System

Abstract

A system is provided.SOLUTION: A system comprising: means for generating a prompt optimized for each company; means for generating a sentence by a AI model based on a human relation in the company and a past message pattern; means for transmitting the generated sentence to a user's device and receiving manual adjustment by the user; means for accumulating feedback from the user as training data and reflecting the feedback in future prompt and sentence generation; and means for transmitting a finally adjusted sentence.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In daily work, much time is spent writing emails and chat messages. However, these communications often contain important content, making it difficult to reduce the time spent on them. Furthermore, each company has its own unique culture and relationships, and generating uniform text without taking these into consideration can hinder effective communication. Therefore, there is a need for an efficient and effective means of writing text that takes into account a company's organizational structure, internal relationships, and past message patterns. [Means for solving the problem]

[0005] This invention provides a system that generates prompts optimized for each company and generates sentences using an AI model, taking into account internal human relationships and past message patterns. Specifically, the system includes a means for generating prompts optimized for each company and a means for the AI ​​model to generate sentences based on internal human relationships and past message patterns. The system also includes a means for sending the generated sentences to a user's device and accepting manual adjustments by the user, a means for accumulating user feedback as learning data and reflecting it in future prompt and sentence generation, and a means for sending the final adjusted sentences, thereby achieving efficient and effective sentence creation. This enables business efficiency and high-quality communication.

[0006] A "prompt" is a sentence generation instruction or guideline that is input to an AI model, containing the information needed to achieve a specific output.

[0007] An "AI model" is an artificial intelligence that analyzes data based on machine learning algorithms and makes predictions and generation decisions.

[0008] "Sentence generation" is the process by which an AI model creates new sentences based on given prompts and data.

[0009] "Feedback" is information that allows the system to learn and improve based on the evaluations and adjustments made by the user to the generated text.

[0010] "Terminal" means an information device used by a user to input text or to review and adjust generated text.

[0011] "Means" refer to the methods or techniques used by a system to achieve a specific function or role.

[0012] "Internal relationships" refers to the roles and relationships between employees within a company, and are factors that influence the tone and content of writing.

[0013] A "message pattern" refers to the tendency of the format and content of text sent and received in the past, and serves as the basis for newly generated text.

[0014] "Training data" refers to data collected and analyzed to continuously improve AI models, and primarily includes feedback information.

[0015] "User" means a person who uses the system to input, review, adjust, and submit text. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11]FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0017] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0018] First, the terms used in the following description will be explained.

[0019] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0020] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0021] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0022] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0023] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0027] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0028] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0029] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0030] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0031] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0034] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0035] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0037] This invention is a system in which AI generates sentences using prompts optimized for each company, taking into account internal human relationships and past message patterns. Below, each component and operation of this system are explained in natural language.

[0038] Software Configuration

[0039] This system is mainly composed of a server, a terminal, and a user. The specific operation of each component is explained below.

[0040] server

[0041] 1. Receiving input data

[0042] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.).

[0043] This data is sent in formats such as JSON or XML.

[0044] 2. Prompt Optimization

[0045] The server generates prompts based on the input data, which are customized for each company and optimized to fit their specific organizational structure and culture.

[0046] For example, this includes the relationship between a specific boss and a subordinate, or past email patterns.

[0047] 3. Sentence generation

[0048] The server inputs the optimized prompts into the AI ​​model and generates sentences according to the specified format.

[0049] The AI ​​model uses machine learning algorithms to generate sentences based on the input prompts.

[0050] 4. Learning Feedback

[0051] It receives manual adjustments and feedback from users and stores them as learning data.

[0052] The feedback will identify areas that need adjustment and will be used to generate future prompts and improve the AI ​​model.

[0053] 5. Sending text

[0054] The final adjusted text is sent to the specified destination.

[0055] Sending is carried out in conjunction with email servers and internal communication tools.

[0056] Terminal

[0057] 1. Data input interface

[0058] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[0059] 2. Displaying the generated text

[0060] The generated text returned from the server is displayed in the interface.

[0061] The generated text is highlighted and annotated as appropriate to make it easier for users to check.

[0062] 3. Manual Adjustment Form

[0063] Provide a form that allows the user to manually adjust the generated text.

[0064] It has the function of sending the adjusted text to the server as feedback.

[0065] User

[0066] 1. Entering data

[0067] The user uses the terminal interface to input information about the content they want to send and the recipient.

[0068] For example, enter information such as "Notification of new project start" and "Team member roles."

[0069] 2. Check and adjust the generated text

[0070] Check the generated text returned by the server and make manual adjustments as necessary.

[0071] Once the adjustments are complete, the final text is sent to the server as feedback.

[0072] 3. Final submission

[0073] Check the final text after adjustments and press the send button to send it to the specified recipient.

[0074] Specific examples

[0075] For example, to send a project launch notice, the user might enter:

[0076] 1. User (Device)

[0077] Enter the project name, start date, and team member information as "Notification of new project start."

[0078] 2. Server

[0079] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[0080] 3. Server

[0081] The generated text is sent to the user's terminal.

[0082] 4. User (Terminal)

[0083] Review the generated text and edit it if necessary.

[0084] 5. User (Terminal)

[0085] The edited text is sent to the server as feedback.

[0086] 6. Server

[0087] Add feedback to the training data to help improve it further.

[0088] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[0089] The processing flow will be explained below.

[0090] Step 1:

[0091] The device displays an interface for the user to input information about what they want to send and who they want to send it to, including detailed project name, start date, team member roles, etc.

[0092] Step 2:

[0093] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[0094] Step 3:

[0095] The server receives the data sent from the device, parses it, and extracts the necessary information, such as the project name, start date, and team members.

[0096] Step 4:

[0097] The server uses the extracted data to generate personalized prompts tailored to the company's culture, organizational structure, and internal relationships.

[0098] Step 5:

[0099] The server then inputs the generated prompts into an AI model, which uses machine learning algorithms to generate sentences based on the prompts, with language and tone that reflects the company's culture.

[0100] Step 6:

[0101] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[0102] Step 7:

[0103] The user reviews the generated text and manually adjusts it as needed, for example by correcting specific job titles or the roles of specific members.

[0104] Step 8:

[0105] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[0106] Step 9:

[0107] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for future prompt optimization and AI model improvement.

[0108] Step 10:

[0109] The user checks and adjusts the final text and sends it from the device. The device then works in conjunction with the mail server and internal communication tools to send the text to the specified destination.

[0110] This series of processing steps enables users to efficiently generate business documents and realize communication that is in line with the unique culture and background of the company.

[0111] Example 1

[0112] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0113] In today's corporate environment, generating sentences that are efficient and reflect the organization's unique culture and interpersonal relationships is extremely important. However, conventional sentence generation systems lack a mechanism for generating sentences using prompts optimized for each company, requiring significant manual adjustments. Furthermore, they lack a means to effectively utilize user feedback to improve the sentence generation algorithm. This reduces the efficiency of the entire sentence generation process and significantly increases the time and effort required.

[0114] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0115] In this invention, the server includes means for receiving input data and generating optimized prompts based on the data, means for inputting the optimized prompts into an AI model and generating sentences based on internal human relationships and past message patterns, means for sending the generated sentences to a user's terminal and accepting manual adjustments by the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, and means for sending the final adjusted sentences to a specified destination. This makes it possible to efficiently generate sentences optimized for each company and significantly reduce the user's effort.

[0116] "Input data" refers to information provided by users to the system, including project overviews and team member information.

[0117] A "prompt" is an instruction or question that is input into an AI model and serves as the basis for generating sentences.

[0118] "Optimized prompts" refer to prompts that are customized based on the company's characteristics, culture, internal relationships, etc.

[0119] An "AI model" is an artificial intelligence program built using machine learning algorithms that generates natural language sentences based on prompts.

[0120] "Generated sentences" refer to sentences generated by the AI ​​model based on optimized prompts.

[0121] "User terminal" refers to a device such as a computer or smartphone used by a user to input data and check generated text.

[0122] "Manual adjustment" refers to the act of a user manually correcting or editing the generated text.

[0123] "Feedback" refers to information that users send back to the system regarding adjusted text and improvements.

[0124] "Training data" is a data set that the system uses to improve its performance, and includes feedback from users.

[0125] "Designated Destination" refers to the address or contact information indicating the specific recipient to whom the final tailored document is to be sent.

[0126] This invention is a system in which a user inputs data using a terminal, generates optimized prompts based on that data, and generates sentences under specific conditions using an AI model. Here, we will explain the specific names of the hardware and software used and the program's processing procedures.

[0127] System configuration

[0128] This system mainly consists of a server, terminals, and users.

[0129] server

[0130] The server has the following roles:

[0131] 1. Receiving input data

[0132] The server receives the data sent by the user from the device. The data is usually sent in a format such as JSON or XML. The received data is stored in memory and used for subsequent processing.

[0133] 2. Prompt optimization

[0134] The server generates prompts based on the input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns.To generate and optimize these prompts, Python scripts and database management systems (e.g., MySQL or PostgreSQL) are used.

[0135] 3. Sentence generation

[0136] The server inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format. The AI ​​model is built using machine learning libraries such as TensorFlow and PyTorch, and generates sentences based on the prompts using natural language processing techniques.

[0137] 4. Sending the generated text

[0138] The server encodes the generated text into JSON format and sends it to the terminal as an HTTP response.

[0139] 5. Learning Feedback

[0140] The server receives user feedback and stores it in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future.

[0141] Terminal

[0142] The terminal has the following roles:

[0143] 1. Providing a data input interface

[0144] The terminal provides the user with a data entry interface that allows them to enter information about the content and audience they wish to send. This interface is built using HTML, CSS, and JavaScript.

[0145] 2. Displaying the generated text

[0146] The terminal displays the generated text returned from the server on its interface, with appropriate highlighting and annotations added to make it easier for the user to check.

[0147] 3. Provide a manual adjustment form

[0148] The terminal provides a form that allows the user to manually adjust the generated text, and sends the adjustment results to the server as feedback.

[0149] User

[0150] The user has the following roles:

[0151] 1. Entering data

[0152] The user uses the device's data entry interface to enter information about the content and recipients of the message, such as the project name, start date, and team member information for "Notice of the start of a new project."

[0153] 2. Check and adjust the generated text

[0154] The user checks the generated sentences sent back from the server and manually adjusts them if necessary. The adjusted sentences are then sent back to the server as feedback.

[0155] Specific examples

[0156] For example, the specific procedure for sending a notification of the start of a new project is as follows.

[0157] 1. User (Device)

[0158] Enter the project name, start date, and team member information as "Notification of new project start." A specific example of how to enter this information is shown below.

[0159] Document Title: Notice of New Project Launch

[0160] Project Name: Project X

[0161] Start date: October 1, 2023

[0162] Team members: Mr. A, Mr. B, Mr. C

[0163] 2. Server

[0164] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[0165] 3. Server

[0166] The generated text is sent to the user's terminal.

[0167] 4. User (Terminal)

[0168] Review the generated text and edit it if necessary.

[0169] 5. User (Terminal)

[0170] The edited text is sent to the server as feedback.

[0171] 6. Server

[0172] Add feedback to the training data to help improve it further.

[0173] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[0174] In this way, the present invention realizes optimal sentence generation adapted to each company.

[0175] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0176] Step 1:

[0177] The user enters the data.

[0178] Specific behavior:

[0179] The user uses the device's data entry interface to enter the content they want to send (e.g., "Notification of the start of a new project") and target information (e.g., project name, start date, team member roles). This input data will be the basis for subsequent processing.

[0180] Input: Project name, start date, team member information

[0181] Output: Data entered by the user

[0182] Step 2:

[0183] The terminal sends the input data to the server.

[0184] Specific behavior:

[0185] The terminal encodes the data entered by the user in JSON format and sends it to the server using the HTTPS protocol, which ensures secure transmission.

[0186] Input: Data entered by the user

[0187] Output: Encoded data in JSON format

[0188] Step 3:

[0189] The server receives the input data.

[0190] Specific behavior:

[0191] The server receives the data sent from the terminal and stores it in memory. It is also possible to temporarily store the data using a database management system (e.g., MySQL or PostgreSQL).

[0192] Input: JSON encoded data

[0193] Output: Data stored in memory

[0194] Step 4:

[0195] The server optimizes the prompt.

[0196] Specific behavior:

[0197] The server generates prompts based on input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns. Python scripts are used to extract relevant information from the database and reflect it in the prompts.

[0198] Input: Data stored in memory

[0199] Output: Optimized prompt

[0200] Step 5:

[0201] The server inputs the prompts into the AI ​​model to generate sentences.

[0202] Specific behavior:

[0203] The server inputs the optimized prompts into a generative AI model, which is built using machine learning libraries such as TensorFlow and PyTorch, and generates natural language sentences based on the prompts.

[0204] Input: Optimized prompts

[0205] Output: Generated sentence

[0206] Step 6:

[0207] The server sends the generated text to the terminal.

[0208] Specific behavior:

[0209] The server encodes the generated text again into JSON format and sends it to the terminal as an HTTP response.

[0210] Input: Generated sentence

[0211] Output: The generated document encoded in JSON format.

[0212] Step 7:

[0213] The terminal displays the generated text to the user.

[0214] Specific behavior:

[0215] The device then displays the generated text in an interface, styled using HTML and CSS to highlight specific feeds and important information.

[0216] Input: The generated text encoded in JSON format

[0217] Output: The generated text displayed to the user

[0218] Step 8:

[0219] The user reviews and adjusts the generated text.

[0220] Specific behavior:

[0221] The user can review the generated text displayed on the device and manually adjust it as needed, for example, by changing specific wording or adjusting the writing style. Adjustments are made using a form on the device.

[0222] Input: The generated text displayed to the user

[0223] Output: User-adjusted text

[0224] Step 9:

[0225] The user sends the adjusted text to the server as feedback.

[0226] Specific behavior:

[0227] The user then sends the adjusted document back to the server from their device, again encoded in JSON and using the HTTPS protocol.

[0228] Input: User-adjusted text

[0229] Output: The adjusted text encoded in JSON format

[0230] Step 10:

[0231] The server takes the feedback as learning.

[0232] Specific behavior:

[0233] The server stores the received feedback in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future. A database management system (e.g., MySQL or PostgreSQL) is used for storage.

[0234] Input: Adjusted text encoded in JSON format

[0235] Output: Feedback data stored in a database

[0236] (Application example 1)

[0237] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0238] The food delivery industry requires fast and accurate customer support, but traditional methods have led to inconsistencies in response time and quality. It has also been difficult to share know-how and make continuous improvements to properly respond to inquiries.

[0239] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0240] In this invention, the server includes means for generating prompts optimized for each company, means for an AI-generated model to generate sentences based on internal human relationships and past message patterns, means for sending the generated sentences to an information processing terminal and accepting manual adjustments from the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, means for sending the final adjusted sentences to another information processing terminal, and means for generating prompts and sentences for customer support inquiries specific to the food delivery industry. This enables fast and accurate customer support responses, standardizes the quality of responses, and enables continuous improvement.

[0241] "Means for generating prompts optimized for each company" refers to methods or functions for automatically creating optimal prompts based on each company's specific organizational structure, culture, and past message patterns.

[0242] "Means for an AI generation model to generate sentences based on internal interpersonal relationships and past message patterns" refers to a method or function that uses input data based on specific internal interpersonal relationships and past message patterns to generate appropriate sentences using an AI generation model.

[0243] "Means for transmitting the generated text to an information processing terminal and accepting manual adjustments by the user" refers to a method or function that transmits text generated by an AI-generated model to a terminal used by the user and enables the user to manually adjust the text.

[0244] "Means for accumulating user feedback as learning data and reflecting it in future prompts and sentence generation" refers to a method or function that accumulates the user's manually adjusted content and feedback in a database and uses it when generating the next prompt or sentence.

[0245] "Means for sending the final adjusted text to another information processing terminal" refers to a method or function for sending the text that has been final checked and adjusted by the user to a designated recipient.

[0246] "Means for generating prompts and generating sentences to respond to customer support inquiries specialized for the food delivery industry" refers to methods and functions for generating prompts and AI-based sentences necessary to respond to customer support inquiries specialized for the content of such inquiries in the food delivery industry.

[0247] The "means for inputting data including inquiry content" refers to a method or function for a user to input data such as customer support inquiry content and customer information.

[0248] The "means for analyzing data and generating an optimized prompt" refers to a method or function for analyzing input data and automatically generating an optimal prompt.

[0249] "Means for analyzing the details of manual adjustments made by the user and storing the details as feedback" refers to a method or function for analyzing the details of manual adjustments made by the user and storing the information as feedback in a database.

[0250] "Means for improving the AI-generated model and prompt-generation algorithm based on accumulated feedback" refers to methods or functions for continuously improving the AI-generated model and prompt-generation algorithm using accumulated feedback data.

[0251] This invention aims to improve the efficiency of customer support in the food delivery industry by using prompts optimized for each company and AI-generated models. This system mainly consists of a server, a terminal, and a user.

[0252] server

[0253] The server performs the following series of processes.

[0254] 1. Receiving input data

[0255] The server receives data (e.g., inquiry details, customer information) entered by the user from the terminal. The data is usually sent in a format such as JSON or XML.

[0256] 2. Prompt Generation

[0257] The server generates prompts based on the input data, which are specific to the food delivery industry and include the following examples:

[0258] "Inquiry: My order hasn't arrived"

[0259] 3. Sentence generation

[0260] The server inputs the optimized prompts into a generative AI model (e.g., GPT-3) to generate appropriate sentences for the query. The generative AI model generates sentences using machine learning algorithms.

[0261] 4. Learning Feedback

[0262] It receives manual adjustments and feedback from users and stores it as training data, identifying areas that need adjustment and helping to generate future prompts and improve the AI ​​model.

[0263] 5. Sending text

[0264] Finally, the adjusted text is sent to the designated recipient, linked to the email server or internal instant communication tool.

[0265] Terminal

[0266] The terminal provides the following features:

[0267] 1. Data input interface

[0268] It provides an interface for users to use and allows them to input necessary information (e.g., inquiry details, customer information, etc.).

[0269] 2. Displaying the generated text

[0270] The generated text sent back from the server is displayed in the interface, and it is also possible to highlight and annotate the generated text for the user's convenience.

[0271] 3. Manual Adjustment Form

[0272] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[0273] User

[0274] Users interact with the system as follows:

[0275] 1. Entering data

[0276] The user uses the terminal interface to input the inquiry and customer information. For example, the user may input "The food I ordered has not arrived."

[0277] 2. Check and adjust the generated text

[0278] Check the generated text sent back from the server and manually adjust it if necessary. For example, adjust the text to something like, "We apologize for the inconvenience. We will recheck it immediately. While you wait, we will provide you with a 1,000 yen discount coupon."

[0279] 3. Providing Feedback

[0280] The adjusted sentences are sent to the server as feedback and stored as learning data in the system.

[0281] This system enables quick and accurate customer support responses, standardizes the quality of responses, and allows for continuous improvement. As a specific example of its use, the server generates the following prompt sentence:

[0282] Inquiry: The food I ordered hasn't arrived. A customer who has made a similar inquiry in the past frequently orders sushi.

[0283] The sentence generated based on this is, "We are sorry to hear that your sushi order has not yet arrived, and we apologize for the inconvenience. We will recheck for you shortly. While you wait, we will provide you with a 1,000 yen discount coupon," which is then adjusted on the device and sent as the final sentence.

[0284] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0285] Step 1:

[0286] The user uses the terminal interface to input the inquiry and customer information. At this time, the inquiry information is entered as "The food I ordered has not arrived." This input data is sent to the server in JSON format.

[0287] Step 2:

[0288] The server analyzes the input data received from the device and generates a prompt optimized for each company. Specifically, it takes into account data such as past inquiry patterns and customer preferences to generate a prompt like the one below.

[0289] "Inquiry: The food I ordered hasn't arrived. This customer has made a similar inquiry in the past and frequently orders sushi."

[0290] Step 3:

[0291] The server that generated the prompt inputs the prompt into a generative AI model (e.g., GPT-3). The AI ​​model generates an appropriate sentence based on the prompt. In doing so, the AI ​​analyzes the input data and uses patterns learned from previous inquiries. An example of a generated sentence might be, "We're sorry to hear that your sushi order hasn't arrived yet, and we apologize for the inconvenience. We'll check again soon. While you wait, we'll provide you with a 1,000 yen discount coupon."

[0292] Step 4:

[0293] The server sends the generated text to the terminal, where the user can review it on the terminal interface. The generated text may also be highlighted or annotated to make it easier for the user to review.

[0294] Step 5:

[0295] The user can review the generated text and manually adjust it if necessary, for example to include more specific instructions or additional information, which is then sent back to the server.

[0296] Step 6:

[0297] The server receives user feedback on the adjustments made, which is stored in a database and used to improve future prompt and sentence generation.

[0298] Step 7:

[0299] The final adjusted text is then sent to the specified destination by the server, enabling quick and accurate customer support, improving the quality of customer support in the food delivery industry.

[0300] This system's series of processes will significantly improve the quality and efficiency of customer support responses in the food delivery industry. For example, it will be able to respond quickly to complex customer needs and continuously improve the quality of responses by utilizing learning data.

[0301] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0302] This invention combines an emotion engine that recognizes the user's emotions with a system in which AI generates sentences using prompts optimized for each company and taking into account internal human relationships and past message patterns. Below, each component and operation of this system is explained in natural language.

[0303] Software Configuration

[0304] This system is primarily composed of a server, a terminal, and a user. By introducing an emotion engine, it becomes possible to generate text that takes into account the user's emotional state, achieving more effective communication. The specific operation of each component is explained below.

[0305] server

[0306] 1. Receiving input data

[0307] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.) This data is sent in a format such as JSON or XML.

[0308] 2. Emotion recognition

[0309] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[0310] 3. Prompt Optimization

[0311] The server generates optimized prompts based on the user's emotion recognition data and customized prompt templates for each company, which are tailored to fit the specific organizational structure and culture.

[0312] 4. Sentence generation

[0313] The server then inputs the optimized prompts into the AI ​​model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[0314] 5. Learning Feedback

[0315] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[0316] 6. Sending text

[0317] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[0318] Terminal

[0319] 1. Data input interface

[0320] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[0321] 2. Displaying the generated text

[0322] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[0323] 3. Manual Adjustment Form

[0324] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[0325] User

[0326] 1. Entering data

[0327] The user uses the device interface to input information about the content and recipients of the message, such as "notification of the start of a new project" or "roles of team members."

[0328] 2. Check and adjust the generated text

[0329] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[0330] 3. Final submission

[0331] Check the final text after adjustments and press the send button to send it to the specified recipient.

[0332] Specific examples

[0333] For example, consider sending a notification about the start of a new project. The specific steps are as follows:

[0334] 1. User (Device)

[0335] Enter the project name, start date, and team member information as "Notification of new project start."

[0336] 2. Server

[0337] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[0338] 3. Emotion recognition

[0339] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state and reflects it in the prompt generation.

[0340] 4. Server

[0341] The generated text is sent to the user's terminal.

[0342] 5. User (Terminal)

[0343] Review the generated text and edit it if necessary.

[0344] 6. User (Terminal)

[0345] The edited text is sent to the server as feedback.

[0346] 7. Server

[0347] The feedback is accumulated as learning data and used to generate future prompts and improve the AI ​​model.

[0348] This system allows users to generate optimal sentences according to their emotional state, enabling effective communication that is in line with the company's unique culture and background.

[0349] The processing flow will be explained below.

[0350] Step 1:

[0351] The terminal displays an interface for the user to enter data: the content they want to send and information about the recipient (e.g., project name, start date, team member roles, etc.).

[0352] Step 2:

[0353] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[0354] Step 3:

[0355] The server receives the data sent from the device, parses it, and extracts information about the project name, start date, and team members.

[0356] Step 4:

[0357] The server uses an emotion engine to recognize the user's emotions, identifying emotions such as joy, anger, and sadness based on the user's input data and past data.

[0358] Step 5:

[0359] The server uses the user's emotional state and a customized prompt template for each company to generate optimized prompts that are tailored to the specific organizational structure and culture.

[0360] Step 6:

[0361] The server inputs the generated prompts into an AI model to generate sentences, which then use machine learning algorithms to generate sentences that reflect the user's emotional state.

[0362] Step 7:

[0363] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[0364] Step 8:

[0365] The user reviews the generated text and makes manual adjustments as needed, for example, correcting specific job titles or member roles.

[0366] Step 9:

[0367] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[0368] Step 10:

[0369] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for use in generating future prompts and improving the AI ​​model.

[0370] Step 11:

[0371] The user then checks the final edited text and presses the send button to send it to the specified recipient. The device then sends the text in conjunction with the email server and internal communication tools.

[0372] Specific examples

[0373] Example: New project start notification

[0374] 1. Step 1:

[0375] The user inputs the project name, start date, and team member information into the terminal interface as a "new project start notification."

[0376] 2. Step 2:

[0377] The terminal sends the input data to the server.

[0378] 3. Step 3:

[0379] The server receives the data and analyzes the content.

[0380] 4. Step 4:

[0381] The server uses an emotion engine to recognize emotions from the user's input. For example, if the user includes many positive comments, the emotion engine will recognize the emotion as "joy."

[0382] 5. Step 5:

[0383] The server generates optimized prompts based on the emotional state and the company's prompt templates.

[0384] 6. Step 6:

[0385] The server inputs prompts into the AI ​​model to generate sentences that reflect the user's emotional state of "joy."

[0386] 7. Step 7:

[0387] The server sends the generated text to the terminal.

[0388] 8. Step 8:

[0389] The user reviews the generated text and manually adjusts it if necessary.

[0390] 9. Step 9:

[0391] The device sends the final text including the adjustments to the server as feedback.

[0392] 10. Step 10:

[0393] The server receives the feedback, analyzes the content, and stores it as learning data.

[0394] 11. Step 11:

[0395] The user then clicks the send button to send the final text that has been adjusted. The device then connects to the mail server and sends the text to the specified recipient.

[0396] Example 2

[0397] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0398] In corporate communication, it is important to generate texts that are optimized for each company's unique culture and organizational structure. At the same time, it is also necessary to consider the user's emotional state when generating texts. However, existing systems often lack sufficient emotion recognition and optimization for each company, hindering effective communication. Furthermore, they lack mechanisms for manual adjustment of generated texts and efficient feedback integration. This leads to problems in improving text quality and reducing user satisfaction.

[0399] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0400] In this invention, the server includes: a means for generating prompts optimized for each company; a means for an AI model to generate sentences based on internal relationships and past message patterns; a means for sending the generated sentences to a user's device and accepting manual adjustments from the user; a means for accumulating user feedback as learning data and reflecting it in future prompt and sentence generation; a means for sending the final adjusted sentences; a means for receiving and preprocessing the sent data; a means for analyzing input data using an emotion engine to recognize the user's emotional state; a means for optimizing prompts based on the recognized emotional data; and a means for inputting the generated prompts into a generative AI model to generate sentences. This enables effective sentence generation that is optimized for each company's unique culture and organizational structure and takes the user's emotional state into consideration. Furthermore, by efficiently incorporating user feedback, the quality of the generated sentences and user satisfaction can be improved.

[0401] The "server" is a device that receives data sent by users, analyzes it, generates appropriate sentences using a generative AI model, and finally transmits them.

[0402] A "terminal" is a device that provides an interface for users to operate, input data, and review and adjust the generated text.

[0403] "Company-optimized prompts" are guidance or instructions that are customized to fit a specific company's culture and organizational structure.

[0404] An "AI model" is an artificial intelligence algorithm that performs natural language processing based on input data and automatically generates sentences in a specified format.

[0405] An "emotion engine" is a program or device that analyzes and recognizes the emotional state of a user from data entered by the user.

[0406] A "prompt" is a predetermined instruction or explanation that is input into an AI model.

[0407] "Feedback" is information provided by a user when conveying corrections or opinions about the generated text to the server.

[0408] "Preprocessing" is a process performed by the server to prepare the transmitted data it receives in a format that is easy to analyze.

[0409] "Optimized prompts" are instructions or explanations that are optimized to take into account the user's emotional state and company-specific factors.

[0410] A "generative AI model" is a system that uses artificial intelligence technology to generate sentences based on user input data and optimized prompts.

[0411] "Sentence generation" is the process by which an AI model creates sentences in natural language based on optimized prompts.

[0412] The present invention provides a sentence generation system that combines an emotion engine and a generative AI model using prompts optimized for each company to effectively communicate within the company. Hereinafter, an embodiment of the present invention will be described in detail.

[0413] System configuration

[0414] server

[0415] A server is a device that performs several major functions:

[0416] 1. Receiving input data

[0417] The server receives data sent from the device (e.g., project name, team member information, etc.) This data is sent in JSON or XML format.

[0418] 2. Emotion recognition

[0419] The server analyzes the received data and uses an emotion engine to identify the user's emotional state. For example, the emotion toward the project name "Next Generation AI Development" is recognized as "joy."

[0420] 3. Prompt Optimization

[0421] The server generates optimized prompts based on the emotion recognition results and company-specific templates. For example, if the emotion is "joy," a positive-toned prompt is used.

[0422] 4. Sentence generation

[0423] The server then inputs the optimized prompts into a generative AI model, specifically OpenAI's GPT-3, to generate the final sentence.

[0424] 5. Learn and incorporate feedback

[0425] The server receives manual adjustments and feedback from users and stores that information as learning data to help generate future prompts and improve the AI ​​model.

[0426] 6. Sending text

[0427] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[0428] Terminal

[0429] The device is directly operated by the user and provides the following functions:

[0430] 1. Providing a data input interface

[0431] It provides an interface for users to enter data (e.g., project name, start date, team member information, etc.).

[0432] 2. Displaying the generated text

[0433] It has the function of displaying the generated text in an appropriate format so that the user can check it.

[0434] 3. Provide a manual adjustment form

[0435] A form is provided for users to manually adjust the generated text, and the adjusted text is sent to the server as feedback.

[0436] User

[0437] The user is the person who operates this system and performs the following operations.

[0438] 1. Entering data

[0439] Use the device to enter information about the content you want to send and the recipient. For example, for a "Notification of the start of a new project," enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki."

[0440] 2. Check and adjust the generated text

[0441] Check the generated text returned by the server and make manual adjustments as necessary. For example, change "Tanaka Ichiro" to "Tanaka Saburo."

[0442] 3. Final submission

[0443] Check the final text after adjustments and press the send button to send it to the specified recipient.

[0444] Specific examples

[0445] For example, to send an announcement about the launch of a new project, the user types the following at the terminal:

[0446] "New project start notification"

[0447] "Project name = Next generation AI development"

[0448] "Start date=October 1, 2023"

[0449] "Team members: Ichiro Tanaka, Jiro Suzuki"

[0450] The server receives this data, and uses the emotion engine to generate prompts based on the emotion it recognizes as "joy," ultimately generating the following sentence:

[0451] "The next-generation AI development project has begun. We appreciate your cooperation."

[0452] The generated text is displayed on the user's device, and after the user confirms and adjusts it, the final text is sent to the specified destination.

[0453] This system can generate sentences that are suited to each company's unique culture and organizational structure, and can also provide effective communication that takes into account the user's emotional state. Furthermore, by efficiently incorporating feedback, the quality of the generated sentences can be improved.

[0454] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0455] Step 1: Data entry

[0456] Subject: User

[0457] Users use the device interface to enter information about the content and recipients they want to send, such as the project name, the start date of the new project, and team member information.

[0458] Input: Project name, start date, team member information

[0459] Output: Input data

[0460] For example, a user might enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki" as a "Notification of the start of a new project."

[0461] Step 2: Submitting input data

[0462] Subject: Terminal

[0463] The terminal sends the data entered by the user to the server in JSON or XML format.

[0464] Input: User-entered data (e.g., project name, start date, team member information)

[0465] Output: Data sent to the server

[0466] Specifically, the terminal converts the user's input data into an appropriate format and transmits it to the server.

[0467] Step 3: Receiving input data

[0468] Subject: Server

[0469] The server receives the data sent from the device. Since the data is in JSON or XML format, it parses it and converts it into a data structure for analysis.

[0470] Input: Data sent from the terminal (JSON or XML)

[0471] Output: Parsable data structure

[0472] The server performs preprocessing to analyze the received data, converting it into a format such as a string or a number, and stores it in an internal data structure.

[0473] Step 4: Emotion Recognition

[0474] Subject: Server

[0475] The server analyzes the received data using an emotion engine to recognize the user's emotional state. It processes the data to identify the emotional state that can be inferred from the input data.

[0476] Input: A parsable data structure

[0477] Output: Emotional state data (e.g., happy, angry, sad)

[0478] The server's emotion engine identifies the user's emotion as "happiness" based on the input data and passes the result to the next processing step.

[0479] Step 5: Prompt optimization

[0480] Subject: Server

[0481] The server generates optimized prompts based on emotion recognition results and templates customized for each company. The template and emotion data are combined to create prompts tailored to the customer's emotional state.

[0482] Input: Emotional state data, company-specific templates

[0483] Output: Optimized prompt

[0484] Specifically, if the emotional state is "joy," a prompt with a positive tone (e.g., "The next-generation AI development project is starting!") is generated.

[0485] Step 6: Sentence generation

[0486] Subject: Server

[0487] The server inputs the optimized prompts into a generative AI model (e.g., OpenAI GPT-3) to generate natural language sentences that are consistent with the company's culture and reflect the user's emotional state.

[0488] Input: Optimized prompts

[0489] Output: Generated sentence

[0490] Specifically, based on optimized prompts, it generates sentences such as, "The next-generation AI development project has begun. We appreciate your cooperation."

[0491] Step 7: Displaying the generated sentences

[0492] Subject: Terminal

[0493] The terminal displays the generated text returned by the server to the user, with appropriate formatting and highlighting.

[0494] Input: Generated text returned by the server

[0495] Output: Text for display

[0496] The displayed text is set up so that it is easy for the user to check, and the content is made easier to understand by highlighting and annotating it.

[0497] Step 8: Manual adjustment

[0498] Subject: User

[0499] The user can review the generated text and manually adjust it if necessary, for example, to correct misspelled names or wording.

[0500] Input: Generated text displayed

[0501] Output: Adjusted text

[0502] For example, the user corrects "Tanaka Ichiro" to "Tanaka Saburo" and sends the content to the next step.

[0503] Step 9: Send your feedback

[0504] Subject: Terminal

[0505] The device sends the user's manual adjustments to the server as feedback, which is then stored as learning data.

[0506] Input: Adjusted text

[0507] Output: Feedback sent to the server

[0508] The device will then send the user's manual adjustments back to the server and use them as feedback.

[0509] Step 10: Learning feedback

[0510] Subject: Server

[0511] The server accumulates the feedback received from the user as learning data and uses it for future sentence generation and prompt optimization. It analyzes the feedback and updates the learning data.

[0512] Input: Adjusted feedback data

[0513] Output: Updated training data

[0514] For example, we will improve our AI models and prompt generation algorithms based on user corrections to prevent similar errors.

[0515] Step 11: Final submission

[0516] Subject: Server

[0517] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[0518] Input: Final, adjusted text

[0519] Output: Send to specified destination

[0520] Specifically, the generated text is sent to a specified email address or internal notification system and delivered to the recipient designated by the user.

[0521] In this way, each step works in conjunction with each other to achieve effective sentence generation that is optimized for each company and takes into account the user's emotional state.

[0522] (Application example 2)

[0523] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0524] In logistics centers and other workplaces, it is important to communicate in a way that takes into account the work situation and the emotional state of employees. However, conventional systems are unable to recognize employees' emotions and generate appropriate messages based on them. This makes it difficult to respond flexibly to their emotional state, which can result in reduced work efficiency and increased employee stress.

[0525] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0526] In this invention, the server includes a means for generating prompts optimized for each company, a means for an AI model to generate sentences based on internal human relationships and past message patterns, a means including an emotion engine that recognizes the user's emotional state, and a means for reflecting emotional data from the user in the generation of prompts, thereby enabling optimal sentence generation according to the user's emotional state.

[0527] "Company-optimized prompts" are prompts that are optimized to fit a specific company's business flow, culture, and communication style.

[0528] "Internal relationships" refer to the connections and relationships between employees working within the same company.

[0529] "Past message patterns" refer to the tendencies, formats, and contents of messages previously exchanged.

[0530] An "AI model" is an algorithm or structure that uses artificial intelligence techniques to perform a specific task.

[0531] "User's emotional state" is information indicating the user's current emotions, including emotions such as joy, anger, and sadness.

[0532] An "emotion engine" is software or algorithm that recognizes emotions from user input data and provides the results.

[0533] "Feedback" refers to the evaluations, opinions, and adjustments that users provide to the system.

[0534] A "prompt generation algorithm" refers to the procedures, methods, and rules for generating optimal prompts, based on which prompts are created.

[0535] The system for implementing this invention consists of a server, a terminal, and a user. This system uses prompts optimized for each company, and an AI model generates sentences based on internal human relationships and past message patterns. Furthermore, by incorporating an emotion engine that recognizes the user's emotions, it becomes possible to generate messages that take into account the user's emotional state. The specific components and operation of the system are described below.

[0536] Server Configuration

[0537] The server includes means to:

[0538] 1. Means of receiving input data

[0539] The server receives data such as project overviews and team member information entered from the device, and sends this data in JSON or XML format.

[0540] 2. Emotion recognition means

[0541] The server analyzes the input data and recognizes the user's emotional state using an emotion engine. It identifies the emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[0542] 3. Prompt Generation Methods

[0543] The server uses the received data and emotion recognition results to generate personalized prompts for each company, which are tailored to fit the specific organizational structure and culture.

[0544] 4. Sentence generation means

[0545] The server then inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[0546] 5. Feedback as a learning tool

[0547] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[0548] 6. Means of sending text

[0549] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[0550] Device configuration

[0551] The terminal includes means for:

[0552] 1. Data input interface

[0553] It provides an interface for users to use and allows them to input the necessary information (e.g., the content they want to send, target information, etc.).

[0554] 2. Display of generated text

[0555] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[0556] 3. Manual Adjustment Form

[0557] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[0558] User operations

[0559] The user uses the system in the following steps:

[0560] 1. Entering data

[0561] Using the device's interface, you enter information about the message and the recipient, such as "Notification of new project start" or "Team member roles."

[0562] 2. Check and adjust the generated text

[0563] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[0564] 3. Final submission

[0565] Check the final text after adjustments and press the send button to send it to the specified recipient.

[0566] Specific examples

[0567] As a concrete example of a server and terminal working together, the following shows a business notification system in a logistics center:

[0568] 1. The administrator types "New shipment is delayed again" into the terminal.

[0569] 2. The server recognizes the user's emotion as "anger."

[0570] 3. The server generates a prompt saying, "Let's try to address any issues calmly and work together to solve them."

[0571] 4. Based on this prompt, the generative AI model generates the sentence, "Let's stay calm and solve the problem that's causing the delay together."

[0572] 5. The administrator reviews the generated text and adjusts it if necessary.

[0573] 6. The adjusted text is fed back to the server and finally sent.

[0574] Example prompt sentence:

[0575] "Let's calmly address the issues that are causing delays and solve them together."

[0576] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0577] Step 1:

[0578] The terminal receives the data sent by the user. Specifically, the user enters a message such as "A new shipment is delayed again," and the terminal sends this information to the server in a data format (e.g., JSON format).

[0579] Input: User-entered message ("New shipment is delayed again")

[0580] Output: Data sent to the server (JSON format)

[0581] Step 2:

[0582] The server receives the data sent from the device and passes it to the emotion recognition engine, which analyzes the user's emotional state from the message and identifies the emotion "anger."

[0583] Input: Data received from the terminal (JSON format message "New shipment is delayed again")

[0584] Output: Analysis result (emotional state "anger")

[0585] Step 3:

[0586] The server uses a prompt generation algorithm based on the emotion recognition results to generate the optimal prompt. In this case, the generated prompt is, "Let's try to address any issues calmly and work together to solve them."

[0587] Input: Emotion recognition result (emotion state "anger")

[0588] Output: Optimized prompt ("Let's try to address any issues calmly and work together to solve them.")

[0589] Step 4:

[0590] The server provides the optimized prompt to the generative AI model, which then generates the optimal sentence based on the prompt. In this case, the sentence generated is, "Let's calmly address the issue that is causing the delay and solve it together."

[0591] Input: Optimized prompts ("Let's try to address any issues calmly and work together to solve them.")

[0592] Output: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[0593] Step 5:

[0594] The server transmits the generated text to the terminal, which displays the generated text to the user.

[0595] Input: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[0596] Output: The text that is displayed to the user

[0597] Step 6:

[0598] The user reviews the displayed text and manually adjusts it if necessary, for example changing "Let's resolve it" to "Let's deal with it."

[0599] Input: The displayed text

[0600] Output: Text after manual adjustment by the user

[0601] Step 7:

[0602] The device sends the user's adjusted sentences as feedback to the server, which stores this feedback as learning data and uses it to generate future prompts and improve the AI ​​model.

[0603] Input: Text after manual adjustment by user ("Let's stay calm and deal with the issues that are causing delays together.")

[0604] Output: Data stored as feedback

[0605] Step 8:

[0606] The server then sends the final adjusted text to the specified destination. The actual communication method is linked to the email server or internal communication tools.

[0607] Input: Final adjusted sentence ("Let's calmly address the issues that are causing delays and work together to resolve them.")

[0608] Output: The text sent to the specified destination

[0609] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0610] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0611] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0612] [Second embodiment]

[0613] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0614] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0615] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0616] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0617] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0618] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0619] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0620] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0621] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0622] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0623] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0624] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0625] This invention is a system in which AI generates sentences using prompts optimized for each company, taking into account internal human relationships and past message patterns. Below, each component and operation of this system are explained in natural language.

[0626] Software Configuration

[0627] This system is mainly composed of a server, a terminal, and a user. The specific operation of each component is explained below.

[0628] server

[0629] 1. Receiving input data

[0630] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.).

[0631] This data is sent in formats such as JSON or XML.

[0632] 2. Prompt Optimization

[0633] The server generates prompts based on the input data, which are customized for each company and optimized to fit their specific organizational structure and culture.

[0634] For example, this includes the relationship between a specific boss and a subordinate, or past email patterns.

[0635] 3. Sentence generation

[0636] The server inputs the optimized prompts into the AI ​​model and generates sentences according to the specified format.

[0637] The AI ​​model uses machine learning algorithms to generate sentences based on the input prompts.

[0638] 4. Learning Feedback

[0639] It receives manual adjustments and feedback from users and stores them as learning data.

[0640] The feedback will identify areas that need adjustment and will be used to generate future prompts and improve the AI ​​model.

[0641] 5. Sending text

[0642] The final adjusted text is sent to the specified destination.

[0643] Sending is carried out in conjunction with email servers and internal communication tools.

[0644] Terminal

[0645] 1. Data input interface

[0646] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[0647] 2. Displaying the generated text

[0648] The generated text returned from the server is displayed in the interface.

[0649] The generated text is highlighted and annotated as appropriate to make it easier for users to check.

[0650] 3. Manual Adjustment Form

[0651] Provide a form that allows the user to manually adjust the generated text.

[0652] It has the function of sending the adjusted text to the server as feedback.

[0653] User

[0654] 1. Entering data

[0655] The user uses the terminal interface to input information about the content they want to send and the recipient.

[0656] For example, enter information such as "Notification of new project start" and "Team member roles."

[0657] 2. Check and adjust the generated text

[0658] Check the generated text returned by the server and make manual adjustments as necessary.

[0659] Once the adjustments are complete, the final text is sent to the server as feedback.

[0660] 3. Final submission

[0661] Check the final text after adjustments and press the send button to send it to the specified recipient.

[0662] Specific examples

[0663] For example, to send a project launch notice, the user might enter:

[0664] 1. User (Device)

[0665] Enter the project name, start date, and team member information as "Notification of new project start."

[0666] 2. Server

[0667] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[0668] 3. Server

[0669] The generated text is sent to the user's terminal.

[0670] 4. User (Terminal)

[0671] Review the generated text and edit it if necessary.

[0672] 5. User (Terminal)

[0673] The edited text is sent to the server as feedback.

[0674] 6. Server

[0675] Add feedback to the training data to help improve it further.

[0676] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[0677] The processing flow will be explained below.

[0678] Step 1:

[0679] The device displays an interface for the user to input information about what they want to send and who they want to send it to, including detailed project name, start date, team member roles, etc.

[0680] Step 2:

[0681] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[0682] Step 3:

[0683] The server receives the data sent from the device, parses it, and extracts the necessary information, such as the project name, start date, and team members.

[0684] Step 4:

[0685] The server uses the extracted data to generate personalized prompts tailored to the company's culture, organizational structure, and internal relationships.

[0686] Step 5:

[0687] The server then inputs the generated prompts into an AI model, which uses machine learning algorithms to generate sentences based on the prompts, with language and tone that reflects the company's culture.

[0688] Step 6:

[0689] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[0690] Step 7:

[0691] The user reviews the generated text and manually adjusts it as needed, for example by correcting specific job titles or the roles of specific members.

[0692] Step 8:

[0693] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[0694] Step 9:

[0695] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for future prompt optimization and AI model improvement.

[0696] Step 10:

[0697] The user checks and adjusts the final text and sends it from the device. The device then works in conjunction with the mail server and internal communication tools to send the text to the specified destination.

[0698] This series of processing steps enables users to efficiently generate business documents and realize communication that is in line with the unique culture and background of the company.

[0699] Example 1

[0700] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0701] In today's corporate environment, generating sentences that are efficient and reflect the organization's unique culture and interpersonal relationships is extremely important. However, conventional sentence generation systems lack a mechanism for generating sentences using prompts optimized for each company, requiring significant manual adjustments. Furthermore, they lack a means to effectively utilize user feedback to improve the sentence generation algorithm. This reduces the efficiency of the entire sentence generation process and significantly increases the time and effort required.

[0702] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0703] In this invention, the server includes means for receiving input data and generating optimized prompts based on the data, means for inputting the optimized prompts into an AI model and generating sentences based on internal human relationships and past message patterns, means for sending the generated sentences to a user's terminal and accepting manual adjustments by the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, and means for sending the final adjusted sentences to a specified destination. This makes it possible to efficiently generate sentences optimized for each company and significantly reduce the user's effort.

[0704] "Input data" refers to information provided by users to the system, including project overviews and team member information.

[0705] A "prompt" is an instruction or question that is input into an AI model and serves as the basis for generating sentences.

[0706] "Optimized prompts" refer to prompts that are customized based on the company's characteristics, culture, internal relationships, etc.

[0707] An "AI model" is an artificial intelligence program built using machine learning algorithms that generates natural language sentences based on prompts.

[0708] "Generated sentences" refer to sentences generated by the AI ​​model based on optimized prompts.

[0709] "User terminal" refers to a device such as a computer or smartphone used by a user to input data and check generated text.

[0710] "Manual adjustment" refers to the act of a user manually correcting or editing the generated text.

[0711] "Feedback" refers to information that users send back to the system regarding adjusted text and improvements.

[0712] "Training data" is a data set that the system uses to improve its performance, and includes feedback from users.

[0713] "Designated Destination" refers to the address or contact information indicating the specific recipient to whom the final tailored document is to be sent.

[0714] This invention is a system in which a user inputs data using a terminal, generates optimized prompts based on that data, and generates sentences under specific conditions using an AI model. Here, we will explain the specific names of the hardware and software used and the program's processing procedures.

[0715] System configuration

[0716] This system mainly consists of a server, terminals, and users.

[0717] server

[0718] The server has the following roles:

[0719] 1. Receiving input data

[0720] The server receives the data sent by the user from the device. The data is usually sent in a format such as JSON or XML. The received data is stored in memory and used for subsequent processing.

[0721] 2. Prompt optimization

[0722] The server generates prompts based on the input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns.To generate and optimize these prompts, Python scripts and database management systems (e.g., MySQL or PostgreSQL) are used.

[0723] 3. Sentence generation

[0724] The server inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format. The AI ​​model is built using machine learning libraries such as TensorFlow and PyTorch, and generates sentences based on the prompts using natural language processing techniques.

[0725] 4. Sending the generated text

[0726] The server encodes the generated text into JSON format and sends it to the terminal as an HTTP response.

[0727] 5. Learning Feedback

[0728] The server receives user feedback and stores it in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future.

[0729] Terminal

[0730] The terminal has the following roles:

[0731] 1. Providing a data input interface

[0732] The terminal provides the user with a data entry interface that allows them to enter information about the content and audience they wish to send. This interface is built using HTML, CSS, and JavaScript.

[0733] 2. Displaying the generated text

[0734] The terminal displays the generated text returned from the server on its interface, with appropriate highlighting and annotations added to make it easier for the user to check.

[0735] 3. Provide a manual adjustment form

[0736] The terminal provides a form that allows the user to manually adjust the generated text, and sends the adjustment results to the server as feedback.

[0737] User

[0738] The user has the following roles:

[0739] 1. Entering data

[0740] The user uses the device's data entry interface to enter information about the content and recipients of the message, such as the project name, start date, and team member information for "Notice of the start of a new project."

[0741] 2. Check and adjust the generated text

[0742] The user checks the generated sentences sent back from the server and manually adjusts them if necessary. The adjusted sentences are then sent back to the server as feedback.

[0743] Specific examples

[0744] For example, the specific procedure for sending a notification of the start of a new project is as follows.

[0745] 1. User (Device)

[0746] Enter the project name, start date, and team member information as "Notification of new project start." A specific example of how to enter this information is shown below.

[0747] Document Title: Notice of New Project Launch

[0748] Project Name: Project X

[0749] Start date: October 1, 2023

[0750] Team members: Mr. A, Mr. B, Mr. C

[0751] 2. Server

[0752] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[0753] 3. Server

[0754] The generated text is sent to the user's terminal.

[0755] 4. User (Terminal)

[0756] Review the generated text and edit it if necessary.

[0757] 5. User (Terminal)

[0758] The edited text is sent to the server as feedback.

[0759] 6. Server

[0760] Add feedback to the training data to help improve it further.

[0761] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[0762] In this way, the present invention realizes optimal sentence generation adapted to each company.

[0763] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0764] Step 1:

[0765] The user enters the data.

[0766] Specific behavior:

[0767] The user uses the device's data entry interface to enter the content they want to send (e.g., "Notification of the start of a new project") and target information (e.g., project name, start date, team member roles). This input data will be the basis for subsequent processing.

[0768] Input: Project name, start date, team member information

[0769] Output: Data entered by the user

[0770] Step 2:

[0771] The terminal sends the input data to the server.

[0772] Specific behavior:

[0773] The terminal encodes the data entered by the user in JSON format and sends it to the server using the HTTPS protocol, which ensures secure transmission.

[0774] Input: Data entered by the user

[0775] Output: Encoded data in JSON format

[0776] Step 3:

[0777] The server receives the input data.

[0778] Specific behavior:

[0779] The server receives the data sent from the terminal and stores it in memory. It is also possible to temporarily store the data using a database management system (e.g., MySQL or PostgreSQL).

[0780] Input: JSON encoded data

[0781] Output: Data stored in memory

[0782] Step 4:

[0783] The server optimizes the prompt.

[0784] Specific behavior:

[0785] The server generates prompts based on input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns. Python scripts are used to extract relevant information from the database and reflect it in the prompts.

[0786] Input: Data stored in memory

[0787] Output: Optimized prompt

[0788] Step 5:

[0789] The server inputs the prompts into the AI ​​model to generate sentences.

[0790] Specific behavior:

[0791] The server inputs the optimized prompts into a generative AI model, which is built using machine learning libraries such as TensorFlow and PyTorch, and generates natural language sentences based on the prompts.

[0792] Input: Optimized prompts

[0793] Output: Generated sentence

[0794] Step 6:

[0795] The server sends the generated text to the terminal.

[0796] Specific behavior:

[0797] The server encodes the generated text again into JSON format and sends it to the terminal as an HTTP response.

[0798] Input: Generated sentence

[0799] Output: The generated document encoded in JSON format.

[0800] Step 7:

[0801] The terminal displays the generated text to the user.

[0802] Specific behavior:

[0803] The device then displays the generated text in an interface, styled using HTML and CSS to highlight specific feeds and important information.

[0804] Input: The generated text encoded in JSON format

[0805] Output: The generated text displayed to the user

[0806] Step 8:

[0807] The user reviews and adjusts the generated text.

[0808] Specific behavior:

[0809] The user can review the generated text displayed on the device and manually adjust it as needed, for example, by changing specific wording or adjusting the writing style. Adjustments are made using a form on the device.

[0810] Input: The generated text displayed to the user

[0811] Output: User-adjusted text

[0812] Step 9:

[0813] The user sends the adjusted text to the server as feedback.

[0814] Specific behavior:

[0815] The user then sends the adjusted document back to the server from their device, again encoded in JSON and using the HTTPS protocol.

[0816] Input: User-adjusted text

[0817] Output: The adjusted text encoded in JSON format

[0818] Step 10:

[0819] The server takes the feedback as learning.

[0820] Specific behavior:

[0821] The server stores the received feedback in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future. A database management system (e.g., MySQL or PostgreSQL) is used for storage.

[0822] Input: Adjusted text encoded in JSON format

[0823] Output: Feedback data stored in a database

[0824] (Application example 1)

[0825] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0826] The food delivery industry requires fast and accurate customer support, but traditional methods have led to inconsistencies in response time and quality. It has also been difficult to share know-how and make continuous improvements to properly respond to inquiries.

[0827] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0828] In this invention, the server includes means for generating prompts optimized for each company, means for an AI-generated model to generate sentences based on internal human relationships and past message patterns, means for sending the generated sentences to an information processing terminal and accepting manual adjustments from the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, means for sending the final adjusted sentences to another information processing terminal, and means for generating prompts and sentences for customer support inquiries specific to the food delivery industry. This enables fast and accurate customer support responses, standardizes the quality of responses, and enables continuous improvement.

[0829] "Means for generating prompts optimized for each company" refers to methods or functions for automatically creating optimal prompts based on each company's specific organizational structure, culture, and past message patterns.

[0830] "Means for an AI generation model to generate sentences based on internal interpersonal relationships and past message patterns" refers to a method or function that uses input data based on specific internal interpersonal relationships and past message patterns to generate appropriate sentences using an AI generation model.

[0831] "Means for transmitting the generated text to an information processing terminal and accepting manual adjustments by the user" refers to a method or function that transmits text generated by an AI-generated model to a terminal used by the user and enables the user to manually adjust the text.

[0832] "Means for accumulating user feedback as learning data and reflecting it in future prompts and sentence generation" refers to a method or function that accumulates the user's manually adjusted content and feedback in a database and uses it when generating the next prompt or sentence.

[0833] "Means for sending the final adjusted text to another information processing terminal" refers to a method or function for sending the text that has been final checked and adjusted by the user to a designated recipient.

[0834] "Means for generating prompts and generating sentences to respond to customer support inquiries specialized for the food delivery industry" refers to methods and functions for generating prompts and AI-based sentences necessary to respond to customer support inquiries specialized for the content of such inquiries in the food delivery industry.

[0835] The "means for inputting data including inquiry content" refers to a method or function for a user to input data such as customer support inquiry content and customer information.

[0836] The "means for analyzing data and generating an optimized prompt" refers to a method or function for analyzing input data and automatically generating an optimal prompt.

[0837] "Means for analyzing the details of manual adjustments made by the user and storing the details as feedback" refers to a method or function for analyzing the details of manual adjustments made by the user and storing the information as feedback in a database.

[0838] "Means for improving the AI-generated model and prompt-generation algorithm based on accumulated feedback" refers to methods or functions for continuously improving the AI-generated model and prompt-generation algorithm using accumulated feedback data.

[0839] This invention aims to improve the efficiency of customer support in the food delivery industry by using prompts optimized for each company and AI-generated models. This system mainly consists of a server, a terminal, and a user.

[0840] server

[0841] The server performs the following series of processes.

[0842] 1. Receiving input data

[0843] The server receives data (e.g., inquiry details, customer information) entered by the user from the terminal. The data is usually sent in a format such as JSON or XML.

[0844] 2. Prompt Generation

[0845] The server generates prompts based on the input data, which are specific to the food delivery industry and include the following examples:

[0846] "Inquiry: My order hasn't arrived"

[0847] 3. Sentence generation

[0848] The server inputs the optimized prompts into a generative AI model (e.g., GPT-3) to generate appropriate sentences for the query. The generative AI model generates sentences using machine learning algorithms.

[0849] 4. Learning Feedback

[0850] It receives manual adjustments and feedback from users and stores it as training data, identifying areas that need adjustment and helping to generate future prompts and improve the AI ​​model.

[0851] 5. Sending text

[0852] Finally, the adjusted text is sent to the designated recipient, linked to the email server or internal instant communication tool.

[0853] Terminal

[0854] The terminal provides the following features:

[0855] 1. Data input interface

[0856] It provides an interface for users to use and allows them to input necessary information (e.g., inquiry details, customer information, etc.).

[0857] 2. Displaying the generated text

[0858] The generated text sent back from the server is displayed in the interface, and it is also possible to highlight and annotate the generated text for the user's convenience.

[0859] 3. Manual Adjustment Form

[0860] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[0861] User

[0862] Users interact with the system as follows:

[0863] 1. Entering data

[0864] The user uses the terminal interface to input the inquiry and customer information. For example, the user may input "The food I ordered has not arrived."

[0865] 2. Check and adjust the generated text

[0866] Check the generated text sent back from the server and manually adjust it if necessary. For example, adjust the text to something like, "We apologize for the inconvenience. We will recheck it immediately. While you wait, we will provide you with a 1,000 yen discount coupon."

[0867] 3. Providing Feedback

[0868] The adjusted sentences are sent to the server as feedback and stored as learning data in the system.

[0869] This system enables quick and accurate customer support responses, standardizes the quality of responses, and allows for continuous improvement. As a specific example of its use, the server generates the following prompt sentence:

[0870] Inquiry: The food I ordered hasn't arrived. A customer who has made a similar inquiry in the past frequently orders sushi.

[0871] The sentence generated based on this is, "We are sorry to hear that your sushi order has not yet arrived, and we apologize for the inconvenience. We will recheck for you shortly. While you wait, we will provide you with a 1,000 yen discount coupon," which is then adjusted on the device and sent as the final sentence.

[0872] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0873] Step 1:

[0874] The user uses the terminal interface to input the inquiry and customer information. At this time, the inquiry information is entered as "The food I ordered has not arrived." This input data is sent to the server in JSON format.

[0875] Step 2:

[0876] The server analyzes the input data received from the device and generates a prompt optimized for each company. Specifically, it takes into account data such as past inquiry patterns and customer preferences to generate a prompt like the one below.

[0877] "Inquiry: The food I ordered hasn't arrived. This customer has made a similar inquiry in the past and frequently orders sushi."

[0878] Step 3:

[0879] The server that generated the prompt inputs the prompt into a generative AI model (e.g., GPT-3). The AI ​​model generates an appropriate sentence based on the prompt. In doing so, the AI ​​analyzes the input data and uses patterns learned from previous inquiries. An example of a generated sentence might be, "We're sorry to hear that your sushi order hasn't arrived yet, and we apologize for the inconvenience. We'll check again soon. While you wait, we'll provide you with a 1,000 yen discount coupon."

[0880] Step 4:

[0881] The server sends the generated text to the terminal, where the user can review it on the terminal interface. The generated text may also be highlighted or annotated to make it easier for the user to review.

[0882] Step 5:

[0883] The user can review the generated text and manually adjust it if necessary, for example to include more specific instructions or additional information, which is then sent back to the server.

[0884] Step 6:

[0885] The server receives user feedback on the adjustments made, which is stored in a database and used to improve future prompt and sentence generation.

[0886] Step 7:

[0887] The final adjusted text is then sent to the specified destination by the server, enabling quick and accurate customer support, improving the quality of customer support in the food delivery industry.

[0888] This system's series of processes will significantly improve the quality and efficiency of customer support responses in the food delivery industry. For example, it will be able to respond quickly to complex customer needs and continuously improve the quality of responses by utilizing learning data.

[0889] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0890] This invention combines an emotion engine that recognizes the user's emotions with a system in which AI generates sentences using prompts optimized for each company and taking into account internal human relationships and past message patterns. Below, each component and operation of this system is explained in natural language.

[0891] Software Configuration

[0892] This system is primarily composed of a server, a terminal, and a user. By introducing an emotion engine, it becomes possible to generate text that takes into account the user's emotional state, achieving more effective communication. The specific operation of each component is explained below.

[0893] server

[0894] 1. Receiving input data

[0895] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.) This data is sent in a format such as JSON or XML.

[0896] 2. Emotion recognition

[0897] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[0898] 3. Prompt Optimization

[0899] The server generates optimized prompts based on the user's emotion recognition data and customized prompt templates for each company, which are tailored to fit the specific organizational structure and culture.

[0900] 4. Sentence generation

[0901] The server then inputs the optimized prompts into the AI ​​model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[0902] 5. Learning Feedback

[0903] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[0904] 6. Sending text

[0905] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[0906] Terminal

[0907] 1. Data input interface

[0908] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[0909] 2. Displaying the generated text

[0910] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[0911] 3. Manual Adjustment Form

[0912] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[0913] User

[0914] 1. Entering data

[0915] The user uses the device interface to input information about the content and recipients of the message, such as "notification of the start of a new project" or "roles of team members."

[0916] 2. Check and adjust the generated text

[0917] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[0918] 3. Final submission

[0919] Check the final text after adjustments and press the send button to send it to the specified recipient.

[0920] Specific examples

[0921] For example, consider sending a notification about the start of a new project. The specific steps are as follows:

[0922] 1. User (Device)

[0923] Enter the project name, start date, and team member information as "Notification of new project start."

[0924] 2. Server

[0925] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[0926] 3. Emotion recognition

[0927] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state and reflects it in the prompt generation.

[0928] 4. Server

[0929] The generated text is sent to the user's terminal.

[0930] 5. User (Terminal)

[0931] Review the generated text and edit it if necessary.

[0932] 6. User (Terminal)

[0933] The edited text is sent to the server as feedback.

[0934] 7. Server

[0935] The feedback is accumulated as learning data and used to generate future prompts and improve the AI ​​model.

[0936] This system allows users to generate optimal sentences according to their emotional state, enabling effective communication that is in line with the company's unique culture and background.

[0937] The processing flow will be explained below.

[0938] Step 1:

[0939] The terminal displays an interface for the user to enter data: the content they want to send and information about the recipient (e.g., project name, start date, team member roles, etc.).

[0940] Step 2:

[0941] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[0942] Step 3:

[0943] The server receives the data sent from the device, parses it, and extracts information about the project name, start date, and team members.

[0944] Step 4:

[0945] The server uses an emotion engine to recognize the user's emotions, identifying emotions such as joy, anger, and sadness based on the user's input data and past data.

[0946] Step 5:

[0947] The server uses the user's emotional state and a customized prompt template for each company to generate optimized prompts that are tailored to the specific organizational structure and culture.

[0948] Step 6:

[0949] The server inputs the generated prompts into an AI model to generate sentences, which then use machine learning algorithms to generate sentences that reflect the user's emotional state.

[0950] Step 7:

[0951] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[0952] Step 8:

[0953] The user reviews the generated text and makes manual adjustments as needed, for example, correcting specific job titles or member roles.

[0954] Step 9:

[0955] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[0956] Step 10:

[0957] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for use in generating future prompts and improving the AI ​​model.

[0958] Step 11:

[0959] The user then checks the final edited text and presses the send button to send it to the specified recipient. The device then sends the text in conjunction with the email server and internal communication tools.

[0960] Specific examples

[0961] Example: New project start notification

[0962] 1. Step 1:

[0963] The user inputs the project name, start date, and team member information into the terminal interface as a "new project start notification."

[0964] 2. Step 2:

[0965] The terminal sends the input data to the server.

[0966] 3. Step 3:

[0967] The server receives the data and analyzes the content.

[0968] 4. Step 4:

[0969] The server uses an emotion engine to recognize emotions from the user's input. For example, if the user includes many positive comments, the emotion engine will recognize the emotion as "joy."

[0970] 5. Step 5:

[0971] The server generates optimized prompts based on the emotional state and the company's prompt templates.

[0972] 6. Step 6:

[0973] The server inputs prompts into the AI ​​model to generate sentences that reflect the user's emotional state of "joy."

[0974] 7. Step 7:

[0975] The server sends the generated text to the terminal.

[0976] 8. Step 8:

[0977] The user reviews the generated text and manually adjusts it if necessary.

[0978] 9. Step 9:

[0979] The device sends the final text including the adjustments to the server as feedback.

[0980] 10. Step 10:

[0981] The server receives the feedback, analyzes the content, and stores it as learning data.

[0982] 11. Step 11:

[0983] The user then clicks the send button to send the final text that has been adjusted. The device then connects to the mail server and sends the text to the specified recipient.

[0984] Example 2

[0985] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0986] In corporate communication, it is important to generate texts that are optimized for each company's unique culture and organizational structure. At the same time, it is also necessary to consider the user's emotional state when generating texts. However, existing systems often lack sufficient emotion recognition and optimization for each company, hindering effective communication. Furthermore, they lack mechanisms for manual adjustment of generated texts and efficient feedback integration. This leads to problems in improving text quality and reducing user satisfaction.

[0987] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0988] In this invention, the server includes: a means for generating prompts optimized for each company; a means for an AI model to generate sentences based on internal relationships and past message patterns; a means for sending the generated sentences to a user's device and accepting manual adjustments from the user; a means for accumulating user feedback as learning data and reflecting it in future prompt and sentence generation; a means for sending the final adjusted sentences; a means for receiving and preprocessing the sent data; a means for analyzing input data using an emotion engine to recognize the user's emotional state; a means for optimizing prompts based on the recognized emotional data; and a means for inputting the generated prompts into a generative AI model to generate sentences. This enables effective sentence generation that is optimized for each company's unique culture and organizational structure and takes the user's emotional state into consideration. Furthermore, by efficiently incorporating user feedback, the quality of the generated sentences and user satisfaction can be improved.

[0989] The "server" is a device that receives data sent by users, analyzes it, generates appropriate sentences using a generative AI model, and finally transmits them.

[0990] A "terminal" is a device that provides an interface for users to operate, input data, and review and adjust the generated text.

[0991] "Company-optimized prompts" are guidance or instructions that are customized to fit a specific company's culture and organizational structure.

[0992] An "AI model" is an artificial intelligence algorithm that performs natural language processing based on input data and automatically generates sentences in a specified format.

[0993] An "emotion engine" is a program or device that analyzes and recognizes the emotional state of a user from data entered by the user.

[0994] A "prompt" is a predetermined instruction or explanation that is input into an AI model.

[0995] "Feedback" is information provided by a user when conveying corrections or opinions about the generated text to the server.

[0996] "Preprocessing" is a process performed by the server to prepare the transmitted data it receives in a format that is easy to analyze.

[0997] "Optimized prompts" are instructions or explanations that are optimized to take into account the user's emotional state and company-specific factors.

[0998] A "generative AI model" is a system that uses artificial intelligence technology to generate sentences based on user input data and optimized prompts.

[0999] "Sentence generation" is the process by which an AI model creates sentences in natural language based on optimized prompts.

[1000] The present invention provides a sentence generation system that combines an emotion engine and a generative AI model using prompts optimized for each company to effectively communicate within the company. Hereinafter, an embodiment of the present invention will be described in detail.

[1001] System configuration

[1002] server

[1003] A server is a device that performs several major functions:

[1004] 1. Receiving input data

[1005] The server receives data sent from the device (e.g., project name, team member information, etc.) This data is sent in JSON or XML format.

[1006] 2. Emotion recognition

[1007] The server analyzes the received data and uses an emotion engine to identify the user's emotional state. For example, the emotion toward the project name "Next Generation AI Development" is recognized as "Joy."

[1008] 3. Prompt Optimization

[1009] The server generates optimized prompts based on the emotion recognition results and company-specific templates. For example, if the emotion is "joy," a positive-toned prompt is used.

[1010] 4. Sentence generation

[1011] The server then inputs the optimized prompts into a generative AI model, specifically OpenAI's GPT-3, to generate the final sentence.

[1012] 5. Learn and incorporate feedback

[1013] The server receives manual adjustments and feedback from users and stores that information as learning data to help generate future prompts and improve the AI ​​model.

[1014] 6. Sending text

[1015] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[1016] Terminal

[1017] The device is directly operated by the user and provides the following functions:

[1018] 1. Providing a data input interface

[1019] It provides an interface for users to enter data (e.g., project name, start date, team member information, etc.).

[1020] 2. Displaying the generated text

[1021] It has the function of displaying the generated text in an appropriate format so that the user can check it.

[1022] 3. Provide a manual adjustment form

[1023] A form is provided for users to manually adjust the generated text, and the adjusted text is sent to the server as feedback.

[1024] User

[1025] The user is the person who operates this system and performs the following operations.

[1026] 1. Entering data

[1027] Use the device to enter information about the content you want to send and the recipient. For example, for "Notice of the start of a new project," enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki."

[1028] 2. Check and adjust the generated text

[1029] Check the generated text returned by the server and make manual adjustments as necessary. For example, change "Tanaka Ichiro" to "Tanaka Saburo."

[1030] 3. Final submission

[1031] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1032] Specific examples

[1033] For example, to send an announcement about the launch of a new project, the user types the following at the terminal:

[1034] "New project start notification"

[1035] "Project name = Next generation AI development"

[1036] "Start date=October 1, 2023"

[1037] "Team members: Ichiro Tanaka, Jiro Suzuki"

[1038] The server receives this data, and uses the emotion engine to generate prompts based on the emotion it recognizes as "joy," ultimately generating the following sentence:

[1039] "The next-generation AI development project has begun. We appreciate your cooperation."

[1040] The generated text is displayed on the user's device, and after the user confirms and adjusts it, the final text is sent to the specified destination.

[1041] This system can generate sentences that are suited to each company's unique culture and organizational structure, and can also provide effective communication that takes into account the user's emotional state. Furthermore, by efficiently incorporating feedback, the quality of the generated sentences can be improved.

[1042] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1043] Step 1: Data entry

[1044] Subject: User

[1045] Users use the device interface to enter information about the content and recipients they want to send, such as the project name, the start date of the new project, and team member information.

[1046] Input: Project name, start date, team member information

[1047] Output: Input data

[1048] For example, a user might enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki" as a "Notification of the start of a new project."

[1049] Step 2: Submitting input data

[1050] Subject: Terminal

[1051] The terminal sends the data entered by the user to the server in JSON or XML format.

[1052] Input: User-entered data (e.g., project name, start date, team member information)

[1053] Output: Data sent to the server

[1054] Specifically, the terminal converts the user's input data into an appropriate format and transmits it to the server.

[1055] Step 3: Receiving input data

[1056] Subject: Server

[1057] The server receives the data sent from the device. Since the data is in JSON or XML format, it parses it and converts it into a data structure for analysis.

[1058] Input: Data sent from the terminal (JSON or XML)

[1059] Output: Parsable data structure

[1060] The server performs preprocessing to analyze the received data, converting it into a format such as a string or a number, and stores it in an internal data structure.

[1061] Step 4: Emotion Recognition

[1062] Subject: Server

[1063] The server analyzes the received data using an emotion engine to recognize the user's emotional state. It processes the data to identify the emotional state that can be inferred from the input data.

[1064] Input: A parsable data structure

[1065] Output: Emotional state data (e.g., happy, angry, sad)

[1066] The server's emotion engine identifies the user's emotion as "happiness" based on the input data and passes the result to the next processing step.

[1067] Step 5: Prompt optimization

[1068] Subject: Server

[1069] The server generates optimized prompts based on emotion recognition results and templates customized for each company. The template and emotion data are combined to create prompts tailored to the customer's emotional state.

[1070] Input: Emotional state data, company-specific templates

[1071] Output: Optimized prompt

[1072] Specifically, if the emotional state is "joy," a prompt with a positive tone (e.g., "The next-generation AI development project is starting!") is generated.

[1073] Step 6: Sentence generation

[1074] Subject: Server

[1075] The server inputs the optimized prompts into a generative AI model (e.g., OpenAI GPT-3) to generate natural language sentences that are consistent with the company's culture and reflect the user's emotional state.

[1076] Input: Optimized prompts

[1077] Output: Generated sentence

[1078] Specifically, based on optimized prompts, it generates sentences such as, "The next-generation AI development project has begun. We appreciate your cooperation."

[1079] Step 7: Displaying the generated sentences

[1080] Subject: Terminal

[1081] The terminal displays the generated text returned by the server to the user, with appropriate formatting and highlighting.

[1082] Input: Generated text returned by the server

[1083] Output: Text for display

[1084] The displayed text is set up so that it is easy for the user to check, and the content is made easier to understand by highlighting and annotating it.

[1085] Step 8: Manual adjustment

[1086] Subject: User

[1087] The user can review the generated text and manually adjust it if necessary, for example, to correct misspelled names or wording.

[1088] Input: Generated text displayed

[1089] Output: Adjusted text

[1090] For example, the user corrects "Tanaka Ichiro" to "Tanaka Saburo" and sends the content to the next step.

[1091] Step 9: Send your feedback

[1092] Subject: Terminal

[1093] The device sends the user's manual adjustments to the server as feedback, which is then stored as learning data.

[1094] Input: Adjusted text

[1095] Output: Feedback sent to the server

[1096] The device will then send the user's manual adjustments back to the server and use them as feedback.

[1097] Step 10: Learning feedback

[1098] Subject: Server

[1099] The server accumulates the feedback received from the user as learning data and uses it for future sentence generation and prompt optimization. It analyzes the feedback and updates the learning data.

[1100] Input: Adjusted feedback data

[1101] Output: Updated training data

[1102] For example, we will improve our AI models and prompt generation algorithms based on user corrections to prevent similar errors.

[1103] Step 11: Final submission

[1104] Subject: Server

[1105] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[1106] Input: Final, adjusted text

[1107] Output: Send to specified destination

[1108] Specifically, the generated text is sent to a specified email address or internal notification system and delivered to the recipient designated by the user.

[1109] In this way, each step works in conjunction with each other to achieve effective sentence generation that is optimized for each company and takes into account the user's emotional state.

[1110] (Application example 2)

[1111] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1112] In logistics centers and other workplaces, it is important to communicate in a way that takes into account the work situation and the emotional state of employees. However, conventional systems are unable to recognize employees' emotions and generate appropriate messages based on them. This makes it difficult to respond flexibly to their emotional state, which can result in reduced work efficiency and increased employee stress.

[1113] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1114] In this invention, the server includes a means for generating prompts optimized for each company, a means for an AI model to generate sentences based on internal human relationships and past message patterns, a means including an emotion engine that recognizes the user's emotional state, and a means for reflecting emotional data from the user in the generation of prompts, thereby enabling optimal sentence generation according to the user's emotional state.

[1115] "Company-optimized prompts" are prompts that are optimized to fit a specific company's business flow, culture, and communication style.

[1116] "Internal relationships" refer to the connections and relationships between employees working within the same company.

[1117] "Past message patterns" refer to the tendencies, formats, and contents of messages previously exchanged.

[1118] An "AI model" is an algorithm or structure that uses artificial intelligence techniques to perform a specific task.

[1119] "User's emotional state" is information indicating the user's current emotions, including emotions such as joy, anger, and sadness.

[1120] An "emotion engine" is software or algorithm that recognizes emotions from user input data and provides the results.

[1121] "Feedback" refers to the evaluations, opinions, and adjustments that users provide to the system.

[1122] A "prompt generation algorithm" refers to the procedures, methods, and rules for generating optimal prompts, based on which prompts are created.

[1123] The system for implementing this invention consists of a server, a terminal, and a user. This system uses prompts optimized for each company, and an AI model generates sentences based on internal human relationships and past message patterns. Furthermore, by incorporating an emotion engine that recognizes the user's emotions, it becomes possible to generate messages that take into account the user's emotional state. The specific components and operation of the system are described below.

[1124] Server Configuration

[1125] The server includes means to:

[1126] 1. Means of receiving input data

[1127] The server receives data such as project overviews and team member information entered from the device, and sends this data in JSON or XML format.

[1128] 2. Emotion recognition means

[1129] The server analyzes the input data and recognizes the user's emotional state using an emotion engine. It identifies the emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[1130] 3. Prompt Generation Methods

[1131] The server uses the received data and emotion recognition results to generate personalized prompts for each company, which are tailored to fit the specific organizational structure and culture.

[1132] 4. Sentence generation means

[1133] The server then inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[1134] 5. Feedback as a learning tool

[1135] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[1136] 6. Means of sending text

[1137] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[1138] Device configuration

[1139] The terminal includes means for:

[1140] 1. Data input interface

[1141] It provides an interface for users to use and allows them to input the necessary information (e.g., the content they want to send, target information, etc.).

[1142] 2. Display of generated text

[1143] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[1144] 3. Manual Adjustment Form

[1145] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[1146] User operations

[1147] The user uses the system in the following steps:

[1148] 1. Entering data

[1149] Using the device's interface, you enter information about the message and the recipient, such as "Notification of new project start" or "Team member roles."

[1150] 2. Check and adjust the generated text

[1151] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[1152] 3. Final submission

[1153] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1154] Specific examples

[1155] As a concrete example of a server and terminal working together, the following shows a business notification system in a logistics center:

[1156] 1. The administrator types "New shipment is delayed again" into the terminal.

[1157] 2. The server recognizes the user's emotion as "anger."

[1158] 3. The server generates a prompt saying, "Let's try to address any issues calmly and work together to solve them."

[1159] 4. Based on this prompt, the generative AI model generates the sentence, "Let's stay calm and solve the problem that's causing the delay together."

[1160] 5. The administrator reviews the generated text and adjusts it if necessary.

[1161] 6. The adjusted text is fed back to the server and finally sent.

[1162] Example prompt sentence:

[1163] "Let's calmly address the issues that are causing delays and solve them together."

[1164] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1165] Step 1:

[1166] The terminal receives the data sent by the user. Specifically, the user enters a message such as "A new shipment is delayed again," and the terminal sends this information to the server in a data format (e.g., JSON format).

[1167] Input: User-entered message ("New shipment is delayed again")

[1168] Output: Data sent to the server (JSON format)

[1169] Step 2:

[1170] The server receives the data sent from the device and passes it to the emotion recognition engine, which analyzes the user's emotional state from the message and identifies the emotion "anger."

[1171] Input: Data received from the terminal (JSON format message "New shipment is delayed again")

[1172] Output: Analysis result (emotional state "anger")

[1173] Step 3:

[1174] The server uses a prompt generation algorithm based on the emotion recognition results to generate the optimal prompt. In this case, the generated prompt is, "Let's try to address any issues calmly and work together to solve them."

[1175] Input: Emotion recognition result (emotion state "anger")

[1176] Output: Optimized prompt ("Let's try to address any issues calmly and work together to solve them.")

[1177] Step 4:

[1178] The server provides the optimized prompt to the generative AI model, which then generates the optimal sentence based on the prompt. In this case, the sentence generated is, "Let's calmly address the issue that is causing the delay and solve it together."

[1179] Input: Optimized prompts ("Let's try to address any issues calmly and work together to solve them.")

[1180] Output: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[1181] Step 5:

[1182] The server transmits the generated text to the terminal, which displays the generated text to the user.

[1183] Input: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[1184] Output: The text that is displayed to the user

[1185] Step 6:

[1186] The user reviews the displayed text and manually adjusts it if necessary, for example changing "Let's resolve it" to "Let's deal with it."

[1187] Input: The displayed text

[1188] Output: Text after manual adjustment by the user

[1189] Step 7:

[1190] The device sends the user's adjusted sentences as feedback to the server, which stores this feedback as learning data and uses it to generate future prompts and improve the AI ​​model.

[1191] Input: Text after manual adjustment by user ("Let's stay calm and deal with the issues that are causing delays together.")

[1192] Output: Data stored as feedback

[1193] Step 8:

[1194] The server then sends the final adjusted text to the specified destination. The actual communication method is linked to the email server or internal communication tools.

[1195] Input: Final adjusted sentence ("Let's calmly address the issues that are causing delays and work together to resolve them.")

[1196] Output: The text sent to the specified destination

[1197] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1198] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1199] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1200] [Third embodiment]

[1201] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1202] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1203] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1204] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1205] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1206] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1207] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1208] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1209] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1210] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1211] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1212] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[1213] This invention is a system in which AI generates sentences using prompts optimized for each company, taking into account internal human relationships and past message patterns. Below, each component and operation of this system are explained in natural language.

[1214] Software Configuration

[1215] This system is mainly composed of a server, a terminal, and a user. The specific operation of each component is explained below.

[1216] server

[1217] 1. Receiving input data

[1218] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.).

[1219] This data is sent in formats such as JSON or XML.

[1220] 2. Prompt Optimization

[1221] The server generates prompts based on the input data, which are customized for each company and optimized to fit their specific organizational structure and culture.

[1222] For example, this includes the relationship between a specific boss and a subordinate, or past email patterns.

[1223] 3. Sentence generation

[1224] The server inputs the optimized prompts into the AI ​​model and generates sentences according to the specified format.

[1225] The AI ​​model uses machine learning algorithms to generate sentences based on the input prompts.

[1226] 4. Learning Feedback

[1227] It receives manual adjustments and feedback from users and stores them as learning data.

[1228] The feedback will identify areas that need adjustment and will be used to generate future prompts and improve the AI ​​model.

[1229] 5. Sending text

[1230] The final adjusted text is sent to the specified destination.

[1231] Sending is carried out in conjunction with email servers and internal communication tools.

[1232] Terminal

[1233] 1. Data input interface

[1234] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[1235] 2. Displaying the generated text

[1236] The generated text returned from the server is displayed in the interface.

[1237] The generated text is highlighted and annotated as appropriate to make it easier for users to check.

[1238] 3. Manual Adjustment Form

[1239] Provide a form that allows the user to manually adjust the generated text.

[1240] It has the function of sending the adjusted text to the server as feedback.

[1241] User

[1242] 1. Entering data

[1243] The user uses the terminal interface to input information about the content they want to send and the recipient.

[1244] For example, enter information such as "Notification of new project start" and "Team member roles."

[1245] 2. Check and adjust the generated text

[1246] Check the generated text returned by the server and make manual adjustments as necessary.

[1247] Once the adjustments are complete, the final text is sent to the server as feedback.

[1248] 3. Final submission

[1249] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1250] Specific examples

[1251] For example, to send a project launch notice, the user might enter:

[1252] 1. User (Device)

[1253] Enter the project name, start date, and team member information as "Notification of new project start."

[1254] 2. Server

[1255] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[1256] 3. Server

[1257] The generated text is sent to the user's terminal.

[1258] 4. User (Terminal)

[1259] Review the generated text and edit it if necessary.

[1260] 5. User (Terminal)

[1261] The edited text is sent to the server as feedback.

[1262] 6. Server

[1263] Add feedback to the training data to help improve it further.

[1264] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[1265] The processing flow will be explained below.

[1266] Step 1:

[1267] The device displays an interface for the user to input information about what they want to send and who they want to send it to, including detailed project name, start date, team member roles, etc.

[1268] Step 2:

[1269] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[1270] Step 3:

[1271] The server receives the data sent from the device, parses it, and extracts the necessary information, such as the project name, start date, and team members.

[1272] Step 4:

[1273] The server uses the extracted data to generate personalized prompts tailored to the company's culture, organizational structure, and internal relationships.

[1274] Step 5:

[1275] The server then inputs the generated prompts into an AI model, which uses machine learning algorithms to generate sentences based on the prompts, with language and tone that reflects the company's culture.

[1276] Step 6:

[1277] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[1278] Step 7:

[1279] The user reviews the generated text and manually adjusts it as needed, for example by correcting specific job titles or the roles of specific members.

[1280] Step 8:

[1281] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[1282] Step 9:

[1283] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for future prompt optimization and AI model improvement.

[1284] Step 10:

[1285] The user checks and adjusts the final text and sends it from the device. The device then works in conjunction with the mail server and internal communication tools to send the text to the specified destination.

[1286] This series of processing steps enables users to efficiently generate business documents and realize communication that is in line with the unique culture and background of the company.

[1287] Example 1

[1288] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1289] In today's corporate environment, generating sentences that are efficient and reflect the organization's unique culture and interpersonal relationships is extremely important. However, conventional sentence generation systems lack a mechanism for generating sentences using prompts optimized for each company, requiring significant manual adjustments. Furthermore, they lack a means to effectively utilize user feedback to improve the sentence generation algorithm. This reduces the efficiency of the entire sentence generation process and significantly increases the time and effort required.

[1290] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1291] In this invention, the server includes means for receiving input data and generating optimized prompts based on the data, means for inputting the optimized prompts into an AI model and generating sentences based on internal human relationships and past message patterns, means for sending the generated sentences to a user's terminal and accepting manual adjustments by the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, and means for sending the final adjusted sentences to a specified destination. This makes it possible to efficiently generate sentences optimized for each company and significantly reduce the user's effort.

[1292] "Input data" refers to information provided by users to the system, including project overviews and team member information.

[1293] A "prompt" is an instruction or question that is input into an AI model and serves as the basis for generating sentences.

[1294] "Optimized prompts" refer to prompts that are customized based on the company's characteristics, culture, internal relationships, etc.

[1295] An "AI model" is an artificial intelligence program built using machine learning algorithms that generates natural language sentences based on prompts.

[1296] "Generated sentences" refer to sentences generated by the AI ​​model based on optimized prompts.

[1297] "User terminal" refers to a device such as a computer or smartphone used by a user to input data and check generated text.

[1298] "Manual adjustment" refers to the act of a user manually correcting or editing the generated text.

[1299] "Feedback" refers to information that users send back to the system regarding adjusted text and improvements.

[1300] "Training data" is a data set that the system uses to improve its performance, and includes feedback from users.

[1301] "Designated Destination" refers to the address or contact information indicating the specific recipient to whom the final tailored document is to be sent.

[1302] This invention is a system in which a user inputs data using a terminal, generates optimized prompts based on that data, and generates sentences under specific conditions using an AI model. Here, we will explain the specific names of the hardware and software used and the program's processing procedures.

[1303] System configuration

[1304] This system mainly consists of a server, terminals, and users.

[1305] server

[1306] The server has the following roles:

[1307] 1. Receiving input data

[1308] The server receives the data sent by the user from the device. The data is usually sent in a format such as JSON or XML. The received data is stored in memory and used for subsequent processing.

[1309] 2. Prompt optimization

[1310] The server generates prompts based on the input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns.To generate and optimize these prompts, Python scripts and database management systems (e.g., MySQL or PostgreSQL) are used.

[1311] 3. Sentence generation

[1312] The server inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format. The AI ​​model is built using machine learning libraries such as TensorFlow and PyTorch, and generates sentences based on the prompts using natural language processing techniques.

[1313] 4. Sending the generated text

[1314] The server encodes the generated text into JSON format and sends it to the terminal as an HTTP response.

[1315] 5. Learning Feedback

[1316] The server receives user feedback and stores it in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future.

[1317] Terminal

[1318] The terminal has the following roles:

[1319] 1. Providing a data input interface

[1320] The terminal provides the user with a data entry interface that allows them to enter information about the content and audience they wish to send. This interface is built using HTML, CSS, and JavaScript.

[1321] 2. Displaying the generated text

[1322] The terminal displays the generated text returned from the server on its interface, with appropriate highlighting and annotations added to make it easier for the user to check.

[1323] 3. Provide a manual adjustment form

[1324] The terminal provides a form that allows the user to manually adjust the generated text, and sends the adjustment results to the server as feedback.

[1325] User

[1326] The user has the following roles:

[1327] 1. Entering data

[1328] The user uses the device's data entry interface to enter information about the content and recipients of the message, such as the project name, start date, and team member information for "Notice of the start of a new project."

[1329] 2. Check and adjust the generated text

[1330] The user checks the generated sentences sent back from the server and manually adjusts them if necessary. The adjusted sentences are then sent back to the server as feedback.

[1331] Specific examples

[1332] For example, the specific procedure for sending a notification of the start of a new project is as follows.

[1333] 1. User (Device)

[1334] Enter the project name, start date, and team member information as "Notification of new project start." A specific example of how to enter this information is shown below.

[1335] Document Title: Notice of New Project Launch

[1336] Project Name: Project X

[1337] Start date: October 1, 2023

[1338] Team members: Mr. A, Mr. B, Mr. C

[1339] 2. Server

[1340] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[1341] 3. Server

[1342] The generated text is sent to the user's terminal.

[1343] 4. User (Terminal)

[1344] Review the generated text and edit it if necessary.

[1345] 5. User (Terminal)

[1346] The edited text is sent to the server as feedback.

[1347] 6. Server

[1348] Add feedback to the training data to help improve it further.

[1349] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[1350] In this way, the present invention realizes optimal sentence generation adapted to each company.

[1351] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1352] Step 1:

[1353] The user enters the data.

[1354] Specific behavior:

[1355] The user uses the device's data entry interface to enter the content they want to send (e.g., "Notification of the start of a new project") and target information (e.g., project name, start date, team member roles). This input data will be the basis for subsequent processing.

[1356] Input: Project name, start date, team member information

[1357] Output: Data entered by the user

[1358] Step 2:

[1359] The terminal sends the input data to the server.

[1360] Specific behavior:

[1361] The terminal encodes the data entered by the user in JSON format and sends it to the server using the HTTPS protocol, which ensures secure transmission.

[1362] Input: Data entered by the user

[1363] Output: Encoded data in JSON format

[1364] Step 3:

[1365] The server receives the input data.

[1366] Specific behavior:

[1367] The server receives the data sent from the terminal and stores it in memory. It is also possible to temporarily store the data using a database management system (e.g., MySQL or PostgreSQL).

[1368] Input: JSON encoded data

[1369] Output: Data stored in memory

[1370] Step 4:

[1371] The server optimizes the prompt.

[1372] Specific behavior:

[1373] The server generates prompts based on input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns. Python scripts are used to extract relevant information from the database and reflect it in the prompts.

[1374] Input: Data stored in memory

[1375] Output: Optimized prompt

[1376] Step 5:

[1377] The server inputs the prompts into the AI ​​model to generate sentences.

[1378] Specific behavior:

[1379] The server inputs the optimized prompts into a generative AI model, which is built using machine learning libraries such as TensorFlow and PyTorch, and generates natural language sentences based on the prompts.

[1380] Input: Optimized prompts

[1381] Output: Generated sentence

[1382] Step 6:

[1383] The server sends the generated text to the terminal.

[1384] Specific behavior:

[1385] The server encodes the generated text again into JSON format and sends it to the terminal as an HTTP response.

[1386] Input: Generated sentence

[1387] Output: The generated document encoded in JSON format.

[1388] Step 7:

[1389] The terminal displays the generated text to the user.

[1390] Specific behavior:

[1391] The device then displays the generated text in an interface, styled using HTML and CSS to highlight specific feeds and important information.

[1392] Input: The generated text encoded in JSON format

[1393] Output: The generated text displayed to the user

[1394] Step 8:

[1395] The user reviews and adjusts the generated text.

[1396] Specific behavior:

[1397] The user can review the generated text displayed on the device and manually adjust it as needed, for example, by changing specific wording or adjusting the writing style. Adjustments are made using a form on the device.

[1398] Input: The generated text displayed to the user

[1399] Output: User-adjusted text

[1400] Step 9:

[1401] The user sends the adjusted text to the server as feedback.

[1402] Specific behavior:

[1403] The user then sends the adjusted document back to the server from their device, again encoded in JSON and using the HTTPS protocol.

[1404] Input: User-adjusted text

[1405] Output: The adjusted text encoded in JSON format

[1406] Step 10:

[1407] The server takes the feedback as learning.

[1408] Specific behavior:

[1409] The server stores the received feedback in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future. A database management system (e.g., MySQL or PostgreSQL) is used for storage.

[1410] Input: Adjusted text encoded in JSON format

[1411] Output: Feedback data stored in a database

[1412] (Application example 1)

[1413] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1414] The food delivery industry requires fast and accurate customer support, but traditional methods have led to inconsistencies in response time and quality. It has also been difficult to share know-how and make continuous improvements to properly respond to inquiries.

[1415] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1416] In this invention, the server includes means for generating prompts optimized for each company, means for an AI-generated model to generate sentences based on internal human relationships and past message patterns, means for sending the generated sentences to an information processing terminal and accepting manual adjustments from the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, means for sending the final adjusted sentences to another information processing terminal, and means for generating prompts and sentences for customer support inquiries specific to the food delivery industry. This enables fast and accurate customer support responses, standardizes the quality of responses, and enables continuous improvement.

[1417] "Means for generating prompts optimized for each company" refers to methods or functions for automatically creating optimal prompts based on each company's specific organizational structure, culture, and past message patterns.

[1418] "Means for an AI generation model to generate sentences based on internal interpersonal relationships and past message patterns" refers to a method or function that uses input data based on specific internal interpersonal relationships and past message patterns to generate appropriate sentences using an AI generation model.

[1419] "Means for transmitting the generated text to an information processing terminal and accepting manual adjustments by the user" refers to a method or function that transmits text generated by an AI-generated model to a terminal used by the user and enables the user to manually adjust the text.

[1420] "Means for accumulating user feedback as learning data and reflecting it in future prompts and sentence generation" refers to a method or function that accumulates the user's manually adjusted content and feedback in a database and uses it when generating the next prompt or sentence.

[1421] "Means for sending the final adjusted text to another information processing terminal" refers to a method or function for sending the text that has been final checked and adjusted by the user to a designated recipient.

[1422] "Means for generating prompts and generating sentences to respond to customer support inquiries specialized for the food delivery industry" refers to methods and functions for generating prompts and AI-based sentences necessary to respond to customer support inquiries specialized for the content of such inquiries in the food delivery industry.

[1423] The "means for inputting data including inquiry content" refers to a method or function for a user to input data such as customer support inquiry content and customer information.

[1424] The "means for analyzing data and generating an optimized prompt" refers to a method or function for analyzing input data and automatically generating an optimal prompt.

[1425] "Means for analyzing the details of manual adjustments made by the user and storing the details as feedback" refers to a method or function for analyzing the details of manual adjustments made by the user and storing the information as feedback in a database.

[1426] "Means for improving the AI-generated model and prompt-generation algorithm based on accumulated feedback" refers to methods or functions for continuously improving the AI-generated model and prompt-generation algorithm using accumulated feedback data.

[1427] This invention aims to improve the efficiency of customer support in the food delivery industry by using prompts optimized for each company and AI-generated models. This system mainly consists of a server, a terminal, and a user.

[1428] server

[1429] The server performs the following series of processes.

[1430] 1. Receiving input data

[1431] The server receives data (e.g., inquiry details, customer information) entered by the user from the terminal. The data is usually sent in a format such as JSON or XML.

[1432] 2. Prompt Generation

[1433] The server generates prompts based on the input data, which are specific to the food delivery industry and include the following examples:

[1434] "Inquiry: My order hasn't arrived"

[1435] 3. Sentence generation

[1436] The server inputs the optimized prompts into a generative AI model (e.g., GPT-3) to generate appropriate sentences for the query. The generative AI model generates sentences using machine learning algorithms.

[1437] 4. Learning Feedback

[1438] It receives manual adjustments and feedback from users and stores it as training data, identifying areas that need adjustment and helping to generate future prompts and improve the AI ​​model.

[1439] 5. Sending text

[1440] Finally, the adjusted text is sent to the designated recipient, linked to the email server or internal instant communication tool.

[1441] Terminal

[1442] The terminal provides the following features:

[1443] 1. Data input interface

[1444] It provides an interface for users to use and allows them to input necessary information (e.g., inquiry details, customer information, etc.).

[1445] 2. Displaying the generated text

[1446] The generated text sent back from the server is displayed in the interface, and it is also possible to highlight and annotate the generated text for the user's convenience.

[1447] 3. Manual Adjustment Form

[1448] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[1449] User

[1450] Users interact with the system as follows:

[1451] 1. Entering data

[1452] The user uses the terminal interface to input the inquiry and customer information. For example, the user may input "The food I ordered has not arrived."

[1453] 2. Check and adjust the generated text

[1454] Check the generated text sent back from the server and manually adjust it if necessary. For example, adjust the text to something like, "We apologize for the inconvenience. We will recheck it immediately. While you wait, we will provide you with a 1,000 yen discount coupon."

[1455] 3. Providing Feedback

[1456] The adjusted sentences are sent to the server as feedback and stored as learning data in the system.

[1457] This system enables quick and accurate customer support responses, standardizes the quality of responses, and allows for continuous improvement. As a specific example of its use, the server generates the following prompt sentence:

[1458] Inquiry: The food I ordered hasn't arrived. A customer who has made a similar inquiry in the past frequently orders sushi.

[1459] The sentence generated based on this is, "We are sorry to hear that your sushi order has not yet arrived, and we apologize for the inconvenience. We will recheck for you shortly. While you wait, we will provide you with a 1,000 yen discount coupon," which is then adjusted on the device and sent as the final sentence.

[1460] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1461] Step 1:

[1462] The user uses the terminal interface to input the inquiry and customer information. At this time, the inquiry information is entered as "The food I ordered has not arrived." This input data is sent to the server in JSON format.

[1463] Step 2:

[1464] The server analyzes the input data received from the device and generates a prompt optimized for each company. Specifically, it takes into account data such as past inquiry patterns and customer preferences to generate a prompt like the one below.

[1465] "Inquiry: The food I ordered hasn't arrived. This customer has made a similar inquiry in the past and frequently orders sushi."

[1466] Step 3:

[1467] The server that generated the prompt inputs the prompt into a generative AI model (e.g., GPT-3). The AI ​​model generates an appropriate sentence based on the prompt. In doing so, the AI ​​analyzes the input data and uses patterns learned from previous inquiries. An example of a generated sentence might be, "We're sorry to hear that your sushi order hasn't arrived yet, and we apologize for the inconvenience. We'll check again soon. While you wait, we'll provide you with a 1,000 yen discount coupon."

[1468] Step 4:

[1469] The server sends the generated text to the terminal, where the user can review it on the terminal interface. The generated text may also be highlighted or annotated to make it easier for the user to review.

[1470] Step 5:

[1471] The user can review the generated text and manually adjust it if necessary, for example to include more specific instructions or additional information, which is then sent back to the server.

[1472] Step 6:

[1473] The server receives user feedback on the adjustments made, which is stored in a database and used to improve future prompt and sentence generation.

[1474] Step 7:

[1475] The final adjusted text is then sent to the specified destination by the server, enabling quick and accurate customer support, improving the quality of customer support in the food delivery industry.

[1476] This system's series of processes will significantly improve the quality and efficiency of customer support responses in the food delivery industry. For example, it will be able to respond quickly to complex customer needs and continuously improve the quality of responses by utilizing learning data.

[1477] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1478] This invention combines an emotion engine that recognizes the user's emotions with a system in which AI generates sentences using prompts optimized for each company and taking into account internal human relationships and past message patterns. Below, each component and operation of this system is explained in natural language.

[1479] Software Configuration

[1480] This system is primarily composed of a server, a terminal, and a user. By introducing an emotion engine, it becomes possible to generate text that takes into account the user's emotional state, achieving more effective communication. The specific operation of each component is explained below.

[1481] server

[1482] 1. Receiving input data

[1483] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.) This data is sent in a format such as JSON or XML.

[1484] 2. Emotion recognition

[1485] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[1486] 3. Prompt Optimization

[1487] The server generates optimized prompts based on the user's emotion recognition data and customized prompt templates for each company, which are tailored to fit the specific organizational structure and culture.

[1488] 4. Sentence generation

[1489] The server then inputs the optimized prompts into the AI ​​model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[1490] 5. Learning Feedback

[1491] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[1492] 6. Sending text

[1493] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[1494] Terminal

[1495] 1. Data input interface

[1496] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[1497] 2. Displaying the generated text

[1498] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[1499] 3. Manual Adjustment Form

[1500] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[1501] User

[1502] 1. Entering data

[1503] The user uses the device interface to input information about the content and recipients of the message, such as "notification of the start of a new project" or "roles of team members."

[1504] 2. Check and adjust the generated text

[1505] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[1506] 3. Final submission

[1507] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1508] Specific examples

[1509] For example, consider sending a notification about the start of a new project. The specific steps are as follows:

[1510] 1. User (Device)

[1511] Enter the project name, start date, and team member information as "Notification of new project start."

[1512] 2. Server

[1513] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[1514] 3. Emotion recognition

[1515] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state and reflects it in the prompt generation.

[1516] 4. Server

[1517] The generated text is sent to the user's terminal.

[1518] 5. User (Terminal)

[1519] Review the generated text and edit it if necessary.

[1520] 6. User (Terminal)

[1521] The edited text is sent to the server as feedback.

[1522] 7. Server

[1523] The feedback is accumulated as learning data and used to generate future prompts and improve the AI ​​model.

[1524] This system allows users to generate optimal sentences according to their emotional state, enabling effective communication that is in line with the company's unique culture and background.

[1525] The processing flow will be explained below.

[1526] Step 1:

[1527] The terminal displays an interface for the user to enter data: the content they want to send and information about the recipient (e.g., project name, start date, team member roles, etc.).

[1528] Step 2:

[1529] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[1530] Step 3:

[1531] The server receives the data sent from the device, parses it, and extracts information about the project name, start date, and team members.

[1532] Step 4:

[1533] The server uses an emotion engine to recognize the user's emotions, identifying emotions such as joy, anger, and sadness based on the user's input data and past data.

[1534] Step 5:

[1535] The server uses the user's emotional state and a customized prompt template for each company to generate optimized prompts that are tailored to the specific organizational structure and culture.

[1536] Step 6:

[1537] The server inputs the generated prompts into an AI model to generate sentences, which then use machine learning algorithms to generate sentences that reflect the user's emotional state.

[1538] Step 7:

[1539] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[1540] Step 8:

[1541] The user reviews the generated text and makes manual adjustments as needed, for example, correcting specific job titles or member roles.

[1542] Step 9:

[1543] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[1544] Step 10:

[1545] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for use in generating future prompts and improving the AI ​​model.

[1546] Step 11:

[1547] The user then checks the final edited text and presses the send button to send it to the specified recipient. The device then sends the text in conjunction with the email server and internal communication tools.

[1548] Specific examples

[1549] Example: New project start notification

[1550] 1. Step 1:

[1551] The user inputs the project name, start date, and team member information into the terminal interface as a "new project start notification."

[1552] 2. Step 2:

[1553] The terminal sends the input data to the server.

[1554] 3. Step 3:

[1555] The server receives the data and analyzes the content.

[1556] 4. Step 4:

[1557] The server uses an emotion engine to recognize emotions from the user's input. For example, if the user includes many positive comments, the emotion engine will recognize the emotion as "joy."

[1558] 5. Step 5:

[1559] The server generates optimized prompts based on the emotional state and the company's prompt templates.

[1560] 6. Step 6:

[1561] The server inputs prompts into the AI ​​model to generate sentences that reflect the user's emotional state of "joy."

[1562] 7. Step 7:

[1563] The server sends the generated text to the terminal.

[1564] 8. Step 8:

[1565] The user reviews the generated text and manually adjusts it if necessary.

[1566] 9. Step 9:

[1567] The device sends the final text including the adjustments to the server as feedback.

[1568] 10. Step 10:

[1569] The server receives the feedback, analyzes the content, and stores it as learning data.

[1570] 11. Step 11:

[1571] The user then clicks the send button to send the final text that has been adjusted. The device then connects to the mail server and sends the text to the specified recipient.

[1572] Example 2

[1573] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1574] In corporate communication, it is important to generate texts that are optimized for each company's unique culture and organizational structure. At the same time, it is also necessary to consider the user's emotional state when generating texts. However, existing systems often lack sufficient emotion recognition and optimization for each company, hindering effective communication. Furthermore, they lack mechanisms for manual adjustment of generated texts and efficient feedback integration. This leads to problems in improving text quality and reducing user satisfaction.

[1575] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1576] In this invention, the server includes: a means for generating prompts optimized for each company; a means for an AI model to generate sentences based on internal relationships and past message patterns; a means for sending the generated sentences to a user's device and accepting manual adjustments from the user; a means for accumulating user feedback as learning data and reflecting it in future prompt and sentence generation; a means for sending the final adjusted sentences; a means for receiving and preprocessing the sent data; a means for analyzing input data using an emotion engine to recognize the user's emotional state; a means for optimizing prompts based on the recognized emotional data; and a means for inputting the generated prompts into a generative AI model to generate sentences. This enables effective sentence generation that is optimized for each company's unique culture and organizational structure and takes the user's emotional state into consideration. Furthermore, by efficiently incorporating user feedback, the quality of the generated sentences and user satisfaction can be improved.

[1577] The "server" is a device that receives data sent by users, analyzes it, generates appropriate sentences using a generative AI model, and finally transmits them.

[1578] A "terminal" is a device that provides an interface for users to operate, input data, and review and adjust the generated text.

[1579] "Company-optimized prompts" are guidance or instructions that are customized to fit a specific company's culture and organizational structure.

[1580] An "AI model" is an artificial intelligence algorithm that performs natural language processing based on input data and automatically generates sentences in a specified format.

[1581] An "emotion engine" is a program or device that analyzes and recognizes the emotional state of a user from data entered by the user.

[1582] A "prompt" is a predetermined instruction or explanation that is input into an AI model.

[1583] "Feedback" is information provided by a user when conveying corrections or opinions about the generated text to the server.

[1584] "Preprocessing" is a process performed by the server to prepare the transmitted data it receives in a format that is easy to analyze.

[1585] "Optimized prompts" are instructions or explanations that are optimized to take into account the user's emotional state and company-specific factors.

[1586] A "generative AI model" is a system that uses artificial intelligence technology to generate sentences based on user input data and optimized prompts.

[1587] "Sentence generation" is the process by which an AI model creates sentences in natural language based on optimized prompts.

[1588] The present invention provides a sentence generation system that combines an emotion engine and a generative AI model using prompts optimized for each company to effectively communicate within the company. Hereinafter, an embodiment of the present invention will be described in detail.

[1589] System configuration

[1590] server

[1591] A server is a device that performs several major functions:

[1592] 1. Receiving input data

[1593] The server receives data sent from the device (e.g., project name, team member information, etc.) This data is sent in JSON or XML format.

[1594] 2. Emotion recognition

[1595] The server analyzes the received data and uses an emotion engine to identify the user's emotional state. For example, the emotion toward the project name "Next Generation AI Development" is recognized as "Joy."

[1596] 3. Prompt Optimization

[1597] The server generates optimized prompts based on the emotion recognition results and company-specific templates. For example, if the emotion is "joy," a positive-toned prompt is used.

[1598] 4. Sentence generation

[1599] The server then inputs the optimized prompts into a generative AI model, specifically OpenAI's GPT-3, to generate the final sentence.

[1600] 5. Learn and incorporate feedback

[1601] The server receives manual adjustments and feedback from users and stores that information as learning data to help generate future prompts and improve the AI ​​model.

[1602] 6. Sending text

[1603] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[1604] Terminal

[1605] The device is directly operated by the user and provides the following functions:

[1606] 1. Providing a data input interface

[1607] It provides an interface for users to enter data (e.g., project name, start date, team member information, etc.).

[1608] 2. Displaying the generated text

[1609] It has the function of displaying the generated text in an appropriate format so that the user can check it.

[1610] 3. Provide a manual adjustment form

[1611] A form is provided for users to manually adjust the generated text, and the adjusted text is sent to the server as feedback.

[1612] User

[1613] The user is the person who operates this system and performs the following operations.

[1614] 1. Entering data

[1615] Use the device to enter information about the content you want to send and the recipient. For example, for "Notice of the start of a new project," enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki."

[1616] 2. Check and adjust the generated text

[1617] Check the generated text returned by the server and make manual adjustments as necessary. For example, change "Tanaka Ichiro" to "Tanaka Saburo."

[1618] 3. Final submission

[1619] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1620] Specific examples

[1621] For example, to send an announcement about the launch of a new project, the user types the following at the terminal:

[1622] "New project start notification"

[1623] "Project name = Next generation AI development"

[1624] "Start date=October 1, 2023"

[1625] "Team members: Ichiro Tanaka, Jiro Suzuki"

[1626] The server receives this data, and uses the emotion engine to generate prompts based on the emotion it recognizes as "joy," ultimately generating the following sentence:

[1627] "The next-generation AI development project has begun. We appreciate your cooperation."

[1628] The generated text is displayed on the user's device, and after the user confirms and adjusts it, the final text is sent to the specified destination.

[1629] This system can generate sentences that are suited to each company's unique culture and organizational structure, and can also provide effective communication that takes into account the user's emotional state. Furthermore, by efficiently incorporating feedback, the quality of the generated sentences can be improved.

[1630] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1631] Step 1: Data entry

[1632] Subject: User

[1633] Users use the device interface to enter information about the content and recipients they want to send, such as the project name, the start date of the new project, and team member information.

[1634] Input: Project name, start date, team member information

[1635] Output: Input data

[1636] For example, a user might enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki" as a "Notification of the start of a new project."

[1637] Step 2: Submitting input data

[1638] Subject: Terminal

[1639] The terminal sends the data entered by the user to the server in JSON or XML format.

[1640] Input: User-entered data (e.g., project name, start date, team member information)

[1641] Output: Data sent to the server

[1642] Specifically, the terminal converts the user's input data into an appropriate format and transmits it to the server.

[1643] Step 3: Receiving input data

[1644] Subject: Server

[1645] The server receives the data sent from the device. Since the data is in JSON or XML format, it parses it and converts it into a data structure for analysis.

[1646] Input: Data sent from the terminal (JSON or XML)

[1647] Output: Parsable data structure

[1648] The server performs preprocessing to analyze the received data, converting it into a format such as a string or a number, and stores it in an internal data structure.

[1649] Step 4: Emotion Recognition

[1650] Subject: Server

[1651] The server analyzes the received data using an emotion engine to recognize the user's emotional state. It processes the data to identify the emotional state that can be inferred from the input data.

[1652] Input: A parsable data structure

[1653] Output: Emotional state data (e.g., happy, angry, sad)

[1654] The server's emotion engine identifies the user's emotion as "happiness" based on the input data and passes the result to the next processing step.

[1655] Step 5: Prompt optimization

[1656] Subject: Server

[1657] The server generates optimized prompts based on emotion recognition results and templates customized for each company. The template and emotion data are combined to create prompts tailored to the customer's emotional state.

[1658] Input: Emotional state data, company-specific templates

[1659] Output: Optimized prompt

[1660] Specifically, if the emotional state is "joy," a prompt with a positive tone (e.g., "The next-generation AI development project is starting!") is generated.

[1661] Step 6: Sentence generation

[1662] Subject: Server

[1663] The server inputs the optimized prompts into a generative AI model (e.g., OpenAI GPT-3) to generate natural language sentences that are consistent with the company's culture and reflect the user's emotional state.

[1664] Input: Optimized prompts

[1665] Output: Generated sentence

[1666] Specifically, based on optimized prompts, it generates sentences such as, "The next-generation AI development project has begun. We appreciate your cooperation."

[1667] Step 7: Displaying the generated sentences

[1668] Subject: Terminal

[1669] The terminal displays the generated text returned by the server to the user, with appropriate formatting and highlighting.

[1670] Input: Generated text returned by the server

[1671] Output: Text for display

[1672] The displayed text is set up so that it is easy for the user to check, and the content is made easier to understand by highlighting and annotating it.

[1673] Step 8: Manual adjustment

[1674] Subject: User

[1675] The user can review the generated text and manually adjust it if necessary, for example, to correct misspelled names or wording.

[1676] Input: Generated text displayed

[1677] Output: Adjusted text

[1678] For example, the user corrects "Tanaka Ichiro" to "Tanaka Saburo" and sends the content to the next step.

[1679] Step 9: Send your feedback

[1680] Subject: Terminal

[1681] The device sends the user's manual adjustments to the server as feedback, which is then stored as learning data.

[1682] Input: Adjusted text

[1683] Output: Feedback sent to the server

[1684] The device will then send the user's manual adjustments back to the server and use them as feedback.

[1685] Step 10: Learning feedback

[1686] Subject: Server

[1687] The server accumulates the feedback received from the user as learning data and uses it for future sentence generation and prompt optimization. It analyzes the feedback and updates the learning data.

[1688] Input: Adjusted feedback data

[1689] Output: Updated training data

[1690] For example, we will improve our AI models and prompt generation algorithms based on user corrections to prevent similar errors.

[1691] Step 11: Final submission

[1692] Subject: Server

[1693] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[1694] Input: Final, adjusted text

[1695] Output: Send to specified destination

[1696] Specifically, the generated text is sent to a specified email address or internal notification system and delivered to the recipient designated by the user.

[1697] In this way, each step works in conjunction with each other to achieve effective sentence generation that is optimized for each company and takes into account the user's emotional state.

[1698] (Application example 2)

[1699] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1700] In logistics centers and other workplaces, it is important to communicate in a way that takes into account the work situation and the emotional state of employees. However, conventional systems are unable to recognize employees' emotions and generate appropriate messages based on them. This makes it difficult to respond flexibly to their emotional state, which can result in reduced work efficiency and increased employee stress.

[1701] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1702] In this invention, the server includes a means for generating prompts optimized for each company, a means for an AI model to generate sentences based on internal human relationships and past message patterns, a means including an emotion engine that recognizes the user's emotional state, and a means for reflecting emotional data from the user in the generation of prompts, thereby enabling optimal sentence generation according to the user's emotional state.

[1703] "Company-optimized prompts" are prompts that are optimized to fit a specific company's business flow, culture, and communication style.

[1704] "Internal relationships" refer to the connections and relationships between employees working within the same company.

[1705] "Past message patterns" refer to the tendencies, formats, and contents of messages previously exchanged.

[1706] An "AI model" is an algorithm or structure that uses artificial intelligence techniques to perform a specific task.

[1707] "User's emotional state" is information indicating the user's current emotions, including emotions such as joy, anger, and sadness.

[1708] An "emotion engine" is software or algorithm that recognizes emotions from user input data and provides the results.

[1709] "Feedback" refers to the evaluations, opinions, and adjustments that users provide to the system.

[1710] A "prompt generation algorithm" refers to the procedures, methods, and rules for generating optimal prompts, based on which prompts are created.

[1711] The system for implementing this invention consists of a server, a terminal, and a user. This system uses prompts optimized for each company, and an AI model generates sentences based on internal human relationships and past message patterns. Furthermore, by incorporating an emotion engine that recognizes the user's emotions, it becomes possible to generate messages that take into account the user's emotional state. The specific components and operation of the system are described below.

[1712] Server Configuration

[1713] The server includes means to:

[1714] 1. Means of receiving input data

[1715] The server receives data such as project overviews and team member information entered from the device, and sends this data in JSON or XML format.

[1716] 2. Emotion recognition means

[1717] The server analyzes the input data and recognizes the user's emotional state using an emotion engine. It identifies the emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[1718] 3. Prompt Generation Methods

[1719] The server uses the received data and emotion recognition results to generate personalized prompts for each company, which are tailored to fit the specific organizational structure and culture.

[1720] 4. Sentence generation means

[1721] The server then inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[1722] 5. Feedback as a learning tool

[1723] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[1724] 6. Means of sending text

[1725] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[1726] Device configuration

[1727] The terminal includes means for:

[1728] 1. Data input interface

[1729] It provides an interface for users to use and allows them to input the necessary information (e.g., the content they want to send, target information, etc.).

[1730] 2. Display of generated text

[1731] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[1732] 3. Manual Adjustment Form

[1733] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[1734] User operations

[1735] The user uses the system in the following steps:

[1736] 1. Entering data

[1737] Using the device's interface, you enter information about the message and the recipient, such as "Notification of new project start" or "Team member roles."

[1738] 2. Check and adjust the generated text

[1739] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[1740] 3. Final submission

[1741] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1742] Specific examples

[1743] As a concrete example of a server and terminal working together, the following shows a business notification system in a logistics center:

[1744] 1. The administrator types "New shipment is delayed again" into the terminal.

[1745] 2. The server recognizes the user's emotion as "anger."

[1746] 3. The server generates a prompt saying, "Let's try to address any issues calmly and work together to solve them."

[1747] 4. Based on this prompt, the generative AI model generates the sentence, "Let's stay calm and solve the problem that's causing the delay together."

[1748] 5. The administrator reviews the generated text and adjusts it if necessary.

[1749] 6. The adjusted text is fed back to the server and finally sent.

[1750] Example prompt sentence:

[1751] "Let's calmly address the issues that are causing delays and solve them together."

[1752] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1753] Step 1:

[1754] The terminal receives the data sent by the user. Specifically, the user enters a message such as "A new shipment is delayed again," and the terminal sends this information to the server in a data format (e.g., JSON format).

[1755] Input: User-entered message ("New shipment is delayed again")

[1756] Output: Data sent to the server (JSON format)

[1757] Step 2:

[1758] The server receives the data sent from the device and passes it to the emotion recognition engine, which analyzes the user's emotional state from the message and identifies the emotion "anger."

[1759] Input: Data received from the terminal (JSON format message "New shipment is delayed again")

[1760] Output: Analysis result (emotional state "anger")

[1761] Step 3:

[1762] The server uses a prompt generation algorithm based on the emotion recognition results to generate the optimal prompt. In this case, the generated prompt is, "Let's try to address any issues calmly and work together to solve them."

[1763] Input: Emotion recognition result (emotion state "anger")

[1764] Output: Optimized prompt ("Let's try to address any issues calmly and work together to solve them.")

[1765] Step 4:

[1766] The server provides the optimized prompt to the generative AI model, which then generates the optimal sentence based on the prompt. In this case, the sentence generated is, "Let's calmly address the issue that is causing the delay and solve it together."

[1767] Input: Optimized prompts ("Let's try to address any issues calmly and work together to solve them.")

[1768] Output: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[1769] Step 5:

[1770] The server transmits the generated text to the terminal, which displays the generated text to the user.

[1771] Input: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[1772] Output: The text that is displayed to the user

[1773] Step 6:

[1774] The user reviews the displayed text and manually adjusts it if necessary, for example changing "Let's resolve it" to "Let's deal with it."

[1775] Input: The displayed text

[1776] Output: Text after manual adjustment by the user

[1777] Step 7:

[1778] The device sends the user's adjusted sentences as feedback to the server, which stores this feedback as learning data and uses it to generate future prompts and improve the AI ​​model.

[1779] Input: Text after manual adjustment by user ("Let's stay calm and deal with the issues that are causing delays together.")

[1780] Output: Data stored as feedback

[1781] Step 8:

[1782] The server then sends the final adjusted text to the specified destination. The actual communication method is linked to the email server or internal communication tools.

[1783] Input: Final adjusted sentence ("Let's calmly address the issues that are causing delays and work together to resolve them.")

[1784] Output: The text sent to the specified destination

[1785] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1786] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1787] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1788] [Fourth embodiment]

[1789] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1790] 7, a 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.

[1791] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1792] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1793] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1794] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1795] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1796] The control object 443 includes a display device, LEDs in the eyes, and motors for driving 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1797] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1798] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1799] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1800] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1801] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1802] This invention is a system in which AI generates sentences using prompts optimized for each company, taking into account internal human relationships and past message patterns. Below, each component and operation of this system are explained in natural language.

[1803] Software Configuration

[1804] This system is mainly composed of a server, a terminal, and a user. The specific operation of each component is explained below.

[1805] server

[1806] 1. Receiving input data

[1807] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.).

[1808] This data is sent in formats such as JSON or XML.

[1809] 2. Prompt Optimization

[1810] The server generates prompts based on the input data, which are customized for each company and optimized to fit their specific organizational structure and culture.

[1811] For example, this includes the relationship between a specific boss and a subordinate, or past email patterns.

[1812] 3. Sentence generation

[1813] The server inputs the optimized prompts into the AI ​​model and generates sentences according to the specified format.

[1814] The AI ​​model uses machine learning algorithms to generate sentences based on the input prompts.

[1815] 4. Learning Feedback

[1816] It receives manual adjustments and feedback from users and stores them as learning data.

[1817] The feedback will identify areas that need adjustment and will be used to generate future prompts and improve the AI ​​model.

[1818] 5. Sending text

[1819] The final adjusted text is sent to the specified destination.

[1820] Sending is carried out in conjunction with email servers and internal communication tools.

[1821] Terminal

[1822] 1. Data input interface

[1823] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[1824] 2. Displaying the generated text

[1825] The generated text returned from the server is displayed in the interface.

[1826] The generated text is highlighted and annotated as appropriate to make it easier for users to check.

[1827] 3. Manual Adjustment Form

[1828] Provide a form that allows the user to manually adjust the generated text.

[1829] It has the function of sending the adjusted text to the server as feedback.

[1830] User

[1831] 1. Entering data

[1832] The user uses the terminal interface to input information about the content they want to send and the recipient.

[1833] For example, enter information such as "Notification of new project start" and "Team member roles."

[1834] 2. Check and adjust the generated text

[1835] Check the generated text returned by the server and make manual adjustments as necessary.

[1836] Once the adjustments are complete, the final text is sent to the server as feedback.

[1837] 3. Final submission

[1838] Check the final text after adjustments and press the send button to send it to the specified recipient.

[1839] Specific examples

[1840] For example, to send a project launch notice, the user might enter:

[1841] 1. User (Device)

[1842] Enter the project name, start date, and team member information as "Notification of new project start."

[1843] 2. Server

[1844] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[1845] 3. Server

[1846] The generated text is sent to the user's terminal.

[1847] 4. User (Terminal)

[1848] Review the generated text and edit it if necessary.

[1849] 5. User (Terminal)

[1850] The edited text is sent to the server as feedback.

[1851] 6. Server

[1852] Add feedback to the training data to help improve it further.

[1853] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[1854] The processing flow will be explained below.

[1855] Step 1:

[1856] The device displays an interface for the user to input information about what they want to send and who they want to send it to, including detailed project name, start date, team member roles, etc.

[1857] Step 2:

[1858] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[1859] Step 3:

[1860] The server receives the data sent from the device, parses it, and extracts the necessary information, such as the project name, start date, and team members.

[1861] Step 4:

[1862] The server uses the extracted data to generate personalized prompts tailored to the company's culture, organizational structure, and internal relationships.

[1863] Step 5:

[1864] The server then inputs the generated prompts into an AI model, which uses machine learning algorithms to generate sentences based on the prompts, with language and tone that reflects the company's culture.

[1865] Step 6:

[1866] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[1867] Step 7:

[1868] The user reviews the generated text and manually adjusts it as needed, for example by correcting specific job titles or the roles of specific members.

[1869] Step 8:

[1870] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[1871] Step 9:

[1872] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for future prompt optimization and AI model improvement.

[1873] Step 10:

[1874] The user checks and adjusts the final text and sends it from the device. The device then works in conjunction with the mail server and internal communication tools to send the text to the specified destination.

[1875] This series of processing steps enables users to efficiently generate business documents and realize communication that is in line with the unique culture and background of the company.

[1876] Example 1

[1877] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1878] In today's corporate environment, generating sentences that are efficient and reflect the organization's unique culture and interpersonal relationships is extremely important. However, conventional sentence generation systems lack a mechanism for generating sentences using prompts optimized for each company, requiring significant manual adjustments. Furthermore, they lack a means to effectively utilize user feedback to improve the sentence generation algorithm. This reduces the efficiency of the entire sentence generation process and significantly increases the time and effort required.

[1879] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1880] In this invention, the server includes means for receiving input data and generating optimized prompts based on the data, means for inputting the optimized prompts into an AI model and generating sentences based on internal human relationships and past message patterns, means for sending the generated sentences to a user's terminal and accepting manual adjustments by the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, and means for sending the final adjusted sentences to a specified destination. This makes it possible to efficiently generate sentences optimized for each company and significantly reduce the user's effort.

[1881] "Input data" refers to information provided by users to the system, including project overviews and team member information.

[1882] A "prompt" is an instruction or question that is input into an AI model and serves as the basis for generating sentences.

[1883] "Optimized prompts" refer to prompts that are customized based on the company's characteristics, culture, internal relationships, etc.

[1884] An "AI model" is an artificial intelligence program built using machine learning algorithms that generates natural language sentences based on prompts.

[1885] "Generated sentences" refer to sentences generated by the AI ​​model based on optimized prompts.

[1886] "User terminal" refers to a device such as a computer or smartphone used by a user to input data and check generated text.

[1887] "Manual adjustment" refers to the act of a user manually correcting or editing the generated text.

[1888] "Feedback" refers to information that users send back to the system regarding adjusted text and improvements.

[1889] "Training data" is a data set that the system uses to improve its performance, and includes feedback from users.

[1890] "Designated Destination" refers to the address or contact information indicating the specific recipient to whom the final tailored document is to be sent.

[1891] This invention is a system in which a user inputs data using a terminal, generates optimized prompts based on that data, and generates sentences under specific conditions using an AI model. Here, we will explain the specific names of the hardware and software used and the program's processing procedures.

[1892] System configuration

[1893] This system mainly consists of a server, terminals, and users.

[1894] server

[1895] The server has the following roles:

[1896] 1. Receiving input data

[1897] The server receives the data sent by the user from the device. The data is usually sent in a format such as JSON or XML. The received data is stored in memory and used for subsequent processing.

[1898] 2. Prompt optimization

[1899] The server generates prompts based on the input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns.To generate and optimize these prompts, Python scripts and database management systems (e.g., MySQL or PostgreSQL) are used.

[1900] 3. Sentence generation

[1901] The server inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format. The AI ​​model is built using machine learning libraries such as TensorFlow and PyTorch, and generates sentences based on the prompts using natural language processing techniques.

[1902] 4. Sending the generated text

[1903] The server encodes the generated text into JSON format and sends it to the terminal as an HTTP response.

[1904] 5. Learning Feedback

[1905] The server receives user feedback and stores it in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future.

[1906] Terminal

[1907] The terminal has the following roles:

[1908] 1. Providing a data input interface

[1909] The terminal provides the user with a data entry interface that allows them to enter information about the content and audience they wish to send. This interface is built using HTML, CSS, and JavaScript.

[1910] 2. Displaying the generated text

[1911] The terminal displays the generated text returned from the server on its interface, with appropriate highlighting and annotations added to make it easier for the user to check.

[1912] 3. Provide a manual adjustment form

[1913] The terminal provides a form that allows the user to manually adjust the generated text, and sends the adjustment results to the server as feedback.

[1914] User

[1915] The user has the following roles:

[1916] 1. Entering data

[1917] The user uses the device's data entry interface to enter information about the content and recipients of the message, such as the project name, start date, and team member information for "Notice of the start of a new project."

[1918] 2. Check and adjust the generated text

[1919] The user checks the generated sentences sent back from the server and manually adjusts them if necessary. The adjusted sentences are then sent back to the server as feedback.

[1920] Specific examples

[1921] For example, the specific procedure for sending a notification of the start of a new project is as follows.

[1922] 1. User (Device)

[1923] Enter the project name, start date, and team member information as "Notification of new project start." A specific example of how to enter this information is shown below.

[1924] Document Title: Notice of New Project Launch

[1925] Project Name: Project X

[1926] Start date: October 1, 2023

[1927] Team members: Mr. A, Mr. B, Mr. C

[1928] 2. Server

[1929] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[1930] 3. Server

[1931] The generated text is sent to the user's terminal.

[1932] 4. User (Terminal)

[1933] Review the generated text and edit it if necessary.

[1934] 5. User (Terminal)

[1935] The edited text is sent to the server as feedback.

[1936] 6. Server

[1937] Add feedback to the training data to help improve it further.

[1938] This process allows users to reduce the effort required for writing documents in their work and quickly and efficiently create documents that suit the unique culture and human relationships of their company.

[1939] In this way, the present invention realizes optimal sentence generation adapted to each company.

[1940] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1941] Step 1:

[1942] The user enters the data.

[1943] Specific behavior:

[1944] The user uses the device's data entry interface to enter the content they want to send (e.g., "Notification of the start of a new project") and target information (e.g., project name, start date, team member roles). This input data will be the basis for subsequent processing.

[1945] Input: Project name, start date, team member information

[1946] Output: Data entered by the user

[1947] Step 2:

[1948] The terminal sends the input data to the server.

[1949] Specific behavior:

[1950] The terminal encodes the data entered by the user in JSON format and sends it to the server using the HTTPS protocol, which ensures secure transmission.

[1951] Input: Data entered by the user

[1952] Output: Encoded data in JSON format

[1953] Step 3:

[1954] The server receives the input data.

[1955] Specific behavior:

[1956] The server receives the data sent from the terminal and stores it in memory. It is also possible to temporarily store the data using a database management system (e.g., MySQL or PostgreSQL).

[1957] Input: JSON encoded data

[1958] Output: Data stored in memory

[1959] Step 4:

[1960] The server optimizes the prompt.

[1961] Specific behavior:

[1962] The server generates prompts based on input data and optimizes them by taking into account the characteristics of each company, internal relationships, and past message patterns. Python scripts are used to extract relevant information from the database and reflect it in the prompts.

[1963] Input: Data stored in memory

[1964] Output: Optimized prompt

[1965] Step 5:

[1966] The server inputs the prompts into the AI ​​model to generate sentences.

[1967] Specific behavior:

[1968] The server inputs the optimized prompts into a generative AI model, which is built using machine learning libraries such as TensorFlow and PyTorch, and generates natural language sentences based on the prompts.

[1969] Input: Optimized prompts

[1970] Output: Generated sentence

[1971] Step 6:

[1972] The server sends the generated text to the terminal.

[1973] Specific behavior:

[1974] The server encodes the generated text again into JSON format and sends it to the terminal as an HTTP response.

[1975] Input: Generated sentence

[1976] Output: The generated document encoded in JSON format.

[1977] Step 7:

[1978] The terminal displays the generated text to the user.

[1979] Specific behavior:

[1980] The device then displays the generated text in an interface, styled using HTML and CSS to highlight specific feeds and important information.

[1981] Input: The generated text encoded in JSON format

[1982] Output: The generated text displayed to the user

[1983] Step 8:

[1984] The user reviews and adjusts the generated text.

[1985] Specific behavior:

[1986] The user can review the generated text displayed on the device and manually adjust it as needed, for example, by changing specific wording or adjusting the writing style. Adjustments are made using a form on the device.

[1987] Input: The generated text displayed to the user

[1988] Output: User-adjusted text

[1989] Step 9:

[1990] The user sends the adjusted text to the server as feedback.

[1991] Specific behavior:

[1992] The user then sends the adjusted document back to the server from their device, again encoded in JSON and using the HTTPS protocol.

[1993] Input: User-adjusted text

[1994] Output: The adjusted text encoded in JSON format

[1995] Step 10:

[1996] The server takes the feedback as learning.

[1997] Specific behavior:

[1998] The server stores the received feedback in a database as learning data, which is used to generate prompts and improve the AI ​​model in the future. A database management system (e.g., MySQL or PostgreSQL) is used for storage.

[1999] Input: Adjusted text encoded in JSON format

[2000] Output: Feedback data stored in a database

[2001] (Application example 1)

[2002] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2003] The food delivery industry requires fast and accurate customer support, but traditional methods have led to inconsistencies in response time and quality. It has also been difficult to share know-how and make continuous improvements to properly respond to inquiries.

[2004] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[2005] In this invention, the server includes means for generating prompts optimized for each company, means for an AI-generated model to generate sentences based on internal human relationships and past message patterns, means for sending the generated sentences to an information processing terminal and accepting manual adjustments from the user, means for accumulating feedback from the user as learning data and reflecting it in future prompt and sentence generation, means for sending the final adjusted sentences to another information processing terminal, and means for generating prompts and sentences for customer support inquiries specific to the food delivery industry. This enables fast and accurate customer support responses, standardizes the quality of responses, and enables continuous improvement.

[2006] "Means for generating prompts optimized for each company" refers to methods or functions for automatically creating optimal prompts based on each company's specific organizational structure, culture, and past message patterns.

[2007] "Means for an AI generation model to generate sentences based on internal interpersonal relationships and past message patterns" refers to a method or function that uses input data based on specific internal interpersonal relationships and past message patterns to generate appropriate sentences using an AI generation model.

[2008] "Means for transmitting the generated text to an information processing terminal and accepting manual adjustments by the user" refers to a method or function that transmits text generated by an AI-generated model to a terminal used by the user and enables the user to manually adjust the text.

[2009] "Means for accumulating user feedback as learning data and reflecting it in future prompts and sentence generation" refers to a method or function that accumulates the user's manually adjusted content and feedback in a database and uses it when generating the next prompt or sentence.

[2010] "Means for sending the final adjusted text to another information processing terminal" refers to a method or function for sending the text that has been final checked and adjusted by the user to a designated recipient.

[2011] "Means for generating prompts and generating sentences to respond to customer support inquiries specialized for the food delivery industry" refers to methods and functions for generating prompts and AI-based sentences necessary to respond to customer support inquiries specialized for the content of such inquiries in the food delivery industry.

[2012] The "means for inputting data including inquiry content" refers to a method or function for a user to input data such as customer support inquiry content and customer information.

[2013] The "means for analyzing data and generating an optimized prompt" refers to a method or function for analyzing input data and automatically generating an optimal prompt.

[2014] "Means for analyzing the details of manual adjustments made by the user and storing the details as feedback" refers to a method or function for analyzing the details of manual adjustments made by the user and storing the information as feedback in a database.

[2015] "Means for improving the AI-generated model and prompt-generation algorithm based on accumulated feedback" refers to methods or functions for continuously improving the AI-generated model and prompt-generation algorithm using accumulated feedback data.

[2016] This invention aims to improve the efficiency of customer support in the food delivery industry by using prompts optimized for each company and AI-generated models. This system mainly consists of a server, a terminal, and a user.

[2017] server

[2018] The server performs the following series of processes.

[2019] 1. Receiving input data

[2020] The server receives data (e.g., inquiry details, customer information) entered by the user from the terminal. The data is usually sent in a format such as JSON or XML.

[2021] 2. Prompt Generation

[2022] The server generates prompts based on the input data, which are specific to the food delivery industry and include the following examples:

[2023] "Inquiry: My order hasn't arrived"

[2024] 3. Sentence generation

[2025] The server inputs the optimized prompts into a generative AI model (e.g., GPT-3) to generate appropriate sentences for the query. The generative AI model generates sentences using machine learning algorithms.

[2026] 4. Learning Feedback

[2027] It receives manual adjustments and feedback from users and stores it as training data, identifying areas that need adjustment and helping to generate future prompts and improve the AI ​​model.

[2028] 5. Sending text

[2029] Finally, the adjusted text is sent to the designated recipient, linked to the email server or internal instant communication tool.

[2030] Terminal

[2031] The terminal provides the following features:

[2032] 1. Data input interface

[2033] It provides an interface for users to use and allows them to input necessary information (e.g., inquiry details, customer information, etc.).

[2034] 2. Displaying the generated text

[2035] The generated text sent back from the server is displayed in the interface, and it is also possible to highlight and annotate the generated text for the user's convenience.

[2036] 3. Manual Adjustment Form

[2037] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[2038] User

[2039] Users interact with the system as follows:

[2040] 1. Entering data

[2041] The user uses the terminal interface to input the inquiry and customer information. For example, the user may input "The food I ordered has not arrived."

[2042] 2. Check and adjust the generated text

[2043] Check the generated text sent back from the server and manually adjust it if necessary. For example, adjust the text to something like, "We apologize for the inconvenience. We will recheck it immediately. While you wait, we will provide you with a 1,000 yen discount coupon."

[2044] 3. Providing Feedback

[2045] The adjusted sentences are sent to the server as feedback and stored as learning data in the system.

[2046] This system enables quick and accurate customer support responses, standardizes the quality of responses, and allows for continuous improvement. As a specific example of its use, the server generates the following prompt sentence:

[2047] Inquiry: The food I ordered hasn't arrived. A customer who has made a similar inquiry in the past frequently orders sushi.

[2048] The sentence generated based on this is, "We are sorry to hear that your sushi order has not yet arrived, and we apologize for the inconvenience. We will recheck for you shortly. While you wait, we will provide you with a 1,000 yen discount coupon," which is then adjusted on the device and sent as the final sentence.

[2049] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[2050] Step 1:

[2051] The user uses the terminal interface to input the inquiry and customer information. At this time, the inquiry information is entered as "The food I ordered has not arrived." This input data is sent to the server in JSON format.

[2052] Step 2:

[2053] The server analyzes the input data received from the device and generates a prompt optimized for each company. Specifically, it takes into account data such as past inquiry patterns and customer preferences to generate a prompt like the one below.

[2054] "Inquiry: The food I ordered hasn't arrived. This customer has made a similar inquiry in the past and frequently orders sushi."

[2055] Step 3:

[2056] The server that generated the prompt inputs the prompt into a generative AI model (e.g., GPT-3). The AI ​​model generates an appropriate sentence based on the prompt. In doing so, the AI ​​analyzes the input data and uses patterns learned from previous inquiries. An example of a generated sentence might be, "We're sorry to hear that your sushi order hasn't arrived yet, and we apologize for the inconvenience. We'll check again soon. While you wait, we'll provide you with a 1,000 yen discount coupon."

[2057] Step 4:

[2058] The server sends the generated text to the terminal, where the user can review it on the terminal interface. The generated text may also be highlighted or annotated to make it easier for the user to review.

[2059] Step 5:

[2060] The user can review the generated text and manually adjust it if necessary, for example to include more specific instructions or additional information, which is then sent back to the server.

[2061] Step 6:

[2062] The server receives user feedback on the adjustments made, which is stored in a database and used to improve future prompt and sentence generation.

[2063] Step 7:

[2064] The final adjusted text is then sent to the specified destination by the server, enabling quick and accurate customer support, improving the quality of customer support in the food delivery industry.

[2065] This system's series of processes will significantly improve the quality and efficiency of customer support responses in the food delivery industry. For example, it will be able to respond quickly to complex customer needs and continuously improve the quality of responses by utilizing learning data.

[2066] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[2067] This invention combines an emotion engine that recognizes the user's emotions with a system in which AI generates sentences using prompts optimized for each company and taking into account internal human relationships and past message patterns. Below, each component and operation of this system is explained in natural language.

[2068] Software Configuration

[2069] This system is primarily composed of a server, a terminal, and a user. By introducing an emotion engine, it becomes possible to generate text that takes into account the user's emotional state, achieving more effective communication. The specific operation of each component is explained below.

[2070] server

[2071] 1. Receiving input data

[2072] The server receives data entered by the user from the terminal (e.g., project name, team member roles, etc.) This data is sent in a format such as JSON or XML.

[2073] 2. Emotion recognition

[2074] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[2075] 3. Prompt Optimization

[2076] The server generates optimized prompts based on the user's emotion recognition data and customized prompt templates for each company, which are tailored to fit the specific organizational structure and culture.

[2077] 4. Sentence generation

[2078] The server then inputs the optimized prompts into the AI ​​model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[2079] 5. Learning Feedback

[2080] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[2081] 6. Sending text

[2082] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[2083] Terminal

[2084] 1. Data input interface

[2085] The terminal provides an interface for the user to use, allowing the user to input necessary information (e.g., the content to be sent, target information, etc.).

[2086] 2. Displaying the generated text

[2087] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[2088] 3. Manual Adjustment Form

[2089] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[2090] User

[2091] 1. Entering data

[2092] The user uses the device interface to input information about the content and recipients of the message, such as "notification of the start of a new project" or "roles of team members."

[2093] 2. Check and adjust the generated text

[2094] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[2095] 3. Final submission

[2096] Check the final text after adjustments and press the send button to send it to the specified recipient.

[2097] Specific examples

[2098] For example, consider sending a notification about the start of a new project. The specific steps are as follows:

[2099] 1. User (Device)

[2100] Enter the project name, start date, and team member information as "Notification of new project start."

[2101] 2. Server

[2102] It generates prompts based on the received data and passes them to an AI model to generate sentences.

[2103] 3. Emotion recognition

[2104] The server analyzes the input data and recognizes the user's emotions using an emotion engine. It identifies the user's emotional state and reflects it in the prompt generation.

[2105] 4. Server

[2106] The generated text is sent to the user's terminal.

[2107] 5. User (Terminal)

[2108] Review the generated text and edit it if necessary.

[2109] 6. User (Terminal)

[2110] The edited text is sent to the server as feedback.

[2111] 7. Server

[2112] The feedback is accumulated as learning data and used to generate future prompts and improve the AI ​​model.

[2113] This system allows users to generate optimal sentences according to their emotional state, enabling effective communication that is in line with the company's unique culture and background.

[2114] The processing flow will be explained below.

[2115] Step 1:

[2116] The terminal displays an interface for the user to enter data: the content they want to send and information about the recipient (e.g., project name, start date, team member roles, etc.).

[2117] Step 2:

[2118] The terminal sends the data entered by the user to the server in JSON or XML format using a secure communication protocol.

[2119] Step 3:

[2120] The server receives the data sent from the device, parses it, and extracts information about the project name, start date, and team members.

[2121] Step 4:

[2122] The server uses an emotion engine to recognize the user's emotions, identifying emotions such as joy, anger, and sadness based on the user's input data and past data.

[2123] Step 5:

[2124] The server uses the user's emotional state and a customized prompt template for each company to generate optimized prompts that are tailored to the specific organizational structure and culture.

[2125] Step 6:

[2126] The server inputs the generated prompts into an AI model to generate sentences, which then use machine learning algorithms to generate sentences that reflect the user's emotional state.

[2127] Step 7:

[2128] The server sends the generated text to the terminal, which provides an interface for displaying the text to the user.

[2129] Step 8:

[2130] The user reviews the generated text and makes manual adjustments as needed, for example, correcting specific job titles or member roles.

[2131] Step 9:

[2132] The device sends the final text edited by the user and the edit details to the server as feedback, including which parts were edited and how.

[2133] Step 10:

[2134] The server receives the feedback, analyzes the adjustments, and stores the feedback as learning data for use in generating future prompts and improving the AI ​​model.

[2135] Step 11:

[2136] The user then checks the final edited text and presses the send button to send it to the specified recipient. The device then sends the text in conjunction with the email server and internal communication tools.

[2137] Specific examples

[2138] Example: New project start notification

[2139] 1. Step 1:

[2140] The user inputs the project name, start date, and team member information into the terminal interface as a "new project start notification."

[2141] 2. Step 2:

[2142] The terminal sends the input data to the server.

[2143] 3. Step 3:

[2144] The server receives the data and analyzes the content.

[2145] 4. Step 4:

[2146] The server uses an emotion engine to recognize emotions from the user's input. For example, if the user includes many positive comments, the emotion engine will recognize the emotion as "joy."

[2147] 5. Step 5:

[2148] The server generates optimized prompts based on the emotional state and the company's prompt templates.

[2149] 6. Step 6:

[2150] The server inputs prompts into the AI ​​model to generate sentences that reflect the user's emotional state of "joy."

[2151] 7. Step 7:

[2152] The server sends the generated text to the terminal.

[2153] 8. Step 8:

[2154] The user reviews the generated text and manually adjusts it if necessary.

[2155] 9. Step 9:

[2156] The device sends the final text including the adjustments to the server as feedback.

[2157] 10. Step 10:

[2158] The server receives the feedback, analyzes the content, and stores it as learning data.

[2159] 11. Step 11:

[2160] The user then clicks the send button to send the final text that has been adjusted. The device then connects to the mail server and sends the text to the specified recipient.

[2161] Example 2

[2162] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2163] In corporate communication, it is important to generate texts that are optimized for each company's unique culture and organizational structure. At the same time, it is also necessary to consider the user's emotional state when generating texts. However, existing systems often lack sufficient emotion recognition and optimization for each company, hindering effective communication. Furthermore, they lack mechanisms for manual adjustment of generated texts and efficient feedback integration. This leads to problems in improving text quality and reducing user satisfaction.

[2164] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[2165] In this invention, the server includes: a means for generating prompts optimized for each company; a means for an AI model to generate sentences based on internal relationships and past message patterns; a means for sending the generated sentences to a user's device and accepting manual adjustments from the user; a means for accumulating user feedback as learning data and reflecting it in future prompt and sentence generation; a means for sending the final adjusted sentences; a means for receiving and preprocessing the sent data; a means for analyzing input data using an emotion engine to recognize the user's emotional state; a means for optimizing prompts based on the recognized emotional data; and a means for inputting the generated prompts into a generative AI model to generate sentences. This enables effective sentence generation that is optimized for each company's unique culture and organizational structure and takes the user's emotional state into consideration. Furthermore, by efficiently incorporating user feedback, the quality of the generated sentences and user satisfaction can be improved.

[2166] The "server" is a device that receives data sent by users, analyzes it, generates appropriate sentences using a generative AI model, and finally transmits them.

[2167] A "terminal" is a device that provides an interface for users to operate, input data, and review and adjust the generated text.

[2168] "Company-optimized prompts" are guidance or instructions that are customized to fit a specific company's culture and organizational structure.

[2169] An "AI model" is an artificial intelligence algorithm that performs natural language processing based on input data and automatically generates sentences in a specified format.

[2170] An "emotion engine" is a program or device that analyzes and recognizes the emotional state of a user from data entered by the user.

[2171] A "prompt" is a predetermined instruction or explanation that is input into an AI model.

[2172] "Feedback" is information provided by a user when conveying corrections or opinions about the generated text to the server.

[2173] "Preprocessing" is a process performed by the server to prepare the transmitted data it receives in a format that is easy to analyze.

[2174] "Optimized prompts" are instructions or explanations that are optimized to take into account the user's emotional state and company-specific factors.

[2175] A "generative AI model" is a system that uses artificial intelligence technology to generate sentences based on user input data and optimized prompts.

[2176] "Sentence generation" is the process by which an AI model creates sentences in natural language based on optimized prompts.

[2177] The present invention provides a sentence generation system that combines an emotion engine and a generative AI model using prompts optimized for each company to effectively communicate within the company. Hereinafter, an embodiment of the present invention will be described in detail.

[2178] System configuration

[2179] server

[2180] A server is a device that performs several major functions:

[2181] 1. Receiving input data

[2182] The server receives data sent from the device (e.g., project name, team member information, etc.) This data is sent in JSON or XML format.

[2183] 2. Emotion recognition

[2184] The server analyzes the received data and uses an emotion engine to identify the user's emotional state. For example, the emotion toward the project name "Next Generation AI Development" is recognized as "Joy."

[2185] 3. Prompt Optimization

[2186] The server generates optimized prompts based on the emotion recognition results and company-specific templates. For example, if the emotion is "joy," a positive-toned prompt is used.

[2187] 4. Sentence generation

[2188] The server then inputs the optimized prompts into a generative AI model, specifically OpenAI's GPT-3, to generate the final sentence.

[2189] 5. Learn and incorporate feedback

[2190] The server receives manual adjustments and feedback from users and stores that information as learning data to help generate future prompts and improve the AI ​​model.

[2191] 6. Sending text

[2192] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[2193] Terminal

[2194] The device is directly operated by the user and provides the following functions:

[2195] 1. Providing a data input interface

[2196] It provides an interface for users to enter data (e.g., project name, start date, team member information, etc.).

[2197] 2. Displaying the generated text

[2198] It has the function of displaying the generated text in an appropriate format so that the user can check it.

[2199] 3. Provide a manual adjustment form

[2200] A form is provided for users to manually adjust the generated text, and the adjusted text is sent to the server as feedback.

[2201] User

[2202] The user is the person who operates this system and performs the following operations.

[2203] 1. Entering data

[2204] Use the device to enter information about the content you want to send and the recipient. For example, for "Notice of the start of a new project," enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki."

[2205] 2. Check and adjust the generated text

[2206] Check the generated text returned by the server and make manual adjustments as necessary. For example, change "Tanaka Ichiro" to "Tanaka Saburo."

[2207] 3. Final submission

[2208] Check the final text after adjustments and press the send button to send it to the specified recipient.

[2209] Specific examples

[2210] For example, to send an announcement about the launch of a new project, the user types the following at the terminal:

[2211] "New project start notification"

[2212] "Project name = Next generation AI development"

[2213] "Start date=October 1, 2023"

[2214] "Team members: Ichiro Tanaka, Jiro Suzuki"

[2215] The server receives this data, and uses the emotion engine to generate prompts based on the emotion it recognizes as "joy," ultimately generating the following sentence:

[2216] "The next-generation AI development project has begun. We appreciate your cooperation."

[2217] The generated text is displayed on the user's device, and after the user confirms and adjusts it, the final text is sent to the specified destination.

[2218] This system can generate sentences that are suited to each company's unique culture and organizational structure, and can also provide effective communication that takes into account the user's emotional state. Furthermore, by efficiently incorporating feedback, the quality of the generated sentences can be improved.

[2219] The flow of the identification process in the second embodiment will be described with reference to FIG.

[2220] Step 1: Data entry

[2221] Subject: User

[2222] Users use the device interface to enter information about the content and recipients they want to send, such as the project name, the start date of the new project, and team member information.

[2223] Input: Project name, start date, team member information

[2224] Output: Input data

[2225] For example, a user might enter the project name "Next Generation AI Development," the start date "October 1, 2023," and the team members "Ichiro Tanaka, Jiro Suzuki" as a "Notification of the start of a new project."

[2226] Step 2: Submitting input data

[2227] Subject: Terminal

[2228] The terminal sends the data entered by the user to the server in JSON or XML format.

[2229] Input: User-entered data (e.g., project name, start date, team member information)

[2230] Output: Data sent to the server

[2231] Specifically, the terminal converts the user's input data into an appropriate format and transmits it to the server.

[2232] Step 3: Receiving input data

[2233] Subject: Server

[2234] The server receives the data sent from the device. Since the data is in JSON or XML format, it parses it and converts it into a data structure for analysis.

[2235] Input: Data sent from the terminal (JSON or XML)

[2236] Output: Parsable data structure

[2237] The server performs preprocessing to analyze the received data, converting it into a format such as a string or a number, and stores it in an internal data structure.

[2238] Step 4: Emotion Recognition

[2239] Subject: Server

[2240] The server analyzes the received data using an emotion engine to recognize the user's emotional state. It processes the data to identify the emotional state that can be inferred from the input data.

[2241] Input: A parsable data structure

[2242] Output: Emotional state data (e.g., happy, angry, sad)

[2243] The server's emotion engine identifies the user's emotion as "happiness" based on the input data and passes the result to the next processing step.

[2244] Step 5: Prompt optimization

[2245] Subject: Server

[2246] The server generates optimized prompts based on emotion recognition results and templates customized for each company. The template and emotion data are combined to create prompts tailored to the customer's emotional state.

[2247] Input: Emotional state data, company-specific templates

[2248] Output: Optimized prompt

[2249] Specifically, if the emotional state is "joy," a prompt with a positive tone (e.g., "The next-generation AI development project is starting!") is generated.

[2250] Step 6: Sentence generation

[2251] Subject: Server

[2252] The server inputs the optimized prompts into a generative AI model (e.g., OpenAI GPT-3) to generate natural language sentences that are consistent with the company's culture and reflect the user's emotional state.

[2253] Input: Optimized prompts

[2254] Output: Generated sentence

[2255] Specifically, based on optimized prompts, it generates sentences such as, "The next-generation AI development project has begun. We appreciate your cooperation."

[2256] Step 7: Displaying the generated sentences

[2257] Subject: Terminal

[2258] The terminal displays the generated text returned by the server to the user, with appropriate formatting and highlighting.

[2259] Input: Generated text returned by the server

[2260] Output: Text for display

[2261] The displayed text is set up so that it is easy for the user to check, and the content is made easier to understand by highlighting and annotating it.

[2262] Step 8: Manual adjustment

[2263] Subject: User

[2264] The user can review the generated text and manually adjust it if necessary, for example, to correct misspelled names or wording.

[2265] Input: Generated text displayed

[2266] Output: Adjusted text

[2267] For example, the user corrects "Tanaka Ichiro" to "Tanaka Saburo" and sends the content to the next step.

[2268] Step 9: Send your feedback

[2269] Subject: Terminal

[2270] The device sends the user's manual adjustments to the server as feedback, which is then stored as learning data.

[2271] Input: Adjusted text

[2272] Output: Feedback sent to the server

[2273] The device will then send the user's manual adjustments back to the server and use them as feedback.

[2274] Step 10: Learning feedback

[2275] Subject: Server

[2276] The server accumulates the feedback received from the user as learning data and uses it for future sentence generation and prompt optimization. It analyzes the feedback and updates the learning data.

[2277] Input: Adjusted feedback data

[2278] Output: Updated training data

[2279] For example, we will improve our AI models and prompt generation algorithms based on user corrections to prevent similar errors.

[2280] Step 11: Final submission

[2281] Subject: Server

[2282] The final, adjusted text is then sent to the specified recipient, in conjunction with the email server and internal communication tools.

[2283] Input: Final, adjusted text

[2284] Output: Send to specified destination

[2285] Specifically, the generated text is sent to a specified email address or internal notification system and delivered to the recipient designated by the user.

[2286] In this way, each step works in conjunction with each other to achieve effective sentence generation that is optimized for each company and takes into account the user's emotional state.

[2287] (Application example 2)

[2288] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2289] In logistics centers and other workplaces, it is important to communicate in a way that takes into account the work situation and the emotional state of employees. However, conventional systems are unable to recognize employees' emotions and generate appropriate messages based on them. This makes it difficult to respond flexibly to their emotional state, which can result in reduced work efficiency and increased employee stress.

[2290] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[2291] In this invention, the server includes a means for generating prompts optimized for each company, a means for an AI model to generate sentences based on internal human relationships and past message patterns, a means including an emotion engine that recognizes the user's emotional state, and a means for reflecting emotional data from the user in the generation of prompts, thereby enabling optimal sentence generation according to the user's emotional state.

[2292] "Company-optimized prompts" are prompts that are optimized to fit a specific company's business flow, culture, and communication style.

[2293] "Internal relationships" refer to the connections and relationships between employees working within the same company.

[2294] "Past message patterns" refer to the tendencies, formats, and contents of messages previously exchanged.

[2295] An "AI model" is an algorithm or structure that uses artificial intelligence techniques to perform a specific task.

[2296] "User's emotional state" is information indicating the user's current emotions, including emotions such as joy, anger, and sadness.

[2297] An "emotion engine" is software or algorithm that recognizes emotions from user input data and provides the results.

[2298] "Feedback" refers to the evaluations, opinions, and adjustments that users provide to the system.

[2299] A "prompt generation algorithm" refers to the procedures, methods, and rules for generating optimal prompts, based on which prompts are created.

[2300] The system for implementing this invention consists of a server, a terminal, and a user. This system uses prompts optimized for each company, and an AI model generates sentences based on internal human relationships and past message patterns. Furthermore, by incorporating an emotion engine that recognizes the user's emotions, it becomes possible to generate messages that take into account the user's emotional state. The specific components and operation of the system are described below.

[2301] Server Configuration

[2302] The server includes means to:

[2303] 1. Means of receiving input data

[2304] The server receives data such as project overviews and team member information entered from the device, and sends this data in JSON or XML format.

[2305] 2. Emotion recognition means

[2306] The server analyzes the input data and recognizes the user's emotional state using an emotion engine. It identifies the emotional state (e.g., joy, anger, sadness, etc.) and uses it as a parameter for generating prompts.

[2307] 3. Prompt Generation Methods

[2308] The server uses the received data and emotion recognition results to generate personalized prompts for each company, which are tailored to fit the specific organizational structure and culture.

[2309] 4. Sentence generation means

[2310] The server then inputs the optimized prompts into a generative AI model, which generates sentences according to the specified format, with expressions and tones that match the user's emotional state.

[2311] 5. Feedback as a learning tool

[2312] It receives manual adjustments and feedback from users and stores it as learning data. The feedback identifies which parts have been adjusted and is used to generate future prompts and improve the AI ​​model.

[2313] 6. Means of sending text

[2314] The final, adjusted text is sent to the specified recipient. This is done in conjunction with the email server and internal communication tools.

[2315] Device configuration

[2316] The terminal includes means for:

[2317] 1. Data input interface

[2318] It provides an interface for users to use and allows them to input the necessary information (e.g., the content they want to send, target information, etc.).

[2319] 2. Display of generated text

[2320] The generated text returned from the server is displayed in the interface, and the generated text is highlighted and annotated as appropriate to make it easier for the user to review.

[2321] 3. Manual Adjustment Form

[2322] It provides a form that allows users to manually adjust the generated text, and has the function of sending the adjusted text to the server as feedback.

[2323] User operations

[2324] The user uses the system in the following steps:

[2325] 1. Entering data

[2326] Using the device's interface, you enter information about the message and the recipient, such as "Notification of new project start" or "Team member roles."

[2327] 2. Check and adjust the generated text

[2328] Review the generated text returned by the server and make manual adjustments as needed, such as correcting specific job titles or specific member roles.

[2329] 3. Final submission

[2330] Check the final text after adjustments and press the send button to send it to the specified recipient.

[2331] Specific examples

[2332] As a concrete example of a server and terminal working together, the following shows a business notification system in a logistics center:

[2333] 1. The administrator types "New shipment is delayed again" into the terminal.

[2334] 2. The server recognizes the user's emotion as "anger."

[2335] 3. The server generates a prompt saying, "Let's try to address any issues calmly and work together to solve them."

[2336] 4. Based on this prompt, the generative AI model generates the sentence, "Let's stay calm and solve the problem that's causing the delay together."

[2337] 5. The administrator reviews the generated text and adjusts it if necessary.

[2338] 6. The adjusted text is fed back to the server and finally sent.

[2339] Example prompt sentence:

[2340] "Let's calmly address the issues that are causing delays and solve them together."

[2341] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2342] Step 1:

[2343] The terminal receives the data sent by the user. Specifically, the user enters a message such as "A new shipment is delayed again," and the terminal sends this information to the server in a data format (e.g., JSON format).

[2344] Input: User-entered message ("New shipment is delayed again")

[2345] Output: Data sent to the server (JSON format)

[2346] Step 2:

[2347] The server receives the data sent from the device and passes it to the emotion recognition engine, which analyzes the user's emotional state from the message and identifies the emotion "anger."

[2348] Input: Data received from the terminal (JSON format message "New shipment is delayed again")

[2349] Output: Analysis result (emotional state "anger")

[2350] Step 3:

[2351] The server uses a prompt generation algorithm based on the emotion recognition results to generate the optimal prompt. In this case, the generated prompt is, "Let's try to address any issues calmly and work together to solve them."

[2352] Input: Emotion recognition result (emotion state "anger")

[2353] Output: Optimized prompt ("Let's try to address any issues calmly and work together to solve them.")

[2354] Step 4:

[2355] The server provides the optimized prompt to the generative AI model, which then generates the optimal sentence based on the prompt. In this case, the sentence generated is, "Let's calmly address the issue that is causing the delay and solve it together."

[2356] Input: Optimized prompts ("Let's try to address any issues calmly and work together to solve them.")

[2357] Output: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[2358] Step 5:

[2359] The server transmits the generated text to the terminal, which displays the generated text to the user.

[2360] Input: Generated sentence ("Let's calmly address the issue that's causing the delay and solve it together.")

[2361] Output: The text that is displayed to the user

[2362] Step 6:

[2363] The user reviews the displayed text and manually adjusts it if necessary, for example changing "Let's resolve it" to "Let's deal with it."

[2364] Input: The displayed text

[2365] Output: Text after manual adjustment by the user

[2366] Step 7:

[2367] The device sends the user's adjusted sentences as feedback to the server, which stores this feedback as learning data and uses it to generate future prompts and improve the AI ​​model.

[2368] Input: Text after manual adjustment by user ("Let's stay calm and deal with the issues that are causing delays together.")

[2369] Output: Data stored as feedback

[2370] Step 8:

[2371] The server then sends the final adjusted text to the specified destination. The actual communication method is linked to the email server or internal communication tools.

[2372] Input: Final adjusted sentence ("Let's calmly address the issues that are causing delays and work together to resolve them.")

[2373] Output: The text sent to the specified destination

[2374] 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 control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[2375] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[2376] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[2377] The emotion identification model 59 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 an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[2378] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[2379] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[2380] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[2381] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[2382] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs 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 a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[2383] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[2384] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[2385] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[2386] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[2387] 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.

[2388] It is not necessary to store all 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 all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[2389] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[2390] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with 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). Also, the hardware resource that executes the specific processing may be a single processor.

[2391] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[2392] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[2393] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[2394] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[2395] The following is further disclosed regarding the above embodiment.

[2396] (Claim 1)

[2397] A means of generating personalized prompts for each company;

[2398] A means for AI models to generate sentences based on internal relationships and past message patterns;

[2399] means for transmitting the generated text to a user's terminal and accepting manual adjustments by the user;

[2400] A means of accumulating user feedback as learning data and reflecting it in future prompts and sentence generation;

[2401] a means for transmitting the final adjusted text;

[2402] A system including:

[2403] (Claim 2)

[2404] A means of inputting data, including project overview and team member information;

[2405] a means for parsing the data to generate optimized prompts;

[2406] 10. The system of claim 1.

[2407] (Claim 3)

[2408] means for analyzing the user's manual adjustments and storing the results as feedback;

[2409] a means for improving the AI ​​model and prompt generation algorithm based on accumulated feedback; and

[2410] 10. The system of claim 1.

[2411] "Example 1"

[2412] (Claim 1)

[2413] means for receiving input data and generating an optimized prompt based on the input data;

[2414] A means of feeding optimized prompts into an AI model to generate sentences based on internal relationships and past messaging patterns; and

[2415] means for transmitting the generated text to a user's terminal and accepting manual adjustments by the user;

[2416] A means of accumulating user feedback as learning data and reflecting it in future prompts and sentence generation;

[2417] means for transmitting the final adjusted text to a designated destination;

[2418] A system including:

[2419] (Claim 2)

[2420] A means of inputting data, including project overview and team member information;

[2421] a means for parsing the data to generate optimized prompts;

[2422] 10. The system of claim 1.

[2423] (Claim 3)

[2424] means for analyzing the user's manual adjustments and storing the results as feedback;

[2425] a means for improving the AI ​​model and prompt generation algorithm based on accumulated feedback; and

[2426] 10. The system of claim 1.

[2427] "Application Example 1"

[2428] (Claim 1)

[2429] A means of generating personalized prompts for each company;

[2430] A means for an AI generative model to generate sentences based on internal relationships and past message patterns;

[2431] means for transmitting the generated text to an information processing terminal and accepting manual adjustments by a user;

[2432] A means of accumulating user feedback as learning data and reflecting it in future prompts and sentence generation;

[2433] means for transmitting the final adjusted text to another information processing terminal;

[2434] A means for generating prompts and sentences for customer support inquiries specific to the food delivery industry;

[2435] A system including:

[2436] (Claim 2)

[2437] a means for inputting data including the nature of the inquiry;

[2438] a means for parsing the data to generate optimized prompts;

[2439] 10. The system of claim 1.

[2440] (Claim 3)

[2441] a means for analyzing the manual adjustments made by the user and storing the results as feedback;

[2442] a means for improving the AI ​​generation model and prompt generation algorithm based on accumulated feedback; and

[2443] 10. The system of claim 1.

[2444] "Example 2: Combining Emotion Engines"

[2445] (Claim 1)

[2446] A means of generating personalized prompts for each company;

[2447] A means for AI models to generate sentences based on internal relationships and past message patterns;

[2448] means for transmitting the generated text to a user's terminal and accepting manual adjustments by the user;

[2449] A means of accumulating user feedback as learning data and reflecting it in future prompts and sentence generation;

[2450] a means for transmitting the final adjusted text;

[2451] means for receiving and pre-processing the transmitted data;

[2452] means for analyzing input data using an emotion engine to recognize the user's emotional state;

[2453] a means for optimizing prompts based on the recognized emotion data; and

[2454] a means for inputting the generated prompts into a generative AI model to generate sentences;

[2455] A system including:

[2456] (Claim 2)

[2457] A means of inputting data, including project overview and team member information;

[2458] a means for parsing the data to generate optimized prompts;

[2459] means for displaying the generated text on a user's terminal and providing an interface for manual adjustment;

[2460] 10. The system of claim 1.

[2461] (Claim 3)

[2462] means for analyzing the user's manual adjustments and storing the results as feedback;

[2463] a means for improving the AI ​​model and prompt generation algorithm based on accumulated feedback; and

[2464] a means for recognizing a user's emotional state using an emotion engine and reflecting the recognition in prompt generation;

[2465] 10. The system of claim 1.

[2466] "Application example 2 when combining emotion engines"

[2467] (Claim 1)

[2468] A means of generating personalized prompts for each company;

[2469] A means for AI models to generate sentences based on internal relationships and past message patterns;

[2470] means including an emotion engine for recognizing an emotional state of a user;

[2471] a means for reflecting emotional data from a user in prompt generation;

[2472] means for transmitting the generated text to a user's terminal and accepting manual adjustments by the user;

[2473] A means of accumulating user feedback as learning data and reflecting it in future prompts and sentence generation;

[2474] a means for transmitting the final adjusted text;

[2475] A system including:

[2476] (Claim 2)

[2477] A means of inputting data, including project overview and team member information;

[2478] a means for parsing the data to generate optimized prompts;

[2479] means for identifying an emotional state of a user based on emotion recognition;

[2480] 10. The system of claim 1.

[2481] (Claim 3)

[2482] means for analyzing the user's manual adjustments and storing the results as feedback;

[2483] a means for improving the AI ​​model and prompt generation algorithm based on accumulated feedback; and

[2484] A means of optimizing prompts by incorporating data from the emotion engine;

[2485] 10. The system of claim 1. [Explanation of symbols]

[2486] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devi...

Claims

1. A means of generating personalized prompts for each company; A means for AI models to generate sentences based on internal relationships and past message patterns; means for transmitting the generated text to a user's terminal and accepting manual adjustments by the user; A means of accumulating user feedback as learning data and reflecting it in future prompts and sentence generation; a means for transmitting the final adjusted text; A system including:

2. A means of inputting data, including project overview and team member information; a means for parsing the data to generate optimized prompts; The system of claim 1 .

3. means for analyzing the user's manual adjustments and storing the results as feedback; a means for improving the AI ​​model and prompt generation algorithm based on accumulated feedback; and The system of claim 1 .

Images