system

Abstract

Provide a system. 【Solution means】 Means for collecting data from in-house emails and intranets, Means for filtering the collected data and extracting useful information, Means for storing the extracted information in a database, Means for searching the database based on the inquiry content and obtaining relevant information, Means for generating a draft of a primary answer using a generative artificial intelligence model based on the obtained information, Means for presenting the generated draft of the primary answer to the user, Means for sending the primary answer modified by the user as the final answer, A system including the above.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, and includes steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a character of the chatbot, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In modern enterprises, each department routinely handles a large number of inquiries, which consumes a lot of labor and becomes a business burden. In addition, the content of the inquiries covers a wide range, and quick and accurate answers are required. However, when the person in charge responds manually, it takes time and effort, and human errors may occur. In such a situation, it is desired to improve the efficiency and accuracy of handling inquiries. In addition, in the future, a system that can also be applied to handling inquiries from outside the company and outsourcing to other companies is required.

Means for Solving the Problems

[0005] This invention provides a system for collecting data from internal email and intranets, filtering that data, and extracting useful information. Specifically, it includes data collection means, filtering means, database storage means, and means for generating a draft initial response using a generative artificial intelligence model. The system presents the user with a draft initial response, which the user can then revise and submit as a final response. The revised final response can also be saved in a database. This improves the efficiency and accuracy of inquiry handling and reduces the workload for each department. Furthermore, it can be applied to handling external inquiries and outsourcing to other companies in the future.

[0006] "Data collection means" refers to a device or program that has the function of periodically or at specified times acquiring necessary data from internal email and intranets.

[0007] "Filtering means" refers to a device or program that has the function of analyzing data acquired by data collection means and extracting useful information.

[0008] "Database storage means" refers to a device or program that has the function of storing useful information extracted by filtering means in a database.

[0009] A "generative artificial intelligence model" is a machine learning model that generates answers to inquiries based on acquired data.

[0010] The "initial response draft" is an initial response generated by a generative artificial intelligence model, and can be modified by the user.

[0011] A "user" refers to an individual or department that operates the system and enters inquiries or modifies responses.

[0012] A "database" is a storage device used to store acquired data and generated responses.

[0013] A "terminal" is a computer or device operated by a user, used for accessing a system and entering data.

[0014] By defining each element of this invention as described above, it becomes possible to accurately interpret the claims. [Brief explanation of the drawing]

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

Mode for Carrying Out the Invention

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

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

[0018] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), APU (Accelerated Processing Unit), etc.

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

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

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

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

[0023] [First Embodiment]

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

[0025] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

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

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

[0028] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

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

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

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

[0032] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

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

[0034] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

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

[0036] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. The generated draft of the initial response is presented to the user, who can modify it as needed and submit it as the final response. The program processing of this system is described below in natural language.

[0037] 1. Data Collection

[0038] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[0039] 2. Data filtering

[0040] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[0041] 3. Database storage

[0042] The server stores filtered, useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[0043] 4. Generating the initial response

[0044] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response. This draft combines the relevant information to create the initial response text.

[0045] 5. Presentation and revision of the initial response.

[0046] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or add information as needed. When making corrections, it is also possible to add specific instructions or comments.

[0047] 6. Submit and save your final response.

[0048] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[0049] Specific example

[0050] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[0051] The generative AI server generates a draft of the initial response as follows:

[0052] If the remote control is not responding, please try the following steps:

[0053] 1. Check the batteries in the remote control and replace them with new ones.

[0054] 2. Make sure the remote control is pointed towards the TV.

[0055] 3. Check that there are no obstacles between the TV and the remote control.

[0056] 4. Clean the light-receiving part of the television.

[0057] If these steps do not resolve the issue, please contact our support center.

[0058] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[0059] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[0060] The following describes the processing flow.

[0061] Data acquisition and filtering

[0062] Step 1:

[0063] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[0064] Step 2:

[0065] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[0066] Step 3:

[0067] The server analyzes and filters the data in temporary storage, extracting useful information (e.g., inquiry details, important notifications).

[0068] Step 4:

[0069] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[0070] Step 5:

[0071] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[0072] Generation of the first response

[0073] Step 6:

[0074] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[0075] Step 7:

[0076] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[0077] Step 8:

[0078] The generative AI server searches the database for information related to the query. The search results are stored.

[0079] Step 9:

[0080] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3®). This draft is generated as an appropriate response to the inquiry.

[0081] Presentation and revision of the initial response

[0082] Step 10:

[0083] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[0084] Step 11:

[0085] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[0086] Submit and save your final response.

[0087] Step 12:

[0088] The user submits their revised initial response as the final response. This confirms the response.

[0089] Step 13:

[0090] The device saves the final answer to the database. This makes the saved answer available for later reference.

[0091] Through the above processing steps, the efficiency and accuracy of inquiry handling are improved. The specific actions of each step enable the entire system to function smoothly.

[0092] (Example 1)

[0093] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0094] In modern companies, email and internal networks are widely used as means of internal communication. However, gathering information and responding to inquiries through these means requires considerable effort, and efficiency improvements are needed. Furthermore, providing quick and accurate answers to inquiries is difficult, making it a challenge to improve customer satisfaction.

[0095] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0096] In this invention, the server includes means for collecting information from internal communication data and the internal network; means for analyzing the collected information and extracting useful data; means for storing the extracted data in a storage device; means for searching the storage device based on the content of the inquiry and obtaining relevant data; means for generating a draft of an initial response using a generative artificial intelligence model based on the obtained data; means for presenting the generated draft of the initial response to the user; and means for sending the initial response modified by the user as the final response. This makes it possible to streamline the inquiry handling process and improve customer satisfaction.

[0097] "Internal communication data" refers to data generated through digital communication methods used within a company, such as email, chat, and intranets.

[0098] An "internal network" refers to a limited network environment used for sharing information and communicating within a company.

[0099] "Means of collecting information" refers to methods using protocols and APIs that allow a server to retrieve data from email servers or intranets.

[0100] "Means of analyzing information and extracting useful data" refers to natural language processing and filtering technologies used to detect specific keywords and patterns from collected communication data and extract necessary information.

[0101] "Storage device" refers to databases and storage systems used to hold, manage, and retrieve data.

[0102] "Inquiry content" refers to questions or requests that users make to the server or system.

[0103] "Generative artificial intelligence models" refer to AI models that generate text using natural language processing and machine learning. Specifically, this includes advanced AI technologies such as GPT and BERT.

[0104] "Draft initial response" refers to the initial response text generated by a generative artificial intelligence model.

[0105] "Means of presentation to the user" refers to interface means that allow the user to review the initial draft of the response displayed on the terminal.

[0106] "Final answer" refers to the answer that has been revised and finalized by the user.

[0107] "Means of transmission" refers to email systems or other means of communication used to send the final response to the recipient.

[0108] This invention is a system that efficiently collects useful information from internal communication data and the company's internal network, and uses that information to generate a draft of an initial response using a generative artificial intelligence model. This streamlines the inquiry response process and enables the provision of quick and accurate answers.

[0109] 1. Data Collection

[0110] The server periodically accesses the company's email server and internal network to retrieve new data. It uses the IMAP protocol to retrieve emails and collects updates from the internal network via a REST API. This ensures that the server consistently collects the latest information, such as inquiries and important notifications.

[0111] 2. Data filtering

[0112] The server uses a natural language processing (NLP) library to filter the acquired data and extract useful information. By utilizing the NLP library, inquiries, important notifications, manual updates, and other relevant information are automatically analyzed and filtered.

[0113] 3. Database storage

[0114] The filtered, useful information is stored in storage by the server. Specifically, the information is stored using a database such as PostgreSQL or MongoDB. This database is categorized and indexed according to the query and answer content.

[0115] 4. Generating the initial response

[0116] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on this retrieved information, it generates a draft of an initial response using generative artificial intelligence models such as OpenAI's GPT model or Google's BERT model.

[0117] 5. Presentation and revision of the initial response.

[0118] The device displays a draft of the initial response to the user. The user reviews this draft and makes corrections or adds information as needed. The response is then optimized, including any additional instructions or comments from the user.

[0119] 6. Submit and save your final response.

[0120] Once the user submits the final, revised response, the device saves the response to its storage. This saved data can be referenced later and used as foundational data to quickly respond to similar inquiries in the future.

[0121] Specific example

[0122] For example, if there is an inquiry about a new remote control not working, the user would enter the inquiry as follows: "My new remote control isn't working, what should I do?". The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working. The generative AI server generates a draft initial response like this:

[0123] If the remote control is not responding, please try the following steps:

[0124] 1. Check the batteries in the remote control and replace them with new ones.

[0125] 2. Make sure the remote control is pointed towards the target device.

[0126] 3. Check that there are no obstacles between the device and the remote control.

[0127] 4. Clean the light-receiving part of the device.

[0128] If these steps do not resolve the issue, please contact our support center.

[0129] Related technologies

[0130] By using the generated AI model and prompt statements, the accuracy and speed of inquiry handling are improved, resulting in reduced workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies.

[0131] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0132] Step 1:

[0133] The server collects information from the company's internal email server and internal network. In this process, the server uses the IMAP protocol to retrieve emails and a REST API to retrieve updates from the internal network. Inputs are new emails and intranet updates, and output is the collected raw data.

[0134] Specifically, the server automatically connects to the mail server at 9 AM every morning and retrieves all new emails from the previous day. Internal network update information is also collected at the same time.

[0135] Step 2:

[0136] The server analyzes the data collected in Step 1 and extracts useful information. Here, a natural language processing (NLP) library is used to filter the data. The input is the collected raw data, and the output is the filtered useful information.

[0137] Specifically, the system analyzes the email content received by the server and extracts emails containing keywords such as "failure," "inquiry," and "support." Within the internal network, it collects documents tagged with "update," "procedure," and "urgent."

[0138] Step 3:

[0139] The server stores filtered, useful information in storage. Specifically, it uses a database such as PostgreSQL or MongoDB to store the information. The input is filtered, useful information, and the output is the data stored in the database.

[0140] Specifically, the server identifies emails inquiring about "remote control malfunction" and stores their contents in the "remote control related inquiries" table in the database.

[0141] Step 4:

[0142] When the terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Next, it uses a generative artificial intelligence model (for example, OpenAI's GPT model or Google's BERT model) to generate a draft of the initial response. The input is the user's inquiry and the relevant information stored in the database, and the output is the draft of the initial response.

[0143] Specifically, when a user inquires that "the new remote control isn't responding," the terminal sends this inquiry to the AI ​​server. The AI ​​server searches its database, retrieves information about the remote control malfunction, and generates a draft message like this: "If the remote control is not responding, try changing the batteries and checking the remote control's orientation and for any obstructions."

[0144] Step 5:

[0145] The terminal displays a draft of the generated initial response to the user. The user reviews this draft and enters corrections or additional information as needed. The input consists of the generated initial response draft and the user's feedback, and the output is the revised initial response draft.

[0146] Specifically, when a user sees the initial response provided and enters an additional request such as "Please specify the type of battery," the device updates the draft to reflect this feedback.

[0147] Step 6:

[0148] When the user submits the final, revised answer, the device saves that answer to its storage device. The input is the final, revised answer made by the user, and the output is the final answer data saved to the storage device.

[0149] Specifically, when a user clicks the "Submit Final Response" button, the device saves that response data to the "Response History" table in the database. This data is then used as a reference when handling future inquiries.

[0150] (Application Example 1)

[0151] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0152] In a factory setting, responding quickly and accurately to machine malfunctions or operational problems requires searching through a vast amount of information to find relevant data and providing it to workers. However, traditional methods involve manually searching for information based on inquiries and generating appropriate responses, which is time-consuming and labor-intensive. Furthermore, if relevant information cannot be found, users may become stuck, leading to work delays. An efficient system is needed to solve these problems.

[0153] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0154] In this invention, the server includes means for collecting data from internal email and intranets, means for filtering the collected data and extracting useful information, means for storing the extracted information in a database, means for searching the database based on the inquiry content and obtaining relevant information, means for generating a draft of a preliminary response using a generative artificial intelligence model based on the obtained information, means for presenting the generated draft of the preliminary response to the user, means for sending the preliminary response modified by the user as the final response, means for inputting the inquiry content using a smart device, means for displaying a default message if relevant information is not found, means for accepting user modifications, and means for saving user input and modifications in the database. This enables quick and accurate responses to troubles and inquiries within the factory, improving work efficiency and speeding up trouble resolution.

[0155] "Inside the company" refers to the internal workings of a company or organization.

[0156] "Email" is a means of sending and receiving text messages and files over the internet.

[0157] An "intranet" is an internal network used within a company; it utilizes internet technology but is isolated from the outside world.

[0158] "Data collection" is the process of systematically gathering information.

[0159] "Filtering" is the process of selecting necessary information and removing unnecessary information.

[0160] "Extraction" is the process of taking out the necessary parts from collected data.

[0161] A "database" is a system for efficiently storing, managing, and retrieving large amounts of data.

[0162] "Inquiry details" refer to the information that users submit to the system, such as questions or requests.

[0163] A "generative artificial intelligence model" is an algorithm that uses artificial intelligence technology to generate new information or answers.

[0164] A "primary response" is a draft answer that is initially generated in response to a user's inquiry.

[0165] A "smart device" is a high-performance portable device that can connect to the internet, such as a smartphone or tablet.

[0166] A "default message" is a standard message that is automatically displayed when certain conditions are not met.

[0167] A "user" is someone who uses a system or device.

[0168] "Means" refer to the methods and techniques used to achieve an objective.

[0169] "Modification" refers to the operation of making changes to a document or data that has been created.

[0170] The "final answer" is the answer that has been finalized after the user has made any revisions.

[0171] This invention is a system for streamlining troubleshooting and inquiry handling within a factory. The system collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft initial response based on that information. Finally, the generated draft initial response is presented to the user, who can revise it as needed and submit it as the final response.

[0172] System Configuration

[0173] hardware

[0174] Server: A server that performs data collection, filtering, and database storage.

[0175] Device: Smart devices such as smartphones and tablets.

[0176] Generative AI server: A server that runs generative AI models.

[0177] software

[0178] Generative artificial intelligence model: GPT-2 or an equivalent generative AI model.

[0179] Transformers Library: A Python library provided by Hugging Face.

[0180] Process Overview

[0181] Data collection:

[0182] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[0183] Data filtering:

[0184] The server filters the retrieved data and extracts useful information (e.g., inquiry details, important notifications, manual updates).

[0185] Database storage:

[0186] The server stores the extracted useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[0187] Generating the initial answer:

[0188] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response.

[0189] Presentation and revision of the initial response:

[0190] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or enter additional information as needed.

[0191] Submit and save your final response:

[0192] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[0193] Specific example

[0194] For example, consider a case where a user has an inquiry about a new machine not working. The user enters the inquiry, "The new machine isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a machine that isn't working.

[0195] The generative AI server generates a draft of the initial response as follows:

[0196] If the machine is not working, try the following steps:

[0197] 1. Check if the power is on.

[0198] 2. Check that the wiring is connected correctly.

[0199] 3. Check that the machine body has been cleaned.

[0200] 4. Check the machine's error messages.

[0201] If these steps do not resolve the issue, please contact our support center.

[0202] The terminal displays this draft to the user. The user enters additional information, such as, "Please tell me the specific steps to take when checking the wiring." Once the revisions are complete, the user submits the final response. The terminal saves this final response to the database so that it can be referenced later if needed.

[0203] In this way, efficient responses become possible when the same inquiry occurs again, and the amount of work required to handle inquiries can be significantly reduced.

[0204] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0205] Step 1:

[0206] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[0207] Input: Internal email and intranet data.

[0208] Data processing: The acquired raw data is analyzed, and each email and update information is organized as metadata.

[0209] Output: Initial dataset for filtering.

[0210] Step 2:

[0211] The server uses a filtering algorithm to filter the acquired data and extract useful information.

[0212] Input: Initial dataset.

[0213] Data processing: Select useful information and remove unnecessary information based on filtering rules.

[0214] Output: Filtered important information (e.g., inquiry details, manual update information).

[0215] Step 3:

[0216] The server stores the extracted useful information in a database.

[0217] Input: Filtered important information.

[0218] Data Storage: Information is appropriately arranged according to the database structure, and indexes are created.

[0219] Output: A database organized in a searchable format.

[0220] Step 4:

[0221] The terminal receives the user's inquiry and sends it to the generation AI server.

[0222] Input: User inquiry details.

[0223] Data transmission: The query content is sent to the generative AI server according to the protocol.

[0224] Output: The query content is forwarded to the AI ​​server.

[0225] Step 5:

[0226] The generative AI server searches the database and retrieves relevant information.

[0227] Input: User inquiry details.

[0228] Data retrieval: Quickly extract relevant information using database indexes.

[0229] Output: A set of related information.

[0230] Step 6:

[0231] The generative AI server uses a generative artificial intelligence model to generate a draft of the initial response based on the acquired information.

[0232] Input: A set of related information.

[0233] Data processing: Input prompt text into a generative AI model (such as GPT-2) and generate a response.

[0234] Output: Draft of the initial response.

[0235] Step 7:

[0236] The terminal presents the user with a draft of the initial response that has been generated.

[0237] Input: Draft of the initial response.

[0238] Data display: The draft of the initial response is displayed in the user interface.

[0239] Output: Displays the initial response that the user can see.

[0240] Step 8:

[0241] The user reviews the initial draft response provided and enters any necessary corrections or additional information.

[0242] Input: Draft of the initial response, user revisions and additional information.

[0243] Data processing: Reflects user-submitted corrections.

[0244] Output: Corrected initial response.

[0245] Step 9:

[0246] The terminal receives the user's corrections and sends the initial response as the final response, saving it to the database.

[0247] Input: Revised initial response.

[0248] Data storage: The final answer is saved in the database for future searches.

[0249] Output: The final answer stored in the database.

[0250] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0251] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions. The program processing of this system is described below in natural language.

[0252] 1. Data collection and filtering

[0253] The server periodically or at designated times accesses the company's email and intranet servers to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[0254] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[0255] The server converts filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database. This facilitates subsequent searching and AI processing.

[0256] 2. Generating the initial response

[0257] The terminal receives user inquiries through an input form. The terminal then sends the inquiry content to the generation AI server.

[0258] The generative AI server searches the database for information related to the query and stores the search results.

[0259] The generative AI server uses a generative AI model (e.g., GPT-3) based on the acquired relevant information to generate a draft of the initial response. This draft is generated as an appropriate response to the inquiry.

[0260] In this process, the generative AI server further utilizes an emotion engine. The emotion engine analyzes the user's inquiry and recognizes their emotions (e.g., anger, frustration, joy). Based on the recognized emotions, the emotion engine adjusts the tone and expression of the initial response.

[0261] 3. Presentation and revision of the initial response.

[0262] The device displays a draft of the initial response to the user, allowing them to review their answer.

[0263] Users can review this initial draft response and make corrections or add information as needed. When making corrections, they can also add specific instructions or comments.

[0264] 4. Submit and save your final response.

[0265] When the user submits their revised initial response as the final response, the generative AI server uses the emotion engine again to analyze the user's final input and confirm whether the user is satisfied.

[0266] The device saves the final answer to a database. This makes the saved answer available for later reference.

[0267] Specific example

[0268] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[0269] The generative AI server generates a draft of the initial response as follows:

[0270] If the remote control is not responding, please try the following steps:

[0271] 1. Check the batteries in the remote control and replace them with new ones.

[0272] 2. Make sure the remote control is pointed towards the TV.

[0273] 3. Check that there are no obstacles between the TV and the remote control.

[0274] 4. Clean the light-receiving part of the television.

[0275] If these steps do not resolve the issue, please contact our support center.

[0276] Additionally, an emotion engine is used to read the user's feelings of dissatisfaction from their inquiry and adjust the tone accordingly:

[0277] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0278] 1. Check the batteries in the remote control and replace them with new ones.

[0279] 2. Make sure the remote control is pointed towards the TV.

[0280] 3. Please check if there is any obstacle between the TV and the remote control.

[0281] 4. Please clean the light-receiving part of the TV.

[0282] If the problem is not solved after trying these, please contact the support center, and we will provide more detailed support.

[0283] The terminal displays this draft to the user. The user enters additional information such as "By the way, what type of battery should I use when replacing the battery?" When the correction is completed, the user sends the final answer. The terminal saves this final answer in the database so that it can be referred to later as needed.

[0284] In this way, the work of inquiry response can be made more efficient, and the man-hours of each department can be reduced. Since the system has high versatility, it can also be applied to external inquiry response and outsourcing for other companies in the future.

[0285] The following describes the processing flow.

[0286] Data collection and filtering

[0287] Step 1:

[0288] The server accesses the in-house email server and intranet server regularly or at a specified time to obtain the latest information.

[0289] Step 2:

[0290] The server saves the newly obtained email and intranet update information in temporary storage to facilitate subsequent processing.

[0291] Step 3:

[0292] The server analyzes the data in temporary storage and extracts useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[0293] Step 4:

[0294] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[0295] Step 5:

[0296] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[0297] Generation of the first response

[0298] Step 6:

[0299] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[0300] Step 7:

[0301] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[0302] Step 8:

[0303] The generative AI server searches the database for information related to the query. The search results are stored.

[0304] Step 9:

[0305] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is then generated as an appropriate response.

[0306] Adjustment by the Emotion Engine

[0307] Step 10:

[0308] When the generative AI server generates a draft of the primary response, it uses the emotion engine to analyze the emotion from the user's inquiry content. As a result, adjustments are made according to the user's emotion.

[0309] Step 11:

[0310] The generative AI server adjusts the tone and expression of the primary response according to the emotion recognized by the emotion engine. For example, when the user feels dissatisfied, expressions of apology and polite language are added.

[0311] Presentation and Revision of the Primary Response

[0312] Step 12:

[0313] The terminal displays the draft of the primary response generated to the user. As a result, the user can confirm the response.

[0314] <00​​​​​​​​​​​​​​​​​​​​​​​​The device saves the final answer to the database. This makes the saved answer available for later reference.

[0321] Specific example

[0322] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[0323] The generative AI server generates a draft of the initial response as follows:

[0324] If the remote control is not responding, please try the following steps:

[0325] 1. Check the batteries in the remote control and replace them with new ones.

[0326] 2. Make sure the remote control is pointed towards the TV.

[0327] 3. Check that there are no obstacles between the TV and the remote control.

[0328] 4. Clean the light-receiving part of the television.

[0329] If these steps do not resolve the issue, please contact our support center.

[0330] Furthermore, it uses an emotion engine to read the user's feelings of dissatisfaction and adjusts the tone as follows:

[0331] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0332] 1. Check the batteries in the remote control and replace them with new ones.

[0333] 2. Make sure the remote control is pointed towards the TV.

[0334] 3. Check that there are no obstacles between the TV and the remote control.

[0335] 4. Clean the light-receiving part of the television.

[0336] If these steps do not resolve the issue, please contact our support center for further assistance.

[0337] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[0338] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[0339] (Example 2)

[0340] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0341] Traditional inquiry handling systems lacked the ability to efficiently extract useful information from the company's vast internal email and intranet, organize and store it in an appropriate format, and automatically generate appropriate responses based on relevant information. In particular, they lacked the function to adjust response content while considering user emotions, which could result in decreased user satisfaction. Against this backdrop, there is a need for a system that can simultaneously improve the efficiency of inquiry handling and enhance user satisfaction.

[0342] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for collecting data from internal electronic messages and the internal network, means for filtering the collected data and extracting useful information, and means for converting the extracted information into a structured data format and storing it in a database. This makes it possible to efficiently organize and store the information necessary for responding to inquiries, and furthermore, to automatically generate responses that provide high user satisfaction using a generative artificial intelligence model and an emotion engine.

[0343] "Internal electronic messages" refer to the content of communications via email and messaging services used within a company or organization.

[0344] An "internal network" refers to an intranet or dedicated network used within a company or organization.

[0345] "Collecting data" means obtaining information from a specific data source.

[0346] "Filtering" is the process of selecting useful information from acquired data and removing unnecessary data.

[0347] A "structured data format" is a format that organizes information systematically and regularly, making it easy to store in a database (e.g., JSON, XML).

[0348] A "database" is a system for storing, managing, and making data searchable.

[0349] "Inquiry content" refers to the text of questions or requests that users submit to the system.

[0350] A "generative artificial intelligence model" is an artificial intelligence that generates natural language based on a large amount of data, such as GPT-3 or similar models.

[0351] A "draft initial response" is the first response proposed by a generative artificial intelligence model.

[0352] An "emotion engine" is a technology that analyzes the user's emotions from text and adjusts the tone and expression of the response based on that analysis.

[0353] A "user" is an individual or group that uses the system.

[0354] A "final answer" is an answer that has been confirmed and revised by the user.

[0355] "Saving" means recording information in a database so that it can be reused or searched later.

[0356] This invention is a system that collects data from internal electronic messages and internal networks, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the content of the inquiry, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions.

[0357] This system includes the following main components:

[0358] 1. Data collection and filtering functions

[0359] 2. Database storage function

[0360] 3. Function to receive inquiries from users

[0361] 4. Search function for related information

[0362] 5. Primary response generation function using generative AI models

[0363] 6. Tone adjustment function using an emotion engine

[0364] 7. Presentation of initial answers and user modification function

[0365] 8. Function to submit and save the final response.

[0366] Hardware and software to be used

[0367] Server: Performs data collection, filtering, database storage, retrieval, initial response generation, sentiment analysis, and final response storage. Protocols and tools used include IMAP (for email collection), HTTP requests (for intranet update collection), natural language processing (NLP) algorithms, generative artificial intelligence models (e.g., GPT-3), sentiment engines (e.g., IBM Watson® Tone Analyzer), and database management systems (e.g., MongoDB or MySQL®).

[0368] Terminal: Receives user inquiries, provides initial responses, and makes corrections. Web forms or desktop applications are used as the user interface.

[0369] Processing details

[0370] Data acquisition and filtering

[0371] The server periodically accesses the company's electronic messaging server and internal network to retrieve new messages and updates. For example, the server accesses the electronic messaging server using the IMAP protocol to retrieve new mail. The retrieved data is analyzed using NLP algorithms, and useful information is filtered.

[0372] Storage in database

[0373] The filtered, useful information is converted to JSON format and stored in a MongoDB or MySQL database. For example, data in the following format is stored:

[0374] json

[0375] {

[0376] "id": "12345",

[0377] "type": "inquiry",

[0378] "content": "New TV remote not responding"

[0379] }

[0380] Generation of the first response

[0381] When a user submits a query through their device, the server searches the database and retrieves relevant information. Based on this information, a generative AI model (e.g., GPT-3) generates a draft of the initial response. For example, the following prompt might be used:

[0382] "My new TV remote isn't working. What should I do?"

[0383] Tone adjustment by the emotion engine

[0384] During the initial response generation process, the emotion engine analyzes the user's inquiry to understand their emotions and appropriately adjusts the tone of the generated response. For example, if anger or frustration is detected, the tone will be adjusted as follows:

[0385] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps...?"

[0386] User modifications and saving of final answers

[0387] The generated initial response is displayed on the terminal, allowing the user to review it and enter corrections or additional information as needed. The revised final response is then sent back to the server and stored in the database.

[0388] Specific example

[0389] For example, if a user submits a request stating that their new TV remote is not working, the server will generate a preliminary response such as:

[0390] If the remote control is not responding, please try the following steps:

[0391] 1. Check the batteries in the remote control and replace them with new ones.

[0392] 2. Make sure the remote control is pointed towards the TV.

[0393] 3. Check that there are no obstacles between the TV and the remote control.

[0394] 4. Clean the light-receiving part of the television.

[0395] If these steps do not resolve the issue, please contact our support center.

[0396] If the emotion engine detects an emotion of dissatisfaction, the tone will be adjusted as follows:

[0397] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0398] 1. Check the batteries in the remote control and replace them with new ones.

[0399] 2. Make sure the remote control is pointed towards the TV.

[0400] 3. Check that there are no obstacles between the TV and the remote control.

[0401] 4. Clean the light-receiving part of the television.

[0402] If these steps do not resolve the issue, please contact our support center for further assistance.

[0403] This system will improve the efficiency of handling inquiries and enhance user satisfaction.

[0404] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0405] Step 1: Data Collection

[0406] The server accesses the company's electronic messaging server and internal network to retrieve new messages and update information.

[0407] Input: Data from electronic message servers or internal networks.

[0408] Output: Retrieved message data and update information.

[0409] Specific operation: The server accesses the email server using the IMAP protocol to retrieve new mail. It also sends HTTP requests to the internal network to retrieve new posts and updates.

[0410] Step 2: Data Filtering

[0411] The server filters the acquired data to extract useful information.

[0412] Input: Retrieved message data and update information.

[0413] Output: Filtered, useful information.

[0414] Specific operation: The server uses natural language processing (NLP) algorithms to extract useful keywords and phrases from text data and filter out the necessary information.

[0415] Step 3: Storing in the database

[0416] The server converts the filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database.

[0417] Input: Filtered, useful information.

[0418] Output: Information converted to a structured data format and stored in a database.

[0419] Specific operation: The server extracts information, converts it to JSON format, and saves it to a database such as MongoDB or MySQL.

[0420] Step 4: Receiving the Inquiry

[0421] The device receives user inquiries through an input form.

[0422] Input: The content of the inquiry entered by the user.

[0423] Output: Sending the query details to the server.

[0424] Specific operation: When a user enters their inquiry into a web form and clicks the "Submit" button, that information is sent to the server.

[0425] Step 5: Search for related information

[0426] The generative AI server searches the database based on the query content and retrieves relevant information.

[0427] Input: Inquiry details and database.

[0428] Output: Retrieve relevant information.

[0429] Specific operation: The generative AI server sends an SQL or NoSQL query to the database and retrieves records related to the query.

[0430] Step 6: Generating the primary answer

[0431] The generative AI server uses a generative AI model (e.g., GPT-3) to generate a draft of the initial response based on the acquired relevant information.

[0432] Input: Related information and a generated AI model.

[0433] Output: Draft of the initial response.

[0434] Specific operation: The generated prompt is sent to the GPT-3 engine to generate a primary response like this:

[0435] "If the remote control is not responding, please try the following steps..."

[0436] Step 7: Adjustment by the Emotional Engine

[0437] The generative AI server analyzes the sentiment behind the inquiry and adjusts the tone and expression of the initial response.

[0438] Input: Inquiry details and a draft of the initial response.

[0439] Output: A primary response adjusted based on emotions.

[0440] Specific actions: Use a sentiment analysis tool (e.g., IBM Watson Tone Analyzer) to analyze the sentiment from the inquiry and adjust the tone of the response based on the results.

[0441] Step 8: Present your initial answer

[0442] The terminal displays a draft of the initial response that has been generated to the user.

[0443] Input: Draft of the generated initial response.

[0444] Output: The initial response displayed to the user.

[0445] Specific operation: The generated response text will be displayed in the user interface of the web page, allowing the user to review and edit it.

[0446] Step 9: User modification

[0447] The user reviews the initial response and enters any necessary corrections or additional information.

[0448] Input: First response.

[0449] Output: Corrected or added information.

[0450] Specific action: The user edits their response on the web form and clicks the "Final Confirmation" or "Submit Revised" button.

[0451] Step 10: Submit your final response

[0452] The user submits their revised initial response as the final response.

[0453] Input: Revised initial response.

[0454] Output: Final answer.

[0455] Specific operation: After the user makes a final confirmation, they press the submit button, and the content is sent again to the generation AI server.

[0456] Step 11: Saving to the database

[0457] The server saves the final answer to the database.

[0458] Input: Final answer.

[0459] Output: Saved final answer.

[0460] Specific operation: Convert the final answer to JSON format and insert it into the answer table in the database.

[0461] (Application Example 2)

[0462] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0463] Traditional customer support systems have the challenge of not being able to provide appropriate answers to inquiries quickly. Furthermore, they fail to adequately improve customer satisfaction because they do not take customer emotions into consideration. This leads to increased workload for customer support and inconsistent service quality.

[0464] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0465] In this invention, the server includes means for collecting data from internal electronic communication means and internal networks; means for filtering the collected data and extracting useful information; means for storing the extracted information in a database; means for searching the database based on the content of the inquiry and obtaining relevant information; means for generating a draft of an initial response using a generative artificial intelligence model; and means for analyzing the user's emotions and adjusting the tone of the initial response. This makes it possible to quickly provide appropriate responses that respond to the user's emotions and improve the efficiency and quality of customer support.

[0466] "Internal electronic communication methods" refer to communication methods used within the company, such as email and chat systems.

[0467] An "internal network" refers to an internal network used within a company or organization, including intranets and dedicated networks.

[0468] "Means of data collection" refers to the processes and mechanisms for obtaining necessary information from electronic communication methods or internal networks.

[0469] "Filtering" is the process of selecting useful information from collected data and eliminating unnecessary data.

[0470] "Useful information" refers to important data that is helpful in handling inquiries and performing tasks, and includes manual information and past inquiry records.

[0471] A "database" is a data storage system that systematically stores collected information, making it easy to search and manage.

[0472] A "generative artificial intelligence model" is an AI model that can generate natural language based on large amounts of data, and models like GPT-3 fall into this category.

[0473] A "draft initial response" is the first proposed answer to an inquiry generated by an AI model.

[0474] "Means of presenting to the user" refers to methods or devices for displaying the generated draft of the initial response so that the user can review it.

[0475] "Methods for analyzing emotions" refer to technologies that identify emotions from user inquiries and adjust the tone and content of responses accordingly.

[0476] "Methods for adjusting tone" refer to the process of appropriately modifying the expression and nuances of the generated response based on the analyzed emotions.

[0477] This invention relates to a system for streamlining customer support operations in physical stores. The system collects data from internal electronic communication methods and networks, filters it, and extracts useful information. The extracted information is stored in a database and searched based on the inquiry content. Using a generative artificial intelligence model, a draft of the initial response is generated based on the acquired information, and the tone of the response is adjusted by analyzing the user's emotions using an emotion engine.

[0478] Description of the system's program processing

[0479] The server periodically or at designated times accesses the company's electronic communication methods (email servers and chat systems) and internal network (intranet) to retrieve new emails and updates. This data is filtered, and useful information is converted into structured data formats (e.g., JSON, XML) and stored in the database.

[0480] The terminal receives user inquiries through an input form and sends the inquiry details to the server. The server searches its database and retrieves relevant information. It then uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial response.

[0481] The generated response uses an emotion engine to analyze the sentiment of the inquiry. For example, it uses TextBlob to identify the sentiment (positive, negative, neutral, etc.) of the text and adjusts the tone and content of the response accordingly. The adjusted initial response draft is then presented to the user via the device.

[0482] The user can review this initial draft response and make corrections or add information as needed. Once the final response is decided, the server saves it to the database for later reference.

[0483] Specific example

[0484] For example, consider a user's inquiry about a new TV remote not working. The user might input, "My new TV remote isn't working, what should I do?" The terminal receives this inquiry and sends it to the server. The server searches its database and retrieves information on how to deal with a non-responsive remote. Then, using a generative artificial intelligence model, it generates a draft initial response like this:

[0485] If the remote control is not responding, please try the following steps:

[0486] 1. Check the batteries in the remote control and replace them with new ones.

[0487] 2. Make sure the remote control is pointed towards the TV.

[0488] 3. Check that there are no obstacles between the TV and the remote control.

[0489] 4. Clean the light-receiving part of the television.

[0490] If these steps do not resolve the issue, please contact our support center.

[0491] Furthermore, using an emotion engine, the system reads the user's feelings of dissatisfaction from their inquiry and adjusts the tone accordingly:

[0492] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0493] 1. Check the batteries in the remote control and replace them with new ones.

[0494] 2. Make sure the remote control is pointed towards the TV.

[0495] 3. Check that there are no obstacles between the TV and the remote control.

[0496] 4. Clean the light-receiving part of the television.

[0497] If these steps do not resolve the issue, please contact our support center for further assistance.

[0498] The terminal displays this draft to the user, who then enters additional information such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer, and the server saves that answer to the database.

[0499] This system enables in-store staff to handle customer interactions efficiently and effectively, thereby improving customer satisfaction.

[0500] Example of a prompt

[0501] "My new TV remote isn't working. What should I do?"

[0502] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0503] Step 1:

[0504] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. Its input consists of unread messages and new information retrieved from email and the intranet; this data is filtered to extract only the useful information. The output is this filtered, useful information.

[0505] Step 2:

[0506] The server converts the filtered data into a structured data format (e.g., JSON, XML). This process organizes the raw input data into an appropriate format so that it can be stored in a unified database. The output is structured data that can be stored in the database.

[0507] Step 3:

[0508] A database stores structured data and manages it so that it can be quickly accessed when needed. This step involves taking structured data as input and performing database operations to efficiently store it. The output is the storage of the most up-to-date data, always accessible.

[0509] Step 4:

[0510] The terminal receives user inquiries through an input form. In this step, the user's inquiry content is received as input and sent to the server. The inquiry content is then returned to the server as output.

[0511] Step 5:

[0512] The server searches the database for information related to the query and retrieves the relevant information. The input is the user's query, and the database is searched based on that content. The output is the information related to the query.

[0513] Step 6:

[0514] The server uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial answer based on the acquired relevant information. In this step, relevant information is taken as input, and the generative artificial intelligence model generates the answer based on it. The output is a draft of the initial answer.

[0515] Step 7:

[0516] The server uses an emotion engine to adjust the tone of the initial response draft based on the emotion of the inquiry. In this step, the emotion contained in the user's inquiry is taken as input, and the emotion engine analyzes it to adjust the tone. The output is the adjusted initial response draft.

[0517] Step 8:

[0518] The terminal presents the user with a draft of the adjusted initial response. In this step, the input is the adjusted initial response, which is displayed to the user. As output, the user can review the draft of the initial response.

[0519] Step 9:

[0520] The user reviews the initial draft response and enters any necessary corrections or additional information. This step includes the user's additional information and corrections as input, and the response is ready to be submitted as the final response. The output is the final response, corrected by the user.

[0521] Step 10:

[0522] The server saves the user's modified final answer to a database for future reference. In this step, the input is the final answer, which is stored in the database. The output is the saved final answer, which can be referenced later.

[0523] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0524] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0525] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0526] [Second Embodiment]

[0527] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0528] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0529] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0530] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0531] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0532] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0533] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0534] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0535] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0536] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0537] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0538] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0539] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. The generated draft of the initial response is presented to the user, who can modify it as needed and submit it as the final response. The program processing of this system is described below in natural language.

[0540] 1. Data Collection

[0541] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[0542] 2. Data filtering

[0543] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[0544] 3. Database storage

[0545] The server stores filtered, useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[0546] 4. Generating the initial response

[0547] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response. This draft combines the relevant information to create the initial response text.

[0548] 5. Presentation and revision of the initial response.

[0549] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or add information as needed. When making corrections, it is also possible to add specific instructions or comments.

[0550] 6. Submit and save your final response.

[0551] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[0552] Specific example

[0553] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[0554] The generative AI server generates a draft of the initial response as follows:

[0555] If the remote control is not responding, please try the following steps:

[0556] 1. Check the batteries in the remote control and replace them with new ones.

[0557] 2. Make sure the remote control is pointed towards the TV.

[0558] 3. Check that there are no obstacles between the TV and the remote control.

[0559] 4. Clean the light-receiving part of the television.

[0560] If these steps do not resolve the issue, please contact our support center.

[0561] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[0562] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[0563] The following describes the processing flow.

[0564] Data acquisition and filtering

[0565] Step 1:

[0566] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[0567] Step 2:

[0568] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[0569] Step 3:

[0570] The server analyzes and filters the data in temporary storage, extracting useful information (e.g., inquiry details, important notifications).

[0571] Step 4:

[0572] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[0573] Step 5:

[0574] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[0575] Generation of the first response

[0576] Step 6:

[0577] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[0578] Step 7:

[0579] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[0580] Step 8:

[0581] The generative AI server searches the database for information related to the query. The search results are stored.

[0582] Step 9:

[0583] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is generated as an appropriate response to the inquiry.

[0584] Presentation and revision of the initial response

[0585] Step 10:

[0586] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[0587] Step 11:

[0588] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[0589] Submit and save your final response.

[0590] Step 12:

[0591] The user submits their revised initial response as the final response. This confirms the response.

[0592] Step 13:

[0593] The device saves the final answer to the database. This makes the saved answer available for later reference.

[0594] Through the above processing steps, the efficiency and accuracy of inquiry handling are improved. The specific actions of each step enable the entire system to function smoothly.

[0595] (Example 1)

[0596] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0597] In modern companies, email and internal networks are widely used as means of internal communication. However, gathering information and responding to inquiries through these means requires considerable effort, and efficiency improvements are needed. Furthermore, providing quick and accurate answers to inquiries is difficult, making it a challenge to improve customer satisfaction.

[0598] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0599] In this invention, the server includes means for collecting information from internal communication data and the internal network; means for analyzing the collected information and extracting useful data; means for storing the extracted data in a storage device; means for searching the storage device based on the content of the inquiry and obtaining relevant data; means for generating a draft of an initial response using a generative artificial intelligence model based on the obtained data; means for presenting the generated draft of the initial response to the user; and means for sending the initial response modified by the user as the final response. This makes it possible to streamline the inquiry handling process and improve customer satisfaction.

[0600] "Internal communication data" refers to data generated through digital communication methods used within a company, such as email, chat, and intranets.

[0601] An "internal network" refers to a limited network environment used for sharing information and communicating within a company.

[0602] "Means of collecting information" refers to methods using protocols and APIs that allow a server to retrieve data from email servers or intranets.

[0603] "Means of analyzing information and extracting useful data" refers to natural language processing and filtering technologies used to detect specific keywords and patterns from collected communication data and extract necessary information.

[0604] "Storage device" refers to databases and storage systems used to hold, manage, and retrieve data.

[0605] "Inquiry content" refers to questions or requests that users make to the server or system.

[0606] "Generative artificial intelligence models" refer to AI models that generate text using natural language processing and machine learning. Specifically, this includes advanced AI technologies such as GPT and BERT.

[0607] "Draft initial response" refers to the initial response text generated by a generative artificial intelligence model.

[0608] "Means of presentation to the user" refers to interface means that allow the user to review the initial draft of the response displayed on the terminal.

[0609] "Final answer" refers to the answer that has been revised and finalized by the user.

[0610] "Means of transmission" refers to email systems or other means of communication used to send the final response to the recipient.

[0611] This invention is a system that efficiently collects useful information from internal communication data and the company's internal network, and uses that information to generate a draft of an initial response using a generative artificial intelligence model. This streamlines the inquiry response process and enables the provision of quick and accurate answers.

[0612] 1. Data Collection

[0613] The server periodically accesses the company's email server and internal network to retrieve new data. It uses the IMAP protocol to retrieve emails and collects updates from the internal network via a REST API. This ensures that the server consistently collects the latest information, such as inquiries and important notifications.

[0614] 2. Data filtering

[0615] The server uses a natural language processing (NLP) library to filter the acquired data and extract useful information. By utilizing the NLP library, inquiries, important notifications, manual updates, and other relevant information are automatically analyzed and filtered.

[0616] 3. Database storage

[0617] The filtered, useful information is stored in storage by the server. Specifically, the information is stored using a database such as PostgreSQL or MongoDB. This database is categorized and indexed according to the query and answer content.

[0618] 4. Generating the initial response

[0619] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on this retrieved information, it uses generative artificial intelligence models such as OpenAI's GPT model or Google's BERT model to generate a draft of the initial response.

[0620] 5. Presentation and revision of the initial response.

[0621] The device displays a draft of the initial response to the user. The user reviews this draft and makes corrections or adds information as needed. The response is then optimized, including any additional instructions or comments from the user.

[0622] 6. Submit and save your final response.

[0623] Once the user submits the final, revised response, the device saves the response to its storage. This saved data can be referenced later and used as foundational data to quickly respond to similar inquiries in the future.

[0624] Specific example

[0625] For example, if there is an inquiry about a new remote control not working, the user would enter the inquiry as follows: "My new remote control isn't working, what should I do?". The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working. The generative AI server generates a draft initial response like this:

[0626] If the remote control is not responding, please try the following steps:

[0627] 1. Check the batteries in the remote control and replace them with new ones.

[0628] 2. Make sure the remote control is pointed towards the target device.

[0629] 3. Check that there are no obstacles between the device and the remote control.

[0630] 4. Clean the light-receiving part of the device.

[0631] If these steps do not resolve the issue, please contact our support center.

[0632] Related technologies

[0633] By using the generated AI model and prompt statements, the accuracy and speed of inquiry handling are improved, resulting in reduced workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies.

[0634] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0635] Step 1:

[0636] The server collects information from the company's internal email server and internal network. In this process, the server uses the IMAP protocol to retrieve emails and a REST API to retrieve updates from the internal network. Inputs are new emails and intranet updates, and output is the collected raw data.

[0637] Specifically, the server automatically connects to the mail server at 9 AM every morning and retrieves all new emails from the previous day. Internal network update information is also collected at the same time.

[0638] Step 2:

[0639] The server analyzes the data collected in Step 1 and extracts useful information. Here, a natural language processing (NLP) library is used to filter the data. The input is the collected raw data, and the output is the filtered useful information.

[0640] Specifically, the system analyzes the email content received by the server and extracts emails containing keywords such as "failure," "inquiry," and "support." Within the internal network, it collects documents tagged with "update," "procedure," and "urgent."

[0641] Step 3:

[0642] The server stores filtered, useful information in storage. Specifically, it uses a database such as PostgreSQL or MongoDB to store the information. The input is filtered, useful information, and the output is the data stored in the database.

[0643] Specifically, the server identifies emails inquiring about "remote control malfunction" and stores their contents in the "remote control related inquiries" table in the database.

[0644] Step 4:

[0645] When the terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Next, it uses a generative artificial intelligence model (for example, OpenAI's GPT model or Google's BERT model) to generate a draft of the initial response. The input is the user's inquiry and the relevant information stored in the database, and the output is the draft of the initial response.

[0646] Specifically, when a user inquires that "the new remote control isn't responding," the terminal sends this inquiry to the AI ​​server. The AI ​​server searches its database, retrieves information about the remote control malfunction, and generates a draft message like this: "If the remote control is not responding, try changing the batteries and checking the remote control's orientation and for any obstructions."

[0647] Step 5:

[0648] The terminal displays a draft of the generated initial response to the user. The user reviews this draft and enters corrections or additional information as needed. The input consists of the generated initial response draft and the user's feedback, and the output is the revised initial response draft.

[0649] Specifically, when a user sees the initial response provided and enters an additional request such as "Please specify the type of battery," the device updates the draft to reflect this feedback.

[0650] Step 6:

[0651] When the user submits the final, revised answer, the device saves that answer to its storage device. The input is the final, revised answer made by the user, and the output is the final answer data saved to the storage device.

[0652] Specifically, when a user clicks the "Submit Final Response" button, the device saves that response data to the "Response History" table in the database. This data is then used as a reference when handling future inquiries.

[0653] (Application Example 1)

[0654] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0655] In a factory setting, responding quickly and accurately to machine malfunctions or operational problems requires searching through a vast amount of information to find relevant data and providing it to workers. However, traditional methods involve manually searching for information based on inquiries and generating appropriate responses, which is time-consuming and labor-intensive. Furthermore, if relevant information cannot be found, users may become stuck, leading to work delays. An efficient system is needed to solve these problems.

[0656] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0657] In this invention, the server includes means for collecting data from internal email and intranets, means for filtering the collected data and extracting useful information, means for storing the extracted information in a database, means for searching the database based on the inquiry content and obtaining relevant information, means for generating a draft of a preliminary response using a generative artificial intelligence model based on the obtained information, means for presenting the generated draft of the preliminary response to the user, means for sending the preliminary response modified by the user as the final response, means for inputting the inquiry content using a smart device, means for displaying a default message if relevant information is not found, means for accepting user modifications, and means for saving user input and modifications in the database. This enables quick and accurate responses to troubles and inquiries within the factory, improving work efficiency and speeding up trouble resolution.

[0658] "Inside the company" refers to the internal workings of a company or organization.

[0659] "Email" is a means of sending and receiving text messages and files over the internet.

[0660] An "intranet" is an internal network used within a company; it utilizes internet technology but is isolated from the outside world.

[0661] "Data collection" is the process of systematically gathering information.

[0662] "Filtering" is the process of selecting necessary information and removing unnecessary information.

[0663] "Extraction" is the process of taking out the necessary parts from collected data.

[0664] A "database" is a system for efficiently storing, managing, and retrieving large amounts of data.

[0665] "Inquiry details" refer to the information that users submit to the system, such as questions or requests.

[0666] A "generative artificial intelligence model" is an algorithm that uses artificial intelligence technology to generate new information or answers.

[0667] A "primary response" is a draft answer that is initially generated in response to a user's inquiry.

[0668] A "smart device" is a high-performance portable device that can connect to the internet, such as a smartphone or tablet.

[0669] A "default message" is a standard message that is automatically displayed when certain conditions are not met.

[0670] A "user" is someone who uses a system or device.

[0671] "Means" refer to the methods and techniques used to achieve an objective.

[0672] "Modification" refers to the operation of making changes to a document or data that has been created.

[0673] The "final answer" is the answer that has been finalized after the user has made any revisions.

[0674] This invention is a system for streamlining troubleshooting and inquiry handling within a factory. The system collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft initial response based on that information. Finally, the generated draft initial response is presented to the user, who can revise it as needed and submit it as the final response.

[0675] System Configuration

[0676] hardware

[0677] Server: A server that performs data collection, filtering, and database storage.

[0678] Device: Smart devices such as smartphones and tablets.

[0679] Generative AI server: A server that runs generative AI models.

[0680] software

[0681] Generative artificial intelligence model: GPT-2 or an equivalent generative AI model.

[0682] Transformers Library: A Python library provided by Hugging Face.

[0683] Process Overview

[0684] Data collection:

[0685] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[0686] Data filtering:

[0687] The server filters the retrieved data and extracts useful information (e.g., inquiry details, important notifications, manual updates).

[0688] Database storage:

[0689] The server stores the extracted useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[0690] Generating the initial answer:

[0691] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response.

[0692] Presentation and revision of the initial response:

[0693] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or enter additional information as needed.

[0694] Submit and save your final response:

[0695] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[0696] Specific example

[0697] For example, consider a case where a user has an inquiry about a new machine not working. The user enters the inquiry, "The new machine isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a machine that isn't working.

[0698] The generative AI server generates a draft of the initial response as follows:

[0699] If the machine is not working, try the following steps:

[0700] 1. Check if the power is on.

[0701] 2. Check that the wiring is connected correctly.

[0702] 3. Check that the machine body has been cleaned.

[0703] 4. Check the machine's error messages.

[0704] If these steps do not resolve the issue, please contact our support center.

[0705] The terminal displays this draft to the user. The user enters additional information, such as, "Please tell me the specific steps to take when checking the wiring." Once the revisions are complete, the user submits the final response. The terminal saves this final response to the database so that it can be referenced later if needed.

[0706] In this way, efficient responses become possible when the same inquiry occurs again, and the amount of work required to handle inquiries can be significantly reduced.

[0707] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0708] Step 1:

[0709] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[0710] Input: Internal email and intranet data.

[0711] Data processing: The acquired raw data is analyzed, and each email and update information is organized as metadata.

[0712] Output: Initial dataset for filtering.

[0713] Step 2:

[0714] The server uses a filtering algorithm to filter the acquired data and extract useful information.

[0715] Input: Initial dataset.

[0716] Data processing: Select useful information and remove unnecessary information based on filtering rules.

[0717] Output: Filtered important information (e.g., inquiry details, manual update information).

[0718] Step 3:

[0719] The server stores the extracted useful information in a database.

[0720] Input: Filtered important information.

[0721] Data Storage: Information is appropriately arranged according to the database structure, and indexes are created.

[0722] Output: A database organized in a searchable format.

[0723] Step 4:

[0724] The terminal receives the user's inquiry and sends it to the generation AI server.

[0725] Input: User inquiry details.

[0726] Data transmission: The query content is sent to the generative AI server according to the protocol.

[0727] Output: The query content is forwarded to the AI ​​server.

[0728] Step 5:

[0729] The generative AI server searches the database and retrieves relevant information.

[0730] Input: User inquiry details.

[0731] Data retrieval: Quickly extract relevant information using database indexes.

[0732] Output: A set of related information.

[0733] Step 6:

[0734] The generative AI server uses a generative artificial intelligence model to generate a draft of the initial response based on the acquired information.

[0735] Input: A set of related information.

[0736] Data processing: Input prompt text into a generative AI model (such as GPT-2) and generate a response.

[0737] Output: Draft of the initial response.

[0738] Step 7:

[0739] The terminal presents the user with a draft of the initial response that has been generated.

[0740] Input: Draft of the initial response.

[0741] Data display: The draft of the initial response is displayed in the user interface.

[0742] Output: Displays the initial response that the user can see.

[0743] Step 8:

[0744] The user reviews the initial draft response provided and enters any necessary corrections or additional information.

[0745] Input: Draft of the initial response, user revisions and additional information.

[0746] Data processing: Reflects user-submitted corrections.

[0747] Output: Corrected initial response.

[0748] Step 9:

[0749] The terminal receives the user's corrections and sends the initial response as the final response, saving it to the database.

[0750] Input: Revised initial response.

[0751] Data storage: The final answer is saved in the database for future searches.

[0752] Output: The final answer stored in the database.

[0753] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0754] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions. The program processing of this system is described below in natural language.

[0755] 1. Data collection and filtering

[0756] The server periodically or at designated times accesses the company's email and intranet servers to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[0757] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[0758] The server converts filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database. This facilitates subsequent searching and AI processing.

[0759] 2. Generating the initial response

[0760] The terminal receives user inquiries through an input form. The terminal then sends the inquiry content to the generation AI server.

[0761] The generative AI server searches the database for information related to the query and stores the search results.

[0762] The generative AI server uses a generative AI model (e.g., GPT-3) based on the acquired relevant information to generate a draft of the initial response. This draft is generated as an appropriate response to the inquiry.

[0763] In this process, the generative AI server further utilizes an emotion engine. The emotion engine analyzes the user's inquiry and recognizes their emotions (e.g., anger, frustration, joy). Based on the recognized emotions, the emotion engine adjusts the tone and expression of the initial response.

[0764] 3. Presentation and revision of the initial response.

[0765] The device displays a draft of the initial response to the user, allowing them to review their answer.

[0766] Users can review this initial draft response and make corrections or add information as needed. When making corrections, they can also add specific instructions or comments.

[0767] 4. Submit and save your final response.

[0768] When the user submits their revised initial response as the final response, the generative AI server uses the emotion engine again to analyze the user's final input and confirm whether the user is satisfied.

[0769] The device saves the final answer to a database. This makes the saved answer available for later reference.

[0770] Specific example

[0771] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[0772] The generative AI server generates a draft of the initial response as follows:

[0773] If the remote control is not responding, please try the following steps:

[0774] 1. Check the batteries in the remote control and replace them with new ones.

[0775] 2. Make sure the remote control is pointed towards the TV.

[0776] 3. Check that there are no obstacles between the TV and the remote control.

[0777] 4. Clean the light-receiving part of the television.

[0778] If these steps do not resolve the issue, please contact our support center.

[0779] Additionally, an emotion engine is used to read the user's feelings of dissatisfaction from their inquiry and adjust the tone accordingly:

[0780] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0781] 1. Check the batteries in the remote control and replace them with new ones.

[0782] 2. Make sure the remote control is pointed towards the TV.

[0783] 3. Check that there are no obstacles between the TV and the remote control.

[0784] 4. Clean the light-receiving part of the television.

[0785] If these steps do not resolve the issue, please contact our support center for further assistance.

[0786] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[0787] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[0788] The following describes the processing flow.

[0789] Data acquisition and filtering

[0790] Step 1:

[0791] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[0792] Step 2:

[0793] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[0794] Step 3:

[0795] The server analyzes the data in temporary storage and extracts useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[0796] Step 4:

[0797] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[0798] Step 5:

[0799] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[0800] Generation of the first response

[0801] Step 6:

[0802] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[0803] Step 7:

[0804] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[0805] Step 8:

[0806] The generative AI server searches the database for information related to the query. The search results are stored.

[0807] Step 9:

[0808] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is then generated as an appropriate response.

[0809] Adjustment by the emotion engine

[0810] Step 10:

[0811] When the generative AI server generates a draft of the initial response, it uses an emotion engine to analyze the user's emotions from the content of their inquiry. This allows for adjustments to be made to match the user's emotions.

[0812] Step 11:

[0813] The generative AI server adjusts the tone and expression of the initial response according to the emotions recognized by the emotion engine. For example, if the user is dissatisfied, expressions of apology and polite language will be added.

[0814] Presentation and revision of the initial response

[0815] Step 12:

[0816] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[0817] Step 13:

[0818] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[0819] Submit and save your final response.

[0820] Step 14:

[0821] The user submits their revised initial response as the final response. At this stage, the generative AI server uses the emotion engine again to analyze the user's final input and maintain the optimal tone.

[0822] Step 15:

[0823] The device saves the final answer to the database. This makes the saved answer available for later reference.

[0824] Specific example

[0825] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[0826] The generative AI server generates a draft of the initial response as follows:

[0827] If the remote control is not responding, please try the following steps:

[0828] 1. Check the batteries in the remote control and replace them with new ones.

[0829] 2. Make sure the remote control is pointed towards the TV.

[0830] 3. Check that there are no obstacles between the TV and the remote control.

[0831] 4. Clean the light-receiving part of the television.

[0832] If these steps do not resolve the issue, please contact our support center.

[0833] Furthermore, it uses an emotion engine to read the user's feelings of dissatisfaction and adjusts the tone as follows:

[0834] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0835] 1. Check the batteries in the remote control and replace them with new ones.

[0836] 2. Make sure the remote control is pointed towards the TV.

[0837] 3. Check that there are no obstacles between the TV and the remote control.

[0838] 4. Clean the light-receiving part of the television.

[0839] If these steps do not resolve the issue, please contact our support center for further assistance.

[0840] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[0841] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[0842] (Example 2)

[0843] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0844] Traditional inquiry handling systems lacked the ability to efficiently extract useful information from the company's vast internal email and intranet, organize and store it in an appropriate format, and automatically generate appropriate responses based on relevant information. In particular, they lacked the function to adjust response content while considering user emotions, which could result in decreased user satisfaction. Against this backdrop, there is a need for a system that can simultaneously improve the efficiency of inquiry handling and enhance user satisfaction.

[0845] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for collecting data from internal electronic messages and the internal network, means for filtering the collected data and extracting useful information, and means for converting the extracted information into a structured data format and storing it in a database. This makes it possible to efficiently organize and store the information necessary for responding to inquiries, and furthermore, to automatically generate responses that provide high user satisfaction using a generative artificial intelligence model and an emotion engine.

[0846] "Internal electronic messages" refer to the content of communications via email and messaging services used within a company or organization.

[0847] An "internal network" refers to an intranet or dedicated network used within a company or organization.

[0848] "Collecting data" means obtaining information from a specific data source.

[0849] "Filtering" is the process of selecting useful information from acquired data and removing unnecessary data.

[0850] A "structured data format" is a format that organizes information systematically and regularly, making it easy to store in a database (e.g., JSON, XML).

[0851] A "database" is a system for storing, managing, and making data searchable.

[0852] "Inquiry content" refers to the text of questions or requests that users submit to the system.

[0853] A "generative artificial intelligence model" is an artificial intelligence that generates natural language based on a large amount of data, such as GPT-3 or similar models.

[0854] A "draft initial response" is the first response proposed by a generative artificial intelligence model.

[0855] An "emotion engine" is a technology that analyzes the user's emotions from text and adjusts the tone and expression of the response based on that analysis.

[0856] A "user" is an individual or group that uses the system.

[0857] A "final answer" is an answer that has been confirmed and revised by the user.

[0858] "Saving" means recording information in a database so that it can be reused or searched later.

[0859] This invention is a system that collects data from internal electronic messages and internal networks, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the content of the inquiry, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions.

[0860] This system includes the following main components:

[0861] 1. Data collection and filtering functions

[0862] 2. Database storage function

[0863] 3. Function to receive inquiries from users

[0864] 4. Search function for related information

[0865] 5. Primary response generation function using generative AI models

[0866] 6. Tone adjustment function using an emotion engine

[0867] 7. Presentation of initial answers and user modification function

[0868] 8. Function to submit and save the final response.

[0869] Hardware and software to be used

[0870] Server: Performs data collection, filtering, database storage, retrieval, initial response generation, sentiment analysis, and final response storage. Protocols and tools used include IMAP (for email collection), HTTP requests (for intranet update collection), natural language processing (NLP) algorithms, generative artificial intelligence models (e.g., GPT-3), sentiment engines (e.g., IBM Watson Tone Analyzer), and database management systems (e.g., MongoDB or MySQL).

[0871] Terminal: Receives user inquiries, provides initial responses, and makes corrections. Web forms or desktop applications are used as the user interface.

[0872] Processing details

[0873] Data acquisition and filtering

[0874] The server periodically accesses the company's electronic messaging server and internal network to retrieve new messages and updates. For example, the server accesses the electronic messaging server using the IMAP protocol to retrieve new mail. The retrieved data is analyzed using NLP algorithms, and useful information is filtered.

[0875] Storage in database

[0876] The filtered, useful information is converted to JSON format and stored in a MongoDB or MySQL database. For example, data in the following format is stored:

[0877] json

[0878] {

[0879] "id": "12345",

[0880] "type": "inquiry",

[0881] "content": "New TV remote not responding"

[0882] }

[0883] Generation of the first response

[0884] When a user submits a query through their device, the server searches the database and retrieves relevant information. Based on this information, a generative AI model (e.g., GPT-3) generates a draft of the initial response. For example, the following prompt might be used:

[0885] "My new TV remote isn't working. What should I do?"

[0886] Tone adjustment by the emotion engine

[0887] During the initial response generation process, the emotion engine analyzes the user's inquiry to understand their emotions and appropriately adjusts the tone of the generated response. For example, if anger or frustration is detected, the tone will be adjusted as follows:

[0888] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps...?"

[0889] User modifications and saving of final answers

[0890] The generated initial response is displayed on the terminal, allowing the user to review it and enter corrections or additional information as needed. The revised final response is then sent back to the server and stored in the database.

[0891] Specific example

[0892] For example, if a user submits a request stating that their new TV remote is not working, the server will generate a preliminary response such as:

[0893] If the remote control is not responding, please try the following steps:

[0894] 1. Check the batteries in the remote control and replace them with new ones.

[0895] 2. Make sure the remote control is pointed towards the TV.

[0896] 3. Check that there are no obstacles between the TV and the remote control.

[0897] 4. Clean the light-receiving part of the television.

[0898] If these steps do not resolve the issue, please contact our support center.

[0899] If the emotion engine detects an emotion of dissatisfaction, the tone will be adjusted as follows:

[0900] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0901] 1. Check the batteries in the remote control and replace them with new ones.

[0902] 2. Make sure the remote control is pointed towards the TV.

[0903] 3. Check that there are no obstacles between the TV and the remote control.

[0904] 4. Clean the light-receiving part of the television.

[0905] If these steps do not resolve the issue, please contact our support center for further assistance.

[0906] This system will improve the efficiency of handling inquiries and enhance user satisfaction.

[0907] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0908] Step 1: Data Collection

[0909] The server accesses the company's electronic messaging server and internal network to retrieve new messages and update information.

[0910] Input: Data from electronic message servers or internal networks.

[0911] Output: Retrieved message data and update information.

[0912] Specific operation: The server accesses the email server using the IMAP protocol to retrieve new mail. It also sends HTTP requests to the internal network to retrieve new posts and updates.

[0913] Step 2: Data Filtering

[0914] The server filters the acquired data to extract useful information.

[0915] Input: Retrieved message data and update information.

[0916] Output: Filtered, useful information.

[0917] Specific operation: The server uses natural language processing (NLP) algorithms to extract useful keywords and phrases from text data and filter out the necessary information.

[0918] Step 3: Storing in the database

[0919] The server converts the filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database.

[0920] Input: Filtered, useful information.

[0921] Output: Information converted to a structured data format and stored in a database.

[0922] Specific operation: The server extracts information, converts it to JSON format, and saves it to a database such as MongoDB or MySQL.

[0923] Step 4: Receiving the Inquiry

[0924] The device receives user inquiries through an input form.

[0925] Input: The content of the inquiry entered by the user.

[0926] Output: Sending the query details to the server.

[0927] Specific operation: When a user enters their inquiry into a web form and clicks the "Submit" button, that information is sent to the server.

[0928] Step 5: Search for related information

[0929] The generative AI server searches the database based on the query content and retrieves relevant information.

[0930] Input: Inquiry details and database.

[0931] Output: Retrieve relevant information.

[0932] Specific operation: The generative AI server sends an SQL or NoSQL query to the database and retrieves records related to the query.

[0933] Step 6: Generating the primary answer

[0934] The generative AI server uses a generative AI model (e.g., GPT-3) to generate a draft of the initial response based on the acquired relevant information.

[0935] Input: Related information and a generated AI model.

[0936] Output: Draft of the initial response.

[0937] Specific operation: The generated prompt is sent to the GPT-3 engine to generate a primary response like this:

[0938] "If the remote control is not responding, please try the following steps..."

[0939] Step 7: Adjustment by the Emotional Engine

[0940] The generative AI server analyzes the sentiment behind the inquiry and adjusts the tone and expression of the initial response.

[0941] Input: Inquiry details and a draft of the initial response.

[0942] Output: A primary response adjusted based on emotions.

[0943] Specific actions: Use a sentiment analysis tool (e.g., IBM Watson Tone Analyzer) to analyze the sentiment from the inquiry and adjust the tone of the response based on the results.

[0944] Step 8: Present your initial answer

[0945] The terminal displays a draft of the initial response that has been generated to the user.

[0946] Input: Draft of the generated initial response.

[0947] Output: The initial response displayed to the user.

[0948] Specific operation: The generated response text will be displayed in the user interface of the web page, allowing the user to review and edit it.

[0949] Step 9: User modification

[0950] The user reviews the initial response and enters any necessary corrections or additional information.

[0951] Input: First response.

[0952] Output: Corrected or added information.

[0953] Specific action: The user edits their response on the web form and clicks the "Final Confirmation" or "Submit Revised" button.

[0954] Step 10: Submit your final response

[0955] The user submits their revised initial response as the final response.

[0956] Input: Revised initial response.

[0957] Output: Final answer.

[0958] Specific operation: After the user makes a final confirmation, they press the submit button, and the content is sent again to the generation AI server.

[0959] Step 11: Saving to the database

[0960] The server saves the final answer to the database.

[0961] Input: Final answer.

[0962] Output: Saved final answer.

[0963] Specific operation: Convert the final answer to JSON format and insert it into the answer table in the database.

[0964] (Application Example 2)

[0965] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0966] Traditional customer support systems have the challenge of not being able to provide appropriate answers to inquiries quickly. Furthermore, they fail to adequately improve customer satisfaction because they do not take customer emotions into consideration. This leads to increased workload for customer support and inconsistent service quality.

[0967] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0968] In this invention, the server includes means for collecting data from internal electronic communication means and internal networks; means for filtering the collected data and extracting useful information; means for storing the extracted information in a database; means for searching the database based on the content of the inquiry and obtaining relevant information; means for generating a draft of an initial response using a generative artificial intelligence model; and means for analyzing the user's emotions and adjusting the tone of the initial response. This makes it possible to quickly provide appropriate responses that respond to the user's emotions and improve the efficiency and quality of customer support.

[0969] "Internal electronic communication methods" refer to communication methods used within the company, such as email and chat systems.

[0970] An "internal network" refers to an internal network used within a company or organization, including intranets and dedicated networks.

[0971] "Means of data collection" refers to the processes and mechanisms for obtaining necessary information from electronic communication methods or internal networks.

[0972] "Filtering" is the process of selecting useful information from collected data and eliminating unnecessary data.

[0973] "Useful information" refers to important data that is helpful in handling inquiries and performing tasks, and includes manual information and past inquiry records.

[0974] A "database" is a data storage system that systematically stores collected information, making it easy to search and manage.

[0975] A "generative artificial intelligence model" is an AI model that can generate natural language based on large amounts of data, and models like GPT-3 fall into this category.

[0976] A "draft initial response" is the first proposed answer to an inquiry generated by an AI model.

[0977] "Means of presenting to the user" refers to methods or devices for displaying the generated draft of the initial response so that the user can review it.

[0978] "Methods for analyzing emotions" refer to technologies that identify emotions from user inquiries and adjust the tone and content of responses accordingly.

[0979] "Methods for adjusting tone" refer to the process of appropriately modifying the expression and nuances of the generated response based on the analyzed emotions.

[0980] This invention relates to a system for streamlining customer support operations in physical stores. The system collects data from internal electronic communication methods and networks, filters it, and extracts useful information. The extracted information is stored in a database and searched based on the inquiry content. Using a generative artificial intelligence model, a draft of the initial response is generated based on the acquired information, and the tone of the response is adjusted by analyzing the user's emotions using an emotion engine.

[0981] Description of the system's program processing

[0982] The server periodically or at designated times accesses the company's electronic communication methods (email servers and chat systems) and internal network (intranet) to retrieve new emails and updates. This data is filtered, and useful information is converted into structured data formats (e.g., JSON, XML) and stored in the database.

[0983] The terminal receives user inquiries through an input form and sends the inquiry details to the server. The server searches its database and retrieves relevant information. It then uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial response.

[0984] The generated response uses an emotion engine to analyze the sentiment of the inquiry. For example, it uses TextBlob to identify the sentiment (positive, negative, neutral, etc.) of the text and adjusts the tone and content of the response accordingly. The adjusted initial response draft is then presented to the user via the device.

[0985] The user can review this initial draft response and make corrections or add information as needed. Once the final response is decided, the server saves it to the database for later reference.

[0986] Specific example

[0987] For example, consider a user's inquiry about a new TV remote not working. The user might input, "My new TV remote isn't working, what should I do?" The terminal receives this inquiry and sends it to the server. The server searches its database and retrieves information on how to deal with a non-responsive remote. Then, using a generative artificial intelligence model, it generates a draft initial response like this:

[0988] If the remote control is not responding, please try the following steps:

[0989] 1. Check the batteries in the remote control and replace them with new ones.

[0990] 2. Make sure the remote control is pointed towards the TV.

[0991] 3. Check that there are no obstacles between the TV and the remote control.

[0992] 4. Clean the light-receiving part of the television.

[0993] If these steps do not resolve the issue, please contact our support center.

[0994] Furthermore, using an emotion engine, the system reads the user's feelings of dissatisfaction from their inquiry and adjusts the tone accordingly:

[0995] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[0996] 1. Check the batteries in the remote control and replace them with new ones.

[0997] 2. Make sure the remote control is pointed towards the TV.

[0998] 3. Check that there are no obstacles between the TV and the remote control.

[0999] 4. Clean the light-receiving part of the television.

[1000] If these steps do not resolve the issue, please contact our support center for further assistance.

[1001] The terminal displays this draft to the user, who then enters additional information such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer, and the server saves that answer to the database.

[1002] This system enables in-store staff to handle customer interactions efficiently and effectively, thereby improving customer satisfaction.

[1003] Example of a prompt

[1004] "My new TV remote isn't working. What should I do?"

[1005] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1006] Step 1:

[1007] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. Its input consists of unread messages and new information retrieved from email and the intranet; this data is filtered to extract only the useful information. The output is this filtered, useful information.

[1008] Step 2:

[1009] The server converts the filtered data into a structured data format (e.g., JSON, XML). This process organizes the raw input data into an appropriate format so that it can be stored in a unified database. The output is structured data that can be stored in the database.

[1010] Step 3:

[1011] A database stores structured data and manages it so that it can be quickly accessed when needed. This step involves taking structured data as input and performing database operations to efficiently store it. The output is the storage of the most up-to-date data, always accessible.

[1012] Step 4:

[1013] The terminal receives user inquiries through an input form. In this step, the user's inquiry content is received as input and sent to the server. The inquiry content is then returned to the server as output.

[1014] Step 5:

[1015] The server searches the database for information related to the query and retrieves the relevant information. The input is the user's query, and the database is searched based on that content. The output is the information related to the query.

[1016] Step 6:

[1017] The server uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial answer based on the acquired relevant information. In this step, relevant information is taken as input, and the generative artificial intelligence model generates the answer based on it. The output is a draft of the initial answer.

[1018] Step 7:

[1019] The server uses an emotion engine to adjust the tone of the initial response draft based on the emotion of the inquiry. In this step, the emotion contained in the user's inquiry is taken as input, and the emotion engine analyzes it to adjust the tone. The output is the adjusted initial response draft.

[1020] Step 8:

[1021] The terminal presents the user with a draft of the adjusted initial response. In this step, the input is the adjusted initial response, which is displayed to the user. As output, the user can review the draft of the initial response.

[1022] Step 9:

[1023] The user reviews the initial draft response and enters any necessary corrections or additional information. This step includes the user's additional information and corrections as input, and the response is ready to be submitted as the final response. The output is the final response, corrected by the user.

[1024] Step 10:

[1025] The server saves the user's modified final answer to a database for future reference. In this step, the input is the final answer, which is stored in the database. The output is the saved final answer, which can be referenced later.

[1026] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1027] The data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One 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">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1028] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[1029] [Third Embodiment]

[1030] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[1031] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1032] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1033] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[1034] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1035] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1036] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1037] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1038] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1039] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1040] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1041] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[1042] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. The generated draft of the initial response is presented to the user, who can modify it as needed and submit it as the final response. The program processing of this system is described below in natural language.

[1043] 1. Data Collection

[1044] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[1045] 2. Data filtering

[1046] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[1047] 3. Database storage

[1048] The server stores filtered, useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[1049] 4. Generating the initial response

[1050] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response. This draft combines the relevant information to create the initial response text.

[1051] 5. Presentation and revision of the initial response.

[1052] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or add information as needed. When making corrections, it is also possible to add specific instructions or comments.

[1053] 6. Submit and save your final response.

[1054] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[1055] Specific example

[1056] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[1057] The generative AI server generates a draft of the initial response as follows:

[1058] If the remote control is not responding, please try the following steps:

[1059] 1. Check the batteries in the remote control and replace them with new ones.

[1060] 2. Make sure the remote control is pointed towards the TV.

[1061] 3. Check that there are no obstacles between the TV and the remote control.

[1062] 4. Clean the light-receiving part of the television.

[1063] If these steps do not resolve the issue, please contact our support center.

[1064] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[1065] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[1066] The following describes the processing flow.

[1067] Data acquisition and filtering

[1068] Step 1:

[1069] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[1070] Step 2:

[1071] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[1072] Step 3:

[1073] The server analyzes and filters the data in temporary storage, extracting useful information (e.g., inquiry details, important notifications).

[1074] Step 4:

[1075] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[1076] Step 5:

[1077] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[1078] Generation of the first response

[1079] Step 6:

[1080] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[1081] Step 7:

[1082] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[1083] Step 8:

[1084] The generative AI server searches the database for information related to the query. The search results are stored.

[1085] Step 9:

[1086] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is generated as an appropriate response to the inquiry.

[1087] Presentation and revision of the initial response

[1088] Step 10:

[1089] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[1090] Step 11:

[1091] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[1092] Submit and save your final response.

[1093] Step 12:

[1094] The user submits their revised initial response as the final response. This confirms the response.

[1095] Step 13:

[1096] The device saves the final answer to the database. This makes the saved answer available for later reference.

[1097] Through the above processing steps, the efficiency and accuracy of inquiry handling are improved. The specific actions of each step enable the entire system to function smoothly.

[1098] (Example 1)

[1099] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1100] In modern companies, email and internal networks are widely used as means of internal communication. However, gathering information and responding to inquiries through these means requires considerable effort, and efficiency improvements are needed. Furthermore, providing quick and accurate answers to inquiries is difficult, making it a challenge to improve customer satisfaction.

[1101] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1102] In this invention, the server includes means for collecting information from internal communication data and the internal network; means for analyzing the collected information and extracting useful data; means for storing the extracted data in a storage device; means for searching the storage device based on the content of the inquiry and obtaining relevant data; means for generating a draft of an initial response using a generative artificial intelligence model based on the obtained data; means for presenting the generated draft of the initial response to the user; and means for sending the initial response modified by the user as the final response. This makes it possible to streamline the inquiry handling process and improve customer satisfaction.

[1103] "Internal communication data" refers to data generated through digital communication methods used within a company, such as email, chat, and intranets.

[1104] An "internal network" refers to a limited network environment used for sharing information and communicating within a company.

[1105] "Means of collecting information" refers to methods using protocols and APIs that allow a server to retrieve data from email servers or intranets.

[1106] "Means of analyzing information and extracting useful data" refers to natural language processing and filtering technologies used to detect specific keywords and patterns from collected communication data and extract necessary information.

[1107] "Storage device" refers to databases and storage systems used to hold, manage, and retrieve data.

[1108] "Inquiry content" refers to questions or requests that users make to the server or system.

[1109] "Generative artificial intelligence models" refer to AI models that generate text using natural language processing and machine learning. Specifically, this includes advanced AI technologies such as GPT and BERT.

[1110] "Draft initial response" refers to the initial response text generated by a generative artificial intelligence model.

[1111] "Means of presentation to the user" refers to interface means that allow the user to review the initial draft of the response displayed on the terminal.

[1112] "Final answer" refers to the answer that has been revised and finalized by the user.

[1113] "Means of transmission" refers to email systems or other means of communication used to send the final response to the recipient.

[1114] This invention is a system that efficiently collects useful information from internal communication data and the company's internal network, and uses that information to generate a draft of an initial response using a generative artificial intelligence model. This streamlines the inquiry response process and enables the provision of quick and accurate answers.

[1115] 1. Data Collection

[1116] The server periodically accesses the company's email server and internal network to retrieve new data. It uses the IMAP protocol to retrieve emails and collects updates from the internal network via a REST API. This ensures that the server consistently collects the latest information, such as inquiries and important notifications.

[1117] 2. Data filtering

[1118] The server uses a natural language processing (NLP) library to filter the acquired data and extract useful information. By utilizing the NLP library, inquiries, important notifications, manual updates, and other relevant information are automatically analyzed and filtered.

[1119] 3. Database storage

[1120] The filtered, useful information is stored in storage by the server. Specifically, the information is stored using a database such as PostgreSQL or MongoDB. This database is categorized and indexed according to the query and answer content.

[1121] 4. Generating the initial response

[1122] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on this retrieved information, it uses generative artificial intelligence models such as OpenAI's GPT model or Google's BERT model to generate a draft of the initial response.

[1123] 5. Presentation and revision of the initial response.

[1124] The device displays a draft of the initial response to the user. The user reviews this draft and makes corrections or adds information as needed. The response is then optimized, including any additional instructions or comments from the user.

[1125] 6. Submit and save your final response.

[1126] Once the user submits the final, revised response, the device saves the response to its storage. This saved data can be referenced later and used as foundational data to quickly respond to similar inquiries in the future.

[1127] Specific example

[1128] For example, if there is an inquiry about a new remote control not working, the user would enter the inquiry as follows: "My new remote control isn't working, what should I do?". The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working. The generative AI server generates a draft initial response like this:

[1129] If the remote control is not responding, please try the following steps:

[1130] 1. Check the batteries in the remote control and replace them with new ones.

[1131] 2. Make sure the remote control is pointed towards the target device.

[1132] 3. Check that there are no obstacles between the device and the remote control.

[1133] 4. Clean the light-receiving part of the device.

[1134] If these steps do not resolve the issue, please contact our support center.

[1135] Related technologies

[1136] By using the generated AI model and prompt statements, the accuracy and speed of inquiry handling are improved, resulting in reduced workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies.

[1137] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1138] Step 1:

[1139] The server collects information from the company's internal email server and internal network. In this process, the server uses the IMAP protocol to retrieve emails and a REST API to retrieve updates from the internal network. Inputs are new emails and intranet updates, and output is the collected raw data.

[1140] Specifically, the server automatically connects to the mail server at 9 AM every morning and retrieves all new emails from the previous day. Internal network update information is also collected at the same time.

[1141] Step 2:

[1142] The server analyzes the data collected in Step 1 and extracts useful information. Here, a natural language processing (NLP) library is used to filter the data. The input is the collected raw data, and the output is the filtered useful information.

[1143] Specifically, the system analyzes the email content received by the server and extracts emails containing keywords such as "failure," "inquiry," and "support." Within the internal network, it collects documents tagged with "update," "procedure," and "urgent."

[1144] Step 3:

[1145] The server stores filtered, useful information in storage. Specifically, it uses a database such as PostgreSQL or MongoDB to store the information. The input is filtered, useful information, and the output is the data stored in the database.

[1146] Specifically, the server identifies emails inquiring about "remote control malfunction" and stores their contents in the "remote control related inquiries" table in the database.

[1147] Step 4:

[1148] When the terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Next, it uses a generative artificial intelligence model (for example, OpenAI's GPT model or Google's BERT model) to generate a draft of the initial response. The input is the user's inquiry and the relevant information stored in the database, and the output is the draft of the initial response.

[1149] Specifically, when a user inquires that "the new remote control isn't responding," the terminal sends this inquiry to the AI ​​server. The AI ​​server searches its database, retrieves information about the remote control malfunction, and generates a draft message like this: "If the remote control is not responding, try changing the batteries and checking the remote control's orientation and for any obstructions."

[1150] Step 5:

[1151] The terminal displays a draft of the generated initial response to the user. The user reviews this draft and enters corrections or additional information as needed. The input consists of the generated initial response draft and the user's feedback, and the output is the revised initial response draft.

[1152] Specifically, when a user sees the initial response provided and enters an additional request such as "Please specify the type of battery," the device updates the draft to reflect this feedback.

[1153] Step 6:

[1154] When the user submits the final, revised answer, the device saves that answer to its storage device. The input is the final, revised answer made by the user, and the output is the final answer data saved to the storage device.

[1155] Specifically, when a user clicks the "Submit Final Response" button, the device saves that response data to the "Response History" table in the database. This data is then used as a reference when handling future inquiries.

[1156] (Application Example 1)

[1157] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1158] In a factory setting, responding quickly and accurately to machine malfunctions or operational problems requires searching through a vast amount of information to find relevant data and providing it to workers. However, traditional methods involve manually searching for information based on inquiries and generating appropriate responses, which is time-consuming and labor-intensive. Furthermore, if relevant information cannot be found, users may become stuck, leading to work delays. An efficient system is needed to solve these problems.

[1159] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1160] In this invention, the server includes means for collecting data from internal email and intranets, means for filtering the collected data and extracting useful information, means for storing the extracted information in a database, means for searching the database based on the inquiry content and obtaining relevant information, means for generating a draft of a preliminary response using a generative artificial intelligence model based on the obtained information, means for presenting the generated draft of the preliminary response to the user, means for sending the preliminary response modified by the user as the final response, means for inputting the inquiry content using a smart device, means for displaying a default message if relevant information is not found, means for accepting user modifications, and means for saving user input and modifications in the database. This enables quick and accurate responses to troubles and inquiries within the factory, improving work efficiency and speeding up trouble resolution.

[1161] "Inside the company" refers to the internal workings of a company or organization.

[1162] "Email" is a means of sending and receiving text messages and files over the internet.

[1163] An "intranet" is an internal network used within a company; it utilizes internet technology but is isolated from the outside world.

[1164] "Data collection" is the process of systematically gathering information.

[1165] "Filtering" is the process of selecting necessary information and removing unnecessary information.

[1166] "Extraction" is the process of taking out the necessary parts from collected data.

[1167] A "database" is a system for efficiently storing, managing, and retrieving large amounts of data.

[1168] "Inquiry details" refer to the information that users submit to the system, such as questions or requests.

[1169] A "generative artificial intelligence model" is an algorithm that uses artificial intelligence technology to generate new information or answers.

[1170] A "primary response" is a draft answer that is initially generated in response to a user's inquiry.

[1171] A "smart device" is a high-performance portable device that can connect to the internet, such as a smartphone or tablet.

[1172] A "default message" is a standard message that is automatically displayed when certain conditions are not met.

[1173] A "user" is someone who uses a system or device.

[1174] "Means" refer to the methods and techniques used to achieve an objective.

[1175] "Modification" refers to the operation of making changes to a document or data that has been created.

[1176] The "final answer" is the answer that has been finalized after the user has made any revisions.

[1177] This invention is a system for streamlining troubleshooting and inquiry handling within a factory. The system collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft initial response based on that information. Finally, the generated draft initial response is presented to the user, who can revise it as needed and submit it as the final response.

[1178] System Configuration

[1179] hardware

[1180] Server: A server that performs data collection, filtering, and database storage.

[1181] Device: Smart devices such as smartphones and tablets.

[1182] Generative AI server: A server that runs generative AI models.

[1183] software

[1184] Generative artificial intelligence model: GPT-2 or an equivalent generative AI model.

[1185] Transformers Library: A Python library provided by Hugging Face.

[1186] Process Overview

[1187] Data collection:

[1188] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[1189] Data filtering:

[1190] The server filters the retrieved data and extracts useful information (e.g., inquiry details, important notifications, manual updates).

[1191] Database storage:

[1192] The server stores the extracted useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[1193] Generating the initial answer:

[1194] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response.

[1195] Presentation and revision of the initial response:

[1196] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or enter additional information as needed.

[1197] Submit and save your final response:

[1198] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[1199] Specific example

[1200] For example, consider a case where a user has an inquiry about a new machine not working. The user enters the inquiry, "The new machine isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a machine that isn't working.

[1201] The generative AI server generates a draft of the initial response as follows:

[1202] If the machine is not working, try the following steps:

[1203] 1. Check if the power is on.

[1204] 2. Check that the wiring is connected correctly.

[1205] 3. Check that the machine body has been cleaned.

[1206] 4. Check the machine's error messages.

[1207] If these steps do not resolve the issue, please contact our support center.

[1208] The terminal displays this draft to the user. The user enters additional information, such as, "Please tell me the specific steps to take when checking the wiring." Once the revisions are complete, the user submits the final response. The terminal saves this final response to the database so that it can be referenced later if needed.

[1209] In this way, efficient responses become possible when the same inquiry occurs again, and the amount of work required to handle inquiries can be significantly reduced.

[1210] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1211] Step 1:

[1212] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[1213] Input: Internal email and intranet data.

[1214] Data processing: The acquired raw data is analyzed, and each email and update information is organized as metadata.

[1215] Output: Initial dataset for filtering.

[1216] Step 2:

[1217] The server uses a filtering algorithm to filter the acquired data and extract useful information.

[1218] Input: Initial dataset.

[1219] Data processing: Select useful information and remove unnecessary information based on filtering rules.

[1220] Output: Filtered important information (e.g., inquiry details, manual update information).

[1221] Step 3:

[1222] The server stores the extracted useful information in a database.

[1223] Input: Filtered important information.

[1224] Data Storage: Information is appropriately arranged according to the database structure, and indexes are created.

[1225] Output: A database organized in a searchable format.

[1226] Step 4:

[1227] The terminal receives the user's inquiry and sends it to the generation AI server.

[1228] Input: User inquiry details.

[1229] Data transmission: The query content is sent to the generative AI server according to the protocol.

[1230] Output: The query content is forwarded to the AI ​​server.

[1231] Step 5:

[1232] The generative AI server searches the database and retrieves relevant information.

[1233] Input: User inquiry details.

[1234] Data retrieval: Quickly extract relevant information using database indexes.

[1235] Output: A set of related information.

[1236] Step 6:

[1237] The generative AI server uses a generative artificial intelligence model to generate a draft of the initial response based on the acquired information.

[1238] Input: A set of related information.

[1239] Data processing: Input prompt text into a generative AI model (such as GPT-2) and generate a response.

[1240] Output: Draft of the initial response.

[1241] Step 7:

[1242] The terminal presents the user with a draft of the initial response that has been generated.

[1243] Input: Draft of the initial response.

[1244] Data display: The draft of the initial response is displayed in the user interface.

[1245] Output: Displays the initial response that the user can see.

[1246] Step 8:

[1247] The user reviews the initial draft response provided and enters any necessary corrections or additional information.

[1248] Input: Draft of the initial response, user revisions and additional information.

[1249] Data processing: Reflects user-submitted corrections.

[1250] Output: Corrected initial response.

[1251] Step 9:

[1252] The terminal receives the user's corrections and sends the initial response as the final response, saving it to the database.

[1253] Input: Revised initial response.

[1254] Data storage: The final answer is saved in the database for future searches.

[1255] Output: The final answer stored in the database.

[1256] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1257] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions. The program processing of this system is described below in natural language.

[1258] 1. Data collection and filtering

[1259] The server periodically or at designated times accesses the company's email and intranet servers to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[1260] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[1261] The server converts filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database. This facilitates subsequent searching and AI processing.

[1262] 2. Generating the initial response

[1263] The terminal receives user inquiries through an input form. The terminal then sends the inquiry content to the generation AI server.

[1264] The generative AI server searches the database for information related to the query and stores the search results.

[1265] The generative AI server uses a generative AI model (e.g., GPT-3) based on the acquired relevant information to generate a draft of the initial response. This draft is generated as an appropriate response to the inquiry.

[1266] In this process, the generative AI server further utilizes an emotion engine. The emotion engine analyzes the user's inquiry and recognizes their emotions (e.g., anger, frustration, joy). Based on the recognized emotions, the emotion engine adjusts the tone and expression of the initial response.

[1267] 3. Presentation and revision of the initial response.

[1268] The device displays a draft of the initial response to the user, allowing them to review their answer.

[1269] Users can review this initial draft response and make corrections or add information as needed. When making corrections, they can also add specific instructions or comments.

[1270] 4. Submit and save your final response.

[1271] When the user submits their revised initial response as the final response, the generative AI server uses the emotion engine again to analyze the user's final input and confirm whether the user is satisfied.

[1272] The device saves the final answer to a database. This makes the saved answer available for later reference.

[1273] Specific example

[1274] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[1275] The generative AI server generates a draft of the initial response as follows:

[1276] If the remote control is not responding, please try the following steps:

[1277] 1. Check the batteries in the remote control and replace them with new ones.

[1278] 2. Make sure the remote control is pointed towards the TV.

[1279] 3. Check that there are no obstacles between the TV and the remote control.

[1280] 4. Clean the light-receiving part of the television.

[1281] If these steps do not resolve the issue, please contact our support center.

[1282] Additionally, an emotion engine is used to read the user's feelings of dissatisfaction from their inquiry and adjust the tone accordingly:

[1283] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1284] 1. Check the batteries in the remote control and replace them with new ones.

[1285] 2. Make sure the remote control is pointed towards the TV.

[1286] 3. Check that there are no obstacles between the TV and the remote control.

[1287] 4. Clean the light-receiving part of the television.

[1288] If these steps do not resolve the issue, please contact our support center for further assistance.

[1289] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[1290] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[1291] The following describes the processing flow.

[1292] Data acquisition and filtering

[1293] Step 1:

[1294] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[1295] Step 2:

[1296] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[1297] Step 3:

[1298] The server analyzes the data in temporary storage and extracts useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[1299] Step 4:

[1300] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[1301] Step 5:

[1302] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[1303] Generation of the first response

[1304] Step 6:

[1305] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[1306] Step 7:

[1307] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[1308] Step 8:

[1309] The generative AI server searches the database for information related to the query. The search results are stored.

[1310] Step 9:

[1311] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is then generated as an appropriate response.

[1312] Adjustment by the emotion engine

[1313] Step 10:

[1314] When the generative AI server generates a draft of the initial response, it uses an emotion engine to analyze the user's emotions from the content of their inquiry. This allows for adjustments to be made to match the user's emotions.

[1315] Step 11:

[1316] The generative AI server adjusts the tone and expression of the initial response according to the emotions recognized by the emotion engine. For example, if the user is dissatisfied, expressions of apology and polite language will be added.

[1317] Presentation and revision of the initial response

[1318] Step 12:

[1319] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[1320] Step 13:

[1321] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[1322] Submit and save your final response.

[1323] Step 14:

[1324] The user submits their revised initial response as the final response. At this stage, the generative AI server uses the emotion engine again to analyze the user's final input and maintain the optimal tone.

[1325] Step 15:

[1326] The device saves the final answer to the database. This makes the saved answer available for later reference.

[1327] Specific example

[1328] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[1329] The generative AI server generates a draft of the initial response as follows:

[1330] If the remote control is not responding, please try the following steps:

[1331] 1. Check the batteries in the remote control and replace them with new ones.

[1332] 2. Make sure the remote control is pointed towards the TV.

[1333] 3. Check that there are no obstacles between the TV and the remote control.

[1334] 4. Clean the light-receiving part of the television.

[1335] If these steps do not resolve the issue, please contact our support center.

[1336] Furthermore, it uses an emotion engine to read the user's feelings of dissatisfaction and adjusts the tone as follows:

[1337] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1338] 1. Check the batteries in the remote control and replace them with new ones.

[1339] 2. Make sure the remote control is pointed towards the TV.

[1340] 3. Check that there are no obstacles between the TV and the remote control.

[1341] 4. Clean the light-receiving part of the television.

[1342] If these steps do not resolve the issue, please contact our support center for further assistance.

[1343] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[1344] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[1345] (Example 2)

[1346] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1347] Traditional inquiry handling systems lacked the ability to efficiently extract useful information from the company's vast internal email and intranet, organize and store it in an appropriate format, and automatically generate appropriate responses based on relevant information. In particular, they lacked the function to adjust response content while considering user emotions, which could result in decreased user satisfaction. Against this backdrop, there is a need for a system that can simultaneously improve the efficiency of inquiry handling and enhance user satisfaction.

[1348] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for collecting data from internal electronic messages and the internal network, means for filtering the collected data and extracting useful information, and means for converting the extracted information into a structured data format and storing it in a database. This makes it possible to efficiently organize and store the information necessary for responding to inquiries, and furthermore, to automatically generate responses that provide high user satisfaction using a generative artificial intelligence model and an emotion engine.

[1349] "Internal electronic messages" refer to the content of communications via email and messaging services used within a company or organization.

[1350] An "internal network" refers to an intranet or dedicated network used within a company or organization.

[1351] "Collecting data" means obtaining information from a specific data source.

[1352] "Filtering" is the process of selecting useful information from acquired data and removing unnecessary data.

[1353] A "structured data format" is a format that organizes information systematically and regularly, making it easy to store in a database (e.g., JSON, XML).

[1354] A "database" is a system for storing, managing, and making data searchable.

[1355] "Inquiry content" refers to the text of questions or requests that users submit to the system.

[1356] A "generative artificial intelligence model" is an artificial intelligence that generates natural language based on a large amount of data, such as GPT-3 or similar models.

[1357] A "draft initial response" is the first response proposed by a generative artificial intelligence model.

[1358] An "emotion engine" is a technology that analyzes the user's emotions from text and adjusts the tone and expression of the response based on that analysis.

[1359] A "user" is an individual or group that uses the system.

[1360] A "final answer" is an answer that has been confirmed and revised by the user.

[1361] "Saving" means recording information in a database so that it can be reused or searched later.

[1362] This invention is a system that collects data from internal electronic messages and internal networks, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the content of the inquiry, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions.

[1363] This system includes the following main components:

[1364] 1. Data collection and filtering functions

[1365] 2. Database storage function

[1366] 3. Function to receive inquiries from users

[1367] 4. Search function for related information

[1368] 5. Primary response generation function using generative AI models

[1369] 6. Tone adjustment function using an emotion engine

[1370] 7. Presentation of initial answers and user modification function

[1371] 8. Function to submit and save the final response.

[1372] Hardware and software to be used

[1373] Server: Performs data collection, filtering, database storage, retrieval, initial response generation, sentiment analysis, and final response storage. Protocols and tools used include IMAP (for email collection), HTTP requests (for intranet update collection), natural language processing (NLP) algorithms, generative artificial intelligence models (e.g., GPT-3), sentiment engines (e.g., IBM Watson Tone Analyzer), and database management systems (e.g., MongoDB or MySQL).

[1374] Terminal: Receives user inquiries, provides initial responses, and makes corrections. Web forms or desktop applications are used as the user interface.

[1375] Processing details

[1376] Data acquisition and filtering

[1377] The server periodically accesses the company's electronic messaging server and internal network to retrieve new messages and updates. For example, the server accesses the electronic messaging server using the IMAP protocol to retrieve new mail. The retrieved data is analyzed using NLP algorithms, and useful information is filtered.

[1378] Storage in database

[1379] The filtered, useful information is converted to JSON format and stored in a MongoDB or MySQL database. For example, data in the following format is stored:

[1380] json

[1381] {

[1382] "id": "12345",

[1383] "type": "inquiry",

[1384] "content": "New TV remote not responding"

[1385] }

[1386] Generation of the first response

[1387] When a user submits a query through their device, the server searches the database and retrieves relevant information. Based on this information, a generative AI model (e.g., GPT-3) generates a draft of the initial response. For example, the following prompt might be used:

[1388] "My new TV remote isn't working. What should I do?"

[1389] Tone adjustment by the emotion engine

[1390] During the initial response generation process, the emotion engine analyzes the user's inquiry to understand their emotions and appropriately adjusts the tone of the generated response. For example, if anger or frustration is detected, the tone will be adjusted as follows:

[1391] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps...?"

[1392] User modifications and saving of final answers

[1393] The generated initial response is displayed on the terminal, allowing the user to review it and enter corrections or additional information as needed. The revised final response is then sent back to the server and stored in the database.

[1394] Specific example

[1395] For example, if a user submits a request stating that their new TV remote is not working, the server will generate a preliminary response such as:

[1396] If the remote control is not responding, please try the following steps:

[1397] 1. Check the batteries in the remote control and replace them with new ones.

[1398] 2. Make sure the remote control is pointed towards the TV.

[1399] 3. Check that there are no obstacles between the TV and the remote control.

[1400] 4. Clean the light-receiving part of the television.

[1401] If these steps do not resolve the issue, please contact our support center.

[1402] If the emotion engine detects an emotion of dissatisfaction, the tone will be adjusted as follows:

[1403] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1404] 1. Check the batteries in the remote control and replace them with new ones.

[1405] 2. Make sure the remote control is pointed towards the TV.

[1406] 3. Check that there are no obstacles between the TV and the remote control.

[1407] 4. Clean the light-receiving part of the television.

[1408] If these steps do not resolve the issue, please contact our support center for further assistance.

[1409] This system will improve the efficiency of handling inquiries and enhance user satisfaction.

[1410] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1411] Step 1: Data Collection

[1412] The server accesses the company's electronic messaging server and internal network to retrieve new messages and update information.

[1413] Input: Data from electronic message servers or internal networks.

[1414] Output: Retrieved message data and update information.

[1415] Specific operation: The server accesses the email server using the IMAP protocol to retrieve new mail. It also sends HTTP requests to the internal network to retrieve new posts and updates.

[1416] Step 2: Data Filtering

[1417] The server filters the acquired data to extract useful information.

[1418] Input: Retrieved message data and update information.

[1419] Output: Filtered, useful information.

[1420] Specific operation: The server uses natural language processing (NLP) algorithms to extract useful keywords and phrases from text data and filter out the necessary information.

[1421] Step 3: Storing in the database

[1422] The server converts the filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database.

[1423] Input: Filtered, useful information.

[1424] Output: Information converted to a structured data format and stored in a database.

[1425] Specific operation: The server extracts information, converts it to JSON format, and saves it to a database such as MongoDB or MySQL.

[1426] Step 4: Receiving the Inquiry

[1427] The device receives user inquiries through an input form.

[1428] Input: The content of the inquiry entered by the user.

[1429] Output: Sending the query details to the server.

[1430] Specific operation: When a user enters their inquiry into a web form and clicks the "Submit" button, that information is sent to the server.

[1431] Step 5: Search for related information

[1432] The generative AI server searches the database based on the query content and retrieves relevant information.

[1433] Input: Inquiry details and database.

[1434] Output: Retrieve relevant information.

[1435] Specific operation: The generative AI server sends an SQL or NoSQL query to the database and retrieves records related to the query.

[1436] Step 6: Generating the primary answer

[1437] The generative AI server uses a generative AI model (e.g., GPT-3) to generate a draft of the initial response based on the acquired relevant information.

[1438] Input: Related information and a generated AI model.

[1439] Output: Draft of the initial response.

[1440] Specific operation: The generated prompt is sent to the GPT-3 engine to generate a primary response like this:

[1441] "If the remote control is not responding, please try the following steps..."

[1442] Step 7: Adjustment by the Emotional Engine

[1443] The generative AI server analyzes the sentiment behind the inquiry and adjusts the tone and expression of the initial response.

[1444] Input: Inquiry details and a draft of the initial response.

[1445] Output: A primary response adjusted based on emotions.

[1446] Specific actions: Use a sentiment analysis tool (e.g., IBM Watson Tone Analyzer) to analyze the sentiment from the inquiry and adjust the tone of the response based on the results.

[1447] Step 8: Present your initial answer

[1448] The terminal displays a draft of the initial response that has been generated to the user.

[1449] Input: Draft of the generated initial response.

[1450] Output: The initial response displayed to the user.

[1451] Specific operation: The generated response text will be displayed in the user interface of the web page, allowing the user to review and edit it.

[1452] Step 9: User modification

[1453] The user reviews the initial response and enters any necessary corrections or additional information.

[1454] Input: First response.

[1455] Output: Corrected or added information.

[1456] Specific action: The user edits their response on the web form and clicks the "Final Confirmation" or "Submit Revised" button.

[1457] Step 10: Submit your final response

[1458] The user submits their revised initial response as the final response.

[1459] Input: Revised initial response.

[1460] Output: Final answer.

[1461] Specific operation: After the user makes a final confirmation, they press the submit button, and the content is sent again to the generation AI server.

[1462] Step 11: Saving to the database

[1463] The server saves the final answer to the database.

[1464] Input: Final answer.

[1465] Output: Saved final answer.

[1466] Specific operation: Convert the final answer to JSON format and insert it into the answer table in the database.

[1467] (Application Example 2)

[1468] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1469] Traditional customer support systems have the challenge of not being able to provide appropriate answers to inquiries quickly. Furthermore, they fail to adequately improve customer satisfaction because they do not take customer emotions into consideration. This leads to increased workload for customer support and inconsistent service quality.

[1470] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1471] In this invention, the server includes means for collecting data from internal electronic communication means and internal networks; means for filtering the collected data and extracting useful information; means for storing the extracted information in a database; means for searching the database based on the content of the inquiry and obtaining relevant information; means for generating a draft of an initial response using a generative artificial intelligence model; and means for analyzing the user's emotions and adjusting the tone of the initial response. This makes it possible to quickly provide appropriate responses that respond to the user's emotions and improve the efficiency and quality of customer support.

[1472] "Internal electronic communication methods" refer to communication methods used within the company, such as email and chat systems.

[1473] An "internal network" refers to an internal network used within a company or organization, including intranets and dedicated networks.

[1474] "Means of data collection" refers to the processes and mechanisms for obtaining necessary information from electronic communication methods or internal networks.

[1475] "Filtering" is the process of selecting useful information from collected data and eliminating unnecessary data.

[1476] "Useful information" refers to important data that is helpful in handling inquiries and performing tasks, and includes manual information and past inquiry records.

[1477] A "database" is a data storage system that systematically stores collected information, making it easy to search and manage.

[1478] A "generative artificial intelligence model" is an AI model that can generate natural language based on large amounts of data, and models like GPT-3 fall into this category.

[1479] A "draft initial response" is the first proposed answer to an inquiry generated by an AI model.

[1480] "Means of presenting to the user" refers to methods or devices for displaying the generated draft of the initial response so that the user can review it.

[1481] "Methods for analyzing emotions" refer to technologies that identify emotions from user inquiries and adjust the tone and content of responses accordingly.

[1482] "Methods for adjusting tone" refer to the process of appropriately modifying the expression and nuances of the generated response based on the analyzed emotions.

[1483] This invention relates to a system for streamlining customer support operations in physical stores. The system collects data from internal electronic communication methods and networks, filters it, and extracts useful information. The extracted information is stored in a database and searched based on the inquiry content. Using a generative artificial intelligence model, a draft of the initial response is generated based on the acquired information, and the tone of the response is adjusted by analyzing the user's emotions using an emotion engine.

[1484] Description of the system's program processing

[1485] The server periodically or at designated times accesses the company's electronic communication methods (email servers and chat systems) and internal network (intranet) to retrieve new emails and updates. This data is filtered, and useful information is converted into structured data formats (e.g., JSON, XML) and stored in the database.

[1486] The terminal receives user inquiries through an input form and sends the inquiry details to the server. The server searches its database and retrieves relevant information. It then uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial response.

[1487] The generated response uses an emotion engine to analyze the sentiment of the inquiry. For example, it uses TextBlob to identify the sentiment (positive, negative, neutral, etc.) of the text and adjusts the tone and content of the response accordingly. The adjusted initial response draft is then presented to the user via the device.

[1488] The user can review this initial draft response and make corrections or add information as needed. Once the final response is decided, the server saves it to the database for later reference.

[1489] Specific example

[1490] For example, consider a user's inquiry about a new TV remote not working. The user might input, "My new TV remote isn't working, what should I do?" The terminal receives this inquiry and sends it to the server. The server searches its database and retrieves information on how to deal with a non-responsive remote. Then, using a generative artificial intelligence model, it generates a draft initial response like this:

[1491] If the remote control is not responding, please try the following steps:

[1492] 1. Check the batteries in the remote control and replace them with new ones.

[1493] 2. Make sure the remote control is pointed towards the TV.

[1494] 3. Check that there are no obstacles between the TV and the remote control.

[1495] 4. Clean the light-receiving part of the television.

[1496] If these steps do not resolve the issue, please contact our support center.

[1497] Furthermore, using an emotion engine, the system reads the user's feelings of dissatisfaction from their inquiry and adjusts the tone accordingly:

[1498] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1499] 1. Check the batteries in the remote control and replace them with new ones.

[1500] 2. Make sure the remote control is pointed towards the TV.

[1501] 3. Check that there are no obstacles between the TV and the remote control.

[1502] 4. Clean the light-receiving part of the television.

[1503] If these steps do not resolve the issue, please contact our support center for further assistance.

[1504] The terminal displays this draft to the user, who then enters additional information such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer, and the server saves that answer to the database.

[1505] This system enables in-store staff to handle customer interactions efficiently and effectively, thereby improving customer satisfaction.

[1506] Example of a prompt

[1507] "My new TV remote isn't working. What should I do?"

[1508] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1509] Step 1:

[1510] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. Its input consists of unread messages and new information retrieved from email and the intranet; this data is filtered to extract only the useful information. The output is this filtered, useful information.

[1511] Step 2:

[1512] The server converts the filtered data into a structured data format (e.g., JSON, XML). This process organizes the raw input data into an appropriate format so that it can be stored in a unified database. The output is structured data that can be stored in the database.

[1513] Step 3:

[1514] A database stores structured data and manages it so that it can be quickly accessed when needed. This step involves taking structured data as input and performing database operations to efficiently store it. The output is the storage of the most up-to-date data, always accessible.

[1515] Step 4:

[1516] The terminal receives user inquiries through an input form. In this step, the user's inquiry content is received as input and sent to the server. The inquiry content is then returned to the server as output.

[1517] Step 5:

[1518] The server searches the database for information related to the query and retrieves the relevant information. The input is the user's query, and the database is searched based on that content. The output is the information related to the query.

[1519] Step 6:

[1520] The server uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial answer based on the acquired relevant information. In this step, relevant information is taken as input, and the generative artificial intelligence model generates the answer based on it. The output is a draft of the initial answer.

[1521] Step 7:

[1522] The server uses an emotion engine to adjust the tone of the initial response draft based on the emotion of the inquiry. In this step, the emotion contained in the user's inquiry is taken as input, and the emotion engine analyzes it to adjust the tone. The output is the adjusted initial response draft.

[1523] Step 8:

[1524] The terminal presents the user with a draft of the adjusted initial response. In this step, the input is the adjusted initial response, which is displayed to the user. As output, the user can review the draft of the initial response.

[1525] Step 9:

[1526] The user reviews the initial draft response and enters any necessary corrections or additional information. This step includes the user's additional information and corrections as input, and the response is ready to be submitted as the final response. The output is the final response, corrected by the user.

[1527] Step 10:

[1528] The server saves the user's modified final answer to a database for future reference. In this step, the input is the final answer, which is stored in the database. The output is the saved final answer, which can be referenced later.

[1529] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1530] The data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One 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">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1531] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1532] [Fourth Embodiment]

[1533] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1534] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1535] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1536] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1537] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1538] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1539] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1540] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1541] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1542] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1543] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1544] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1545] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1546] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. The generated draft of the initial response is presented to the user, who can modify it as needed and submit it as the final response. The program processing of this system is described below in natural language.

[1547] 1. Data Collection

[1548] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[1549] 2. Data filtering

[1550] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[1551] 3. Database storage

[1552] The server stores filtered, useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[1553] 4. Generating the initial response

[1554] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response. This draft combines the relevant information to create the initial response text.

[1555] 5. Presentation and revision of the initial response.

[1556] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or add information as needed. When making corrections, it is also possible to add specific instructions or comments.

[1557] 6. Submit and save your final response.

[1558] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[1559] Specific example

[1560] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[1561] The generative AI server generates a draft of the initial response as follows:

[1562] If the remote control is not responding, please try the following steps:

[1563] 1. Check the batteries in the remote control and replace them with new ones.

[1564] 2. Make sure the remote control is pointed towards the TV.

[1565] 3. Check that there are no obstacles between the TV and the remote control.

[1566] 4. Clean the light-receiving part of the television.

[1567] If these steps do not resolve the issue, please contact our support center.

[1568] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[1569] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[1570] The following describes the processing flow.

[1571] Data acquisition and filtering

[1572] Step 1:

[1573] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[1574] Step 2:

[1575] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[1576] Step 3:

[1577] The server analyzes and filters the data in temporary storage, extracting useful information (e.g., inquiry details, important notifications).

[1578] Step 4:

[1579] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[1580] Step 5:

[1581] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[1582] Generation of the first response

[1583] Step 6:

[1584] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[1585] Step 7:

[1586] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[1587] Step 8:

[1588] The generative AI server searches the database for information related to the query. The search results are stored.

[1589] Step 9:

[1590] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is generated as an appropriate response to the inquiry.

[1591] Presentation and revision of the initial response

[1592] Step 10:

[1593] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[1594] Step 11:

[1595] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[1596] Submit and save your final response.

[1597] Step 12:

[1598] The user submits their revised initial response as the final response. This confirms the response.

[1599] Step 13:

[1600] The device saves the final answer to the database. This makes the saved answer available for later reference.

[1601] Through the above processing steps, the efficiency and accuracy of inquiry handling are improved. The specific actions of each step enable the entire system to function smoothly.

[1602] (Example 1)

[1603] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1604] In modern companies, email and internal networks are widely used as means of internal communication. However, gathering information and responding to inquiries through these means requires considerable effort, and efficiency improvements are needed. Furthermore, providing quick and accurate answers to inquiries is difficult, making it a challenge to improve customer satisfaction.

[1605] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1606] In this invention, the server includes means for collecting information from internal communication data and the internal network; means for analyzing the collected information and extracting useful data; means for storing the extracted data in a storage device; means for searching the storage device based on the content of the inquiry and obtaining relevant data; means for generating a draft of an initial response using a generative artificial intelligence model based on the obtained data; means for presenting the generated draft of the initial response to the user; and means for sending the initial response modified by the user as the final response. This makes it possible to streamline the inquiry handling process and improve customer satisfaction.

[1607] "Internal communication data" refers to data generated through digital communication methods used within a company, such as email, chat, and intranets.

[1608] An "internal network" refers to a limited network environment used for sharing information and communicating within a company.

[1609] "Means of collecting information" refers to methods using protocols and APIs that allow a server to retrieve data from email servers or intranets.

[1610] "Means of analyzing information and extracting useful data" refers to natural language processing and filtering technologies used to detect specific keywords and patterns from collected communication data and extract necessary information.

[1611] "Storage device" refers to databases and storage systems used to hold, manage, and retrieve data.

[1612] "Inquiry content" refers to questions or requests that users make to the server or system.

[1613] "Generative artificial intelligence models" refer to AI models that generate text using natural language processing and machine learning. Specifically, this includes advanced AI technologies such as GPT and BERT.

[1614] "Draft initial response" refers to the initial response text generated by a generative artificial intelligence model.

[1615] "Means of presentation to the user" refers to interface means that allow the user to review the initial draft of the response displayed on the terminal.

[1616] "Final answer" refers to the answer that has been revised and finalized by the user.

[1617] "Means of transmission" refers to email systems or other means of communication used to send the final response to the recipient.

[1618] This invention is a system that efficiently collects useful information from internal communication data and the company's internal network, and uses that information to generate a draft of an initial response using a generative artificial intelligence model. This streamlines the inquiry response process and enables the provision of quick and accurate answers.

[1619] 1. Data Collection

[1620] The server periodically accesses the company's email server and internal network to retrieve new data. It uses the IMAP protocol to retrieve emails and collects updates from the internal network via a REST API. This ensures that the server consistently collects the latest information, such as inquiries and important notifications.

[1621] 2. Data filtering

[1622] The server uses a natural language processing (NLP) library to filter the acquired data and extract useful information. By utilizing the NLP library, inquiries, important notifications, manual updates, and other relevant information are automatically analyzed and filtered.

[1623] 3. Database storage

[1624] The filtered, useful information is stored in storage by the server. Specifically, the information is stored using a database such as PostgreSQL or MongoDB. This database is categorized and indexed according to the query and answer content.

[1625] 4. Generating the initial response

[1626] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on this retrieved information, it uses generative artificial intelligence models such as OpenAI's GPT model or Google's BERT model to generate a draft of the initial response.

[1627] 5. Presentation and revision of the initial response.

[1628] The device displays a draft of the initial response to the user. The user reviews this draft and makes corrections or adds information as needed. The response is then optimized, including any additional instructions or comments from the user.

[1629] 6. Submit and save your final response.

[1630] Once the user submits the final, revised response, the device saves the response to its storage. This saved data can be referenced later and used as foundational data to quickly respond to similar inquiries in the future.

[1631] Specific example

[1632] For example, if there is an inquiry about a new remote control not working, the user would enter the inquiry as follows: "My new remote control isn't working, what should I do?". The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working. The generative AI server generates a draft initial response like this:

[1633] If the remote control is not responding, please try the following steps:

[1634] 1. Check the batteries in the remote control and replace them with new ones.

[1635] 2. Make sure the remote control is pointed towards the target device.

[1636] 3. Check that there are no obstacles between the device and the remote control.

[1637] 4. Clean the light-receiving part of the device.

[1638] If these steps do not resolve the issue, please contact our support center.

[1639] Related technologies

[1640] By using the generated AI model and prompt statements, the accuracy and speed of inquiry handling are improved, resulting in reduced workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies.

[1641] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1642] Step 1:

[1643] The server collects information from the company's internal email server and internal network. In this process, the server uses the IMAP protocol to retrieve emails and a REST API to retrieve updates from the internal network. Inputs are new emails and intranet updates, and output is the collected raw data.

[1644] Specifically, the server automatically connects to the mail server at 9 AM every morning and retrieves all new emails from the previous day. Internal network update information is also collected at the same time.

[1645] Step 2:

[1646] The server analyzes the data collected in Step 1 and extracts useful information. Here, a natural language processing (NLP) library is used to filter the data. The input is the collected raw data, and the output is the filtered useful information.

[1647] Specifically, the system analyzes the email content received by the server and extracts emails containing keywords such as "failure," "inquiry," and "support." Within the internal network, it collects documents tagged with "update," "procedure," and "urgent."

[1648] Step 3:

[1649] The server stores filtered, useful information in storage. Specifically, it uses a database such as PostgreSQL or MongoDB to store the information. The input is filtered, useful information, and the output is the data stored in the database.

[1650] Specifically, the server identifies emails inquiring about "remote control malfunction" and stores their contents in the "remote control related inquiries" table in the database.

[1651] Step 4:

[1652] When the terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Next, it uses a generative artificial intelligence model (for example, OpenAI's GPT model or Google's BERT model) to generate a draft of the initial response. The input is the user's inquiry and the relevant information stored in the database, and the output is the draft of the initial response.

[1653] Specifically, when a user inquires that "the new remote control isn't responding," the terminal sends this inquiry to the AI ​​server. The AI ​​server searches its database, retrieves information about the remote control malfunction, and generates a draft message like this: "If the remote control is not responding, try changing the batteries and checking the remote control's orientation and for any obstructions."

[1654] Step 5:

[1655] The terminal displays a draft of the generated initial response to the user. The user reviews this draft and enters corrections or additional information as needed. The input consists of the generated initial response draft and the user's feedback, and the output is the revised initial response draft.

[1656] Specifically, when a user sees the initial response provided and enters an additional request such as "Please specify the type of battery," the device updates the draft to reflect this feedback.

[1657] Step 6:

[1658] When the user submits the final, revised answer, the device saves that answer to its storage device. The input is the final, revised answer made by the user, and the output is the final answer data saved to the storage device.

[1659] Specifically, when a user clicks the "Submit Final Response" button, the device saves that response data to the "Response History" table in the database. This data is then used as a reference when handling future inquiries.

[1660] (Application Example 1)

[1661] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1662] In a factory setting, responding quickly and accurately to machine malfunctions or operational problems requires searching through a vast amount of information to find relevant data and providing it to workers. However, traditional methods involve manually searching for information based on inquiries and generating appropriate responses, which is time-consuming and labor-intensive. Furthermore, if relevant information cannot be found, users may become stuck, leading to work delays. An efficient system is needed to solve these problems.

[1663] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1664] In this invention, the server includes means for collecting data from internal email and intranets, means for filtering the collected data and extracting useful information, means for storing the extracted information in a database, means for searching the database based on the inquiry content and obtaining relevant information, means for generating a draft of a preliminary response using a generative artificial intelligence model based on the obtained information, means for presenting the generated draft of the preliminary response to the user, means for sending the preliminary response modified by the user as the final response, means for inputting the inquiry content using a smart device, means for displaying a default message if relevant information is not found, means for accepting user modifications, and means for saving user input and modifications in the database. This enables quick and accurate responses to troubles and inquiries within the factory, improving work efficiency and speeding up trouble resolution.

[1665] "Inside the company" refers to the internal workings of a company or organization.

[1666] "Email" is a means of sending and receiving text messages and files over the internet.

[1667] An "intranet" is an internal network used within a company; it utilizes internet technology but is isolated from the outside world.

[1668] "Data collection" is the process of systematically gathering information.

[1669] "Filtering" is the process of selecting necessary information and removing unnecessary information.

[1670] "Extraction" is the process of taking out the necessary parts from collected data.

[1671] A "database" is a system for efficiently storing, managing, and retrieving large amounts of data.

[1672] "Inquiry details" refer to the information that users submit to the system, such as questions or requests.

[1673] A "generative artificial intelligence model" is an algorithm that uses artificial intelligence technology to generate new information or answers.

[1674] A "primary response" is a draft answer that is initially generated in response to a user's inquiry.

[1675] A "smart device" is a high-performance portable device that can connect to the internet, such as a smartphone or tablet.

[1676] A "default message" is a standard message that is automatically displayed when certain conditions are not met.

[1677] A "user" is someone who uses a system or device.

[1678] "Means" refer to the methods and techniques used to achieve an objective.

[1679] "Modification" refers to the operation of making changes to a document or data that has been created.

[1680] The "final answer" is the answer that has been finalized after the user has made any revisions.

[1681] This invention is a system for streamlining troubleshooting and inquiry handling within a factory. The system collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft initial response based on that information. Finally, the generated draft initial response is presented to the user, who can revise it as needed and submit it as the final response.

[1682] System Configuration

[1683] hardware

[1684] Server: A server that performs data collection, filtering, and database storage.

[1685] Device: Smart devices such as smartphones and tablets.

[1686] Generative AI server: A server that runs generative AI models.

[1687] software

[1688] Generative artificial intelligence model: GPT-2 or an equivalent generative AI model.

[1689] Transformers Library: A Python library provided by Hugging Face.

[1690] Process Overview

[1691] Data collection:

[1692] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[1693] Data filtering:

[1694] The server filters the retrieved data and extracts useful information (e.g., inquiry details, important notifications, manual updates).

[1695] Database storage:

[1696] The server stores the extracted useful information in a database. This makes it easier for generative artificial intelligence models to access the data later.

[1697] Generating the initial answer:

[1698] When a terminal receives a user inquiry, the generative AI server searches the database and retrieves relevant information. Based on the retrieved information, the generative artificial intelligence model generates a draft of the initial response.

[1699] Presentation and revision of the initial response:

[1700] The terminal displays a draft of the initial response that has been generated. The user can review this draft and make corrections or enter additional information as needed.

[1701] Submit and save your final response:

[1702] Once the user submits their revised final answer, the device saves that answer to a database. This saved data can be referenced later and will be useful if similar inquiries arise.

[1703] Specific example

[1704] For example, consider a case where a user has an inquiry about a new machine not working. The user enters the inquiry, "The new machine isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a machine that isn't working.

[1705] The generative AI server generates a draft of the initial response as follows:

[1706] If the machine is not working, try the following steps:

[1707] 1. Check if the power is on.

[1708] 2. Check that the wiring is connected correctly.

[1709] 3. Check that the machine body has been cleaned.

[1710] 4. Check the machine's error messages.

[1711] If these steps do not resolve the issue, please contact our support center.

[1712] The terminal displays this draft to the user. The user enters additional information, such as, "Please tell me the specific steps to take when checking the wiring." Once the revisions are complete, the user submits the final response. The terminal saves this final response to the database so that it can be referenced later if needed.

[1713] In this way, efficient responses become possible when the same inquiry occurs again, and the amount of work required to handle inquiries can be significantly reduced.

[1714] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1715] Step 1:

[1716] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates.

[1717] Input: Internal email and intranet data.

[1718] Data processing: The acquired raw data is analyzed, and each email and update information is organized as metadata.

[1719] Output: Initial dataset for filtering.

[1720] Step 2:

[1721] The server uses a filtering algorithm to filter the acquired data and extract useful information.

[1722] Input: Initial dataset.

[1723] Data processing: Select useful information and remove unnecessary information based on filtering rules.

[1724] Output: Filtered important information (e.g., inquiry details, manual update information).

[1725] Step 3:

[1726] The server stores the extracted useful information in a database.

[1727] Input: Filtered important information.

[1728] Data Storage: Information is appropriately arranged according to the database structure, and indexes are created.

[1729] Output: A database organized in a searchable format.

[1730] Step 4:

[1731] The terminal receives the user's inquiry and sends it to the generation AI server.

[1732] Input: User inquiry details.

[1733] Data transmission: The query content is sent to the generative AI server according to the protocol.

[1734] Output: The query content is forwarded to the AI ​​server.

[1735] Step 5:

[1736] The generative AI server searches the database and retrieves relevant information.

[1737] Input: User inquiry details.

[1738] Data retrieval: Quickly extract relevant information using database indexes.

[1739] Output: A set of related information.

[1740] Step 6:

[1741] The generative AI server uses a generative artificial intelligence model to generate a draft of the initial response based on the acquired information.

[1742] Input: A set of related information.

[1743] Data processing: Input prompt text into a generative AI model (such as GPT-2) and generate a response.

[1744] Output: Draft of the initial response.

[1745] Step 7:

[1746] The terminal presents the user with a draft of the initial response that has been generated.

[1747] Input: Draft of the initial response.

[1748] Data display: The draft of the initial response is displayed in the user interface.

[1749] Output: Displays the initial response that the user can see.

[1750] Step 8:

[1751] The user reviews the initial draft response provided and enters any necessary corrections or additional information.

[1752] Input: Draft of the initial response, user revisions and additional information.

[1753] Data processing: Reflects user-submitted corrections.

[1754] Output: Corrected initial response.

[1755] Step 9:

[1756] The terminal receives the user's corrections and sends the initial response as the final response, saving it to the database.

[1757] Input: Revised initial response.

[1758] Data storage: The final answer is saved in the database for future searches.

[1759] Output: The final answer stored in the database.

[1760] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1761] This invention is a system that collects data from internal email and intranet sources, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the inquiry content, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions. The program processing of this system is described below in natural language.

[1762] 1. Data collection and filtering

[1763] The server periodically or at designated times accesses the company's email and intranet servers to retrieve new emails and intranet updates. This ensures that the latest inquiries and relevant information are collected without fail.

[1764] The server filters the retrieved data to extract useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[1765] The server converts filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database. This facilitates subsequent searching and AI processing.

[1766] 2. Generating the initial response

[1767] The terminal receives user inquiries through an input form. The terminal then sends the inquiry content to the generation AI server.

[1768] The generative AI server searches the database for information related to the query and stores the search results.

[1769] The generative AI server uses a generative AI model (e.g., GPT-3) based on the acquired relevant information to generate a draft of the initial response. This draft is generated as an appropriate response to the inquiry.

[1770] In this process, the generative AI server further utilizes an emotion engine. The emotion engine analyzes the user's inquiry and recognizes their emotions (e.g., anger, frustration, joy). Based on the recognized emotions, the emotion engine adjusts the tone and expression of the initial response.

[1771] 3. Presentation and revision of the initial response.

[1772] The device displays a draft of the initial response to the user, allowing them to review their answer.

[1773] Users can review this initial draft response and make corrections or add information as needed. When making corrections, they can also add specific instructions or comments.

[1774] 4. Submit and save your final response.

[1775] When the user submits their revised initial response as the final response, the generative AI server uses the emotion engine again to analyze the user's final input and confirm whether the user is satisfied.

[1776] The device saves the final answer to a database. This makes the saved answer available for later reference.

[1777] Specific example

[1778] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[1779] The generative AI server generates a draft of the initial response as follows:

[1780] If the remote control is not responding, please try the following steps:

[1781] 1. Check the batteries in the remote control and replace them with new ones.

[1782] 2. Make sure the remote control is pointed towards the TV.

[1783] 3. Check that there are no obstacles between the TV and the remote control.

[1784] 4. Clean the light-receiving part of the television.

[1785] If these steps do not resolve the issue, please contact our support center.

[1786] Additionally, an emotion engine is used to read the user's feelings of dissatisfaction from their inquiry and adjust the tone accordingly:

[1787] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1788] 1. Check the batteries in the remote control and replace them with new ones.

[1789] 2. Make sure the remote control is pointed towards the TV.

[1790] 3. Check that there are no obstacles between the TV and the remote control.

[1791] 4. Clean the light-receiving part of the television.

[1792] If these steps do not resolve the issue, please contact our support center for further assistance.

[1793] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[1794] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[1795] The following describes the processing flow.

[1796] Data acquisition and filtering

[1797] Step 1:

[1798] The server accesses the company's email server and intranet server periodically or at specified times, ensuring that the latest information is available.

[1799] Step 2:

[1800] The server saves newly acquired emails and intranet update information to temporary storage. This makes subsequent processing easier.

[1801] Step 3:

[1802] The server analyzes the data in temporary storage and extracts useful information. For example, inquiry details, important notifications, and manual update information are filtered.

[1803] Step 4:

[1804] The server converts filtered, useful information into a structured data format (e.g., JSON, XML). This makes it easier to store the information in a database.

[1805] Step 5:

[1806] The server stores structured data in a database, which facilitates subsequent searching and AI processing.

[1807] Generation of the first response

[1808] Step 6:

[1809] The terminal receives user inquiries through an input form. This transmits the inquiry details to the system.

[1810] Step 7:

[1811] The terminal sends the user's inquiry to the generation AI server. This allows the process to proceed to the next step.

[1812] Step 8:

[1813] The generative AI server searches the database for information related to the query. The search results are stored.

[1814] Step 9:

[1815] Based on the relevant information acquired by the generative AI server, a draft of the initial response is generated using a generative AI model (e.g., GPT-3). This draft is then generated as an appropriate response.

[1816] Adjustment by the emotion engine

[1817] Step 10:

[1818] When the generative AI server generates a draft of the initial response, it uses an emotion engine to analyze the user's emotions from the content of their inquiry. This allows for adjustments to be made to match the user's emotions.

[1819] Step 11:

[1820] The generative AI server adjusts the tone and expression of the initial response according to the emotions recognized by the emotion engine. For example, if the user is dissatisfied, expressions of apology and polite language will be added.

[1821] Presentation and revision of the initial response

[1822] Step 12:

[1823] The terminal displays a draft of the initial response to the user, allowing the user to review their answer.

[1824] Step 13:

[1825] Users review their initial draft responses and make corrections or add information as needed. This optimizes their responses.

[1826] Submit and save your final response.

[1827] Step 14:

[1828] The user submits their revised initial response as the final response. At this stage, the generative AI server uses the emotion engine again to analyze the user's final input and maintain the optimal tone.

[1829] Step 15:

[1830] The device saves the final answer to the database. This makes the saved answer available for later reference.

[1831] Specific example

[1832] For example, consider an inquiry about a new TV remote control not working. The user enters the inquiry, "My new TV remote control isn't working, what should I do?" The terminal receives this inquiry and sends it to the generative AI server. The generative AI server searches its database and retrieves information on how to deal with a remote control that is not working.

[1833] The generative AI server generates a draft of the initial response as follows:

[1834] If the remote control is not responding, please try the following steps:

[1835] 1. Check the batteries in the remote control and replace them with new ones.

[1836] 2. Make sure the remote control is pointed towards the TV.

[1837] 3. Check that there are no obstacles between the TV and the remote control.

[1838] 4. Clean the light-receiving part of the television.

[1839] If these steps do not resolve the issue, please contact our support center.

[1840] Furthermore, it uses an emotion engine to read the user's feelings of dissatisfaction and adjusts the tone as follows:

[1841] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1842] 1. Check the batteries in the remote control and replace them with new ones.

[1843] 2. Make sure the remote control is pointed towards the TV.

[1844] 3. Check that there are no obstacles between the TV and the remote control.

[1845] 4. Clean the light-receiving part of the television.

[1846] If these steps do not resolve the issue, please contact our support center for further assistance.

[1847] The device displays this draft to the user. The user enters additional information, such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer. The device saves this final answer to its database so that it can be referenced later if needed.

[1848] In this way, the inquiry handling process is streamlined, reducing the workload for each department. Due to the system's high versatility, it can also be applied to handling external inquiries and outsourcing to other companies in the future.

[1849] (Example 2)

[1850] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1851] Traditional inquiry handling systems lacked the ability to efficiently extract useful information from the company's vast internal email and intranet, organize and store it in an appropriate format, and automatically generate appropriate responses based on relevant information. In particular, they lacked the function to adjust response content while considering user emotions, which could result in decreased user satisfaction. Against this backdrop, there is a need for a system that can simultaneously improve the efficiency of inquiry handling and enhance user satisfaction.

[1852] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for collecting data from internal electronic messages and the internal network, means for filtering the collected data and extracting useful information, and means for converting the extracted information into a structured data format and storing it in a database. This makes it possible to efficiently organize and store the information necessary for responding to inquiries, and furthermore, to automatically generate responses that provide high user satisfaction using a generative artificial intelligence model and an emotion engine.

[1853] "Internal electronic messages" refer to the content of communications via email and messaging services used within a company or organization.

[1854] An "internal network" refers to an intranet or dedicated network used within a company or organization.

[1855] "Collecting data" means obtaining information from a specific data source.

[1856] "Filtering" is the process of selecting useful information from acquired data and removing unnecessary data.

[1857] A "structured data format" is a format that organizes information systematically and regularly, making it easy to store in a database (e.g., JSON, XML).

[1858] A "database" is a system for storing, managing, and making data searchable.

[1859] "Inquiry content" refers to the text of questions or requests that users submit to the system.

[1860] A "generative artificial intelligence model" is an artificial intelligence that generates natural language based on a large amount of data, such as GPT-3 or similar models.

[1861] A "draft initial response" is the first response proposed by a generative artificial intelligence model.

[1862] An "emotion engine" is a technology that analyzes the user's emotions from text and adjusts the tone and expression of the response based on that analysis.

[1863] A "user" is an individual or group that uses the system.

[1864] A "final answer" is an answer that has been confirmed and revised by the user.

[1865] "Saving" means recording information in a database so that it can be reused or searched later.

[1866] This invention is a system that collects data from internal electronic messages and internal networks, filters that data to extract useful information, and stores it in a database. Furthermore, it searches the database based on the content of the inquiry, retrieves relevant information, and uses a generative artificial intelligence model to generate a draft of the initial response based on that information. In the initial response generation process, an emotion engine is incorporated to provide an appropriate response that corresponds to the user's emotions.

[1867] This system includes the following main components:

[1868] 1. Data collection and filtering functions

[1869] 2. Database storage function

[1870] 3. Function to receive inquiries from users

[1871] 4. Search function for related information

[1872] 5. Primary response generation function using generative AI models

[1873] 6. Tone adjustment function using an emotion engine

[1874] 7. Presentation of initial answers and user modification function

[1875] 8. Function to submit and save the final response.

[1876] Hardware and software to be used

[1877] Server: Performs data collection, filtering, database storage, retrieval, initial response generation, sentiment analysis, and final response storage. Protocols and tools used include IMAP (for email collection), HTTP requests (for intranet update collection), natural language processing (NLP) algorithms, generative artificial intelligence models (e.g., GPT-3), sentiment engines (e.g., IBM Watson Tone Analyzer), and database management systems (e.g., MongoDB or MySQL).

[1878] Terminal: Receives user inquiries, provides initial responses, and makes corrections. Web forms or desktop applications are used as the user interface.

[1879] Processing details

[1880] Data acquisition and filtering

[1881] The server periodically accesses the company's electronic messaging server and internal network to retrieve new messages and updates. For example, the server accesses the electronic messaging server using the IMAP protocol to retrieve new mail. The retrieved data is analyzed using NLP algorithms, and useful information is filtered.

[1882] Storage in database

[1883] The filtered, useful information is converted to JSON format and stored in a MongoDB or MySQL database. For example, data in the following format is stored:

[1884] json

[1885] {

[1886] "id": "12345",

[1887] "type": "inquiry",

[1888] "content": "New TV remote not responding"

[1889] }

[1890] Generation of the first response

[1891] When a user submits a query through their device, the server searches the database and retrieves relevant information. Based on this information, a generative AI model (e.g., GPT-3) generates a draft of the initial response. For example, the following prompt might be used:

[1892] "My new TV remote isn't working. What should I do?"

[1893] Tone adjustment by the emotion engine

[1894] During the initial response generation process, the emotion engine analyzes the user's inquiry to understand their emotions and appropriately adjusts the tone of the generated response. For example, if anger or frustration is detected, the tone will be adjusted as follows:

[1895] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps...?"

[1896] User modifications and saving of final answers

[1897] The generated initial response is displayed on the terminal, allowing the user to review it and enter corrections or additional information as needed. The revised final response is then sent back to the server and stored in the database.

[1898] Specific example

[1899] For example, if a user submits a request stating that their new TV remote is not working, the server will generate a preliminary response such as:

[1900] If the remote control is not responding, please try the following steps:

[1901] 1. Check the batteries in the remote control and replace them with new ones.

[1902] 2. Make sure the remote control is pointed towards the TV.

[1903] 3. Check that there are no obstacles between the TV and the remote control.

[1904] 4. Clean the light-receiving part of the television.

[1905] If these steps do not resolve the issue, please contact our support center.

[1906] If the emotion engine detects an emotion of dissatisfaction, the tone will be adjusted as follows:

[1907] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[1908] 1. Check the batteries in the remote control and replace them with new ones.

[1909] 2. Make sure the remote control is pointed towards the TV.

[1910] 3. Check that there are no obstacles between the TV and the remote control.

[1911] 4. Clean the light-receiving part of the television.

[1912] If these steps do not resolve the issue, please contact our support center for further assistance.

[1913] This system will improve the efficiency of handling inquiries and enhance user satisfaction.

[1914] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1915] Step 1: Data Collection

[1916] The server accesses the company's electronic messaging server and internal network to retrieve new messages and update information.

[1917] Input: Data from electronic message servers or internal networks.

[1918] Output: Retrieved message data and update information.

[1919] Specific operation: The server accesses the email server using the IMAP protocol to retrieve new mail. It also sends HTTP requests to the internal network to retrieve new posts and updates.

[1920] Step 2: Data Filtering

[1921] The server filters the acquired data to extract useful information.

[1922] Input: Retrieved message data and update information.

[1923] Output: Filtered, useful information.

[1924] Specific operation: The server uses natural language processing (NLP) algorithms to extract useful keywords and phrases from text data and filter out the necessary information.

[1925] Step 3: Storing in the database

[1926] The server converts the filtered, useful information into a structured data format (e.g., JSON, XML) and stores it in the database.

[1927] Input: Filtered, useful information.

[1928] Output: Information converted to a structured data format and stored in a database.

[1929] Specific operation: The server extracts information, converts it to JSON format, and saves it to a database such as MongoDB or MySQL.

[1930] Step 4: Receiving the Inquiry

[1931] The device receives user inquiries through an input form.

[1932] Input: The content of the inquiry entered by the user.

[1933] Output: Sending the query details to the server.

[1934] Specific operation: When a user enters their inquiry into a web form and clicks the "Submit" button, that information is sent to the server.

[1935] Step 5: Search for related information

[1936] The generative AI server searches the database based on the query content and retrieves relevant information.

[1937] Input: Inquiry details and database.

[1938] Output: Retrieve relevant information.

[1939] Specific operation: The generative AI server sends an SQL or NoSQL query to the database and retrieves records related to the query.

[1940] Step 6: Generating the primary answer

[1941] The generative AI server uses a generative AI model (e.g., GPT-3) to generate a draft of the initial response based on the acquired relevant information.

[1942] Input: Related information and a generated AI model.

[1943] Output: Draft of the initial response.

[1944] Specific operation: The generated prompt is sent to the GPT-3 engine to generate a primary response like this:

[1945] "If the remote control is not responding, please try the following steps..."

[1946] Step 7: Adjustment by the Emotional Engine

[1947] The generative AI server analyzes the sentiment behind the inquiry and adjusts the tone and expression of the initial response.

[1948] Input: Inquiry details and a draft of the initial response.

[1949] Output: A primary response adjusted based on emotions.

[1950] Specific actions: Use a sentiment analysis tool (e.g., IBM Watson Tone Analyzer) to analyze the sentiment from the inquiry and adjust the tone of the response based on the results.

[1951] Step 8: Present your initial answer

[1952] The terminal displays a draft of the initial response that has been generated to the user.

[1953] Input: Draft of the generated initial response.

[1954] Output: The initial response displayed to the user.

[1955] Specific operation: The generated response text will be displayed in the user interface of the web page, allowing the user to review and edit it.

[1956] Step 9: User modification

[1957] The user reviews the initial response and enters any necessary corrections or additional information.

[1958] Input: First response.

[1959] Output: Corrected or added information.

[1960] Specific action: The user edits their response on the web form and clicks the "Final Confirmation" or "Submit Revised" button.

[1961] Step 10: Submit your final response

[1962] The user submits their revised initial response as the final response.

[1963] Input: Revised initial response.

[1964] Output: Final answer.

[1965] Specific operation: After the user makes a final confirmation, they press the submit button, and the content is sent again to the generation AI server.

[1966] Step 11: Saving to the database

[1967] The server saves the final answer to the database.

[1968] Input: Final answer.

[1969] Output: Saved final answer.

[1970] Specific operation: Convert the final answer to JSON format and insert it into the answer table in the database.

[1971] (Application Example 2)

[1972] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1973] Traditional customer support systems have the challenge of not being able to provide appropriate answers to inquiries quickly. Furthermore, they fail to adequately improve customer satisfaction because they do not take customer emotions into consideration. This leads to increased workload for customer support and inconsistent service quality.

[1974] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1975] In this invention, the server includes means for collecting data from internal electronic communication means and internal networks; means for filtering the collected data and extracting useful information; means for storing the extracted information in a database; means for searching the database based on the content of the inquiry and obtaining relevant information; means for generating a draft of an initial response using a generative artificial intelligence model; and means for analyzing the user's emotions and adjusting the tone of the initial response. This makes it possible to quickly provide appropriate responses that respond to the user's emotions and improve the efficiency and quality of customer support.

[1976] "Internal electronic communication methods" refer to communication methods used within the company, such as email and chat systems.

[1977] An "internal network" refers to an internal network used within a company or organization, including intranets and dedicated networks.

[1978] "Means of data collection" refers to the processes and mechanisms for obtaining necessary information from electronic communication methods or internal networks.

[1979] "Filtering" is the process of selecting useful information from collected data and eliminating unnecessary data.

[1980] "Useful information" refers to important data that is helpful in handling inquiries and performing tasks, and includes manual information and past inquiry records.

[1981] A "database" is a data storage system that systematically stores collected information, making it easy to search and manage.

[1982] A "generative artificial intelligence model" is an AI model that can generate natural language based on large amounts of data, and models like GPT-3 fall into this category.

[1983] A "draft initial response" is the first proposed answer to an inquiry generated by an AI model.

[1984] "Means of presenting to the user" refers to methods or devices for displaying the generated draft of the initial response so that the user can review it.

[1985] "Methods for analyzing emotions" refer to technologies that identify emotions from user inquiries and adjust the tone and content of responses accordingly.

[1986] "Methods for adjusting tone" refer to the process of appropriately modifying the expression and nuances of the generated response based on the analyzed emotions.

[1987] This invention relates to a system for streamlining customer support operations in physical stores. The system collects data from internal electronic communication methods and networks, filters it, and extracts useful information. The extracted information is stored in a database and searched based on the inquiry content. Using a generative artificial intelligence model, a draft of the initial response is generated based on the acquired information, and the tone of the response is adjusted by analyzing the user's emotions using an emotion engine.

[1988] Description of the system's program processing

[1989] The server periodically or at designated times accesses the company's electronic communication methods (email servers and chat systems) and internal network (intranet) to retrieve new emails and updates. This data is filtered, and useful information is converted into structured data formats (e.g., JSON, XML) and stored in the database.

[1990] The terminal receives user inquiries through an input form and sends the inquiry details to the server. The server searches its database and retrieves relevant information. It then uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial response.

[1991] The generated response uses an emotion engine to analyze the sentiment of the inquiry. For example, it uses TextBlob to identify the sentiment (positive, negative, neutral, etc.) of the text and adjusts the tone and content of the response accordingly. The adjusted initial response draft is then presented to the user via the device.

[1992] The user can review this initial draft response and make corrections or add information as needed. Once the final response is decided, the server saves it to the database for later reference.

[1993] Specific example

[1994] For example, consider a user's inquiry about a new TV remote not working. The user might input, "My new TV remote isn't working, what should I do?" The terminal receives this inquiry and sends it to the server. The server searches its database and retrieves information on how to deal with a non-responsive remote. Then, using a generative artificial intelligence model, it generates a draft initial response like this:

[1995] If the remote control is not responding, please try the following steps:

[1996] 1. Check the batteries in the remote control and replace them with new ones.

[1997] 2. Make sure the remote control is pointed towards the TV.

[1998] 3. Check that there are no obstacles between the TV and the remote control.

[1999] 4. Clean the light-receiving part of the television.

[2000] If these steps do not resolve the issue, please contact our support center.

[2001] Furthermore, using an emotion engine, the system reads the user's feelings of dissatisfaction from their inquiry and adjusts the tone accordingly:

[2002] "We apologize for the inconvenience caused by the unresponsive remote control. Could you please try the following steps?"

[2003] 1. Check the batteries in the remote control and replace them with new ones.

[2004] 2. Make sure the remote control is pointed towards the TV.

[2005] 3. Check that there are no obstacles between the TV and the remote control.

[2006] 4. Clean the light-receiving part of the television.

[2007] If these steps do not resolve the issue, please contact our support center for further assistance.

[2008] The terminal displays this draft to the user, who then enters additional information such as, "By the way, what type of battery should I use when replacing the battery?" Once the revisions are complete, the user submits the final answer, and the server saves that answer to the database.

[2009] This system enables in-store staff to handle customer interactions efficiently and effectively, thereby improving customer satisfaction.

[2010] Example of a prompt

[2011] "My new TV remote isn't working. What should I do?"

[2012] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[2013] Step 1:

[2014] The server periodically accesses the company's email server and intranet server to retrieve new emails and intranet updates. Its input consists of unread messages and new information retrieved from email and the intranet; this data is filtered to extract only the useful information. The output is this filtered, useful information.

[2015] Step 2:

[2016] The server converts the filtered data into a structured data format (e.g., JSON, XML). This process organizes the raw input data into an appropriate format so that it can be stored in a unified database. The output is structured data that can be stored in the database.

[2017] Step 3:

[2018] A database stores structured data and manages it so that it can be quickly accessed when needed. This step involves taking structured data as input and performing database operations to efficiently store it. The output is the storage of the most up-to-date data, always accessible.

[2019] Step 4:

[2020] The terminal receives user inquiries through an input form. In this step, the user's inquiry content is received as input and sent to the server. The inquiry content is then returned to the server as output.

[2021] Step 5:

[2022] The server searches the database for information related to the query and retrieves the relevant information. The input is the user's query, and the database is searched based on that content. The output is the information related to the query.

[2023] Step 6:

[2024] The server uses a generative artificial intelligence model (e.g., GPT-3) to generate a draft of the initial answer based on the acquired relevant information. In this step, relevant information is taken as input, and the generative artificial intelligence model generates the answer based on it. The output is a draft of the initial answer.

[2025] Step 7:

[2026] The server uses an emotion engine to adjust the tone of the initial response draft based on the emotion of the inquiry. In this step, the emotion contained in the user's inquiry is taken as input, and the emotion engine analyzes it to adjust the tone. The output is the adjusted initial response draft.

[2027] Step 8:

[2028] The terminal presents the user with a draft of the adjusted initial response. In this step, the input is the adjusted initial response, which is displayed to the user. As output, the user can review the draft of the initial response.

[2029] Step 9:

[2030] The user reviews the initial draft response and enters any necessary corrections or additional information. This step includes the user's additional information and corrections as input, and the response is ready to be submitted as the final response. The output is the final response, corrected by the user.

[2031] Step 10:

[2032] The server saves the user's modified final answer to a database for future reference. In this step, the input is the final answer, which is stored in the database. The output is the saved final answer, which can be referenced later.

[2033] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[2034] The data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One 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">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[2035] 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 this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[2036] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[2037] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[2038] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[2039] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[2040] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[2041] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[2042] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[2043] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[2044] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[2045] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[2046] 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.

[2047] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[2048] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[2049] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[2050] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[2051] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[2052] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[2053] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[2054] The following is further disclosed regarding the embodiments described above.

[2055] (Claim 1)

[2056] Means of collecting data from internal email and intranet,

[2057] A means of filtering collected data and extracting useful information,

[2058] A means of storing the extracted information in a database,

[2059] A means of searching a database based on the query content and obtaining relevant information,

[2060] A means of generating a draft of an initial response using a generative artificial intelligence model based on acquired information,

[2061] A means of presenting the generated initial draft to the user,

[2062] A means of submitting a user-corrected initial response as the final response,

[2063] A system that includes this.

[2064] (Claim 2)

[2065] The system according to claim 1, wherein data collection is performed at a specific time or periodically.

[2066] (Claim 3)

[2067] The system according to claim 1, wherein the generated initial draft response is modifiable according to the user's request.

[2068] (Claim 4)

[2069] The system according to claim 1, further comprising means for recording the final response modified by the user and storing it in a database.

[2070] "Example 1"

[2071] (Claim 1)

[2072] Means for collecting information from internal communication data and the internal network,

[2073] A means of analyzing the collected information and extracting useful data,

[2074] Means for storing the extracted data in a storage device,

[2075] A means of searching a storage device based on the inquiry content and obtaining related data,

[2076] A means of generating a draft of an initial response using a generative artificial intelligence model based on acquired data,

[2077] A means of presenting the generated initial draft to the user,

[2078] A means of submitting a user-corrected initial response as the final response,

[2079] A system that includes this.

[2080] (Claim 2)

[2081] The system according to claim 1, in which information is collected at a specific time or periodically.

[2082] (Claim 3)

[2083] The system according to claim 1, wherein the generated initial draft response is modifiable according to the user's request.

[2084] "Application Example 1"

[2085] (Claim 1)

[2086] Means of collecting data from internal email and intranet,

[2087] A means of filtering collected data and extracting useful information,

[2088] A means of storing the extracted information in a database,

[2089] A means of searching a database based on the query content and obtaining relevant information,

[2090] A means of generating a draft of an initial response using a generative artificial intelligence model based on acquired information,

[2091] A means of presenting the generated initial draft to the user,

[2092] A means of submitting a user-corrected initial response as the final response,

[2093] A means of entering inquiry details using a smart device,

[2094] A means of displaying a default message when no relevant information is found,

[2095] A means of accepting user corrections,

[2096] A means of saving user input and modifications to a database,

[2097] A system that includes this.

[2098] (Claim 2)

[2099] The system according to claim 1, wherein data collection is performed at a specific time or periodically.

[2100] (Claim 3)

[2101] The sy...

Claims

1. Means of collecting data from internal email and intranet, A means of filtering collected data and extracting useful information, A means of storing the extracted information in a database, A means of searching a database based on the query content and obtaining relevant information, A means of generating a draft of an initial response using a generative artificial intelligence model based on acquired information, A means of presenting the generated initial draft to the user, A means of submitting a user-corrected initial response as the final response, A system that includes this.

2. The system according to claim 1, wherein data collection is performed at a specific time or periodically.

3. The system according to claim 1, wherein the generated draft of the initial response can be modified according to the user's request.

4. The system according to claim 1, further comprising means for recording the user's final corrected answer and storing it in a database.

Images