry reception:

[0592] The terminal receives user inquiries in real time via PCs and smartphones. Users input their questions in natural language.

[0593] 2. Providing the answer:

[0594] The device displays the responses sent from the server to the user in real time.

[0595] user

[0596] 1. Enter your inquiry:

[0597] Users input questions in natural language via their devices.

[0598] 2. Providing feedback:

[0599] Feedback on the provided answers is entered into the terminal, and that feedback is sent to the server.

[0600] Specific example

[0601] Example 1: Data Collection

[0602] The server collects emails and chat logs sent and received by employee A over the past six months.

[0603] Example 2: Data Analysis

[0604] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[0605] Example 3: Training an AI model

[0606] The server trains an AI model based on the collected data, learning appropriate response patterns based on specific keywords and contexts.

[0607] Example 4: Real-time inquiry reception

[0608] User B uses their device to type "Please tell me how to apply for a new project" and sends it.

[0609] Example 5: Analysis of inquiry content

[0610] The server extracts and analyzes the keywords "new project" and "application method".

[0611] Example 6: Generating an answer

[0612] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0613] Example 7: Providing an answer

[0614] The generated response is displayed on User B's terminal screen.

[0615] Example 8: Gathering feedback and improving the AI ​​model

[0616] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it to improve the AI ​​model.

[0617] Example of a prompt

[0618] "Please tell me how to apply for a new project."

[0619] "What is the procedure for applying for leave?"

[0620] "Could you show me the minutes of the recent meeting?"

[0621] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[0622] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0623] Step 1:

[0624] Data collection

[0625] input:

[0626] All employee communication records (emails, chat logs), and business-related information (guidelines, manuals)

[0627] Specific actions:

[0628] The server uses the Microsoft Exchange Server API and the Slack API to collect employee emails and chat logs. It also retrieves business-related documents (guidelines, manuals) from the company's file server.

[0629] output:

[0630] Collected communication records and work-related information of all employees

[0631] Specific example:

[0632] The server collects employee A's emails and chat logs from the past six months, and simultaneously retrieves guidelines and manuals.

[0633] Step 2:

[0634] Data analysis and index creation

[0635] input:

[0636] Collected communication records and work-related information of all employees

[0637] Specific actions:

[0638] The server analyzes the collected data using Apache Lucene and Elasticsearch and creates an index. This involves using natural language processing techniques to extract and classify keywords.

[0639] output:

[0640] Indexed dataset

[0641] Specific example:

[0642] The server analyzes the data for keywords such as "how to apply for a new project" and "guidelines," and creates an index.

[0643] Step 3:

[0644] AI model training

[0645] input:

[0646] Indexed dataset

[0647] Specific actions:

[0648] The server uses TensorFlow or PyTorch to train an AI model based on indexed data. This process uses a large amount of text data to learn question and associated answer patterns.

[0649] output:

[0650] Trained AI model

[0651] Specific example:

[0652] The server trains an AI model based on indexed data, learning appropriate response patterns regarding "how to submit a new project application."

[0653] Step 4:

[0654] Real-time inquiry reception

[0655] input:

[0656] User inquiries (questions entered in natural language)

[0657] Specific actions:

[0658] The user sends a question to the system using a terminal. The terminal sends this inquiry to the server in real time.

[0659] output:

[0660] Inquiry content sent to the server

[0661] Specific example:

[0662] User B uses their device and types and sends the message, "Please tell me how to apply for a new project."

[0663] Step 5:

[0664] Analysis of inquiry content

[0665] input:

[0666] Inquiry content sent from the device

[0667] Specific actions:

[0668] The server analyzes queries using natural language processing techniques such as BERT and GPT-3, and extracts key keywords and contextual information.

[0669] output:

[0670] Analyzed query content (main keywords and contextual information)

[0671] Specific example:

[0672] The server extracts and analyzes the keywords "new project" and "application method" from user B's inquiry.

[0673] Step 6:

[0674] Answer generation

[0675] input:

[0676] Analyzed inquiry content

[0677] Specific actions:

[0678] The server uses a trained AI model to generate appropriate answers based on the analyzed query. Links to relevant internal documents and manuals are included as needed.

[0679] output:

[0680] Generated answer

[0681] Specific example:

[0682] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0683] Step 7:

[0684] Providing an answer

[0685] input:

[0686] Generated answer

[0687] Specific actions:

[0688] The device displays the responses sent from the server to the user in real time.

[0689] output:

[0690] User-displayed response

[0691] Specific example:

[0692] The generated response is displayed on User B's device.

[0693] Step 8:

[0694] Gathering feedback and improving the AI ​​model

[0695] input:

[0696] User Feedback

[0697] Specific actions:

[0698] The user enters feedback on the answer into the device. The device sends this feedback to the server, which adds it to the AI ​​model's training data to make improvements.

[0699] output:

[0700] Improved AI model

[0701] Specific example:

[0702] User B sends feedback saying "The answer was helpful," and the server uses that feedback to improve the accuracy of the AI ​​model.

[0703] (Application Example 1)

[0704] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0705] There is a problem in factories where information regarding business processes, operating procedures, and troubleshooting is ambiguous, making it difficult for employees to quickly obtain this information. This often leads to decreased operational efficiency and delays in troubleshooting. This invention aims to solve these problems and improve operational efficiency within factories by providing an AI-powered system.

[0706] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0707] In this invention, the server includes means for collecting communication records and related documents of all employees, means for analyzing the collected data and creating an index, and means for training an AI model based on the analyzed data. This makes it possible to respond quickly and accurately to user inquiries. Furthermore, by adding means for factory workers to make inquiries using tablet devices or robots, and means for supporting questions regarding factory operating procedures and troubleshooting, it becomes possible to speed up and streamline information acquisition within the factory.

[0708] "Communication records" refer to records of digital communications, such as emails and chat logs, conducted by all employees within the company.

[0709] "Related documents" refer to documents containing information such as guidelines, manuals, and operational procedures used within a company.

[0710] An "index" is a list or database structure that organizes analyzed data into a searchable format.

[0711] An "AI model" is an artificial intelligence model generated using machine learning algorithms based on collected and analyzed data.

[0712] An "inquiry" refers to a question or request that a user enters into a system.

[0713] "Answer" refers to the information generated by the AI ​​model and provided in response to a user's inquiry.

[0714] "Feedback" refers to evaluations and opinions given to responses provided by users.

[0715] A "factory worker" refers to an employee who performs tasks such as production, management, and maintenance within a factory.

[0716] A "tablet device" is a portable computer device equipped with a touchscreen.

[0717] A "robot" is a machine that operates automatically and performs tasks within a factory.

[0718] "Operating procedures" refer to step-by-step instructions or methods for performing a specific task or operation.

[0719] "Troubleshooting" refers to the procedures and methods used to resolve problems that occur in a system or machine.

[0720] "To support" means to provide the necessary information and means to perform a specific task or operation.

[0721] This invention is a system for efficiently acquiring information on operating procedures and troubleshooting within a factory. The system functions by collecting communication records and related documents from all employees, analyzing this data to create an index, and training an AI model. It also includes a process of receiving user inquiries in real time, analyzing them, generating and presenting answers, and collecting user feedback to improve the AI ​​model.

[0722] The server first collects communication records and related documents, such as emails and chat logs, from all employees. Next, it analyzes the collected data using natural language processing technology and creates an index in a searchable format. The analyzed data is used to train an AI model, which is built to generate appropriate responses.

[0723] Users (factory workers) can make inquiries using tablet devices or robots. For example, if a worker asks "How do I configure machine A?" or "How do I troubleshoot error code E123?", the device sends this inquiry to the server in real time. The server analyzes the received inquiry, extracts relevant keywords and contextual information, generates an appropriate answer based on the analysis results, and presents the answer to the device.

[0724] The generated responses may include links to relevant internal documents and manuals. This allows workers to quickly obtain the necessary information and improve work efficiency. User feedback is also important; by collecting feedback on whether the provided responses were helpful, the server can continuously improve the AI ​​model.

[0725] This system is implemented using hardware such as NVIDIA GPUs, cloud instances from AWS and Azure, and software such as Hugging Face's Transformers library. This enables high-performance data processing and analysis.

[0726] Specific example

[0727] For example, a factory worker might type "How do I troubleshoot error code E123 on machine B?" into a tablet. In this case, the system would generate the following response: "Error code E123 requires you to first turn off the power and restart the machine. Detailed troubleshooting information can be found at the following link."

[0728] Example of a prompt

[0729] "Please tell me how to troubleshoot error code E123 on machine B."

[0730] In this way, this system can streamline business processes within the factory and improve worker productivity.

[0731] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0732] Step 1:

[0733] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from all employees. The input is the communication records and related documents of all employees, and the output is a collection of the collected data. This data is used for subsequent analysis and indexing.

[0734] Step 2:

[0735] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. The input is the data collected in step 1, and the output is the analyzed data and its index. In this process, the meaning of the text is understood and converted into a searchable format.

[0736] Step 3:

[0737] The server trains an AI model based on the analyzed data. The input is the indexed data obtained in step 2, and the output is the trained AI model. This gives the AI ​​the ability to generate appropriate answers to queries.

[0738] Step 4:

[0739] The user enters the inquiry in natural language using a tablet or robot. The input is the user's inquiry statement, and the output is the inquiry request. This request is sent to the server in real time.

[0740] Step 5:

[0741] The server analyzes the received inquiry. The input is the inquiry request sent in step 4, and the output is the analyzed inquiry content. Keywords and contextual information are extracted and used as basic data to generate appropriate responses.

[0742] Step 6:

[0743] The server generates an appropriate response based on the analysis results. The input is the query analyzed in step 5 and the trained AI model, and the output is the generated response. Links to relevant documents and manuals are included as needed.

[0744] Step 7:

[0745] The terminal presents the generated answer to the user. The input is the answer generated in step 6, and the output is the answer displayed to the user. The user can then use this answer to solve the problem.

[0746] Step 8:

[0747] The user provides feedback on the provided answers. The input is the user's feedback, and the output is a feedback request. This request is sent to the server and used to improve the AI ​​model.

[0748] Step 9:

[0749] The server improves the AI ​​model based on user feedback. The input is the feedback request submitted in step 8, and the output is the improved AI model. This improves the accuracy of answers to future inquiries.

[0750] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0751] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[0752] System Overview

[0753] This system operates using the following procedure:

[0754] 1. Data Collection

[0755] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[0756] 2. Data analysis and index creation

[0757] The server uses an NLP engine to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[0758] 3. Training the AI ​​model

[0759] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[0760] 4. Real-time inquiry reception

[0761] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the device's interface.

[0762] 5. Emotion recognition by an emotion engine

[0763] The server uses an emotion engine to recognize emotions from the user's inquiry. This allows it to understand the user's emotional state and use the data to generate more appropriate responses.

[0764] 6. Analysis of the inquiry content

[0765] The server re-analyzes the query content, taking sentiment information into account, and then extracts keywords and contextual information.

[0766] 7. Generating the answer

[0767] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[0768] 8. Providing the answer

[0769] The device displays the generated response to the user. The user can view the generated response and associated links on the device screen.

[0770] 9. Gathering feedback and improving the AI ​​model

[0771] The device collects user feedback and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[0772] Specific example

[0773] Example 1: Data Collection

[0774] The server collects emails and chat logs sent and received by employee A over the past six months.

[0775] Example 2: Data analysis and index creation

[0776] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[0777] Example 3: Training an AI model

[0778] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[0779] Example 4: Real-time inquiry reception

[0780] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[0781] Example 5: Emotion recognition by an emotion engine

[0782] The server uses an emotion engine to recognize that the user is confused, based on the user's inquiry, "How do I apply for a new project?"

[0783] Example 6: Analysis of inquiry content

[0784] The server extracts keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[0785] Example 7: Generating an answer

[0786] The server generates a response stating, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below," and provides it in an easily understandable format to alleviate user confusion.

[0787] Example 8: Providing an answer

[0788] The generated response is displayed on User B's terminal screen.

[0789] Example 9: Gathering feedback and improving the AI ​​model

[0790] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[0791] In this way, the system analyzes the content of inquiries and the user's emotions to generate more appropriate and effective responses. This leads to increased operational efficiency and improved employee productivity.

[0792] The following describes the processing flow.

[0793] Step 1: Data Collection

[0794] The server connects to all employees' mail and chat servers to download emails and chat logs from the past six months. It also accesses the document management system to collect relevant documents such as company guidelines, manuals, and application procedures.

[0795] Step 2: Data Analysis and Indexing

[0796] The server analyzes the collected data using an NLP engine. Specifically, it extracts important keywords and contextual information from each document and message. Then, it creates an index based on the analysis results and stores it in a database.

[0797] Step 3: Training the AI ​​model

[0798] The server trains an AI model using a supervised learning algorithm based on the analyzed data. The training data includes pairs of past queries and their appropriate answers. This gives the model the ability to generate optimal answers to a variety of queries.

[0799] Step 4: Real-time inquiry reception

[0800] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions using natural language.

[0801] Step 5: Emotion recognition by the emotion engine

[0802] The server receives the user's inquiry and uses an emotion engine to analyze the user's emotions. The emotion engine uses NLP techniques to identify emotions from the inquiry text.

[0803] Step 6: Analyzing the inquiry content

[0804] The server takes sentiment information into account and re-analyzes the query using an NLP engine. This extracts keywords and contextual information, and adds sentiment information.

[0805] Step 7: Generating the answer

[0806] Based on the analysis results, the server generates the most appropriate response for the user's emotional state and inquiry. Links to relevant internal documents and manuals are also added as needed.

[0807] Step 8: Present your answer

[0808] The device displays the generated answers to the user. The user then views the provided answers and associated links on the device screen.

[0809] Step 9: Gathering Feedback and Improving the AI ​​Model

[0810] The device collects user feedback and sends it to the server. The server analyzes this feedback and adds it to the AI ​​model as new training data. This improves the accuracy of the AI ​​model.

[0811] In this way, the system comprehensively analyzes the user's inquiry and emotions, enabling a process that generates the optimal response. This leads to increased operational efficiency and improved employee productivity.

[0812] (Example 2)

[0813] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0814] Traditional inquiry handling systems struggled to provide 24 / 7 support and often suffered from low-quality responses due to their failure to consider the user's emotional state. Furthermore, they were inefficient in handling inquiries because they couldn't effectively utilize all employee communication records and related documents. A system is needed to solve these problems and improve the efficiency of internal operations.

[0815] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0816] In this invention, the server includes means for collecting communication records and related documents of all employees; means for analyzing the collected data using a natural language processing engine and creating an index; means for training a machine learning model based on the analyzed data; means for receiving user inquiries in real time; means for recognizing the user's emotions from the received inquiry content using an emotion engine; means for analyzing the inquiry content taking emotional information into account; means for generating appropriate answers based on the analysis results; means for presenting the generated answers to the user; and means for collecting user feedback and improving the machine learning model. This enables 24 / 7 inquiry support and the provision of high-quality answers that take user emotions into consideration.

[0817] "All employee communication records" refers to digital communication data such as emails and chat logs used by company employees for work purposes.

[0818] "Related documents" refer to documents related to business operations, such as guidelines and manuals, that the company possesses.

[0819] A "natural language processing engine" is software that analyzes text data to extract keywords and perform sentiment analysis.

[0820] A "machine learning model" is a collection of algorithms that learn from data and perform pattern recognition and prediction.

[0821] A "means of receiving in real time" refers to a system or interface that can immediately receive inquiries from users.

[0822] An "emotion engine" is a tool or software that recognizes a user's emotional state from text data.

[0823] "Index creation" refers to techniques that generate lists of keywords and topics to enable rapid searching of large amounts of data.

[0824] "Means for analyzing inquiry content" refers to technologies for understanding the questions and requests entered by users and extracting the necessary information.

[0825] "Means for generating appropriate answers" refers to the process of creating the optimal response to a user's inquiry based on analyzed data.

[0826] "Means of presenting to the user" refers to a system for displaying generated answers and information to the user.

[0827] "Means of collecting feedback" refers to a system for gathering evaluations and opinions from users and using them to improve the system.

[0828] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[0829] System Overview

[0830] Data collection

[0831] The server collects communication records (emails, chat logs) of all employees using Microsoft Exchange Server and the Slack API. It also collects relevant documents such as guidelines and manuals from the company's file server and document management system.

[0832] Data analysis and index creation

[0833] The server uses natural language processing engines such as Google NLP and IBM Watson to analyze the collected data and extract key keywords and topics. This creates an efficient data index, enabling rapid information retrieval.

[0834] AI model training

[0835] The server uses libraries such as Hugging Face's Transformers library to train a machine learning model based on the analyzed data. During the training process, the data is split into training and validation sets, and the model parameters are adjusted. The trained AI model is then saved and prepared to generate the optimal response to queries.

[0836] Real-time inquiry reception

[0837] The device (PC or smartphone) receives user inquiries in real time through a web interface built with React or Vue.js. Users input and submit questions in natural language.

[0838] Emotion recognition by an emotion engine

[0839] The server uses emotion engines such as Affectiva or IBM Watson Tone Analyzer to recognize emotions from user inquiries. This data is then used to understand the user's emotional state and generate appropriate responses.

[0840] Analysis of inquiry content

[0841] The server re-analyzes the query content using the spaCy library, taking sentiment information into account. It extracts key keywords and contextual information and performs a detailed analysis based on the user's emotional state.

[0842] Answer generation

[0843] The server uses models such as OpenAI's GPT-3 to generate appropriate responses based on the analysis results. Links to relevant documents and manuals are included as needed. For example, it might generate a response like, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0844] Providing an answer

[0845] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[0846] Gathering feedback and improving the AI ​​model

[0847] The device collects user feedback and sends it to the server. The server analyzes the feedback, adds it as training data for the model, and improves the accuracy of the AI ​​model by retraining it.

[0848] Specific example

[0849] Data collection

[0850] The server uses the Microsoft Exchange Server API to retrieve employee A's email data and the Slack API to download employee A's chat logs.

[0851] Data analysis and index creation

[0852] The server uses the Google NLP API to analyze keywords such as "how to apply for a new project" and "project guidelines" and creates an index.

[0853] AI model training

[0854] The server uses the Hugging Face Transformers library to train an AI model based on the analyzed data, splitting it into training data and validation data for further training.

[0855] Real-time inquiry reception

[0856] The user enters "How do I submit a new project?" into a web interface built with React and submits it.

[0857] Emotion recognition by an emotion engine

[0858] The server uses IBM Watson Tone Analyzer to recognize the emotion of "confusion" from the user's inquiry.

[0859] Analysis of inquiry content

[0860] The server uses the spaCy library to extract keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[0861] Answer generation

[0862] The server uses GPT-3 to generate the response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[0863] Providing an answer

[0864] The device displays the generated answer and related links on the user's screen.

[0865] Gathering feedback and improving the AI ​​model

[0866] The user enters feedback such as "The answer was helpful" into the React interface, and the device sends this to the server. The server analyzes the feedback data and uses it as training data for the AI ​​model.

[0867] Example of a prompt

[0868] "I'd like to know how to apply for a new project. Please provide a link to the application form on the portal site."

[0869] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0870] Step 1:

[0871] Data collection

[0872] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals) from all employees. Specifically, it uses Microsoft Exchange Server and the Slack API to retrieve data. Input data includes emails and chat messages, while output data includes all communication records and related documents. The server stores this data in a database for later analysis.

[0873] Step 2:

[0874] Data analysis and index creation

[0875] The server analyzes the collected data using a natural language processing (NLP) engine. Specifically, it uses Google NLP or IBM Watson to extract key keywords and topics. The input data consists of collected emails, chat logs, and related documents, while the output data is an index based on keywords and topics. The server creates this index to enable efficient data retrieval.

[0876] Step 3:

[0877] AI model training

[0878] The server trains a machine learning model using the analyzed data. Specifically, it uses the Hugging Face Transformers library to create the model. The input data consists of the analyzed data and extracted indices, while the output data is the trained machine learning model. The server stores this model so that it can generate the best possible answers to queries.

[0879] Step 4:

[0880] Real-time inquiry reception

[0881] The terminal receives user inquiries in real time. Users enter their inquiries through the terminal's web interface (built with React or Vue.js). The input data is the inquiry text entered by the user, and the output data is the transmission of that inquiry text to the server.

[0882] Step 5:

[0883] Emotion recognition by an emotion engine

[0884] The server recognizes emotions from the content of the received inquiry. Specifically, it uses Affectiva or IBM Watson Tone Analyzer. The input data is the user's inquiry, and the output data is the user's emotional state (e.g., confused, angry, happy). The server uses this emotional information for subsequent processing.

[0885] Step 6:

[0886] Analysis of inquiry content

[0887] The server re-analyzes the query content, taking sentiment information into account. Specifically, it uses the spaCy library to extract key keywords and contextual information. The input data consists of the user's query and sentiment information, while the output data consists of the analyzed keywords and contextual information.

[0888] Step 7:

[0889] Answer generation

[0890] The server generates appropriate answers based on the analysis results. Specifically, it uses models such as OpenAI's GPT-3. The input data consists of analyzed keywords and contextual information, and the output data is the generated answer. For example, an answer such as "To apply for a new project, please use the application form on the portal site" might be generated.

[0891] Step 8:

[0892] Providing an answer

[0893] The terminal displays the response sent from the server to the user. The input data is the generated response sent from the server, and the output data is the response displayed on the user's terminal. The user can view the response and related links on the terminal screen.

[0894] Step 9:

[0895] Gathering feedback and improving the AI ​​model

[0896] The device collects user feedback and sends it to the server. The input data is the user's feedback, and the output data is the transmission of that feedback to the server. The server analyzes the feedback data and adds it as training data for a machine learning model. This improves the model's accuracy and the quality of its responses.

[0897] (Application Example 2)

[0898] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0899] Traditional customer support systems often fail to provide adequate support because they generate answers without considering the user's emotions. This is particularly problematic for inquiries involving feelings of confusion or frustration, which can lead to decreased user satisfaction. Furthermore, there is a lack of mechanisms to effectively utilize user feedback to improve AI models, making it difficult to enhance system quality. Therefore, there is a need for an intelligent customer support system that provides high-quality answers and improves user satisfaction.

[0900] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0901] In this invention, the server includes means for collecting communication records and related documents from the entire organization, means for analyzing the collected data and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users, means for analyzing the content of the received inquiries, means for recognizing the emotional state of the user using an emotion recognition engine, means for generating an appropriate response based on the analysis results and emotional state, means for presenting the generated response to the user, and means for collecting feedback from the user and improving the artificial intelligence model. This enables the generation of high-quality responses that take into account the user's emotions in real time, improving the quality of inquiry handling and increasing user satisfaction.

[0902] "Communication records of the entire organization" refers to the history of digital communications, such as emails and chat logs, conducted within an organization such as a company or group.

[0903] "Related documents" refer to documents within an organization that are related to business operations, such as guidelines, manuals, memos, and reports.

[0904] "Means of collection" refers to systems that collect necessary information using databases, cloud storage, log management systems, etc.

[0905] "Means of analysis" refer to methods that use natural language processing techniques and machine learning algorithms to convert collected data into a format that can be understood and used.

[0906] "Creating an index" refers to a method of organizing data by adding keywords and metadata so that it can be quickly searched and referenced.

[0907] An "artificial intelligence model" is a computer program that uses machine learning and deep learning algorithms to learn knowledge from data and has the ability to perform specific tasks.

[0908] "Means of receiving inquiries from users" refers to interfaces or applications that receive questions and requests that users make to the system.

[0909] "Means for analyzing received inquiries" refers to methods for interpreting inquiries received from users and extracting information necessary to generate appropriate responses.

[0910] An "emotion recognition engine" is an artificial intelligence technology that identifies a user's emotional state from data such as text and audio.

[0911] "Emotional state" refers to the psychological state a user is in when making an inquiry, and includes, for example, confusion, frustration, or joy.

[0912] "Means for generating appropriate answers" refers to methods for creating specific and helpful answers to user inquiries based on collected and analyzed data.

[0913] "Means of presenting generated answers to users" refers to interfaces or applications that display system-generated answers in a user-friendly format.

[0914] "Means of collecting feedback" refers to methods of receiving opinions and evaluations provided by users and processing them within the system.

[0915] "Methods for improving artificial intelligence models" refer to methods of improving the accuracy and performance of existing artificial intelligence models using collected feedback data.

[0916] This invention provides an intelligent inquiry response system that collects communication records and related documents from the entire organization, trains an artificial intelligence model based on them, and provides appropriate answers to user inquiries. In particular, this system can improve user satisfaction by recognizing the user's emotional state using an emotion recognition engine and generating answers based on that state.

[0917] System Configuration

[0918] 1. Data Collection

[0919] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from the entire organization via cloud storage (e.g., Amazon S3). This data is used for analysis and indexing, as described later.

[0920] 2. Data analysis and index creation

[0921] The server uses a natural language processing engine (e.g., Amazon Comprehend) to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[0922] 3. Training the AI ​​model

[0923] The server trains an artificial intelligence model (e.g., TensorFlow, PyTorch) based on the analyzed data. This model has the ability to generate the best possible answers to user inquiries.

[0924] 4. Real-time inquiry reception

[0925] The device (e.g., a smartphone application) receives inquiries from users in real time. Users input and submit questions in natural language using the device's interface.

[0926] 5. Emotion recognition

[0927] The server analyzes the received inquiry using an emotion recognition engine (e.g., Amazon Rekognition, IBM Watson Tone Analyzer) to recognize the user's emotional state. This emotional state is a crucial factor in generating the response.

[0928] 6. Answer generation

[0929] The server generates appropriate responses based on the analysis results and recognized emotional states. By using a generative AI model (e.g., OpenAI's GPT-4), it provides highly accurate responses. These responses may include links to relevant internal documents and instructions.

[0930] 7. Answer presentation

[0931] The device displays the generated response to the user. The user can check the response on the device screen.

[0932] 8. Gathering feedback and improving the AI ​​model

[0933] The device collects feedback from users and sends the feedback data to a server. The server uses this feedback to retrain the artificial intelligence model and improve the overall accuracy of the system.

[0934] Specific example

[0935] For example, consider a scenario where a user inquires, "My product hasn't arrived yet." This system can recognize the user's feelings of confusion or frustration and generate a response that includes a gentler tone and specific solutions. An example of an input prompt for the AI ​​model in this case is shown below.

[0936] Example of a prompt:

[0937] User inquiry: "My item hasn't arrived yet."

[0938] Emotional state: "Confused", "Irritated"

[0939] In this way, the present invention makes it possible to generate high-quality responses that take into account the user's emotions in real time, thereby improving the quality of inquiry handling and increasing user satisfaction.

[0940] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0941] Step 1:

[0942] The server collects communication records and related documents from the entire organization's cloud storage. The input to this process is the location where the communication records and related documents are stored (e.g., the cloud storage bucket ID), and the output is the collected data set. This data primarily includes emails, chat logs, guidelines, manuals, etc. The server stores this data in a database for use in the next analysis step.

[0943] Step 2:

[0944] The server analyzes the collected data using a natural language processing engine and creates an index. The input to this process is the collected dataset, and the output is the analyzed data and its index information. The server performs keyword extraction and contextual analysis to generate an index that enables rapid data retrieval.

[0945] Step 3:

[0946] The server trains an artificial intelligence model based on the analysis results. The input to this process is the analyzed data and index information, and the output is the trained AI model. Machine learning libraries (e.g., TensorFlow, PyTorch) are used to train the model. The server tunes the model to optimal performance, improving its ability to generate answers to queries.

[0947] Step 4:

[0948] The terminal receives inquiries from users in real time. The input in this process is a natural language question from the user, and the output is a notification that the inquiry has been received. The terminal sends the text entered by the user to the server for preparation for analysis.

[0949] Step 5:

[0950] The server analyzes the received inquiry using an emotion recognition engine to recognize the user's emotional state. The input for this process is the user's inquiry text, and the output is the recognized emotional state (e.g., confused, irritated). The server utilizes the emotion recognition engine to analyze the emotional tone.

[0951] Step 6:

[0952] The server generates an appropriate response based on the inquiry content and the recognized emotional state. The input to this process is the inquiry content and emotional state, and the output is the generated response. The server uses a generative AI model (e.g., OpenAI's GPT-4) to create specific and helpful responses.

[0953] Step 7:

[0954] The terminal presents the generated response to the user. The input in this process is the generated response sent from the server, and the output is the response displayed in a format visible to the user. The terminal provides the response to the user through a GUI (Graphical User Interface).

[0955] Step 8:

[0956] The terminal collects feedback from users and sends it to the server. The input in this process is the feedback provided by the user, and the output is the recording of the feedback and its transfer to the server. The server uses this feedback as retraining data for machine learning to continuously improve the artificial intelligence model.

[0957] Thus, the system of the present invention can provide highly accurate answers to user inquiries in real time using emotion recognition and an AI model.

[0958] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0959] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0960] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0961] [Third Embodiment]

[0962] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0963] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0964] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0965] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0966] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0967] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0968] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0969] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0970] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0971] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0972] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0973] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0974] This invention is a system designed to solve a business problem where companies face the issue of unclear internal system application procedures, rules, guidelines, and manuals, which require significant time to research. This system utilizes AI to provide intelligent inquiry support available 24 / 7, aiming to improve operational efficiency and optimize resources.

[0975] System Overview

[0976] This system operates using the following procedure:

[0977] 1. Data Collection

[0978] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[0979] 2. Data analysis and index creation

[0980] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. This allows for the rapid extraction of appropriate information in response to queries.

[0981] 3. Training the AI ​​model

[0982] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[0983] 4. Real-time inquiry reception

[0984] The device (PC or smartphone) receives user inquiries in real time. Users input their questions in natural language.

[0985] 5. Analysis of the inquiry content

[0986] The server analyzes the received inquiry and extracts relevant keywords and contextual information.

[0987] 6. Generating the answer

[0988] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[0989] 7. Providing the answer

[0990] The device presents the generated answer to the user.

[0991] 8. Gathering feedback and improving the AI ​​model

[0992] The device collects user feedback and sends it to the server. The server uses this feedback to improve the accuracy of the AI ​​model.

[0993] Specific example

[0994] Example 1: Data Collection

[0995] The server collects emails and chat logs sent and received by employee A over the past six months.

[0996] Example 2: Data analysis and index creation

[0997] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[0998] Example 3: Training an AI model

[0999] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[1000] Example 4: Real-time inquiry reception

[1001] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[1002] Example 5: Analysis of inquiry content

[1003] The server extracts and analyzes the keywords "new project" and "application method".

[1004] Example 6: Generating an answer

[1005] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1006] Example 7: Providing an answer

[1007] The generated response is displayed on User B's terminal screen.

[1008] Example 8: Gathering feedback and improving the AI ​​model

[1009] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[1010] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[1011] The following describes the processing flow.

[1012] Step 1: Data Collection

[1013] The server connects to all employees' email and chat servers and downloads communication records for the past six months. It also accesses the document management system to retrieve relevant documents such as company guidelines, manuals, and application procedures.

[1014] Step 2: Data Analysis and Indexing

[1015] The server uses a natural language processing (NLP) engine to analyze the collected data. Specifically, it extracts keywords from each document and message and understands their contextual information. Then, based on the analysis results, it creates an index and stores it in a database.

[1016] Step 3: Training the AI ​​model

[1017] The server uses the analyzed data to train an AI model using a supervised learning algorithm. The model is designed to generate the best possible answers to user inquiries. The training process involves numerous question-and-answer pairs.

[1018] Step 4: Real-time inquiry reception

[1019] The terminal (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the terminal's interface.

[1020] Step 5: Analyzing the inquiry content

[1021] The server receives inquiries from users and uses a natural language processing engine to analyze the question content. The analysis extracts important keywords and contextual information from the question and compares them against an index database.

[1022] Step 6: Generating the answer

[1023] The server generates an appropriate response based on the query analysis results. If necessary, it also retrieves links to relevant internal documents and manuals from the index database and includes them in the response.

[1024] Step 7: Present your answer

[1025] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[1026] Step 8: Gathering feedback and improving the AI ​​model

[1027] The device receives feedback provided by the user and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[1028] (Example 1)

[1029] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1030] Traditional corporate systems often lacked clear internal rules, guidelines, and manuals, resulting in employees spending a significant amount of time searching for this information. This led to decreased employee efficiency and a substantial decline in productivity.

[1031] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1032] In this invention, the server includes means for collecting communication records and business-related information, means for analyzing the collected data using natural language processing technology and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users in real time, means for analyzing the content of the received inquiries using natural language processing technology, means for generating appropriate answers based on the analysis results, means for presenting the generated answers to users, and means for collecting feedback from users and improving the artificial intelligence model. As a result, users can obtain information quickly and appropriately, enabling increased efficiency in internal operations and optimization of resources.

[1033] "Communication records" refer to records of communication between employees within a company, such as emails and chat logs.

[1034] "Business-related information" refers to documentary information such as guidelines, manuals, and procedural documents necessary for carrying out business operations within a company.

[1035] "Natural language processing technology" refers to technologies that enable computers to recognize, interpret, and generate human language, and specifically includes text analysis and contextual understanding.

[1036] An "artificial intelligence model" is an algorithm that learns from large amounts of data, extracts patterns and knowledge, and automatically generates the optimal answer to an inquiry.

[1037] An "index" refers to a systematic organization of keywords and contextual information used to quickly retrieve specific information from a large amount of data.

[1038] "Users" refers to company employees who use this system to make internal inquiries.

[1039] "Real-time" refers to a timeframe in which processing and responses occur almost instantly, meaning that a response is provided immediately after a user submits a question.

[1040] "Feedback" refers to information collected from users who provide evaluations and opinions on the answers they receive, in order to help improve the system.

[1041] "Analysis results" refer to data containing important keywords and contextual information obtained after analyzing the content of a query using natural language processing technology.

[1042] "Answer" refers to the information and instructions provided in response to a user's inquiry, generated based on the analysis results.

[1043] A "link" is a means of reference, such as a URL, that contains direct reference information to internal documents or manuals related to the generated response.

[1044] The system according to this invention collects communication records and business-related information, analyzes and utilizes it to provide prompt and appropriate responses to user inquiries. This system significantly improves operational efficiency within a company and optimizes resources.

[1045] Hardware and software configuration

[1046] server

[1047] 1. Data collection:

[1048] The server collects communication records such as employee emails and chat logs, as well as business-related information such as guidelines and manuals. This utilizes the Microsoft Exchange Server API and the Slack API.

[1049] 2. Data Analysis:

[1050] The server analyzes the collected data using natural language processing (NLP) techniques. During this process, it creates indexes using Apache Lucene or Elasticsearch.

[1051] 3. AI model training:

[1052] Based on the analyzed and indexed data, AI models are trained using machine learning frameworks such as TensorFlow and PyTorch.

[1053] 4. Analysis of the inquiry content:

[1054] The server analyzes user inquiries using advanced NLP models such as BERT and GPT-3.

[1055] 5. Generating the answer:

[1056] Based on the analysis results, generate appropriate answers. Include links to relevant internal documents and manuals as needed.

[1057] 6. Gathering feedback and improving the AI ​​model:

[1058] We collect user feedback and use it to continuously improve our AI models.

[1059] terminal

[1060] 1. Real-time inquiry reception:

[1061] The terminal receives user inquiries in real time via PCs and smartphones. Users input their questions in natural language.

[1062] 2. Providing the answer:

[1063] The device displays the responses sent from the server to the user in real time.

[1064] user

[1065] 1. Enter your inquiry:

[1066] Users input questions in natural language via their devices.

[1067] 2. Providing feedback:

[1068] Feedback on the provided answers is entered into the terminal, and that feedback is sent to the server.

[1069] Specific example

[1070] Example 1: Data Collection

[1071] The server collects emails and chat logs sent and received by employee A over the past six months.

[1072] Example 2: Data Analysis

[1073] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[1074] Example 3: Training an AI model

[1075] The server trains an AI model based on the collected data, learning appropriate response patterns based on specific keywords and contexts.

[1076] Example 4: Real-time inquiry reception

[1077] User B uses their device to type "Please tell me how to apply for a new project" and sends it.

[1078] Example 5: Analysis of inquiry content

[1079] The server extracts and analyzes the keywords "new project" and "application method".

[1080] Example 6: Generating an answer

[1081] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1082] Example 7: Providing an answer

[1083] The generated response is displayed on User B's terminal screen.

[1084] Example 8: Gathering feedback and improving the AI ​​model

[1085] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it to improve the AI ​​model.

[1086] Example of a prompt

[1087] "Please tell me how to apply for a new project."

[1088] "What is the procedure for applying for leave?"

[1089] "Could you show me the minutes of the recent meeting?"

[1090] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[1091] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1092] Step 1:

[1093] Data collection

[1094] input:

[1095] All employee communication records (emails, chat logs), and business-related information (guidelines, manuals)

[1096] Specific actions:

[1097] The server uses the Microsoft Exchange Server API and the Slack API to collect employee emails and chat logs. It also retrieves business-related documents (guidelines, manuals) from the company's file server.

[1098] output:

[1099] Collected communication records and work-related information of all employees

[1100] Specific example:

[1101] The server collects employee A's emails and chat logs from the past six months, and simultaneously retrieves guidelines and manuals.

[1102] Step 2:

[1103] Data analysis and index creation

[1104] input:

[1105] Collected communication records and work-related information of all employees

[1106] Specific actions:

[1107] The server analyzes the collected data using Apache Lucene and Elasticsearch and creates an index. This involves using natural language processing techniques to extract and classify keywords.

[1108] output:

[1109] Indexed dataset

[1110] Specific example:

[1111] The server analyzes the data for keywords such as "how to apply for a new project" and "guidelines," and creates an index.

[1112] Step 3:

[1113] AI model training

[1114] input:

[1115] Indexed dataset

[1116] Specific actions:

[1117] The server uses TensorFlow or PyTorch to train an AI model based on indexed data. This process uses a large amount of text data to learn question and associated answer patterns.

[1118] output:

[1119] Trained AI model

[1120] Specific example:

[1121] The server trains an AI model based on indexed data, learning appropriate response patterns regarding "how to submit a new project application."

[1122] Step 4:

[1123] Real-time inquiry reception

[1124] input:

[1125] User inquiries (questions entered in natural language)

[1126] Specific actions:

[1127] The user sends a question to the system using a terminal. The terminal sends this inquiry to the server in real time.

[1128] output:

[1129] Inquiry content sent to the server

[1130] Specific example:

[1131] User B uses their device and types and sends the message, "Please tell me how to apply for a new project."

[1132] Step 5:

[1133] Analysis of inquiry content

[1134] input:

[1135] Inquiry content sent from the device

[1136] Specific actions:

[1137] The server analyzes queries using natural language processing techniques such as BERT and GPT-3, and extracts key keywords and contextual information.

[1138] output:

[1139] Analyzed query content (main keywords and contextual information)

[1140] Specific example:

[1141] The server extracts and analyzes the keywords "new project" and "application method" from user B's inquiry.

[1142] Step 6:

[1143] Answer generation

[1144] input:

[1145] Analyzed inquiry content

[1146] Specific actions:

[1147] The server uses a trained AI model to generate appropriate answers based on the analyzed query. Links to relevant internal documents and manuals are included as needed.

[1148] output:

[1149] Generated answer

[1150] Specific example:

[1151] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1152] Step 7:

[1153] Providing an answer

[1154] input:

[1155] Generated answer

[1156] Specific actions:

[1157] The device displays the responses sent from the server to the user in real time.

[1158] output:

[1159] User-displayed response

[1160] Specific example:

[1161] The generated response is displayed on User B's device.

[1162] Step 8:

[1163] Gathering feedback and improving the AI ​​model

[1164] input:

[1165] User Feedback

[1166] Specific actions:

[1167] The user enters feedback on the answer into the device. The device sends this feedback to the server, which adds it to the AI ​​model's training data to make improvements.

[1168] output:

[1169] Improved AI model

[1170] Specific example:

[1171] User B sends feedback saying "The answer was helpful," and the server uses that feedback to improve the accuracy of the AI ​​model.

[1172] (Application Example 1)

[1173] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1174] There is a problem in factories where information regarding business processes, operating procedures, and troubleshooting is ambiguous, making it difficult for employees to quickly obtain this information. This often leads to decreased operational efficiency and delays in troubleshooting. This invention aims to solve these problems and improve operational efficiency within factories by providing an AI-powered system.

[1175] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1176] In this invention, the server includes means for collecting communication records and related documents of all employees, means for analyzing the collected data and creating an index, and means for training an AI model based on the analyzed data. This makes it possible to respond quickly and accurately to user inquiries. Furthermore, by adding means for factory workers to make inquiries using tablet devices or robots, and means for supporting questions regarding factory operating procedures and troubleshooting, it becomes possible to speed up and streamline information acquisition within the factory.

[1177] "Communication records" refer to records of digital communications, such as emails and chat logs, conducted by all employees within the company.

[1178] "Related documents" refer to documents containing information such as guidelines, manuals, and operational procedures used within a company.

[1179] An "index" is a list or database structure that organizes analyzed data into a searchable format.

[1180] An "AI model" is an artificial intelligence model generated using machine learning algorithms based on collected and analyzed data.

[1181] An "inquiry" refers to a question or request that a user enters into a system.

[1182] "Answer" refers to the information generated by the AI ​​model and provided in response to a user's inquiry.

[1183] "Feedback" refers to evaluations and opinions given to responses provided by users.

[1184] A "factory worker" refers to an employee who performs tasks such as production, management, and maintenance within a factory.

[1185] A "tablet device" is a portable computer device equipped with a touchscreen.

[1186] A "robot" is a machine that operates automatically and performs tasks within a factory.

[1187] "Operating procedures" refer to step-by-step instructions or methods for performing a specific task or operation.

[1188] "Troubleshooting" refers to the procedures and methods used to resolve problems that occur in a system or machine.

[1189] "To support" means to provide the necessary information and means to perform a specific task or operation.

[1190] This invention is a system for efficiently acquiring information on operating procedures and troubleshooting within a factory. The system functions by collecting communication records and related documents from all employees, analyzing this data to create an index, and training an AI model. It also includes a process of receiving user inquiries in real time, analyzing them, generating and presenting answers, and collecting user feedback to improve the AI ​​model.

[1191] The server first collects communication records and related documents, such as emails and chat logs, from all employees. Next, it analyzes the collected data using natural language processing technology and creates an index in a searchable format. The analyzed data is used to train an AI model, which is built to generate appropriate responses.

[1192] Users (factory workers) can make inquiries using tablet devices or robots. For example, if a worker asks "How do I configure machine A?" or "How do I troubleshoot error code E123?", the device sends this inquiry to the server in real time. The server analyzes the received inquiry, extracts relevant keywords and contextual information, generates an appropriate answer based on the analysis results, and presents the answer to the device.

[1193] The generated responses may include links to relevant internal documents and manuals. This allows workers to quickly obtain the necessary information and improve work efficiency. User feedback is also important; by collecting feedback on whether the provided responses were helpful, the server can continuously improve the AI ​​model.

[1194] This system is implemented using hardware such as NVIDIA GPUs, cloud instances from AWS and Azure, and software such as Hugging Face's Transformers library. This enables high-performance data processing and analysis.

[1195] Specific example

[1196] For example, a factory worker might type "How do I troubleshoot error code E123 on machine B?" into a tablet. In this case, the system would generate the following response: "Error code E123 requires you to first turn off the power and restart the machine. Detailed troubleshooting information can be found at the following link."

[1197] Example of a prompt

[1198] "Please tell me how to troubleshoot error code E123 on machine B."

[1199] In this way, this system can streamline business processes within the factory and improve worker productivity.

[1200] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1201] Step 1:

[1202] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from all employees. The input is the communication records and related documents of all employees, and the output is a collection of the collected data. This data is used for subsequent analysis and indexing.

[1203] Step 2:

[1204] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. The input is the data collected in step 1, and the output is the analyzed data and its index. In this process, the meaning of the text is understood and converted into a searchable format.

[1205] Step 3:

[1206] The server trains an AI model based on the analyzed data. The input is the indexed data obtained in step 2, and the output is the trained AI model. This gives the AI ​​the ability to generate appropriate answers to queries.

[1207] Step 4:

[1208] The user enters the inquiry in natural language using a tablet or robot. The input is the user's inquiry statement, and the output is the inquiry request. This request is sent to the server in real time.

[1209] Step 5:

[1210] The server analyzes the received inquiry. The input is the inquiry request sent in step 4, and the output is the analyzed inquiry content. Keywords and contextual information are extracted and used as basic data to generate appropriate responses.

[1211] Step 6:

[1212] The server generates an appropriate response based on the analysis results. The input is the query analyzed in step 5 and the trained AI model, and the output is the generated response. Links to relevant documents and manuals are included as needed.

[1213] Step 7:

[1214] The terminal presents the generated answer to the user. The input is the answer generated in step 6, and the output is the answer displayed to the user. The user can then use this answer to solve the problem.

[1215] Step 8:

[1216] The user provides feedback on the provided answers. The input is the user's feedback, and the output is a feedback request. This request is sent to the server and used to improve the AI ​​model.

[1217] Step 9:

[1218] The server improves the AI ​​model based on user feedback. The input is the feedback request submitted in step 8, and the output is the improved AI model. This improves the accuracy of answers to future inquiries.

[1219] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1220] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[1221] System Overview

[1222] This system operates using the following procedure:

[1223] 1. Data Collection

[1224] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[1225] 2. Data analysis and index creation

[1226] The server uses an NLP engine to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[1227] 3. Training the AI ​​model

[1228] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[1229] 4. Real-time inquiry reception

[1230] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the device's interface.

[1231] 5. Emotion recognition by an emotion engine

[1232] The server uses an emotion engine to recognize emotions from the user's inquiry. This allows it to understand the user's emotional state and use the data to generate more appropriate responses.

[1233] 6. Analysis of the inquiry content

[1234] The server re-analyzes the query content, taking sentiment information into account, and then extracts keywords and contextual information.

[1235] 7. Generating the answer

[1236] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[1237] 8. Providing the answer

[1238] The device displays the generated response to the user. The user can view the generated response and associated links on the device screen.

[1239] 9. Gathering feedback and improving the AI ​​model

[1240] The device collects user feedback and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[1241] Specific example

[1242] Example 1: Data Collection

[1243] The server collects emails and chat logs sent and received by employee A over the past six months.

[1244] Example 2: Data analysis and index creation

[1245] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[1246] Example 3: Training an AI model

[1247] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[1248] Example 4: Real-time inquiry reception

[1249] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[1250] Example 5: Emotion recognition by an emotion engine

[1251] The server uses an emotion engine to recognize that the user is confused, based on the user's inquiry, "How do I apply for a new project?"

[1252] Example 6: Analysis of inquiry content

[1253] The server extracts keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[1254] Example 7: Generating an answer

[1255] The server generates a response stating, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below," and provides it in an easily understandable format to alleviate user confusion.

[1256] Example 8: Providing an answer

[1257] The generated response is displayed on User B's terminal screen.

[1258] Example 9: Gathering feedback and improving the AI ​​model

[1259] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[1260] In this way, the system analyzes the content of inquiries and the user's emotions to generate more appropriate and effective responses. This leads to increased operational efficiency and improved employee productivity.

[1261] The following describes the processing flow.

[1262] Step 1: Data Collection

[1263] The server connects to all employees' mail and chat servers to download emails and chat logs from the past six months. It also accesses the document management system to collect relevant documents such as company guidelines, manuals, and application procedures.

[1264] Step 2: Data Analysis and Indexing

[1265] The server analyzes the collected data using an NLP engine. Specifically, it extracts important keywords and contextual information from each document and message. Then, it creates an index based on the analysis results and stores it in a database.

[1266] Step 3: Training the AI ​​model

[1267] The server trains an AI model using a supervised learning algorithm based on the analyzed data. The training data includes pairs of past queries and their appropriate answers. This gives the model the ability to generate optimal answers to a variety of queries.

[1268] Step 4: Real-time inquiry reception

[1269] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions using natural language.

[1270] Step 5: Emotion recognition by the emotion engine

[1271] The server receives the user's inquiry and uses an emotion engine to analyze the user's emotions. The emotion engine uses NLP techniques to identify emotions from the inquiry text.

[1272] Step 6: Analyzing the inquiry content

[1273] The server takes sentiment information into account and re-analyzes the query using an NLP engine. This extracts keywords and contextual information, and adds sentiment information.

[1274] Step 7: Generating the answer

[1275] Based on the analysis results, the server generates the most appropriate response for the user's emotional state and inquiry. Links to relevant internal documents and manuals are also added as needed.

[1276] Step 8: Present your answer

[1277] The device displays the generated answers to the user. The user then views the provided answers and associated links on the device screen.

[1278] Step 9: Gathering Feedback and Improving the AI ​​Model

[1279] The device collects user feedback and sends it to the server. The server analyzes this feedback and adds it to the AI ​​model as new training data. This improves the accuracy of the AI ​​model.

[1280] In this way, the system comprehensively analyzes the user's inquiry and emotions, enabling a process that generates the optimal response. This leads to increased operational efficiency and improved employee productivity.

[1281] (Example 2)

[1282] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1283] Traditional inquiry handling systems struggled to provide 24 / 7 support and often suffered from low-quality responses due to their failure to consider the user's emotional state. Furthermore, they were inefficient in handling inquiries because they couldn't effectively utilize all employee communication records and related documents. A system is needed to solve these problems and improve the efficiency of internal operations.

[1284] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1285] In this invention, the server includes means for collecting communication records and related documents of all employees; means for analyzing the collected data using a natural language processing engine and creating an index; means for training a machine learning model based on the analyzed data; means for receiving user inquiries in real time; means for recognizing the user's emotions from the received inquiry content using an emotion engine; means for analyzing the inquiry content taking emotional information into account; means for generating appropriate answers based on the analysis results; means for presenting the generated answers to the user; and means for collecting user feedback and improving the machine learning model. This enables 24 / 7 inquiry support and the provision of high-quality answers that take user emotions into consideration.

[1286] "All employee communication records" refers to digital communication data such as emails and chat logs used by company employees for work purposes.

[1287] "Related documents" refer to documents related to business operations, such as guidelines and manuals, that the company possesses.

[1288] A "natural language processing engine" is software that analyzes text data to extract keywords and perform sentiment analysis.

[1289] A "machine learning model" is a collection of algorithms that learn from data and perform pattern recognition and prediction.

[1290] A "means of receiving in real time" refers to a system or interface that can immediately receive inquiries from users.

[1291] An "emotion engine" is a tool or software that recognizes a user's emotional state from text data.

[1292] "Index creation" refers to techniques that generate lists of keywords and topics to enable rapid searching of large amounts of data.

[1293] "Means for analyzing inquiry content" refers to technologies for understanding the questions and requests entered by users and extracting the necessary information.

[1294] "Means for generating appropriate answers" refers to the process of creating the optimal response to a user's inquiry based on analyzed data.

[1295] "Means of presenting to the user" refers to a system for displaying generated answers and information to the user.

[1296] "Means of collecting feedback" refers to a system for gathering evaluations and opinions from users and using them to improve the system.

[1297] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[1298] System Overview

[1299] Data collection

[1300] The server collects communication records (emails, chat logs) of all employees using Microsoft Exchange Server and the Slack API. It also collects relevant documents such as guidelines and manuals from the company's file server and document management system.

[1301] Data analysis and index creation

[1302] The server uses natural language processing engines such as Google NLP and IBM Watson to analyze the collected data and extract key keywords and topics. This creates an efficient data index, enabling rapid information retrieval.

[1303] AI model training

[1304] The server uses libraries such as Hugging Face's Transformers library to train a machine learning model based on the analyzed data. During the training process, the data is split into training and validation sets, and the model parameters are adjusted. The trained AI model is then saved and prepared to generate the optimal response to queries.

[1305] Real-time inquiry reception

[1306] The device (PC or smartphone) receives user inquiries in real time through a web interface built with React or Vue.js. Users input and submit questions in natural language.

[1307] Emotion recognition by an emotion engine

[1308] The server uses emotion engines such as Affectiva or IBM Watson Tone Analyzer to recognize emotions from user inquiries. This data is then used to understand the user's emotional state and generate appropriate responses.

[1309] Analysis of inquiry content

[1310] The server re-analyzes the query content using the spaCy library, taking sentiment information into account. It extracts key keywords and contextual information and performs a detailed analysis based on the user's emotional state.

[1311] Answer generation

[1312] The server uses models such as OpenAI's GPT-3 to generate appropriate responses based on the analysis results. Links to relevant documents and manuals are included as needed. For example, it might generate a response like, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1313] Providing an answer

[1314] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[1315] Gathering feedback and improving the AI ​​model

[1316] The device collects user feedback and sends it to the server. The server analyzes the feedback, adds it as training data for the model, and improves the accuracy of the AI ​​model by retraining it.

[1317] Specific example

[1318] Data collection

[1319] The server uses the Microsoft Exchange Server API to retrieve employee A's email data and the Slack API to download employee A's chat logs.

[1320] Data analysis and index creation

[1321] The server uses the Google NLP API to analyze keywords such as "how to apply for a new project" and "project guidelines" and creates an index.

[1322] AI model training

[1323] The server uses the Hugging Face Transformers library to train an AI model based on the analyzed data, splitting it into training data and validation data for further training.

[1324] Real-time inquiry reception

[1325] The user enters "How do I submit a new project?" into a web interface built with React and submits it.

[1326] Emotion recognition by an emotion engine

[1327] The server uses IBM Watson Tone Analyzer to recognize the emotion of "confusion" from the user's inquiry.

[1328] Analysis of inquiry content

[1329] The server uses the spaCy library to extract keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[1330] Answer generation

[1331] The server uses GPT-3 to generate the response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1332] Providing an answer

[1333] The device displays the generated answer and related links on the user's screen.

[1334] Gathering feedback and improving the AI ​​model

[1335] The user enters feedback such as "The answer was helpful" into the React interface, and the device sends this to the server. The server analyzes the feedback data and uses it as training data for the AI ​​model.

[1336] Example of a prompt

[1337] "I'd like to know how to apply for a new project. Please provide a link to the application form on the portal site."

[1338] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1339] Step 1:

[1340] Data collection

[1341] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals) from all employees. Specifically, it uses Microsoft Exchange Server and the Slack API to retrieve data. Input data includes emails and chat messages, while output data includes all communication records and related documents. The server stores this data in a database for later analysis.

[1342] Step 2:

[1343] Data analysis and index creation

[1344] The server analyzes the collected data using a natural language processing (NLP) engine. Specifically, it uses Google NLP or IBM Watson to extract key keywords and topics. The input data consists of collected emails, chat logs, and related documents, while the output data is an index based on keywords and topics. The server creates this index to enable efficient data retrieval.

[1345] Step 3:

[1346] AI model training

[1347] The server trains a machine learning model using the analyzed data. Specifically, it uses the Hugging Face Transformers library to create the model. The input data consists of the analyzed data and extracted indices, while the output data is the trained machine learning model. The server stores this model so that it can generate the best possible answers to queries.

[1348] Step 4:

[1349] Real-time inquiry reception

[1350] The terminal receives user inquiries in real time. Users enter their inquiries through the terminal's web interface (built with React or Vue.js). The input data is the inquiry text entered by the user, and the output data is the transmission of that inquiry text to the server.

[1351] Step 5:

[1352] Emotion recognition by an emotion engine

[1353] The server recognizes emotions from the content of the received inquiry. Specifically, it uses Affectiva or IBM Watson Tone Analyzer. The input data is the user's inquiry, and the output data is the user's emotional state (e.g., confused, angry, happy). The server uses this emotional information for subsequent processing.

[1354] Step 6:

[1355] Analysis of inquiry content

[1356] The server re-analyzes the query content, taking sentiment information into account. Specifically, it uses the spaCy library to extract key keywords and contextual information. The input data consists of the user's query and sentiment information, while the output data consists of the analyzed keywords and contextual information.

[1357] Step 7:

[1358] Answer generation

[1359] The server generates appropriate answers based on the analysis results. Specifically, it uses models such as OpenAI's GPT-3. The input data consists of analyzed keywords and contextual information, and the output data is the generated answer. For example, an answer such as "To apply for a new project, please use the application form on the portal site" might be generated.

[1360] Step 8:

[1361] Providing an answer

[1362] The terminal displays the response sent from the server to the user. The input data is the generated response sent from the server, and the output data is the response displayed on the user's terminal. The user can view the response and related links on the terminal screen.

[1363] Step 9:

[1364] Gathering feedback and improving the AI ​​model

[1365] The device collects user feedback and sends it to the server. The input data is the user's feedback, and the output data is the transmission of that feedback to the server. The server analyzes the feedback data and adds it as training data for a machine learning model. This improves the model's accuracy and the quality of its responses.

[1366] (Application Example 2)

[1367] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1368] Traditional customer support systems often fail to provide adequate support because they generate answers without considering the user's emotions. This is particularly problematic for inquiries involving feelings of confusion or frustration, which can lead to decreased user satisfaction. Furthermore, there is a lack of mechanisms to effectively utilize user feedback to improve AI models, making it difficult to enhance system quality. Therefore, there is a need for an intelligent customer support system that provides high-quality answers and improves user satisfaction.

[1369] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1370] In this invention, the server includes means for collecting communication records and related documents from the entire organization, means for analyzing the collected data and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users, means for analyzing the content of the received inquiries, means for recognizing the emotional state of the user using an emotion recognition engine, means for generating an appropriate response based on the analysis results and emotional state, means for presenting the generated response to the user, and means for collecting feedback from the user and improving the artificial intelligence model. This enables the generation of high-quality responses that take into account the user's emotions in real time, improving the quality of inquiry handling and increasing user satisfaction.

[1371] "Communication records of the entire organization" refers to the history of digital communications, such as emails and chat logs, conducted within an organization such as a company or group.

[1372] "Related documents" refer to documents within an organization that are related to business operations, such as guidelines, manuals, memos, and reports.

[1373] "Means of collection" refers to systems that collect necessary information using databases, cloud storage, log management systems, etc.

[1374] "Means of analysis" refer to methods that use natural language processing techniques and machine learning algorithms to convert collected data into a format that can be understood and used.

[1375] "Creating an index" refers to a method of organizing data by adding keywords and metadata so that it can be quickly searched and referenced.

[1376] An "artificial intelligence model" is a computer program that uses machine learning and deep learning algorithms to learn knowledge from data and has the ability to perform specific tasks.

[1377] "Means of receiving inquiries from users" refers to interfaces or applications that receive questions and requests that users make to the system.

[1378] "Means for analyzing received inquiries" refers to methods for interpreting inquiries received from users and extracting information necessary to generate appropriate responses.

[1379] An "emotion recognition engine" is an artificial intelligence technology that identifies a user's emotional state from data such as text and audio.

[1380] "Emotional state" refers to the psychological state a user is in when making an inquiry, and includes, for example, confusion, frustration, or joy.

[1381] "Means for generating appropriate answers" refers to methods for creating specific and helpful answers to user inquiries based on collected and analyzed data.

[1382] "Means of presenting generated answers to users" refers to interfaces or applications that display system-generated answers in a user-friendly format.

[1383] "Means of collecting feedback" refers to methods of receiving opinions and evaluations provided by users and processing them within the system.

[1384] "Methods for improving artificial intelligence models" refer to methods of improving the accuracy and performance of existing artificial intelligence models using collected feedback data.

[1385] This invention provides an intelligent inquiry response system that collects communication records and related documents from the entire organization, trains an artificial intelligence model based on them, and provides appropriate answers to user inquiries. In particular, this system can improve user satisfaction by recognizing the user's emotional state using an emotion recognition engine and generating answers based on that state.

[1386] System Configuration

[1387] 1. Data Collection

[1388] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from the entire organization via cloud storage (e.g., Amazon S3). This data is used for analysis and indexing, as described later.

[1389] 2. Data analysis and index creation

[1390] The server uses a natural language processing engine (e.g., Amazon Comprehend) to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[1391] 3. Training the AI ​​model

[1392] The server trains an artificial intelligence model (e.g., TensorFlow, PyTorch) based on the analyzed data. This model has the ability to generate the best possible answers to user inquiries.

[1393] 4. Real-time inquiry reception

[1394] The device (e.g., a smartphone application) receives inquiries from users in real time. Users input and submit questions in natural language using the device's interface.

[1395] 5. Emotion recognition

[1396] The server analyzes the received inquiry using an emotion recognition engine (e.g., Amazon Rekognition, IBM Watson Tone Analyzer) to recognize the user's emotional state. This emotional state is a crucial factor in generating the response.

[1397] 6. Answer generation

[1398] The server generates appropriate responses based on the analysis results and recognized emotional states. By using a generative AI model (e.g., OpenAI's GPT-4), it provides highly accurate responses. These responses may include links to relevant internal documents and instructions.

[1399] 7. Answer presentation

[1400] The device displays the generated response to the user. The user can check the response on the device screen.

[1401] 8. Gathering feedback and improving the AI ​​model

[1402] The device collects feedback from users and sends the feedback data to a server. The server uses this feedback to retrain the artificial intelligence model and improve the overall accuracy of the system.

[1403] Specific example

[1404] For example, consider a scenario where a user inquires, "My product hasn't arrived yet." This system can recognize the user's feelings of confusion or frustration and generate a response that includes a gentler tone and specific solutions. An example of an input prompt for the AI ​​model in this case is shown below.

[1405] Example of a prompt:

[1406] User inquiry: "My item hasn't arrived yet."

[1407] Emotional state: "Confused", "Irritated"

[1408] In this way, the present invention makes it possible to generate high-quality responses that take into account the user's emotions in real time, thereby improving the quality of inquiry handling and increasing user satisfaction.

[1409] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1410] Step 1:

[1411] The server collects communication records and related documents from the entire organization's cloud storage. The input to this process is the location where the communication records and related documents are stored (e.g., the cloud storage bucket ID), and the output is the collected data set. This data primarily includes emails, chat logs, guidelines, manuals, etc. The server stores this data in a database for use in the next analysis step.

[1412] Step 2:

[1413] The server analyzes the collected data using a natural language processing engine and creates an index. The input to this process is the collected dataset, and the output is the analyzed data and its index information. The server performs keyword extraction and contextual analysis to generate an index that enables rapid data retrieval.

[1414] Step 3:

[1415] The server trains an artificial intelligence model based on the analysis results. The input to this process is the analyzed data and index information, and the output is the trained AI model. Machine learning libraries (e.g., TensorFlow, PyTorch) are used to train the model. The server tunes the model to optimal performance, improving its ability to generate answers to queries.

[1416] Step 4:

[1417] The terminal receives inquiries from users in real time. The input in this process is a natural language question from the user, and the output is a notification that the inquiry has been received. The terminal sends the text entered by the user to the server for preparation for analysis.

[1418] Step 5:

[1419] The server analyzes the received inquiry using an emotion recognition engine to recognize the user's emotional state. The input for this process is the user's inquiry text, and the output is the recognized emotional state (e.g., confused, irritated). The server utilizes the emotion recognition engine to analyze the emotional tone.

[1420] Step 6:

[1421] The server generates an appropriate response based on the inquiry content and the recognized emotional state. The input to this process is the inquiry content and emotional state, and the output is the generated response. The server uses a generative AI model (e.g., OpenAI's GPT-4) to create specific and helpful responses.

[1422] Step 7:

[1423] The terminal presents the generated response to the user. The input in this process is the generated response sent from the server, and the output is the response displayed in a format visible to the user. The terminal provides the response to the user through a GUI (Graphical User Interface).

[1424] Step 8:

[1425] The terminal collects feedback from users and sends it to the server. The input in this process is the feedback provided by the user, and the output is the recording of the feedback and its transfer to the server. The server uses this feedback as retraining data for machine learning to continuously improve the artificial intelligence model.

[1426] Thus, the system of the present invention can provide highly accurate answers to user inquiries in real time using emotion recognition and an AI model.

[1427] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1428] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1429] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1430] [Fourth Embodiment]

[1431] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1432] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1433] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1434] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1435] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1436] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1437] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1438] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1439] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1440] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[1441] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1442] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1443] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1444] This invention is a system designed to solve a business problem where companies face the issue of unclear internal system application procedures, rules, guidelines, and manuals, which require significant time to research. This system utilizes AI to provide intelligent inquiry support available 24 / 7, aiming to improve operational efficiency and optimize resources.

[1445] System Overview

[1446] This system operates using the following procedure:

[1447] 1. Data Collection

[1448] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[1449] 2. Data analysis and index creation

[1450] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. This allows for the rapid extraction of appropriate information in response to queries.

[1451] 3. Training the AI ​​model

[1452] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[1453] 4. Real-time inquiry reception

[1454] The device (PC or smartphone) receives user inquiries in real time. Users input their questions in natural language.

[1455] 5. Analysis of the inquiry content

[1456] The server analyzes the received inquiry and extracts relevant keywords and contextual information.

[1457] 6. Generating the answer

[1458] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[1459] 7. Providing the answer

[1460] The device presents the generated answer to the user.

[1461] 8. Gathering feedback and improving the AI ​​model

[1462] The device collects user feedback and sends it to the server. The server uses this feedback to improve the accuracy of the AI ​​model.

[1463] Specific example

[1464] Example 1: Data Collection

[1465] The server collects emails and chat logs sent and received by employee A over the past six months.

[1466] Example 2: Data analysis and index creation

[1467] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[1468] Example 3: Training an AI model

[1469] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[1470] Example 4: Real-time inquiry reception

[1471] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[1472] Example 5: Analysis of inquiry content

[1473] The server extracts and analyzes the keywords "new project" and "application method".

[1474] Example 6: Generating an answer

[1475] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1476] Example 7: Providing an answer

[1477] The generated response is displayed on User B's terminal screen.

[1478] Example 8: Gathering feedback and improving the AI ​​model

[1479] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[1480] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[1481] The following describes the processing flow.

[1482] Step 1: Data Collection

[1483] The server connects to all employees' email and chat servers and downloads communication records for the past six months. It also accesses the document management system to retrieve relevant documents such as company guidelines, manuals, and application procedures.

[1484] Step 2: Data Analysis and Indexing

[1485] The server uses a natural language processing (NLP) engine to analyze the collected data. Specifically, it extracts keywords from each document and message and understands their contextual information. Then, based on the analysis results, it creates an index and stores it in a database.

[1486] Step 3: Training the AI ​​model

[1487] The server uses the analyzed data to train an AI model using a supervised learning algorithm. The model is designed to generate the best possible answers to user inquiries. The training process involves numerous question-and-answer pairs.

[1488] Step 4: Real-time inquiry reception

[1489] The terminal (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the terminal's interface.

[1490] Step 5: Analyzing the inquiry content

[1491] The server receives inquiries from users and uses a natural language processing engine to analyze the question content. The analysis extracts important keywords and contextual information from the question and compares them against an index database.

[1492] Step 6: Generating the answer

[1493] The server generates an appropriate response based on the query analysis results. If necessary, it also retrieves links to relevant internal documents and manuals from the index database and includes them in the response.

[1494] Step 7: Present your answer

[1495] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[1496] Step 8: Gathering feedback and improving the AI ​​model

[1497] The device receives feedback provided by the user and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[1498] (Example 1)

[1499] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1500] Traditional corporate systems often lacked clear internal rules, guidelines, and manuals, resulting in employees spending a significant amount of time searching for this information. This led to decreased employee efficiency and a substantial decline in productivity.

[1501] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1502] In this invention, the server includes means for collecting communication records and business-related information, means for analyzing the collected data using natural language processing technology and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users in real time, means for analyzing the content of the received inquiries using natural language processing technology, means for generating appropriate answers based on the analysis results, means for presenting the generated answers to users, and means for collecting feedback from users and improving the artificial intelligence model. As a result, users can obtain information quickly and appropriately, enabling increased efficiency in internal operations and optimization of resources.

[1503] "Communication records" refer to records of communication between employees within a company, such as emails and chat logs.

[1504] "Business-related information" refers to documentary information such as guidelines, manuals, and procedural documents necessary for carrying out business operations within a company.

[1505] "Natural language processing technology" refers to technologies that enable computers to recognize, interpret, and generate human language, and specifically includes text analysis and contextual understanding.

[1506] An "artificial intelligence model" is an algorithm that learns from large amounts of data, extracts patterns and knowledge, and automatically generates the optimal answer to an inquiry.

[1507] An "index" refers to a systematic organization of keywords and contextual information used to quickly retrieve specific information from a large amount of data.

[1508] "Users" refers to company employees who use this system to make internal inquiries.

[1509] "Real-time" refers to a timeframe in which processing and responses occur almost instantly, meaning that a response is provided immediately after a user submits a question.

[1510] "Feedback" refers to information collected from users who provide evaluations and opinions on the answers they receive, in order to help improve the system.

[1511] "Analysis results" refer to data containing important keywords and contextual information obtained after analyzing the content of a query using natural language processing technology.

[1512] "Answer" refers to the information and instructions provided in response to a user's inquiry, generated based on the analysis results.

[1513] A "link" is a means of reference, such as a URL, that contains direct reference information to internal documents or manuals related to the generated response.

[1514] The system according to this invention collects communication records and business-related information, analyzes and utilizes it to provide prompt and appropriate responses to user inquiries. This system significantly improves operational efficiency within a company and optimizes resources.

[1515] Hardware and software configuration

[1516] server

[1517] 1. Data collection:

[1518] The server collects communication records such as employee emails and chat logs, as well as business-related information such as guidelines and manuals. This utilizes the Microsoft Exchange Server API and the Slack API.

[1519] 2. Data Analysis:

[1520] The server analyzes the collected data using natural language processing (NLP) techniques. During this process, it creates indexes using Apache Lucene or Elasticsearch.

[1521] 3. AI model training:

[1522] Based on the analyzed and indexed data, AI models are trained using machine learning frameworks such as TensorFlow and PyTorch.

[1523] 4. Analysis of the inquiry content:

[1524] The server analyzes user inquiries using advanced NLP models such as BERT and GPT-3.

[1525] 5. Generating the answer:

[1526] Based on the analysis results, generate appropriate answers. Include links to relevant internal documents and manuals as needed.

[1527] 6. Gathering feedback and improving the AI ​​model:

[1528] We collect user feedback and use it to continuously improve our AI models.

[1529] terminal

[1530] 1. Real-time inquiry reception:

[1531] The terminal receives user inquiries in real time via PCs and smartphones. Users input their questions in natural language.

[1532] 2. Providing the answer:

[1533] The device displays the responses sent from the server to the user in real time.

[1534] user

[1535] 1. Enter your inquiry:

[1536] Users input questions in natural language via their devices.

[1537] 2. Providing feedback:

[1538] Feedback on the provided answers is entered into the terminal, and that feedback is sent to the server.

[1539] Specific example

[1540] Example 1: Data Collection

[1541] The server collects emails and chat logs sent and received by employee A over the past six months.

[1542] Example 2: Data Analysis

[1543] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[1544] Example 3: Training an AI model

[1545] The server trains an AI model based on the collected data, learning appropriate response patterns based on specific keywords and contexts.

[1546] Example 4: Real-time inquiry reception

[1547] User B uses their device to type "Please tell me how to apply for a new project" and sends it.

[1548] Example 5: Analysis of inquiry content

[1549] The server extracts and analyzes the keywords "new project" and "application method".

[1550] Example 6: Generating an answer

[1551] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1552] Example 7: Providing an answer

[1553] The generated response is displayed on User B's terminal screen.

[1554] Example 8: Gathering feedback and improving the AI ​​model

[1555] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it to improve the AI ​​model.

[1556] Example of a prompt

[1557] "Please tell me how to apply for a new project."

[1558] "What is the procedure for applying for leave?"

[1559] "Could you show me the minutes of the recent meeting?"

[1560] In this way, this system can streamline the handling of internal business inquiries and improve employee productivity.

[1561] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1562] Step 1:

[1563] Data collection

[1564] input:

[1565] All employee communication records (emails, chat logs), and business-related information (guidelines, manuals)

[1566] Specific actions:

[1567] The server uses the Microsoft Exchange Server API and the Slack API to collect employee emails and chat logs. It also retrieves business-related documents (guidelines, manuals) from the company's file server.

[1568] output:

[1569] Collected communication records and work-related information of all employees

[1570] Specific example:

[1571] The server collects employee A's emails and chat logs from the past six months, and simultaneously retrieves guidelines and manuals.

[1572] Step 2:

[1573] Data analysis and index creation

[1574] input:

[1575] Collected communication records and work-related information of all employees

[1576] Specific actions:

[1577] The server analyzes the collected data using Apache Lucene and Elasticsearch and creates an index. This involves using natural language processing techniques to extract and classify keywords.

[1578] output:

[1579] Indexed dataset

[1580] Specific example:

[1581] The server analyzes the data for keywords such as "how to apply for a new project" and "guidelines," and creates an index.

[1582] Step 3:

[1583] AI model training

[1584] input:

[1585] Indexed dataset

[1586] Specific actions:

[1587] The server uses TensorFlow or PyTorch to train an AI model based on indexed data. This process uses a large amount of text data to learn question and associated answer patterns.

[1588] output:

[1589] Trained AI model

[1590] Specific example:

[1591] The server trains an AI model based on indexed data, learning appropriate response patterns regarding "how to submit a new project application."

[1592] Step 4:

[1593] Real-time inquiry reception

[1594] input:

[1595] User inquiries (questions entered in natural language)

[1596] Specific actions:

[1597] The user sends a question to the system using a terminal. The terminal sends this inquiry to the server in real time.

[1598] output:

[1599] Inquiry content sent to the server

[1600] Specific example:

[1601] User B uses their device and types and sends the message, "Please tell me how to apply for a new project."

[1602] Step 5:

[1603] Analysis of inquiry content

[1604] input:

[1605] Inquiry content sent from the device

[1606] Specific actions:

[1607] The server analyzes queries using natural language processing techniques such as BERT and GPT-3, and extracts key keywords and contextual information.

[1608] output:

[1609] Analyzed query content (main keywords and contextual information)

[1610] Specific example:

[1611] The server extracts and analyzes the keywords "new project" and "application method" from user B's inquiry.

[1612] Step 6:

[1613] Answer generation

[1614] input:

[1615] Analyzed inquiry content

[1616] Specific actions:

[1617] The server uses a trained AI model to generate appropriate answers based on the analyzed query. Links to relevant internal documents and manuals are included as needed.

[1618] output:

[1619] Generated answer

[1620] Specific example:

[1621] The server generates the following response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1622] Step 7:

[1623] Providing an answer

[1624] input:

[1625] Generated answer

[1626] Specific actions:

[1627] The device displays the responses sent from the server to the user in real time.

[1628] output:

[1629] User-displayed response

[1630] Specific example:

[1631] The generated response is displayed on User B's device.

[1632] Step 8:

[1633] Gathering feedback and improving the AI ​​model

[1634] input:

[1635] User Feedback

[1636] Specific actions:

[1637] The user enters feedback on the answer into the device. The device sends this feedback to the server, which adds it to the AI ​​model's training data to make improvements.

[1638] output:

[1639] Improved AI model

[1640] Specific example:

[1641] User B sends feedback saying "The answer was helpful," and the server uses that feedback to improve the accuracy of the AI ​​model.

[1642] (Application Example 1)

[1643] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1644] There is a problem in factories where information regarding business processes, operating procedures, and troubleshooting is ambiguous, making it difficult for employees to quickly obtain this information. This often leads to decreased operational efficiency and delays in troubleshooting. This invention aims to solve these problems and improve operational efficiency within factories by providing an AI-powered system.

[1645] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1646] In this invention, the server includes means for collecting communication records and related documents of all employees, means for analyzing the collected data and creating an index, and means for training an AI model based on the analyzed data. This makes it possible to respond quickly and accurately to user inquiries. Furthermore, by adding means for factory workers to make inquiries using tablet devices or robots, and means for supporting questions regarding factory operating procedures and troubleshooting, it becomes possible to speed up and streamline information acquisition within the factory.

[1647] "Communication records" refer to records of digital communications, such as emails and chat logs, conducted by all employees within the company.

[1648] "Related documents" refer to documents containing information such as guidelines, manuals, and operational procedures used within a company.

[1649] An "index" is a list or database structure that organizes analyzed data into a searchable format.

[1650] An "AI model" is an artificial intelligence model generated using machine learning algorithms based on collected and analyzed data.

[1651] An "inquiry" refers to a question or request that a user enters into a system.

[1652] "Answer" refers to the information generated by the AI ​​model and provided in response to a user's inquiry.

[1653] "Feedback" refers to evaluations and opinions given to responses provided by users.

[1654] A "factory worker" refers to an employee who performs tasks such as production, management, and maintenance within a factory.

[1655] A "tablet device" is a portable computer device equipped with a touchscreen.

[1656] A "robot" is a machine that operates automatically and performs tasks within a factory.

[1657] "Operating procedures" refer to step-by-step instructions or methods for performing a specific task or operation.

[1658] "Troubleshooting" refers to the procedures and methods used to resolve problems that occur in a system or machine.

[1659] "To support" means to provide the necessary information and means to perform a specific task or operation.

[1660] This invention is a system for efficiently acquiring information on operating procedures and troubleshooting within a factory. The system functions by collecting communication records and related documents from all employees, analyzing this data to create an index, and training an AI model. It also includes a process of receiving user inquiries in real time, analyzing them, generating and presenting answers, and collecting user feedback to improve the AI ​​model.

[1661] The server first collects communication records and related documents, such as emails and chat logs, from all employees. Next, it analyzes the collected data using natural language processing technology and creates an index in a searchable format. The analyzed data is used to train an AI model, which is built to generate appropriate responses.

[1662] Users (factory workers) can make inquiries using tablet devices or robots. For example, if a worker asks "How do I configure machine A?" or "How do I troubleshoot error code E123?", the device sends this inquiry to the server in real time. The server analyzes the received inquiry, extracts relevant keywords and contextual information, generates an appropriate answer based on the analysis results, and presents the answer to the device.

[1663] The generated responses may include links to relevant internal documents and manuals. This allows workers to quickly obtain the necessary information and improve work efficiency. User feedback is also important; by collecting feedback on whether the provided responses were helpful, the server can continuously improve the AI ​​model.

[1664] This system is implemented using hardware such as NVIDIA GPUs, cloud instances from AWS and Azure, and software such as Hugging Face's Transformers library. This enables high-performance data processing and analysis.

[1665] Specific example

[1666] For example, a factory worker might type "How do I troubleshoot error code E123 on machine B?" into a tablet. In this case, the system would generate the following response: "Error code E123 requires you to first turn off the power and restart the machine. Detailed troubleshooting information can be found at the following link."

[1667] Example of a prompt

[1668] "Please tell me how to troubleshoot error code E123 on machine B."

[1669] In this way, this system can streamline business processes within the factory and improve worker productivity.

[1670] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1671] Step 1:

[1672] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from all employees. The input is the communication records and related documents of all employees, and the output is a collection of the collected data. This data is used for subsequent analysis and indexing.

[1673] Step 2:

[1674] The server analyzes the collected data using natural language processing (NLP) techniques and creates an index. The input is the data collected in step 1, and the output is the analyzed data and its index. In this process, the meaning of the text is understood and converted into a searchable format.

[1675] Step 3:

[1676] The server trains an AI model based on the analyzed data. The input is the indexed data obtained in step 2, and the output is the trained AI model. This gives the AI ​​the ability to generate appropriate answers to queries.

[1677] Step 4:

[1678] The user enters the inquiry in natural language using a tablet or robot. The input is the user's inquiry statement, and the output is the inquiry request. This request is sent to the server in real time.

[1679] Step 5:

[1680] The server analyzes the received inquiry. The input is the inquiry request sent in step 4, and the output is the analyzed inquiry content. Keywords and contextual information are extracted and used as basic data to generate appropriate responses.

[1681] Step 6:

[1682] The server generates an appropriate response based on the analysis results. The input is the query analyzed in step 5 and the trained AI model, and the output is the generated response. Links to relevant documents and manuals are included as needed.

[1683] Step 7:

[1684] The terminal presents the generated answer to the user. The input is the answer generated in step 6, and the output is the answer displayed to the user. The user can then use this answer to solve the problem.

[1685] Step 8:

[1686] The user provides feedback on the provided answers. The input is the user's feedback, and the output is a feedback request. This request is sent to the server and used to improve the AI ​​model.

[1687] Step 9:

[1688] The server improves the AI ​​model based on user feedback. The input is the feedback request submitted in step 8, and the output is the improved AI model. This improves the accuracy of answers to future inquiries.

[1689] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1690] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[1691] System Overview

[1692] This system operates using the following procedure:

[1693] 1. Data Collection

[1694] The server collects all employees' communication records (emails, chat logs) and related documents (guidelines, manuals, etc.).

[1695] 2. Data analysis and index creation

[1696] The server uses an NLP engine to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[1697] 3. Training the AI ​​model

[1698] The server trains an AI model based on the analyzed data. This AI model can generate the optimal response to user inquiries.

[1699] 4. Real-time inquiry reception

[1700] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions in natural language using the device's interface.

[1701] 5. Emotion recognition by an emotion engine

[1702] The server uses an emotion engine to recognize emotions from the user's inquiry. This allows it to understand the user's emotional state and use the data to generate more appropriate responses.

[1703] 6. Analysis of the inquiry content

[1704] The server re-analyzes the query content, taking sentiment information into account, and then extracts keywords and contextual information.

[1705] 7. Generating the answer

[1706] The server generates appropriate answers based on the analysis results, including links to relevant documents and manuals as needed.

[1707] 8. Providing the answer

[1708] The device displays the generated response to the user. The user can view the generated response and associated links on the device screen.

[1709] 9. Gathering feedback and improving the AI ​​model

[1710] The device collects user feedback and sends it to the server. The server analyzes this feedback and uses it to improve the AI ​​model. Specifically, it adds the feedback as training data and retrains the model.

[1711] Specific example

[1712] Example 1: Data Collection

[1713] The server collects emails and chat logs sent and received by employee A over the past six months.

[1714] Example 2: Data analysis and index creation

[1715] The server analyzes keywords such as "how to apply for a new project" and "guidelines" and creates an index.

[1716] Example 3: Training an AI model

[1717] The server uses the collected data to train an AI model, learning appropriate response patterns based on keywords and context.

[1718] Example 4: Real-time inquiry reception

[1719] User B uses a terminal and types "Please tell me how to apply for a new project," then sends the message.

[1720] Example 5: Emotion recognition by an emotion engine

[1721] The server uses an emotion engine to recognize that the user is confused, based on the user's inquiry, "How do I apply for a new project?"

[1722] Example 6: Analysis of inquiry content

[1723] The server extracts keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[1724] Example 7: Generating an answer

[1725] The server generates a response stating, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below," and provides it in an easily understandable format to alleviate user confusion.

[1726] Example 8: Providing an answer

[1727] The generated response is displayed on User B's terminal screen.

[1728] Example 9: Gathering feedback and improving the AI ​​model

[1729] When User B provides feedback stating "The answer was helpful," the device sends this to the server, which then uses it as training data for the AI ​​model.

[1730] In this way, the system analyzes the content of inquiries and the user's emotions to generate more appropriate and effective responses. This leads to increased operational efficiency and improved employee productivity.

[1731] The following describes the processing flow.

[1732] Step 1: Data Collection

[1733] The server connects to all employees' mail and chat servers to download emails and chat logs from the past six months. It also accesses the document management system to collect relevant documents such as company guidelines, manuals, and application procedures.

[1734] Step 2: Data Analysis and Indexing

[1735] The server analyzes the collected data using an NLP engine. Specifically, it extracts important keywords and contextual information from each document and message. Then, it creates an index based on the analysis results and stores it in a database.

[1736] Step 3: Training the AI ​​model

[1737] The server trains an AI model using a supervised learning algorithm based on the analyzed data. The training data includes pairs of past queries and their appropriate answers. This gives the model the ability to generate optimal answers to a variety of queries.

[1738] Step 4: Real-time inquiry reception

[1739] The device (PC or smartphone) receives user inquiries in real time. Users input and submit questions using natural language.

[1740] Step 5: Emotion recognition by the emotion engine

[1741] The server receives the user's inquiry and uses an emotion engine to analyze the user's emotions. The emotion engine uses NLP techniques to identify emotions from the inquiry text.

[1742] Step 6: Analyzing the inquiry content

[1743] The server takes sentiment information into account and re-analyzes the query using an NLP engine. This extracts keywords and contextual information, and adds sentiment information.

[1744] Step 7: Generating the answer

[1745] Based on the analysis results, the server generates the most appropriate response for the user's emotional state and inquiry. Links to relevant internal documents and manuals are also added as needed.

[1746] Step 8: Present your answer

[1747] The device displays the generated answers to the user. The user then views the provided answers and associated links on the device screen.

[1748] Step 9: Gathering Feedback and Improving the AI ​​Model

[1749] The device collects user feedback and sends it to the server. The server analyzes this feedback and adds it to the AI ​​model as new training data. This improves the accuracy of the AI ​​model.

[1750] In this way, the system comprehensively analyzes the user's inquiry and emotions, enabling a process that generates the optimal response. This leads to increased operational efficiency and improved employee productivity.

[1751] (Example 2)

[1752] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1753] Traditional inquiry handling systems struggled to provide 24 / 7 support and often suffered from low-quality responses due to their failure to consider the user's emotional state. Furthermore, they were inefficient in handling inquiries because they couldn't effectively utilize all employee communication records and related documents. A system is needed to solve these problems and improve the efficiency of internal operations.

[1754] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1755] In this invention, the server includes means for collecting communication records and related documents of all employees; means for analyzing the collected data using a natural language processing engine and creating an index; means for training a machine learning model based on the analyzed data; means for receiving user inquiries in real time; means for recognizing the user's emotions from the received inquiry content using an emotion engine; means for analyzing the inquiry content taking emotional information into account; means for generating appropriate answers based on the analysis results; means for presenting the generated answers to the user; and means for collecting user feedback and improving the machine learning model. This enables 24 / 7 inquiry support and the provision of high-quality answers that take user emotions into consideration.

[1756] "All employee communication records" refers to digital communication data such as emails and chat logs used by company employees for work purposes.

[1757] "Related documents" refer to documents related to business operations, such as guidelines and manuals, that the company possesses.

[1758] A "natural language processing engine" is software that analyzes text data to extract keywords and perform sentiment analysis.

[1759] A "machine learning model" is a collection of algorithms that learn from data and perform pattern recognition and prediction.

[1760] A "means of receiving in real time" refers to a system or interface that can immediately receive inquiries from users.

[1761] An "emotion engine" is a tool or software that recognizes a user's emotional state from text data.

[1762] "Index creation" refers to techniques that generate lists of keywords and topics to enable rapid searching of large amounts of data.

[1763] "Means for analyzing inquiry content" refers to technologies for understanding the questions and requests entered by users and extracting the necessary information.

[1764] "Means for generating appropriate answers" refers to the process of creating the optimal response to a user's inquiry based on analyzed data.

[1765] "Means of presenting to the user" refers to a system for displaying generated answers and information to the user.

[1766] "Means of collecting feedback" refers to a system for gathering evaluations and opinions from users and using them to improve the system.

[1767] This invention is an intelligent inquiry response system that combines an emotion engine to recognize user emotions. Designed to improve the efficiency of internal operations, this system provides 24 / 7 inquiry support. Furthermore, by analyzing user emotions, it can enhance the quality of responses.

[1768] System Overview

[1769] Data collection

[1770] The server collects communication records (emails, chat logs) of all employees using Microsoft Exchange Server and the Slack API. It also collects relevant documents such as guidelines and manuals from the company's file server and document management system.

[1771] Data analysis and index creation

[1772] The server uses natural language processing engines such as Google NLP and IBM Watson to analyze the collected data and extract key keywords and topics. This creates an efficient data index, enabling rapid information retrieval.

[1773] AI model training

[1774] The server uses libraries such as Hugging Face's Transformers library to train a machine learning model based on the analyzed data. During the training process, the data is split into training and validation sets, and the model parameters are adjusted. The trained AI model is then saved and prepared to generate the optimal response to queries.

[1775] Real-time inquiry reception

[1776] The device (PC or smartphone) receives user inquiries in real time through a web interface built with React or Vue.js. Users input and submit questions in natural language.

[1777] Emotion recognition by an emotion engine

[1778] The server uses emotion engines such as Affectiva or IBM Watson Tone Analyzer to recognize emotions from user inquiries. This data is then used to understand the user's emotional state and generate appropriate responses.

[1779] Analysis of inquiry content

[1780] The server re-analyzes the query content using the spaCy library, taking sentiment information into account. It extracts key keywords and contextual information and performs a detailed analysis based on the user's emotional state.

[1781] Answer generation

[1782] The server uses models such as OpenAI's GPT-3 to generate appropriate responses based on the analysis results. Links to relevant documents and manuals are included as needed. For example, it might generate a response like, "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1783] Providing an answer

[1784] The device displays the response sent from the server to the user. The user can view the generated response and associated links on the device screen.

[1785] Gathering feedback and improving the AI ​​model

[1786] The device collects user feedback and sends it to the server. The server analyzes the feedback, adds it as training data for the model, and improves the accuracy of the AI ​​model by retraining it.

[1787] Specific example

[1788] Data collection

[1789] The server uses the Microsoft Exchange Server API to retrieve employee A's email data and the Slack API to download employee A's chat logs.

[1790] Data analysis and index creation

[1791] The server uses the Google NLP API to analyze keywords such as "how to apply for a new project" and "project guidelines" and creates an index.

[1792] AI model training

[1793] The server uses the Hugging Face Transformers library to train an AI model based on the analyzed data, splitting it into training data and validation data for further training.

[1794] Real-time inquiry reception

[1795] The user enters "How do I submit a new project?" into a web interface built with React and submits it.

[1796] Emotion recognition by an emotion engine

[1797] The server uses IBM Watson Tone Analyzer to recognize the emotion of "confusion" from the user's inquiry.

[1798] Analysis of inquiry content

[1799] The server uses the spaCy library to extract keywords such as "new project" and "application method," and then performs a deeper analysis based on sentiment information.

[1800] Answer generation

[1801] The server uses GPT-3 to generate the response: "To apply for a new project, please use the application form on the portal site. For detailed instructions, please refer to the link below."

[1802] Providing an answer

[1803] The device displays the generated answer and related links on the user's screen.

[1804] Gathering feedback and improving the AI ​​model

[1805] The user enters feedback such as "The answer was helpful" into the React interface, and the device sends this to the server. The server analyzes the feedback data and uses it as training data for the AI ​​model.

[1806] Example of a prompt

[1807] "I'd like to know how to apply for a new project. Please provide a link to the application form on the portal site."

[1808] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1809] Step 1:

[1810] Data collection

[1811] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals) from all employees. Specifically, it uses Microsoft Exchange Server and the Slack API to retrieve data. Input data includes emails and chat messages, while output data includes all communication records and related documents. The server stores this data in a database for later analysis.

[1812] Step 2:

[1813] Data analysis and index creation

[1814] The server analyzes the collected data using a natural language processing (NLP) engine. Specifically, it uses Google NLP or IBM Watson to extract key keywords and topics. The input data consists of collected emails, chat logs, and related documents, while the output data is an index based on keywords and topics. The server creates this index to enable efficient data retrieval.

[1815] Step 3:

[1816] AI model training

[1817] The server trains a machine learning model using the analyzed data. Specifically, it uses the Hugging Face Transformers library to create the model. The input data consists of the analyzed data and extracted indices, while the output data is the trained machine learning model. The server stores this model so that it can generate the best possible answers to queries.

[1818] Step 4:

[1819] Real-time inquiry reception

[1820] The terminal receives user inquiries in real time. Users enter their inquiries through the terminal's web interface (built with React or Vue.js). The input data is the inquiry text entered by the user, and the output data is the transmission of that inquiry text to the server.

[1821] Step 5:

[1822] Emotion recognition by an emotion engine

[1823] The server recognizes emotions from the content of the received inquiry. Specifically, it uses Affectiva or IBM Watson Tone Analyzer. The input data is the user's inquiry, and the output data is the user's emotional state (e.g., confused, angry, happy). The server uses this emotional information for subsequent processing.

[1824] Step 6:

[1825] Analysis of inquiry content

[1826] The server re-analyzes the query content, taking sentiment information into account. Specifically, it uses the spaCy library to extract key keywords and contextual information. The input data consists of the user's query and sentiment information, while the output data consists of the analyzed keywords and contextual information.

[1827] Step 7:

[1828] Answer generation

[1829] The server generates appropriate answers based on the analysis results. Specifically, it uses models such as OpenAI's GPT-3. The input data consists of analyzed keywords and contextual information, and the output data is the generated answer. For example, an answer such as "To apply for a new project, please use the application form on the portal site" might be generated.

[1830] Step 8:

[1831] Providing an answer

[1832] The terminal displays the response sent from the server to the user. The input data is the generated response sent from the server, and the output data is the response displayed on the user's terminal. The user can view the response and related links on the terminal screen.

[1833] Step 9:

[1834] Gathering feedback and improving the AI ​​model

[1835] The device collects user feedback and sends it to the server. The input data is the user's feedback, and the output data is the transmission of that feedback to the server. The server analyzes the feedback data and adds it as training data for a machine learning model. This improves the model's accuracy and the quality of its responses.

[1836] (Application Example 2)

[1837] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1838] Traditional customer support systems often fail to provide adequate support because they generate answers without considering the user's emotions. This is particularly problematic for inquiries involving feelings of confusion or frustration, which can lead to decreased user satisfaction. Furthermore, there is a lack of mechanisms to effectively utilize user feedback to improve AI models, making it difficult to enhance system quality. Therefore, there is a need for an intelligent customer support system that provides high-quality answers and improves user satisfaction.

[1839] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1840] In this invention, the server includes means for collecting communication records and related documents from the entire organization, means for analyzing the collected data and creating an index, means for training an artificial intelligence model based on the analyzed data, means for receiving inquiries from users, means for analyzing the content of the received inquiries, means for recognizing the emotional state of the user using an emotion recognition engine, means for generating an appropriate response based on the analysis results and emotional state, means for presenting the generated response to the user, and means for collecting feedback from the user and improving the artificial intelligence model. This enables the generation of high-quality responses that take into account the user's emotions in real time, improving the quality of inquiry handling and increasing user satisfaction.

[1841] "Communication records of the entire organization" refers to the history of digital communications, such as emails and chat logs, conducted within an organization such as a company or group.

[1842] "Related documents" refer to documents within an organization that are related to business operations, such as guidelines, manuals, memos, and reports.

[1843] "Means of collection" refers to systems that collect necessary information using databases, cloud storage, log management systems, etc.

[1844] "Means of analysis" refer to methods that use natural language processing techniques and machine learning algorithms to convert collected data into a format that can be understood and used.

[1845] "Creating an index" refers to a method of organizing data by adding keywords and metadata so that it can be quickly searched and referenced.

[1846] An "artificial intelligence model" is a computer program that uses machine learning and deep learning algorithms to learn knowledge from data and has the ability to perform specific tasks.

[1847] "Means of receiving inquiries from users" refers to interfaces or applications that receive questions and requests that users make to the system.

[1848] "Means for analyzing received inquiries" refers to methods for interpreting inquiries received from users and extracting information necessary to generate appropriate responses.

[1849] An "emotion recognition engine" is an artificial intelligence technology that identifies a user's emotional state from data such as text and audio.

[1850] "Emotional state" refers to the psychological state a user is in when making an inquiry, and includes, for example, confusion, frustration, or joy.

[1851] "Means for generating appropriate answers" refers to methods for creating specific and helpful answers to user inquiries based on collected and analyzed data.

[1852] "Means of presenting generated answers to users" refers to interfaces or applications that display system-generated answers in a user-friendly format.

[1853] "Means of collecting feedback" refers to methods of receiving opinions and evaluations provided by users and processing them within the system.

[1854] "Methods for improving artificial intelligence models" refer to methods of improving the accuracy and performance of existing artificial intelligence models using collected feedback data.

[1855] This invention provides an intelligent inquiry response system that collects communication records and related documents from the entire organization, trains an artificial intelligence model based on them, and provides appropriate answers to user inquiries. In particular, this system can improve user satisfaction by recognizing the user's emotional state using an emotion recognition engine and generating answers based on that state.

[1856] System Configuration

[1857] 1. Data Collection

[1858] The server collects communication records (emails, chat logs) and related documents (guidelines, manuals, etc.) from the entire organization via cloud storage (e.g., Amazon S3). This data is used for analysis and indexing, as described later.

[1859] 2. Data analysis and index creation

[1860] The server uses a natural language processing engine (e.g., Amazon Comprehend) to analyze the collected data and create an index. This allows for the rapid extraction of relevant information in response to queries.

[1861] 3. Training the AI ​​model

[1862] The server trains an artificial intelligence model (e.g., TensorFlow, PyTorch) based on the analyzed data. This model has the ability to generate the best possible answers to user inquiries.

[1863] 4. Real-time inquiry reception

[1864] The device (e.g., a smartphone application) receives inquiries from users in real time. Users input and submit questions in natural language using the device's interface.

[1865] 5. Emotion recognition

[1866] The server analyzes the received inquiry using an emotion recognition engine (e.g., Amazon Rekognition, IBM Watson Tone Analyzer) to recognize the user's emotional state. This emotional state is a crucial factor in generating the response.

[1867] 6. Answer generation

[1868] The server generates appropriate responses based on the analysis results and recognized emotional states. By using a generative AI model (e.g., OpenAI's GPT-4), it provides highly accurate responses. These responses may include links to relevant internal documents and instructions.

[1869] 7. Answer presentation

[1870] The device displays the generated response to the user. The user can check the response on the device screen.

[1871] 8. Gathering feedback and improving the AI ​​model

[1872] The device collects feedback from users and sends the feedback data to a server. The server uses this feedback to retrain the artificial intelligence model and improve the overall accuracy of the system.

[1873] Specific example

[1874] For example, consider a scenario where a user inquires, "My product hasn't arrived yet." This system can recognize the user's feelings of confusion or frustration and generate a response that includes a gentler tone and specific solutions. An example of an input prompt for the AI ​​model in this case is shown below.

[1875] Example of a prompt:

[1876] User inquiry: "My item hasn't arrived yet."

[1877] Emotional state: "Confused", "Irritated"

[1878] In this way, the present invention makes it possible to generate high-quality responses that take into account the user's emotions in real time, thereby improving the quality of inquiry handling and increasing user satisfaction.

[1879] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1880] Step 1:

[1881] The server collects communication records and related documents from the entire organization's cloud storage. The input to this process is the location where the communication records and related documents are stored (e.g., the cloud storage bucket ID), and the output is the collected data set. This data primarily includes emails, chat logs, guidelines, manuals, etc. The server stores this data in a database for use in the next analysis step.

[1882] Step 2:

[1883] The server analyzes the collected data using a natural language processing engine and creates an index. The input to this process is the collected dataset, and the output is the analyzed data and its index information. The server performs keyword extraction and contextual analysis to generate an index that enables rapid data retrieval.

[1884] Step 3:

[1885] The server trains an artificial intelligence model based on the analysis results. The input to this process is the analyzed data and index information, and the output is the trained AI model. Machine learning libraries (e.g., TensorFlow, PyTorch) are used to train the model. The server tunes the model to optimal performance, improving its ability to generate answers to queries.

[1886] Step 4:

[1887] The terminal receives inquiries from users in real time. The input in this process is a natural language question from the user, and the output is a notification that the inquiry has been received. The terminal sends the text entered by the user to the server for preparation for analysis.

[1888] Step 5:

[1889] The server analyzes the received inquiry using an emotion recognition engine to recognize the user's emotional state. The input for this process is the user's inquiry text, and the output is the recognized emotional state (e.g., confused, irritated). The server utilizes the emotion recognition engine to analyze the emotional tone.

[1890] Step 6:

[1891] The server generates an appropriate response based on the inquiry content and the recognized emotional state. The input to this process is the inquiry content and emotional state, and the output is the generated response. The server uses a generative AI model (e.g., OpenAI's GPT-4) to create specific and helpful responses.

[1892] Step 7:

[1893] The terminal presents the generated response to the user. The input in this process is the generated response sent from the server, and the output is the response displayed in a format visible to the user. The terminal provides the response to the user through a GUI (Graphical User Interface).

[1894] Step 8:

[1895] The terminal collects feedback from users and sends it to the server. The input in this process is the feedback provided by the user, and the output is the recording of the feedback and its transfer to the server. The server uses this feedback as retraining data for machine learning to continuously improve the artificial intelligence model.

[1896] Thus, the system of the present invention can provide highly accurate answers to user inquiries in real time using emotion recognition and an AI model.

[1897] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1898] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1899] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1900] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1901] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1902] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1903] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1904] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1905] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1906] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1907] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1908] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1909] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1910] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1911] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1912] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1913] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1914] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1915] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1916] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1917] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[1918] The following is further disclosed regarding the embodiments described above.

[1919] (Claim 1)

[1920] Means for collecting communication records and related documents of all employees,

[1921] A means of analyzing the collected data and creating an index,

[1922] A method for training an AI model based on the analyzed data,

[1923] A means of receiving inquiries from users,

[1924] A means of analyzing the content of the inquiries received,

[1925] A means for generating an appropriate answer based on the analysis results,

[1926] A means of presenting the generated answer to the user,

[1927] A means of collecting user feedback and improving the AI ​​model,

[1928] A system that includes this.

[1929] (Claim 2)

[1930] The system according to claim 1, characterized in that the communication records of all employees include emails and chat logs.

[1931] (Claim 3)

[1932] The system according to claim 1, characterized in that it includes a link to an internal document or manual related to the generated response.

[1933] "Example 1"

[1934] (Claim 1)

[1935] Means for collecting communication records and business-related information,

[1936] A means of analyzing the collected data using natural language processing technology and creating an index,

[1937] A method for training an artificial intelligence model based on the analyzed data,

[1938] A means of receiving inquiries from users in real time,

[1939] A means of analyzing the content of received inquiries using natural language processing technology,

[1940] A means for generating an appropriate answer based on the analysis results,

[1941] A means of presenting the generated response to the user,

[1942] A means of collecting user feedback to improve artificial intelligence models,

[1943] A system that includes this.

[1944] (Claim 2)

[1945] The system according to claim 1, characterized in that the communication records include emails and chat logs.

[1946] (Claim 3)

[1947] The system according to claim 1, characterized in that it includes a link to an internal document or manual related to the generated response.

[1948] "Application Example 1"

[1949] (Claim 1)

[1950] Means for collecting communication records and related documents of all employees,

[1951] A means of analyzing the collected data and creating an index,

[1952] A method for training an AI model based on the analyzed data,

[1953] A means of receiving inquiries from users,

[1954] A means of analyzing the content of the inquiries received,

[1955] A means for generating an appropriate answer based on the analysis results,

[1956] A means of presenting the generated answer to the user,

[1957] A means of collecting user feedback and improving the AI ​​model,

[1958] A means for factory workers to make inquiries using tablet devices or robots,

[1959] A means of supporting questions regarding factory operating procedures and troubleshooting,

[1960] A system that includes this.

[1961] (Claim 2)

[1962] The system according to claim 1, characterized in that the communication records of all employees include emails and chat logs.

[1963] (Claim 3)

[1964] The system according to claim 1, characterized in that it includes a link to an internal document or manual related to the generated response.

[1965] "Example 2 of combining an emotion engine"

[1966] (Claim 1)

[1967] Means for collecting communication records and related documents of all employees,

[1968] A means of analyzing the collected data using a natural language processing engine and creating an index,

[1969] A method for training a machine learning model based on the analyzed data,

[1970] A means of receiving inquiries from users in real time,

[1971] A means of recognizing a user's emotions from the content of an inquiry received using an emotion engine,

[1972] A method for analyzing inquiry content while taking emotional information into account,

[1973] A means for generating an appropriate answer based on the analysis results,

[1974] A means of presenting the generated answer to the user,

[1975] A means of collecting user feedback and improving machine learning models,

[1976] A system that includes this.

[1977] (Claim 2)

[1978] The system according to claim 1, characterized in that the communication records of all employees include emails and chat logs.

[1979] (Claim 3)

[1980] The system according to claim 1, characterized in that it includes a link to an internal document or manual related to the generated response.

[1981] "Application example 2 when combining with an emotional engine"

[1982] (Claim 1)

[1983] Means for collecting communication records and related documents from the entire organization,

[1984] A means of analyzing the collected data and creating an index,

[1985] A method for training an artificial intelligence model based on the analyzed data,

[1986] A means of receiving inquiries from users,

[1987] A means of analyzing the content of the inquiries received,

[1988] A means of recognizing the user's emotional state using an emotion recognition engine,

[1989] A means for generating an appropriate response based on the analysis results and emotional state,

[1990] A means of presenting the generated response to the user,

[1991] A means of collecting user feedback and improving artificial intelligence models,

[1992] A system that includes this.

[1993] (Claim 2)

[1994] The system according to claim 1, characterized in that the communication records of the entire organization include emails and chat logs.

[1995] (Claim 3)

[1996] The system according to claim 1, characterized in that it includes a link to internal documents or instructions related to the generated response. [Explanation of symbols]

[1997] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. Means for collecting communication records and related documents of all employees, A means of analyzing the collected data and creating an index, A method for training an AI model based on the analyzed data, A means of receiving inquiries from users, A means of analyzing the content of the inquiries received, A means for generating an appropriate answer based on the analysis results, A means of presenting the generated answer to the user, A means of collecting user feedback and improving the AI ​​model, A system that includes this.

2. The system according to claim 1, characterized in that the communication records of all employees include emails and chat logs.

3. The system according to claim 1, characterized in that it includes a link to an internal document or manual related to the generated response.

Images