system

Abstract

We provide the system. [Solution] A means of receiving questions entered by the user, A means of inputting a received question into a generative model and generating an appropriate answer, A means of sending the generated response to the user, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Current customer support systems have the problem of lacking a quick and efficient response to customer questions and problems. In conventional systems, there are limitations in human resources, response time, and variations in support quality, which lead to a decrease in customer satisfaction and inefficiency in support operations. Therefore, there is a need for a system that can provide consistent high-quality support especially during non-business hours and holidays.

Means for Solving the Problems

[0005] The present invention solves the above problems by providing a system that includes means for receiving questions entered by a user, means for inputting the received questions into a generative model and generating appropriate answers, and means for sending the generated answers to the user. With this system, users can receive 24-hour support and are provided with consistent, high-quality answers by utilizing a generative AI model. In addition, the generative AI model can handle a wide range of topics, enabling quick and efficient responses to diverse user questions and problems. This makes it possible to optimize human resources and improve customer satisfaction.

[0006] A "user" refers to an individual or organization that uses the system to input questions or problems.

[0007] A "question" refers to text data of doubts or inquiries that a user enters and sends to the system.

[0008] "Means of receiving" refers to an interface or module that allows the system to acquire and process questions entered by the user.

[0009] A "generative model" refers to an algorithm or software that uses natural language processing technology to analyze input text data and generate appropriate responses.

[0010] "Answer" refers to the text data of the answer that the generative model generates in response to the user's question.

[0011] "Means of transmission" refers to an interface or module for communicating the generated response to the user.

[0012] A "system" refers to a set of processes and components that combine these means to provide answers to user questions. [Brief explanation of the drawing]

[0013] [Figure 1]This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14] This is a sequence diagram showing the processing flow of the data processing system in Application Example 2, which combines an emotion engine. [Modes for carrying out the invention]

[0014] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0015] First, the terms used in the following description will be explained.

[0016] 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 a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), etc.

[0017] 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.

[0018] In the following embodiments, the numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0019] In the following embodiments, the numbered communication I / F (Interface) is an interface including a communication processor and an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark), etc.

[0020] 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."

[0021] [First Embodiment]

[0022] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0023] 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.

[0024] 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).

[0025] 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.

[0026] 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.

[0027] 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.

[0028] 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.

[0029] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] 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.

[0031] 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.

[0032] 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.

[0033] 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".

[0034] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0035] Program Processing Overview

[0036] 1. User input and submission of questions

[0037] The user types their question into the customer support chat window and presses the send button. This action causes the user's device to retrieve the entered text data and prepare to send it to the server.

[0038] 2. Server receives and analyzes the question.

[0039] The server receives text data sent from the terminal. The server then performs preprocessing on the received text data to input it into the generative model. This preprocessing includes text data cleansing and format conversion.

[0040] 3. Generating answers using generative models

[0041] A generative model (e.g., GPT-3) receives pre-processed text data, performs analysis, and generates an appropriate response. The generated response is then returned to the server.

[0042] 4. The server sends the response.

[0043] The server receives the generated response, reformats or corrects it if necessary, and then sends it to the terminal.

[0044] 5. Displaying responses via the device

[0045] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[0046] Specific example

[0047] Example 1: Questions about the product return policy

[0048] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[0049] The device retrieves this text data and sends it to the server.

[0050] The server receives text data, inputs it into a generative model, and performs analysis.

[0051] The generative model generates the response, "Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[0052] The server receives the generated response, formats it, and sends it to the terminal.

[0053] The device displays the answer in the user's chat window.

[0054] Users can review the displayed answers and ask further questions if necessary.

[0055] This system allows users to receive fast and consistent support 24 / 7. The use of generative AI models enables support across a wide range of topics, leading to consistent support quality and improved customer satisfaction. This, in turn, enables the efficient use of human resources in customer support.

[0056] The following describes the processing flow.

[0057] Step 1:

[0058] The user types their question into the customer support chat window and presses the submit button.

[0059] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[0060] Step 2:

[0061] The terminal retrieves the user's input and formats that text data.

[0062] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[0063] Step 3:

[0064] The terminal sends formatted text data to the server.

[0065] Specific operation: Generate an HTTP POST request and send the formatted text to the server's API endpoint.

[0066] Step 4:

[0067] The server receives text data from the terminal.

[0068] Specific operation: The server's API receives an HTTP POST request and extracts text data as input for analysis.

[0069] Step 5:

[0070] The server preprocesses the received text data before inputting it into the generative model.

[0071] Specific operations: Performs text data cleansing (removes unnecessary characters and spaces) and tokenization (breaks down words and sentences into tokens).

[0072] Step 6:

[0073] The generative model receives pre-processed text data and generates an appropriate response.

[0074] Specific operation: A generative AI model (e.g., GPT-3) analyzes the text data and generates the response text, "Our return policy allows returns within 30 days of receiving the product."

[0075] Step 7:

[0076] The server receives the generated response and formats or corrects it as needed.

[0077] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[0078] Step 8:

[0079] The server sends the formatted response to the terminal.

[0080] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[0081] Step 9:

[0082] The terminal displays the response received from the server in the user's chat window.

[0083] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[0084] Step 10:

[0085] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[0086] Specific actions: Read the response text, enter a new question if necessary, and submit again.

[0087] As a result, users will be able to receive customer support quickly and efficiently. This system is particularly capable of 24 / 7 support and provides consistently high-quality support. This leads to improved customer satisfaction and optimized human resources.

[0088] (Example 1)

[0089] 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."

[0090] Traditional customer support systems have a problem in that they struggle to provide quick and consistent answers to user inquiries. Furthermore, providing 24 / 7 support requires significant human resources, resulting in high operational costs. Additionally, the inconsistency in support quality due to different personnel handling inquiries is another challenge. It is necessary to improve this situation and achieve efficient and high-quality customer support.

[0091] 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.

[0092] In this invention, the server includes means for receiving questions entered by the user, means for cleansing the received text data and converting it into a prompt sentence, means for inputting the pre-processed text data into a generative model and generating an appropriate answer, and means for formatting the generated answer and sending it to the user. This makes it possible to provide rapid and consistent answers to user inquiries 24 hours a day. Furthermore, by using a generative AI model, it becomes possible to handle a wide range of topics, and it is expected that the uniformity of support quality and the improvement of customer satisfaction can be expected.

[0093] A "user" refers to anyone who uses this system to contact customer support.

[0094] A "question" is text data containing the content of a user's inquiry to customer support.

[0095] "Receiving" refers to the act of a server acquiring data sent from a terminal.

[0096] "Cleansing" refers to the process of removing unnecessary spaces and symbols from text data and standardizing the data format.

[0097] A "prompt statement" is a query statement generated to be input into a generative model in order to obtain a response.

[0098] A "generative model" refers to an artificial intelligence algorithm that analyzes input text data and automatically generates appropriate responses.

[0099] "Formatting" refers to the process of adding line breaks and symbols, or standardizing the writing style, to make the generated response text easier to read.

[0100] "Sending" refers to the act of the server sending back the response obtained from the generative model to the terminal.

[0101] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0102] The user enters a question into a customer support chat window and presses the submit button. At this point, the device retrieves the entered text data and prepares to send it to the server. Specifically, the device converts the entered data into JSON format and sends it to the server using an HTTP request. Hardware used includes the user's device (e.g., a PC or smartphone). Software used includes a web browser or mobile app that provides the chat window.

[0103] The server receives HTTP requests sent from terminals and extracts the text data portion. At this time, the server performs preprocessing on the received text data. This preprocessing includes text data cleansing, such as removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. The preprocessed data is then converted into prompt statements to be passed to the generating AI model. The server contains a program for receiving HTTP requests and processing text data.

[0104] The server sends the generated prompt to a generation AI model (for example, OpenAI's GPT-3). The generation model analyzes the prompt and generates an appropriate response. This generated response text is returned to the server. Specific examples of prompts include the following sentences:

[0105] "A user is asking, 'What is the return policy for this product?' Please provide an appropriate answer."

[0106] The server receives the response returned from the generative model and performs formatting if necessary, such as adding line breaks, correcting symbols, and unifying the writing style. The formatted response text is then converted back to JSON format and sent to the terminal as an HTTP response. The server includes a program to appropriately format and resend the output from the generative model.

[0107] The terminal parses the HTTP response received from the server and retrieves the response text. It can then display this retrieved response text in the user's chat window. The user can review the displayed response and ask additional questions as needed.

[0108] Specific examples are given below.

[0109] Example: Questions about the product return policy

[0110] 1. The user types "Please tell me about the return policy for this product" in the chat window.

[0111] 2. The terminal converts the entered text data into JSON format and sends an HTTP request to the server.

[0112] 3. The server receives the HTTP request, cleanses the text data, and generates a prompt message.

[0113] 4. The server sends the generated prompt message to the generation model.

[0114] 5. The generative model generates the response "Our return policy allows returns within 30 days of receiving the product" and returns it to the server.

[0115] 6. The server formats the generated response and sends it to the terminal.

[0116] 7. The device receives the response and displays it in the user's chat window.

[0117] 8. The user reviews the displayed answers. If necessary, they can enter additional questions and resubmit.

[0118] This embodiment improves the quality of customer support and enables the provision of fast and consistent service. Furthermore, by utilizing generative AI models, it becomes possible to address a wide range of topics, leading to improved customer satisfaction and more efficient use of human resources.

[0119] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0120] Step 1:

[0121] User input and submission of questions

[0122] The user enters a question into the customer support chat window and presses the submit button. The user enters the question, "Please tell me about the return policy for this product." The terminal receives this input text. Since the input is the user's text data, the terminal converts it to JSON format and sends it to the server in the format "{"question": "Please tell me about the return policy for this product"}". Specifically, it sends the data to the server using an HTTP request.

[0123] Step 2:

[0124] Server receives and analyzes questions.

[0125] The server receives an HTTP request sent from the terminal. The received data is input in JSON format, namely "{"question": "Please tell me about the return policy for this product"}". The server parses this data and extracts the text "Please tell me about the return policy for this product". After extraction, the server cleanses this text data. Specifically, this involves removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. After processing, the pre-processed text data "Please tell me about the return policy for this product" is obtained.

[0126] Step 3:

[0127] Server-driven generation of prompt messages

[0128] Based on the pre-processed text data, "Please tell me about the return policy for this product," the server generates a prompt. This prompt is then input into the generative model. Specifically, it is generated in the format, "The user is asking, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." This prompt becomes the input for the next step.

[0129] Step 4:

[0130] Generating answers using generative models

[0131] The server sends the generated prompt to a generating AI model (e.g., GPT-3). The input is the generated prompt "The user asks, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." The generating AI model analyzes this prompt and generates an appropriate answer: "Our return policy allows returns within 30 days of receiving the product." This generated answer is returned to the server, which then receives it.

[0132] Step 5:

[0133] Server formatting and sending of responses

[0134] The server receives the response "Our return policy allows returns within 30 days of receiving the product." from the generative model. The input is the response text from the generative model. If necessary, formatting is performed on the response, such as line breaks, symbol corrections, and consistency of writing style. The formatted response text is converted back into JSON format and sent to the terminal in the format "{"answer": "Our return policy allows returns within 30 days of receiving the product." The server sends the data to the terminal using an HTTP response.

[0135] Step 6:

[0136] Displaying responses via device

[0137] The terminal parses the HTTP response received from the server and obtains the response text: "Our return policy allows returns within 30 days of receiving the product." The input is the response data from the server. The terminal displays this response text in the user's chat window. The user can review the displayed response and ask additional questions if necessary. In this case, the process is repeated from step 1.

[0138] (Application Example 1)

[0139] 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."

[0140] In e-commerce sites, responding quickly and accurately to a wide variety of user inquiries is crucial for maintaining customer satisfaction. Traditionally, customer support has relied on human resources, making 24 / 7 support and immediate responses difficult. This increased the burden on departments and could lead to inconsistent support quality. This invention aims to solve these problems and achieve efficient and consistent support by using a generative AI model.

[0141] 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.

[0142] In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generative model and generating an appropriate answer, means for sending the generated answer to the user, means for cleansing the question as text data and performing format conversion, and means for formatting the generated answer and displaying it on the user's terminal. This enables a quick and accurate response to user questions, realizing 24-hour support and consistent response quality.

[0143] A "user" refers to someone who operates the system, inputs a question, and obtains an answer.

[0144] A "question" refers to text data entered by a user and analyzed by a generative model.

[0145] A "generative model" refers to AI technology that generates responses based on input text data.

[0146] "Cleansing" refers to the process of removing noise from text data and preparing it in a format that is easy to analyze.

[0147] "Format conversion" refers to the process of converting text data into a format that a generative model can understand.

[0148] "Formatting" refers to the process of converting the answers generated by a generative model into a format that is easy for the user to read.

[0149] A "terminal" refers to a device used by a user to access the system, enter questions, and check answers.

[0150] A "server" refers to a computer system that receives questions from users, generates answers using a generative model, and sends them back to the user.

[0151] A "chat interface" refers to the user interface of a system that allows users to input questions and view generated answers.

[0152] This invention provides a system for automatically answering user questions in customer support on e-commerce websites. The system works as follows: the user inputs a question, an AI model analyzes it to generate a response, and then sends the response back to the user. The detailed configuration and processing are described below.

[0153] The server includes means for receiving questions entered by the user, means for inputting the received questions into a generative model and generating appropriate answers, means for sending the generated answers to the user, means for cleansing the questions as text data and performing format conversion, and means for formatting the generated answers and displaying them on the user's terminal.

[0154] The system will utilize smartphones and servers as hardware. Users will operate the system via an application on their smartphones. The software used will include a generative AI model (OpenAI's GPT-3), a text analysis tool (NLTK Natural Language Processing Toolkit), and an application server (Node.js).

[0155] When the server receives text data sent by the user, it first cleanses the text data. Specifically, it uses NLTK to remove unnecessary spaces and symbols. Next, it formats the cleaned text data, converting it into a format that the generative AI model can understand. This is done by receiving the text data as plain text and generating appropriate prompt sentences. The generative model analyzes these prompt sentences and generates an appropriate response. The generated response is returned to the server and formatted as needed. The formatted response is sent to the user's smartphone and displayed in the chat window.

[0156] As a concrete example, a user enters the question "Please tell me about the return policy for this product" and sends it to the server from their smartphone. An example of the prompt text is shown below.

[0157] "Please provide the best answer to the following question: What is the return policy for this product?"

[0158] The server cleanses and formats this text data before inputting it into a generating AI model. This results in the output: "Our return policy allows returns within 30 days of receiving the product." This output is then formatted and displayed on the user's smartphone. This allows the user to receive quick and accurate responses and access 24 / 7 support.

[0159] In this way, the present invention makes it possible to improve the efficiency and quality of customer support on e-commerce sites.

[0160] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0161] Step 1:

[0162] The user enters a question.

[0163] The user opens a chat application on their smartphone, enters their question in text format, and presses the send button. The entered question is saved on the device as text data.

[0164] Step 2:

[0165] The terminal sends the question to the server.

[0166] The user's device sends the entered text data to the server as an HTTP request. The input data includes the user's question.

[0167] Step 3:

[0168] The server receives the question.

[0169] The server receives the HTTP request sent from the terminal and retrieves the question as text data. The received data is stored in a variable for subsequent processing.

[0170] Step 4:

[0171] The server cleanses the text data.

[0172] The server uses NLTK to cleanse the received text data, removing unnecessary spaces and symbols. The cleansed text data is then output.

[0173] Step 5:

[0174] The server converts the text data to a different format.

[0175] The cleansed text data is converted into a format that can be input into the AI ​​model. The server generates a prompt in plain text format. This prompt is as follows: "Provide the best answer to the following question: [cleansed text]".

[0176] Step 6:

[0177] The server inputs prompt messages into the AI ​​model and generates responses.

[0178] The server sends the generated prompt text as input to the generative AI model. The generative AI model (GPT-3) analyzes the prompt text and generates the appropriate response text. The generated response text is obtained as output.

[0179] Step 7:

[0180] The server formats the generated response.

[0181] The server formats the response text obtained from the generating AI model into a user-friendly format. It adds necessary line breaks and highlighting to match the specific format. The formatted text data is then output.

[0182] Step 8:

[0183] The server sends the formatted response to the user's device.

[0184] The server sends the formatted response text data as an HTTP response to the user's device. The device receives this data.

[0185] Step 9:

[0186] The device will display the answer.

[0187] The user's device displays the received response text data in the chat window. The user can review the displayed response and enter further questions if necessary.

[0188] 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.

[0189] This invention relates to a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0190] Program Processing Overview

[0191] 1. User input and submission of questions

[0192] The user types their question into the customer support chat window and presses the send button. At this point, the device retrieves the entered text data and prepares to send it to the server.

[0193] 2. Emotion recognition by the emotion engine

[0194] After the terminal receives the question entered by the user, the emotion engine analyzes the text data and recognizes the emotion. The recognized emotion information is sent to the server along with the text data.

[0195] 3. Receiving questions and sentiments from the server

[0196] The server receives text data and sentiment information sent from the terminal. The server then performs preprocessing on the received text data and sentiment information to input it into the generative model. This preprocessing includes formatting the sentiment information and cleaning the text data.

[0197] 4. Generating answers using generative models

[0198] The generative model receives pre-processed text data and sentiment information, performs analysis, and generates appropriate responses. The generated responses are expressed in a way that takes the user's emotions into consideration, both in content and tone.

[0199] 5. Server sends response

[0200] The server receives the generated response, formats and corrects it as needed, and then sends it to the terminal.

[0201] 6. Displaying responses via the device

[0202] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[0203] Specific example

[0204] Example 1: Questions about the product return policy

[0205] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[0206] The device retrieves this text data, and the emotion engine recognizes the emotion that "the user appears anxious."

[0207] The device sends text data and sentiment information to the server.

[0208] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[0209] The generative model generates the response, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[0210] The server receives the generated response, formats it, and sends it to the terminal.

[0211] The device displays the answer in the user's chat window.

[0212] The user checks the displayed answer and feels reassured.

[0213] Example 2: A user's question expressing dissatisfaction with support.

[0214] The user types "Why is the support so slow?" and submits it.

[0215] The device acquires text data, and the emotion engine recognizes "anger."

[0216] The device sends text data and information about the emotion of anger to the server.

[0217] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[0218] The generative model generates the response, "We apologize for the inconvenience. We are currently working to resolve this as quickly as possible," and returns it to the server.

[0219] The server receives the generated response, formats it, and sends it to the terminal.

[0220] The device displays the answer in the user's chat window.

[0221] The user reviews the displayed answer and is somewhat convinced.

[0222] This system allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[0223] The following describes the processing flow.

[0224] Step 1:

[0225] The user types their question into the customer support chat window and presses the submit button.

[0226] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[0227] Step 2:

[0228] The terminal retrieves the user's input and formats that text data.

[0229] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[0230] Step 3:

[0231] The device sends formatted text data to the emotion engine, which then recognizes the user's emotions.

[0232] Specific operation: Text data is passed to the emotion engine, which then reads emotions such as "anxiety" from the text.

[0233] Step 4:

[0234] The emotion engine returns the recognized emotion information to the device.

[0235] Specific operation: The emotion engine performs text analysis and returns the identified emotion information to the device.

[0236] Step 5:

[0237] The device sends text data and sentiment information to the server.

[0238] Specific operation: Generate an HTTP POST request containing text data and sentiment information, and send it to the server's API endpoint.

[0239] Step 6:

[0240] The server receives text data and sentiment information from the terminal.

[0241] Specific operation: The server's API receives an HTTP POST request and extracts text data and sentiment information as input for analysis.

[0242] Step 7:

[0243] The server preprocesses the received text data and sentiment information.

[0244] Specific operations: Cleans text data (removes unnecessary characters and spaces), tokenizes (breaks down words and sentences into tokens), and converts the format of sentiment information.

[0245] Step 8:

[0246] The generative model receives pre-processed text data and sentiment information and generates an appropriate response.

[0247] Specific operation: A generative AI model (e.g., GPT-3) analyzes the input data and generates response text that takes sentiment into account.

[0248] Step 9:

[0249] The server receives the generated response and formats or corrects it as needed.

[0250] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[0251] Step 10:

[0252] The server sends the formatted response to the terminal.

[0253] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[0254] Step 11:

[0255] The terminal displays the response received from the server in the user's chat window.

[0256] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[0257] Step 12:

[0258] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[0259] Specific operation: The user reads the displayed answer text, enters a new question if necessary, and resubmits.

[0260] This allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[0261] (Example 2)

[0262] 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".

[0263] Traditional customer support systems struggled to accurately recognize and respond to user emotions, sometimes resulting in decreased user satisfaction and delays in problem resolution. Furthermore, efficiently providing 24 / 7 support placed a significant burden on human resources.

[0264] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0265] In this invention, the server includes means for receiving a question entered by a user, means for generating emotional information from the received question using an emotional analysis engine, means for inputting the question along with the generated emotional information into a generation model to generate an appropriate answer, means for formatting the generated answer in a way that takes the user's emotions into consideration, and means for sending the formatted answer to the user. This makes it possible to provide high-quality support that takes the user's emotions into consideration 24 hours a day.

[0266] "Means for receiving user-entered questions" refers to a mechanism for acquiring text data entered by users through customer support chat windows or other interfaces and sending it to the server.

[0267] "Means for generating emotional information using an emotion analysis engine" refers to a mechanism that analyzes text data entered by a user, recognizes the user's emotional state from its content, and generates appropriate emotional information.

[0268] "A means of inputting questions along with generated sentiment information into a generative model to generate appropriate answers" refers to a mechanism for inputting recognized sentiment information and user questions into a generative model-trained generative algorithm to generate appropriate answers that take the user's emotions into consideration.

[0269] "Methods for shaping generated responses to reflect user emotions" refers to a mechanism that processes responses generated by a generative model into a format and phrasing that aligns with the user's emotions, making them easy to understand and considerate for the user.

[0270] "Means for sending formatted responses to users" refers to a mechanism for sending formatted responses from a server to the user's terminal and displaying them in a chat window or other interface.

[0271] This system combines a generative AI model and an emotion analysis engine to generate responses that take user emotions into consideration, in order to provide effective customer support. The specific configuration and operation of the system are described below.

[0272] Hardware and software to use

[0273] Hardware:

[0274] Server: Cloud service (e.g., AWS EC2)

[0275] Devices: User devices such as PCs, smartphones, and tablets.

[0276] software:

[0277] Chat window platform: Customer support system (e.g., Intercom, Zendesk)

[0278] Sentiment analysis engine: Text analysis library (e.g., Affective Computing SDK)

[0279] Generative AI models: Natural language generation models (e.g., OpenAI GPT-3)

[0280] System response processing

[0281] The server has a component that first receives the question entered by the user. The user enters the question in the chat window and presses the send button. For example, when the user enters "Please tell me about the return policy of this product", this text data is obtained by the terminal. The terminal prepares to send this text data to the server.

[0282] The server further has a component that analyzes the received question using a sentiment analysis engine and generates sentiment information. The sentiment analysis engine analyzes the text entered by the user and generates sentiment information. For example, when the emotion of "uneasiness" is recognized from the user's speech, emotion information of "uneasiness" is generated.

[0283] Next, the server inputs the question together with the generated sentiment information into a generation AI model to generate an appropriate answer. The generation model receives the preprocessed data, performs analysis, and generates an answer. This answer has content and a tone that takes into account the user's emotions. For example, an answer such as "You can rest assured because there is a refund guarantee. The return policy allows returns within 30 days after the product arrives." may be generated.

[0284] The generated answer is then formatted with content and a tone that takes into account the user's emotions. This process makes the answer more acceptable to the user. The formatted answer is sent by the server to the terminal again.

[0285] The terminal displays the received answer in the chat window. The user can check the displayed answer and determine whether the problem has been solved.

[0286] Specific operation example

[0287] The following shows a specific example of the actual system operation.

[0288] Example of a prompt sentence:

[0289] 1. "Please tell me about the return policy for this product."

[0290] 2. "Why is the support so slow?"

[0291] For example, if a user types "Please tell me about the return policy for this product," the device receives this input, and its sentiment analysis engine recognizes "anxiety." The server then inputs this text data and sentiment information into a generating AI model, which generates and formats a response such as "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and sends it back to the device. The device then displays this response in the user's chat window, allowing the user to see it and feel reassured.

[0292] In this way, the system provides high-quality customer support that takes user emotions into consideration, improving user satisfaction and streamlining support operations. Furthermore, it enables 24 / 7 support, contributing to a reduction in the burden on human resources.

[0293] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0294] Step 1:

[0295] The user types their question into the customer support chat window and presses the send button. This is then sent to the device as text data.

[0296] Input: Text data entered by the user in the chat window

[0297] Output: Text data is sent to the terminal.

[0298] Specific action: The user types "Please tell me about the return policy for this product" into the chat window and presses the send button.

[0299] Step 2:

[0300] The terminal acquires the input text data and holds it in memory. Also, it prepares to transmit the text data.

[0301] Input: Text data input by the user

[0302] Output: Text data held in the terminal

[0303] Specific operation: The terminal detects the user's input and saves the text data in temporary storage.

[0304] Step 3:

[0305] The terminal causes the text data to be analyzed by the sentiment analysis engine. From this analysis, the user's sentiment information is extracted from the text data.

[0306] Input: Text data held in the terminal

[0307] Output: Sentiment information generated based on the text data

[0308] Specific operation: The terminal's sentiment analysis engine analyzes the text data and recognizes that the user is feeling "uneasy".

[0309] Step 4:

[0310] The terminal transmits the text data and the sentiment information to the server.

[0311] Input: Text data and sentiment information

[0312] Output: Text data and sentiment information transmitted to the server

[0313] Specific operation: The terminal transmits the text data and the sentiment information to the server in packet format.

[0314] Step 5:

[0315] The server receives text data and sentiment information sent from the terminal and performs preprocessing. Preprocessing includes text data cleansing and sentiment information format conversion.

[0316] Input: Text data and sentiment information received from the device.

[0317] Output: Preprocessed text data and sentiment information

[0318] Specific operation: The server cleanses the received data and converts it to the required format.

[0319] Step 6:

[0320] The server inputs pre-processed text data and sentiment information into a generating AI model to generate responses.

[0321] Input: Preprocessed text data and sentiment information

[0322] Output: Answers generated by the generative AI model

[0323] Specific operation: The server inputs data into the generated AI model and generates the response, "Rest assured, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[0324] Step 7:

[0325] The server receives the generated responses and performs formatting and necessary corrections. This formatting ensures the consistency of the responses.

[0326] Input: Answers generated by the generative AI model

[0327] Output: Formatted answer

[0328] Specific operation: The server inspects the response and corrects grammar and tone.

[0329] Step 8:

[0330] The server sends the formatted response to the terminal.

[0331] Input: Formatted answer

[0332] Output: Formatted response sent to the terminal

[0333] Specific operation: The server sends the formatted response to the terminal.

[0334] Step 9:

[0335] The device displays the received response in the user's chat window.

[0336] Input: Formatted response received from the server

[0337] Output: Answers displayed in the chat window

[0338] Specific operation: The device displays the received response in the chat window and notifies the user.

[0339] Step 10:

[0340] The user reviews the displayed answers and asks additional questions as needed.

[0341] Input: Answer displayed in the chat window

[0342] Output: Additional questions (if necessary)

[0343] Specific action: The user reads the displayed answer and enters whether they are satisfied or ask further questions.

[0344] (Application Example 2)

[0345] 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".

[0346] Traditional customer support systems often struggle to respond to user emotions and provide only uniform answers, which can lead to decreased customer satisfaction. Furthermore, in physical stores, situations may arise where staff are unavailable, resulting in customers not receiving prompt and appropriate support.

[0347] 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. In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generation model and generating an appropriate answer, means for sending the generated answer to the user, means for recognizing the user's emotions using an emotion engine, means for inputting the recognized emotion information into the generation model and generating an answer, and means for displaying or playing the generated answer in voice or text. This enables the provision of answers that correspond to the user's emotions, and allows for quick and appropriate support even in physical stores.

[0348] A "user" refers to a person who enters questions or information into a system.

[0349] "Means for receiving questions" refers to a part of the system that retrieves and processes questions entered by users.

[0350] A "generative model" refers to an algorithm used to generate appropriate answers based on received questions and data.

[0351] "Means for generating answers" refers to a function that uses a generative model to generate answers to received questions.

[0352] An "emotion engine" refers to an algorithm that analyzes and recognizes emotions from user input data.

[0353] "Means of recognizing emotions" refers to a function that uses an emotion engine to identify the user's emotions.

[0354] "Emotional information" refers to information about emotions recognized from user input data.

[0355] "Means of submitting responses" refers to a part of the system used to communicate generated responses to users.

[0356] "Means of displaying or playing" refers to a function that provides the generated response to the user visually or audibly in the form of audio or text.

[0357] This invention is a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions, enabling fast and emotionally responsive customer service in physical stores. The system components include a server, a terminal (smart device), an emotion engine, a generative AI model, and communication means.

[0358] System Configuration

[0359] 1. Voice input processing

[0360] Hardware: Smart glasses

[0361] Example: The general term is "smart glasses device".

[0362] Software: Speech Recognition API

[0363] Example: "Speech-to-text conversion service" as a general term.

[0364] Process: The user wears smart glasses and inputs a question by voice. The smart glasses convert the voice input into text data.

[0365] 2. Emotion Recognition Engine Processing

[0366] Software: Natural Language Processing Library

[0367] Example: "Natural language processing engine" as a general term

[0368] Processing: Text data converted from speech is input into the emotion recognition engine to identify the customer's emotion. For example, if a customer asks, "What are the features of this product?", the emotion engine identifies "interest." This allows the customer's emotion information to be input into the generative model.

[0369] 3. Data transmission process

[0370] Software: HTTP communication library

[0371] Example: "Communication Library" as a general term

[0372] Processing: Text data and sentiment information are sent to the server. The server receives this data and inputs it into a generative model to generate a response.

[0373] 4. Generative Model Processing

[0374] Hardware: Server environment

[0375] Example: "Cloud Server" as a general term

[0376] Software: Generative AI Model

[0377] Example: "Generative algorithm" as a general term

[0378] Processing: The server inputs the received text data and sentiment information into a generative model to generate an appropriate response. The generated response is expressed in a tone and content that is sensitive to the customer's emotions.

[0379] 5. Sending the response

[0380] Software: HTTP communication library

[0381] Example: "Communication Library" as a general term

[0382] Processing: The generated response is sent back to the smart glasses.

[0383] 6. Displaying the answer

[0384] Hardware: Smart glasses

[0385] Example: "Smart device" as a general term

[0386] Software: Display Management Library

[0387] Example: "Display management software" as a general term

[0388] Processing: The response is displayed as audio or text on the smart glasses worn by the customer, allowing the user to confirm it.

[0389] Specific example

[0390] In a physical store, if a customer asks a question to smart glasses by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the customer's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[0391] Example of a prompt

[0392] A customer asked the following question: "What are the features of this product?" Emotion: Interest

[0393] The generating AI model should consider the customer's interest in this question and respond as follows: "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design."

[0394] This form allows for a concrete understanding of how the invention is implemented and provides a detailed process that can be reproduced by anyone.

[0395] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0396] Step 1:

[0397] The user wears smart glasses and inputs questions by voice. The input voice data is captured through the microphone of the smart device.

[0398] Step 2:

[0399] The device sends the acquired audio data to a speech recognition API. The speech recognition API converts the audio data into text data. The input is audio data, and the output is text data.

[0400] Step 3:

[0401] The device inputs the converted text data into the emotion recognition engine. The emotion recognition engine analyzes the text data and identifies the user's emotions. The input is text data, and the output is emotion information.

[0402] Step 4:

[0403] The terminal sends text data and recognized sentiment information to the server via a communication library. The input is text data and sentiment information, and the output is the data sent to the server.

[0404] Step 5:

[0405] The server preprocesses the received text data and sentiment information. This includes data cleaning and format conversion to prepare the data for input into the generative model. The input is the received data, and the output is the preprocessed data.

[0406] Step 6:

[0407] The server inputs pre-processed data into a generative AI model, which then generates appropriate responses. The generative model analyzes the data and generates responses that take the user's emotions into consideration. The input is pre-processed data, and the output is the generated response.

[0408] Step 7:

[0409] The server receives the generated response and performs further formatting and editing as needed. The formatted response is then sent back to the terminal via the communication library. The input is the generated response, and the output is the formatted response.

[0410] Step 8:

[0411] The device displays the received response to the user using the smart glasses' display management library. Audio output may also be provided. The input is a formatted response, and the output is presented to the user as audio or text.

[0412] Specific example

[0413] For example, if a user asks the smart glasses a question by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the user's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[0414] 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.

[0415] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0416] 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.

[0417] [Second Embodiment]

[0418] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0419] 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.

[0420] 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).

[0421] 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.

[0422] 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.

[0423] 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).

[0424] 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.

[0425] 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.

[0426] 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.

[0427] 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.

[0428] 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.

[0429] 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".

[0430] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0431] Program Processing Overview

[0432] 1. User input and submission of questions

[0433] The user types their question into the customer support chat window and presses the send button. This action causes the user's device to retrieve the entered text data and prepare to send it to the server.

[0434] 2. Server receives and analyzes the question.

[0435] The server receives text data sent from the terminal. The server then performs preprocessing on the received text data to input it into the generative model. This preprocessing includes text data cleansing and format conversion.

[0436] 3. Generating answers using generative models

[0437] A generative model (e.g., GPT-3) receives pre-processed text data, performs analysis, and generates an appropriate response. The generated response is then returned to the server.

[0438] 4. The server sends the response.

[0439] The server receives the generated response, reformats or corrects it if necessary, and then sends it to the terminal.

[0440] 5. Displaying responses via the device

[0441] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[0442] Specific example

[0443] Example 1: Questions about the product return policy

[0444] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[0445] The device retrieves this text data and sends it to the server.

[0446] The server receives text data, inputs it into a generative model, and performs analysis.

[0447] The generative model generates the response, "Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[0448] The server receives the generated response, formats it, and sends it to the terminal.

[0449] The device displays the answer in the user's chat window.

[0450] Users can review the displayed answers and ask further questions if necessary.

[0451] This system allows users to receive fast and consistent support 24 / 7. The use of generative AI models enables support across a wide range of topics, leading to consistent support quality and improved customer satisfaction. This, in turn, enables the efficient use of human resources in customer support.

[0452] The following describes the processing flow.

[0453] Step 1:

[0454] The user types their question into the customer support chat window and presses the submit button.

[0455] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[0456] Step 2:

[0457] The terminal retrieves the user's input and formats that text data.

[0458] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[0459] Step 3:

[0460] The terminal sends formatted text data to the server.

[0461] Specific operation: Generate an HTTP POST request and send the formatted text to the server's API endpoint.

[0462] Step 4:

[0463] The server receives text data from the terminal.

[0464] Specific operation: The server's API receives an HTTP POST request and extracts text data as input for analysis.

[0465] Step 5:

[0466] The server preprocesses the received text data before inputting it into the generative model.

[0467] Specific operations: Performs text data cleansing (removes unnecessary characters and spaces) and tokenization (breaks down words and sentences into tokens).

[0468] Step 6:

[0469] The generative model receives pre-processed text data and generates an appropriate response.

[0470] Specific operation: A generative AI model (e.g., GPT-3) analyzes the text data and generates the response text, "Our return policy allows returns within 30 days of receiving the product."

[0471] Step 7:

[0472] The server receives the generated response and formats or corrects it as needed.

[0473] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[0474] Step 8:

[0475] The server sends the formatted response to the terminal.

[0476] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[0477] Step 9:

[0478] The terminal displays the response received from the server in the user's chat window.

[0479] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[0480] Step 10:

[0481] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[0482] Specific actions: Read the response text, enter a new question if necessary, and submit again.

[0483] As a result, users will be able to receive customer support quickly and efficiently. This system is particularly capable of 24 / 7 support and provides consistently high-quality support. This leads to improved customer satisfaction and optimized human resources.

[0484] (Example 1)

[0485] 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."

[0486] Traditional customer support systems have a problem in that they struggle to provide quick and consistent answers to user inquiries. Furthermore, providing 24 / 7 support requires significant human resources, resulting in high operational costs. Additionally, the inconsistency in support quality due to different personnel handling inquiries is another challenge. It is necessary to improve this situation and achieve efficient and high-quality customer support.

[0487] 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.

[0488] In this invention, the server includes means for receiving questions entered by the user, means for cleansing the received text data and converting it into a prompt sentence, means for inputting the pre-processed text data into a generative model and generating an appropriate answer, and means for formatting the generated answer and sending it to the user. This makes it possible to provide rapid and consistent answers to user inquiries 24 hours a day. Furthermore, by using a generative AI model, it becomes possible to handle a wide range of topics, and it is expected that the uniformity of support quality and the improvement of customer satisfaction can be expected.

[0489] A "user" refers to anyone who uses this system to contact customer support.

[0490] A "question" is text data containing the content of a user's inquiry to customer support.

[0491] "Receiving" refers to the act of a server acquiring data sent from a terminal.

[0492] "Cleansing" refers to the process of removing unnecessary spaces and symbols from text data and standardizing the data format.

[0493] A "prompt statement" is a query statement generated to be input into a generative model in order to obtain a response.

[0494] A "generative model" refers to an artificial intelligence algorithm that analyzes input text data and automatically generates appropriate responses.

[0495] "Formatting" refers to the process of adding line breaks and symbols, or standardizing the writing style, to make the generated response text easier to read.

[0496] "Sending" refers to the act of the server sending back the response obtained from the generative model to the terminal.

[0497] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0498] The user enters a question into a customer support chat window and presses the submit button. At this point, the device retrieves the entered text data and prepares to send it to the server. Specifically, the device converts the entered data into JSON format and sends it to the server using an HTTP request. Hardware used includes the user's device (e.g., a PC or smartphone). Software used includes a web browser or mobile app that provides the chat window.

[0499] The server receives HTTP requests sent from terminals and extracts the text data portion. At this time, the server performs preprocessing on the received text data. This preprocessing includes text data cleansing, such as removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. The preprocessed data is then converted into prompt statements to be passed to the generating AI model. The server contains a program for receiving HTTP requests and processing text data.

[0500] The server sends the generated prompt to a generation AI model (for example, OpenAI's GPT-3). The generation model analyzes the prompt and generates an appropriate response. This generated response text is returned to the server. Specific examples of prompts include the following sentences:

[0501] "A user is asking, 'What is the return policy for this product?' Please provide an appropriate answer."

[0502] The server receives the response returned from the generative model and performs formatting if necessary, such as adding line breaks, correcting symbols, and unifying the writing style. The formatted response text is then converted back to JSON format and sent to the terminal as an HTTP response. The server includes a program to appropriately format and resend the output from the generative model.

[0503] The terminal parses the HTTP response received from the server and retrieves the response text. It can then display this retrieved response text in the user's chat window. The user can review the displayed response and ask additional questions as needed.

[0504] Specific examples are given below.

[0505] Example: Questions about the product return policy

[0506] 1. The user types "Please tell me about the return policy for this product" in the chat window.

[0507] 2. The terminal converts the entered text data into JSON format and sends an HTTP request to the server.

[0508] 3. The server receives the HTTP request, cleanses the text data, and generates a prompt message.

[0509] 4. The server sends the generated prompt message to the generation model.

[0510] 5. The generative model generates the response "Our return policy allows returns within 30 days of receiving the product" and returns it to the server.

[0511] 6. The server formats the generated response and sends it to the terminal.

[0512] 7. The device receives the response and displays it in the user's chat window.

[0513] 8. The user reviews the displayed answers. If necessary, they can enter additional questions and resubmit.

[0514] This embodiment improves the quality of customer support and enables the provision of fast and consistent service. Furthermore, by utilizing generative AI models, it becomes possible to address a wide range of topics, leading to improved customer satisfaction and more efficient use of human resources.

[0515] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0516] Step 1:

[0517] User input and submission of questions

[0518] The user enters a question into the customer support chat window and presses the submit button. The user enters the question, "Please tell me about the return policy for this product." The terminal receives this input text. Since the input is the user's text data, the terminal converts it to JSON format and sends it to the server in the format "{"question": "Please tell me about the return policy for this product"}". Specifically, it sends the data to the server using an HTTP request.

[0519] Step 2:

[0520] Server receives and analyzes questions.

[0521] The server receives an HTTP request sent from the terminal. The received data is input in JSON format, namely "{"question": "Please tell me about the return policy for this product"}". The server parses this data and extracts the text "Please tell me about the return policy for this product". After extraction, the server cleanses this text data. Specifically, this involves removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. After processing, the pre-processed text data "Please tell me about the return policy for this product" is obtained.

[0522] Step 3:

[0523] Server-driven generation of prompt messages

[0524] Based on the pre-processed text data, "Please tell me about the return policy for this product," the server generates a prompt. This prompt is then input into the generative model. Specifically, it is generated in the format, "The user is asking, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." This prompt becomes the input for the next step.

[0525] Step 4:

[0526] Generating answers using generative models

[0527] The server sends the generated prompt to a generating AI model (e.g., GPT-3). The input is the generated prompt "The user asks, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." The generating AI model analyzes this prompt and generates an appropriate answer: "Our return policy allows returns within 30 days of receiving the product." This generated answer is returned to the server, which then receives it.

[0528] Step 5:

[0529] Server formatting and sending of responses

[0530] The server receives the response "Our return policy allows returns within 30 days of receiving the product." from the generative model. The input is the response text from the generative model. If necessary, formatting is performed on the response, such as line breaks, symbol corrections, and consistency of writing style. The formatted response text is converted back into JSON format and sent to the terminal in the format "{"answer": "Our return policy allows returns within 30 days of receiving the product." The server sends the data to the terminal using an HTTP response.

[0531] Step 6:

[0532] Displaying responses via device

[0533] The terminal parses the HTTP response received from the server and obtains the response text: "Our return policy allows returns within 30 days of receiving the product." The input is the response data from the server. The terminal displays this response text in the user's chat window. The user can review the displayed response and ask additional questions if necessary. In this case, the process is repeated from step 1.

[0534] (Application Example 1)

[0535] 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."

[0536] In e-commerce sites, responding quickly and accurately to a wide variety of user inquiries is crucial for maintaining customer satisfaction. Traditionally, customer support has relied on human resources, making 24 / 7 support and immediate responses difficult. This increased the burden on departments and could lead to inconsistent support quality. This invention aims to solve these problems and achieve efficient and consistent support by using a generative AI model.

[0537] 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.

[0538] In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generative model and generating an appropriate answer, means for sending the generated answer to the user, means for cleansing the question as text data and performing format conversion, and means for formatting the generated answer and displaying it on the user's terminal. This enables a quick and accurate response to user questions, realizing 24-hour support and consistent response quality.

[0539] A "user" refers to someone who operates the system, inputs a question, and obtains an answer.

[0540] A "question" refers to text data entered by a user and analyzed by a generative model.

[0541] A "generative model" refers to AI technology that generates responses based on input text data.

[0542] "Cleansing" refers to the process of removing noise from text data and preparing it in a format that is easy to analyze.

[0543] "Format conversion" refers to the process of converting text data into a format that a generative model can understand.

[0544] "Formatting" refers to the process of converting the answers generated by a generative model into a format that is easy for the user to read.

[0545] A "terminal" refers to a device used by a user to access the system, enter questions, and check answers.

[0546] A "server" refers to a computer system that receives questions from users, generates answers using a generative model, and sends them back to the user.

[0547] A "chat interface" refers to the user interface of a system that allows users to input questions and view generated answers.

[0548] This invention provides a system for automatically answering user questions in customer support on e-commerce websites. The system works as follows: the user inputs a question, an AI model analyzes it to generate a response, and then sends the response back to the user. The detailed configuration and processing are described below.

[0549] The server includes means for receiving questions entered by the user, means for inputting the received questions into a generative model and generating appropriate answers, means for sending the generated answers to the user, means for cleansing the questions as text data and performing format conversion, and means for formatting the generated answers and displaying them on the user's terminal.

[0550] The system will utilize smartphones and servers as hardware. Users will operate the system via an application on their smartphones. The software used will include a generative AI model (OpenAI's GPT-3), a text analysis tool (NLTK Natural Language Processing Toolkit), and an application server (Node.js).

[0551] When the server receives text data sent by the user, it first cleanses the text data. Specifically, it uses NLTK to remove unnecessary spaces and symbols. Next, it formats the cleaned text data, converting it into a format that the generative AI model can understand. This is done by receiving the text data as plain text and generating appropriate prompt sentences. The generative model analyzes these prompt sentences and generates an appropriate response. The generated response is returned to the server and formatted as needed. The formatted response is sent to the user's smartphone and displayed in the chat window.

[0552] As a concrete example, a user enters the question "Please tell me about the return policy for this product" and sends it to the server from their smartphone. An example of the prompt text is shown below.

[0553] "Please provide the best answer to the following question: What is the return policy for this product?"

[0554] The server cleanses and formats this text data before inputting it into a generating AI model. This results in the output: "Our return policy allows returns within 30 days of receiving the product." This output is then formatted and displayed on the user's smartphone. This allows the user to receive quick and accurate responses and access 24 / 7 support.

[0555] In this way, the present invention makes it possible to improve the efficiency and quality of customer support on e-commerce sites.

[0556] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0557] Step 1:

[0558] The user enters a question.

[0559] The user opens a chat application on their smartphone, enters their question in text format, and presses the send button. The entered question is saved on the device as text data.

[0560] Step 2:

[0561] The terminal sends the question to the server.

[0562] The user's device sends the entered text data to the server as an HTTP request. The input data includes the user's question.

[0563] Step 3:

[0564] The server receives the question.

[0565] The server receives the HTTP request sent from the terminal and retrieves the question as text data. The received data is stored in a variable for subsequent processing.

[0566] Step 4:

[0567] The server cleanses the text data.

[0568] The server uses NLTK to cleanse the received text data, removing unnecessary spaces and symbols. The cleansed text data is then output.

[0569] Step 5:

[0570] The server converts the text data to a different format.

[0571] The cleansed text data is converted into a format that can be input into the AI ​​model. The server generates a prompt in plain text format. This prompt is as follows: "Provide the best answer to the following question: [cleansed text]".

[0572] Step 6:

[0573] The server inputs prompt messages into the AI ​​model and generates responses.

[0574] The server sends the generated prompt text as input to the generative AI model. The generative AI model (GPT-3) analyzes the prompt text and generates the appropriate response text. The generated response text is obtained as output.

[0575] Step 7:

[0576] The server formats the generated response.

[0577] The server formats the response text obtained from the generating AI model into a user-friendly format. It adds necessary line breaks and highlighting to match the specific format. The formatted text data is then output.

[0578] Step 8:

[0579] The server sends the formatted response to the user's device.

[0580] The server sends the formatted response text data as an HTTP response to the user's device. The device receives this data.

[0581] Step 9:

[0582] The device will display the answer.

[0583] The user's device displays the received response text data in the chat window. The user can review the displayed response and enter further questions if necessary.

[0584] 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.

[0585] This invention relates to a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0586] Program Processing Overview

[0587] 1. User input and submission of questions

[0588] The user types their question into the customer support chat window and presses the send button. At this point, the device retrieves the entered text data and prepares to send it to the server.

[0589] 2. Emotion recognition by the emotion engine

[0590] After the terminal receives the question entered by the user, the emotion engine analyzes the text data and recognizes the emotion. The recognized emotion information is sent to the server along with the text data.

[0591] 3. Receiving questions and sentiments from the server

[0592] The server receives text data and sentiment information sent from the terminal. The server then performs preprocessing on the received text data and sentiment information to input it into the generative model. This preprocessing includes formatting the sentiment information and cleaning the text data.

[0593] 4. Generating answers using generative models

[0594] The generative model receives pre-processed text data and sentiment information, performs analysis, and generates appropriate responses. The generated responses are expressed in a way that takes the user's emotions into consideration, both in content and tone.

[0595] 5. Server sends response

[0596] The server receives the generated response, formats and corrects it as needed, and then sends it to the terminal.

[0597] 6. Displaying responses via the device

[0598] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[0599] Specific example

[0600] Example 1: Questions about the product return policy

[0601] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[0602] The device retrieves this text data, and the emotion engine recognizes the emotion that "the user appears anxious."

[0603] The device sends text data and sentiment information to the server.

[0604] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[0605] The generative model generates the response, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[0606] The server receives the generated response, formats it, and sends it to the terminal.

[0607] The device displays the answer in the user's chat window.

[0608] The user checks the displayed answer and feels reassured.

[0609] Example 2: A user's question expressing dissatisfaction with support.

[0610] The user types "Why is the support so slow?" and submits it.

[0611] The device acquires text data, and the emotion engine recognizes "anger."

[0612] The device sends text data and information about the emotion of anger to the server.

[0613] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[0614] The generative model generates the response, "We apologize for the inconvenience. We are currently working to resolve this as quickly as possible," and returns it to the server.

[0615] The server receives the generated response, formats it, and sends it to the terminal.

[0616] The device displays the answer in the user's chat window.

[0617] The user reviews the displayed answer and is somewhat convinced.

[0618] This system allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[0619] The following describes the processing flow.

[0620] Step 1:

[0621] The user types their question into the customer support chat window and presses the submit button.

[0622] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[0623] Step 2:

[0624] The terminal retrieves the user's input and formats that text data.

[0625] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[0626] Step 3:

[0627] The device sends formatted text data to the emotion engine, which then recognizes the user's emotions.

[0628] Specific operation: Text data is passed to the emotion engine, which then reads emotions such as "anxiety" from the text.

[0629] Step 4:

[0630] The emotion engine returns the recognized emotion information to the device.

[0631] Specific operation: The emotion engine performs text analysis and returns the identified emotion information to the device.

[0632] Step 5:

[0633] The device sends text data and sentiment information to the server.

[0634] Specific operation: Generate an HTTP POST request containing text data and sentiment information, and send it to the server's API endpoint.

[0635] Step 6:

[0636] The server receives text data and sentiment information from the terminal.

[0637] Specific operation: The server's API receives an HTTP POST request and extracts text data and sentiment information as input for analysis.

[0638] Step 7:

[0639] The server preprocesses the received text data and sentiment information.

[0640] Specific operations: Cleans text data (removes unnecessary characters and spaces), tokenizes (breaks down words and sentences into tokens), and converts the format of sentiment information.

[0641] Step 8:

[0642] The generative model receives pre-processed text data and sentiment information and generates an appropriate response.

[0643] Specific operation: A generative AI model (e.g., GPT-3) analyzes the input data and generates response text that takes sentiment into account.

[0644] Step 9:

[0645] The server receives the generated response and formats or corrects it as needed.

[0646] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[0647] Step 10:

[0648] The server sends the formatted response to the terminal.

[0649] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[0650] Step 11:

[0651] The terminal displays the response received from the server in the user's chat window.

[0652] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[0653] Step 12:

[0654] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[0655] Specific operation: The user reads the displayed answer text, enters a new question if necessary, and resubmits.

[0656] This allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[0657] (Example 2)

[0658] 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".

[0659] Traditional customer support systems struggled to accurately recognize and respond to user emotions, sometimes resulting in decreased user satisfaction and delays in problem resolution. Furthermore, efficiently providing 24 / 7 support placed a significant burden on human resources.

[0660] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0661] In this invention, the server includes means for receiving a question entered by a user, means for generating emotional information from the received question using an emotional analysis engine, means for inputting the question along with the generated emotional information into a generation model to generate an appropriate answer, means for formatting the generated answer in a way that takes the user's emotions into consideration, and means for sending the formatted answer to the user. This makes it possible to provide high-quality support that takes the user's emotions into consideration 24 hours a day.

[0662] "Means for receiving user-entered questions" refers to a mechanism for acquiring text data entered by users through customer support chat windows or other interfaces and sending it to the server.

[0663] "Means for generating emotional information using an emotion analysis engine" refers to a mechanism that analyzes text data entered by a user, recognizes the user's emotional state from its content, and generates appropriate emotional information.

[0664] "A means of inputting questions along with generated sentiment information into a generative model to generate appropriate answers" refers to a mechanism for inputting recognized sentiment information and user questions into a generative model-trained generative algorithm to generate appropriate answers that take the user's emotions into consideration.

[0665] "Methods for shaping generated responses to reflect user emotions" refers to a mechanism that processes responses generated by a generative model into a format and phrasing that aligns with the user's emotions, making them easy to understand and considerate for the user.

[0666] "Means for sending formatted responses to users" refers to a mechanism for sending formatted responses from a server to the user's terminal and displaying them in a chat window or other interface.

[0667] This system combines a generative AI model and an emotion analysis engine to generate responses that take user emotions into consideration, in order to provide effective customer support. The specific configuration and operation of the system are described below.

[0668] Hardware and software to use

[0669] Hardware:

[0670] Server: Cloud service (e.g., AWS EC2)

[0671] Devices: User devices such as PCs, smartphones, and tablets.

[0672] software:

[0673] Chat window platform: Customer support system (e.g., Intercom, Zendesk)

[0674] Sentiment analysis engine: Text analysis library (e.g., Affective Computing SDK)

[0675] Generative AI models: Natural language generation models (e.g., OpenAI GPT-3)

[0676] System response processing

[0677] The server first has a component that receives questions entered by the user. The user types their question in the chat window and presses the send button. For example, if the user types "Please tell me about the return policy for this product," this text data is captured by the terminal. The terminal then prepares to send this text data to the server.

[0678] The server also has a component that analyzes received questions using an emotion analysis engine and generates emotion information. The emotion analysis engine analyzes the text entered by the user and generates emotion information. For example, if it recognizes the emotion of "anxiety" from the user's statement, it will generate emotion information for "anxiety".

[0679] Next, the server inputs the question, along with the generated sentiment information, into the generating AI model to produce an appropriate answer. The generating model receives the pre-processed data, analyzes it, and generates the answer. This answer will have content and tone that takes the user's emotions into consideration. For example, it might generate an answer such as, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[0680] The generated response is then formatted to reflect the user's emotions and tone. This process makes the response more acceptable to the user. The formatted response is then sent back to the terminal by the server.

[0681] The device displays the received response in the chat window. The user can review the displayed response and determine whether the problem has been resolved.

[0682] Specific examples of operation

[0683] The following is a concrete example of actual system operation.

[0684] Example of a prompt:

[0685] 1. "Please tell me about the return policy for this product."

[0686] 2. "Why is the support so slow?"

[0687] For example, if a user types "Please tell me about the return policy for this product," the device receives this input, and its sentiment analysis engine recognizes "anxiety." The server then inputs this text data and sentiment information into a generating AI model, which generates and formats a response such as "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and sends it back to the device. The device then displays this response in the user's chat window, allowing the user to see it and feel reassured.

[0688] In this way, the system provides high-quality customer support that takes user emotions into consideration, improving user satisfaction and streamlining support operations. Furthermore, it enables 24 / 7 support, contributing to a reduction in the burden on human resources.

[0689] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0690] Step 1:

[0691] The user types their question into the customer support chat window and presses the send button. This is then sent to the device as text data.

[0692] Input: Text data entered by the user in the chat window

[0693] Output: Text data is sent to the terminal.

[0694] Specific action: The user types "Please tell me about the return policy for this product" into the chat window and presses the send button.

[0695] Step 2:

[0696] The terminal receives the input text data and stores it in memory. It also prepares the text data for transmission.

[0697] Input: Text data entered by the user

[0698] Output: Text data stored within the device

[0699] Specific operation: The device detects user input and saves that text data to temporary storage.

[0700] Step 3:

[0701] The device uses an emotion analysis engine to analyze text data. This analysis extracts user emotion information from the text data.

[0702] Input: Text data stored on the device

[0703] Output: Sentiment information generated based on text data

[0704] Specific operation: The device's emotion analysis engine analyzes the text data and recognizes that the user is feeling "anxious".

[0705] Step 4:

[0706] The device sends text data and sentiment information to the server.

[0707] Input: Text data and sentiment information

[0708] Output: Text data and sentiment information sent to the server

[0709] Specific operation: The terminal sends text data and sentiment information to the server in packet format.

[0710] Step 5:

[0711] The server receives text data and sentiment information sent from the terminal and performs preprocessing. Preprocessing includes text data cleansing and sentiment information format conversion.

[0712] Input: Text data and sentiment information received from the device.

[0713] Output: Preprocessed text data and sentiment information

[0714] Specific operation: The server cleanses the received data and converts it to the required format.

[0715] Step 6:

[0716] The server inputs pre-processed text data and sentiment information into a generating AI model to generate responses.

[0717] Input: Preprocessed text data and sentiment information

[0718] Output: Answers generated by the generative AI model

[0719] Specific operation: The server inputs data into the generated AI model and generates the response, "Rest assured, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[0720] Step 7:

[0721] The server receives the generated responses and performs formatting and necessary corrections. This formatting ensures the consistency of the responses.

[0722] Input: Answers generated by the generative AI model

[0723] Output: Formatted answer

[0724] Specific operation: The server inspects the response and corrects grammar and tone.

[0725] Step 8:

[0726] The server sends the formatted response to the terminal.

[0727] Input: Formatted answer

[0728] Output: Formatted response sent to the terminal

[0729] Specific operation: The server sends the formatted response to the terminal.

[0730] Step 9:

[0731] The device displays the received response in the user's chat window.

[0732] Input: Formatted response received from the server

[0733] Output: Answers displayed in the chat window

[0734] Specific operation: The device displays the received response in the chat window and notifies the user.

[0735] Step 10:

[0736] The user reviews the displayed answers and asks additional questions as needed.

[0737] Input: Answer displayed in the chat window

[0738] Output: Additional questions (if necessary)

[0739] Specific action: The user reads the displayed answer and enters whether they are satisfied or ask further questions.

[0740] (Application Example 2)

[0741] 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."

[0742] Traditional customer support systems often struggle to respond to user emotions and provide only uniform answers, which can lead to decreased customer satisfaction. Furthermore, in physical stores, situations may arise where staff are unavailable, resulting in customers not receiving prompt and appropriate support.

[0743] 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. In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generation model and generating an appropriate answer, means for sending the generated answer to the user, means for recognizing the user's emotions using an emotion engine, means for inputting the recognized emotion information into the generation model and generating an answer, and means for displaying or playing the generated answer in voice or text. This enables the provision of answers that correspond to the user's emotions, and allows for quick and appropriate support even in physical stores.

[0744] A "user" refers to a person who enters questions or information into a system.

[0745] "Means for receiving questions" refers to a part of the system that retrieves and processes questions entered by users.

[0746] A "generative model" refers to an algorithm used to generate appropriate answers based on received questions and data.

[0747] "Means for generating answers" refers to a function that uses a generative model to generate answers to received questions.

[0748] An "emotion engine" refers to an algorithm that analyzes and recognizes emotions from user input data.

[0749] "Means of recognizing emotions" refers to a function that uses an emotion engine to identify the user's emotions.

[0750] "Emotional information" refers to information about emotions recognized from user input data.

[0751] "Means of submitting responses" refers to a part of the system used to communicate generated responses to users.

[0752] "Means of displaying or playing" refers to a function that provides the generated response to the user visually or audibly in the form of audio or text.

[0753] This invention is a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions, enabling fast and emotionally responsive customer service in physical stores. The system components include a server, a terminal (smart device), an emotion engine, a generative AI model, and communication means.

[0754] System Configuration

[0755] 1. Voice input processing

[0756] Hardware: Smart glasses

[0757] Example: The general term is "smart glasses device".

[0758] Software: Speech Recognition API

[0759] Example: "Speech-to-text conversion service" as a general term.

[0760] Process: The user wears smart glasses and inputs a question by voice. The smart glasses convert the voice input into text data.

[0761] 2. Emotion Recognition Engine Processing

[0762] Software: Natural Language Processing Library

[0763] Example: "Natural language processing engine" as a general term

[0764] Processing: Text data converted from speech is input into the emotion recognition engine to identify the customer's emotion. For example, if a customer asks, "What are the features of this product?", the emotion engine identifies "interest." This allows the customer's emotion information to be input into the generative model.

[0765] 3. Data transmission process

[0766] Software: HTTP communication library

[0767] Example: "Communication Library" as a general term

[0768] Processing: Text data and sentiment information are sent to the server. The server receives this data and inputs it into a generative model to generate a response.

[0769] 4. Generative Model Processing

[0770] Hardware: Server environment

[0771] Example: "Cloud Server" as a general term

[0772] Software: Generative AI Model

[0773] Example: "Generative algorithm" as a general term

[0774] Processing: The server inputs the received text data and sentiment information into a generative model to generate an appropriate response. The generated response is expressed in a tone and content that is sensitive to the customer's emotions.

[0775] 5. Sending the response

[0776] Software: HTTP communication library

[0777] Example: "Communication Library" as a general term

[0778] Processing: The generated response is sent back to the smart glasses.

[0779] 6. Displaying the answer

[0780] Hardware: Smart glasses

[0781] Example: "Smart device" as a general term

[0782] Software: Display Management Library

[0783] Example: "Display management software" as a general term

[0784] Processing: The response is displayed as audio or text on the smart glasses worn by the customer, allowing the user to confirm it.

[0785] Specific example

[0786] In a physical store, if a customer asks a question to smart glasses by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the customer's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[0787] Example of a prompt

[0788] A customer asked the following question: "What are the features of this product?" Emotion: Interest

[0789] The generating AI model should consider the customer's interest in this question and respond as follows: "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design."

[0790] This form allows for a concrete understanding of how the invention is implemented and provides a detailed process that can be reproduced by anyone.

[0791] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0792] Step 1:

[0793] The user wears smart glasses and inputs questions by voice. The input voice data is captured through the microphone of the smart device.

[0794] Step 2:

[0795] The device sends the acquired audio data to a speech recognition API. The speech recognition API converts the audio data into text data. The input is audio data, and the output is text data.

[0796] Step 3:

[0797] The device inputs the converted text data into the emotion recognition engine. The emotion recognition engine analyzes the text data and identifies the user's emotions. The input is text data, and the output is emotion information.

[0798] Step 4:

[0799] The terminal sends text data and recognized sentiment information to the server via a communication library. The input is text data and sentiment information, and the output is the data sent to the server.

[0800] Step 5:

[0801] The server preprocesses the received text data and sentiment information. This includes data cleaning and format conversion to prepare the data for input into the generative model. The input is the received data, and the output is the preprocessed data.

[0802] Step 6:

[0803] The server inputs pre-processed data into a generative AI model, which then generates appropriate responses. The generative model analyzes the data and generates responses that take the user's emotions into consideration. The input is pre-processed data, and the output is the generated response.

[0804] Step 7:

[0805] The server receives the generated response and performs further formatting and editing as needed. The formatted response is then sent back to the terminal via the communication library. The input is the generated response, and the output is the formatted response.

[0806] Step 8:

[0807] The device displays the received response to the user using the smart glasses' display management library. Audio output may also be provided. The input is a formatted response, and the output is presented to the user as audio or text.

[0808] Specific example

[0809] For example, if a user asks the smart glasses a question by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the user's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[0810] 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.

[0811] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0812] 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.

[0813] [Third Embodiment]

[0814] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0815] 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.

[0816] 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).

[0817] 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.

[0818] 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.

[0819] 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).

[0820] 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.

[0821] 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.

[0822] 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.

[0823] 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.

[0824] 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.

[0825] 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".

[0826] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0827] Program Processing Overview

[0828] 1. User input and submission of questions

[0829] The user types their question into the customer support chat window and presses the send button. This action causes the user's device to retrieve the entered text data and prepare to send it to the server.

[0830] 2. Server receives and analyzes the question.

[0831] The server receives text data sent from the terminal. The server then performs preprocessing on the received text data to input it into the generative model. This preprocessing includes text data cleansing and format conversion.

[0832] 3. Generating answers using generative models

[0833] A generative model (e.g., GPT-3) receives pre-processed text data, performs analysis, and generates an appropriate response. The generated response is then returned to the server.

[0834] 4. The server sends the response.

[0835] The server receives the generated response, reformats or corrects it if necessary, and then sends it to the terminal.

[0836] 5. Displaying responses via the device

[0837] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[0838] Specific example

[0839] Example 1: Questions about the product return policy

[0840] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[0841] The device retrieves this text data and sends it to the server.

[0842] The server receives text data, inputs it into a generative model, and performs analysis.

[0843] The generative model generates the response, "Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[0844] The server receives the generated response, formats it, and sends it to the terminal.

[0845] The device displays the answer in the user's chat window.

[0846] Users can review the displayed answers and ask further questions if necessary.

[0847] This system allows users to receive fast and consistent support 24 / 7. The use of generative AI models enables support across a wide range of topics, leading to consistent support quality and improved customer satisfaction. This, in turn, enables the efficient use of human resources in customer support.

[0848] The following describes the processing flow.

[0849] Step 1:

[0850] The user types their question into the customer support chat window and presses the submit button.

[0851] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[0852] Step 2:

[0853] The terminal retrieves the user's input and formats that text data.

[0854] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[0855] Step 3:

[0856] The terminal sends formatted text data to the server.

[0857] Specific operation: Generate an HTTP POST request and send the formatted text to the server's API endpoint.

[0858] Step 4:

[0859] The server receives text data from the terminal.

[0860] Specific operation: The server's API receives an HTTP POST request and extracts text data as input for analysis.

[0861] Step 5:

[0862] The server preprocesses the received text data before inputting it into the generative model.

[0863] Specific operations: Performs text data cleansing (removes unnecessary characters and spaces) and tokenization (breaks down words and sentences into tokens).

[0864] Step 6:

[0865] The generative model receives pre-processed text data and generates an appropriate response.

[0866] Specific operation: A generative AI model (e.g., GPT-3) analyzes the text data and generates the response text, "Our return policy allows returns within 30 days of receiving the product."

[0867] Step 7:

[0868] The server receives the generated response and formats or corrects it as needed.

[0869] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[0870] Step 8:

[0871] The server sends the formatted response to the terminal.

[0872] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[0873] Step 9:

[0874] The terminal displays the response received from the server in the user's chat window.

[0875] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[0876] Step 10:

[0877] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[0878] Specific actions: Read the response text, enter a new question if necessary, and submit again.

[0879] As a result, users will be able to receive customer support quickly and efficiently. This system is particularly capable of 24 / 7 support and provides consistently high-quality support. This leads to improved customer satisfaction and optimized human resources.

[0880] (Example 1)

[0881] 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."

[0882] Traditional customer support systems have a problem in that they struggle to provide quick and consistent answers to user inquiries. Furthermore, providing 24 / 7 support requires significant human resources, resulting in high operational costs. Additionally, the inconsistency in support quality due to different personnel handling inquiries is another challenge. It is necessary to improve this situation and achieve efficient and high-quality customer support.

[0883] 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.

[0884] In this invention, the server includes means for receiving questions entered by the user, means for cleansing the received text data and converting it into a prompt sentence, means for inputting the pre-processed text data into a generative model and generating an appropriate answer, and means for formatting the generated answer and sending it to the user. This makes it possible to provide rapid and consistent answers to user inquiries 24 hours a day. Furthermore, by using a generative AI model, it becomes possible to handle a wide range of topics, and it is expected that the uniformity of support quality and the improvement of customer satisfaction can be expected.

[0885] A "user" refers to anyone who uses this system to contact customer support.

[0886] A "question" is text data containing the content of a user's inquiry to customer support.

[0887] "Receiving" refers to the act of a server acquiring data sent from a terminal.

[0888] "Cleansing" refers to the process of removing unnecessary spaces and symbols from text data and standardizing the data format.

[0889] A "prompt statement" is a query statement generated to be input into a generative model in order to obtain a response.

[0890] A "generative model" refers to an artificial intelligence algorithm that analyzes input text data and automatically generates appropriate responses.

[0891] "Formatting" refers to the process of adding line breaks and symbols, or standardizing the writing style, to make the generated response text easier to read.

[0892] "Sending" refers to the act of the server sending back the response obtained from the generative model to the terminal.

[0893] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0894] The user enters a question into a customer support chat window and presses the submit button. At this point, the device retrieves the entered text data and prepares to send it to the server. Specifically, the device converts the entered data into JSON format and sends it to the server using an HTTP request. Hardware used includes the user's device (e.g., a PC or smartphone). Software used includes a web browser or mobile app that provides the chat window.

[0895] The server receives HTTP requests sent from terminals and extracts the text data portion. At this time, the server performs preprocessing on the received text data. This preprocessing includes text data cleansing, such as removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. The preprocessed data is then converted into prompt statements to be passed to the generating AI model. The server contains a program for receiving HTTP requests and processing text data.

[0896] The server sends the generated prompt to a generation AI model (for example, OpenAI's GPT-3). The generation model analyzes the prompt and generates an appropriate response. This generated response text is returned to the server. Specific examples of prompts include the following sentences:

[0897] "A user is asking, 'What is the return policy for this product?' Please provide an appropriate answer."

[0898] The server receives the response returned from the generative model and performs formatting if necessary, such as adding line breaks, correcting symbols, and unifying the writing style. The formatted response text is then converted back to JSON format and sent to the terminal as an HTTP response. The server includes a program to appropriately format and resend the output from the generative model.

[0899] The terminal parses the HTTP response received from the server and retrieves the response text. It can then display this retrieved response text in the user's chat window. The user can review the displayed response and ask additional questions as needed.

[0900] Specific examples are given below.

[0901] Example: Questions about the product return policy

[0902] 1. The user types "Please tell me about the return policy for this product" in the chat window.

[0903] 2. The terminal converts the entered text data into JSON format and sends an HTTP request to the server.

[0904] 3. The server receives the HTTP request, cleanses the text data, and generates a prompt message.

[0905] 4. The server sends the generated prompt message to the generation model.

[0906] 5. The generative model generates the response "Our return policy allows returns within 30 days of receiving the product" and returns it to the server.

[0907] 6. The server formats the generated response and sends it to the terminal.

[0908] 7. The device receives the response and displays it in the user's chat window.

[0909] 8. The user reviews the displayed answers. If necessary, they can enter additional questions and resubmit.

[0910] This embodiment improves the quality of customer support and enables the provision of fast and consistent service. Furthermore, by utilizing generative AI models, it becomes possible to address a wide range of topics, leading to improved customer satisfaction and more efficient use of human resources.

[0911] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0912] Step 1:

[0913] User input and submission of questions

[0914] The user enters a question into the customer support chat window and presses the submit button. The user enters the question, "Please tell me about the return policy for this product." The terminal receives this input text. Since the input is the user's text data, the terminal converts it to JSON format and sends it to the server in the format "{"question": "Please tell me about the return policy for this product"}". Specifically, it sends the data to the server using an HTTP request.

[0915] Step 2:

[0916] Server receives and analyzes questions.

[0917] The server receives an HTTP request sent from the terminal. The received data is input in JSON format, namely "{"question": "Please tell me about the return policy for this product"}". The server parses this data and extracts the text "Please tell me about the return policy for this product". After extraction, the server cleanses this text data. Specifically, this involves removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. After processing, the pre-processed text data "Please tell me about the return policy for this product" is obtained.

[0918] Step 3:

[0919] Server-driven generation of prompt messages

[0920] Based on the pre-processed text data, "Please tell me about the return policy for this product," the server generates a prompt. This prompt is then input into the generative model. Specifically, it is generated in the format, "The user is asking, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." This prompt becomes the input for the next step.

[0921] Step 4:

[0922] Generating answers using generative models

[0923] The server sends the generated prompt to a generating AI model (e.g., GPT-3). The input is the generated prompt "The user asks, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." The generating AI model analyzes this prompt and generates an appropriate answer: "Our return policy allows returns within 30 days of receiving the product." This generated answer is returned to the server, which then receives it.

[0924] Step 5:

[0925] Server formatting and sending of responses

[0926] The server receives the response "Our return policy allows returns within 30 days of receiving the product." from the generative model. The input is the response text from the generative model. If necessary, formatting is performed on the response, such as line breaks, symbol corrections, and consistency of writing style. The formatted response text is converted back into JSON format and sent to the terminal in the format "{"answer": "Our return policy allows returns within 30 days of receiving the product." The server sends the data to the terminal using an HTTP response.

[0927] Step 6:

[0928] Displaying responses via device

[0929] The terminal parses the HTTP response received from the server and obtains the response text: "Our return policy allows returns within 30 days of receiving the product." The input is the response data from the server. The terminal displays this response text in the user's chat window. The user can review the displayed response and ask additional questions if necessary. In this case, the process is repeated from step 1.

[0930] (Application Example 1)

[0931] 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."

[0932] In e-commerce sites, responding quickly and accurately to a wide variety of user inquiries is crucial for maintaining customer satisfaction. Traditionally, customer support has relied on human resources, making 24 / 7 support and immediate responses difficult. This increased the burden on departments and could lead to inconsistent support quality. This invention aims to solve these problems and achieve efficient and consistent support by using a generative AI model.

[0933] 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.

[0934] In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generative model and generating an appropriate answer, means for sending the generated answer to the user, means for cleansing the question as text data and performing format conversion, and means for formatting the generated answer and displaying it on the user's terminal. This enables a quick and accurate response to user questions, realizing 24-hour support and consistent response quality.

[0935] A "user" refers to someone who operates the system, inputs a question, and obtains an answer.

[0936] A "question" refers to text data entered by a user and analyzed by a generative model.

[0937] A "generative model" refers to AI technology that generates responses based on input text data.

[0938] "Cleansing" refers to the process of removing noise from text data and preparing it in a format that is easy to analyze.

[0939] "Format conversion" refers to the process of converting text data into a format that a generative model can understand.

[0940] "Formatting" refers to the process of converting the answers generated by a generative model into a format that is easy for the user to read.

[0941] A "terminal" refers to a device used by a user to access the system, enter questions, and check answers.

[0942] A "server" refers to a computer system that receives questions from users, generates answers using a generative model, and sends them back to the user.

[0943] A "chat interface" refers to the user interface of a system that allows users to input questions and view generated answers.

[0944] This invention provides a system for automatically answering user questions in customer support on e-commerce websites. The system works as follows: the user inputs a question, an AI model analyzes it to generate a response, and then sends the response back to the user. The detailed configuration and processing are described below.

[0945] The server includes means for receiving questions entered by the user, means for inputting the received questions into a generative model and generating appropriate answers, means for sending the generated answers to the user, means for cleansing the questions as text data and performing format conversion, and means for formatting the generated answers and displaying them on the user's terminal.

[0946] The system will utilize smartphones and servers as hardware. Users will operate the system via an application on their smartphones. The software used will include a generative AI model (OpenAI's GPT-3), a text analysis tool (NLTK Natural Language Processing Toolkit), and an application server (Node.js).

[0947] When the server receives text data sent by the user, it first cleanses the text data. Specifically, it uses NLTK to remove unnecessary spaces and symbols. Next, it formats the cleaned text data, converting it into a format that the generative AI model can understand. This is done by receiving the text data as plain text and generating appropriate prompt sentences. The generative model analyzes these prompt sentences and generates an appropriate response. The generated response is returned to the server and formatted as needed. The formatted response is sent to the user's smartphone and displayed in the chat window.

[0948] As a concrete example, a user enters the question "Please tell me about the return policy for this product" and sends it to the server from their smartphone. An example of the prompt text is shown below.

[0949] "Please provide the best answer to the following question: What is the return policy for this product?"

[0950] The server cleanses and formats this text data before inputting it into a generating AI model. This results in the output: "Our return policy allows returns within 30 days of receiving the product." This output is then formatted and displayed on the user's smartphone. This allows the user to receive quick and accurate responses and access 24 / 7 support.

[0951] In this way, the present invention makes it possible to improve the efficiency and quality of customer support on e-commerce sites.

[0952] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0953] Step 1:

[0954] The user enters a question.

[0955] The user opens a chat application on their smartphone, enters their question in text format, and presses the send button. The entered question is saved on the device as text data.

[0956] Step 2:

[0957] The terminal sends the question to the server.

[0958] The user's device sends the entered text data to the server as an HTTP request. The input data includes the user's question.

[0959] Step 3:

[0960] The server receives the question.

[0961] The server receives the HTTP request sent from the terminal and retrieves the question as text data. The received data is stored in a variable for subsequent processing.

[0962] Step 4:

[0963] The server cleanses the text data.

[0964] The server uses NLTK to cleanse the received text data, removing unnecessary spaces and symbols. The cleansed text data is then output.

[0965] Step 5:

[0966] The server converts the text data to a different format.

[0967] The cleansed text data is converted into a format that can be input into the AI ​​model. The server generates a prompt in plain text format. This prompt is as follows: "Provide the best answer to the following question: [cleansed text]".

[0968] Step 6:

[0969] The server inputs prompt messages into the AI ​​model and generates responses.

[0970] The server sends the generated prompt text as input to the generative AI model. The generative AI model (GPT-3) analyzes the prompt text and generates the appropriate response text. The generated response text is obtained as output.

[0971] Step 7:

[0972] The server formats the generated response.

[0973] The server formats the response text obtained from the generating AI model into a user-friendly format. It adds necessary line breaks and highlighting to match the specific format. The formatted text data is then output.

[0974] Step 8:

[0975] The server sends the formatted response to the user's device.

[0976] The server sends the formatted response text data as an HTTP response to the user's device. The device receives this data.

[0977] Step 9:

[0978] The device will display the answer.

[0979] The user's device displays the received response text data in the chat window. The user can review the displayed response and enter further questions if necessary.

[0980] 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.

[0981] This invention relates to a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[0982] Program Processing Overview

[0983] 1. User input and submission of questions

[0984] The user types their question into the customer support chat window and presses the send button. At this point, the device retrieves the entered text data and prepares to send it to the server.

[0985] 2. Emotion recognition by the emotion engine

[0986] After the terminal receives the question entered by the user, the emotion engine analyzes the text data and recognizes the emotion. The recognized emotion information is sent to the server along with the text data.

[0987] 3. Receiving questions and sentiments from the server

[0988] The server receives text data and sentiment information sent from the terminal. The server then performs preprocessing on the received text data and sentiment information to input it into the generative model. This preprocessing includes formatting the sentiment information and cleaning the text data.

[0989] 4. Generating answers using generative models

[0990] The generative model receives pre-processed text data and sentiment information, performs analysis, and generates appropriate responses. The generated responses are expressed in a way that takes the user's emotions into consideration, both in content and tone.

[0991] 5. Server sends response

[0992] The server receives the generated response, formats and corrects it as needed, and then sends it to the terminal.

[0993] 6. Displaying responses via the device

[0994] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[0995] Specific example

[0996] Example 1: Questions about the product return policy

[0997] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[0998] The device retrieves this text data, and the emotion engine recognizes the emotion that "the user appears anxious."

[0999] The device sends text data and sentiment information to the server.

[1000] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[1001] The generative model generates the response, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[1002] The server receives the generated response, formats it, and sends it to the terminal.

[1003] The device displays the answer in the user's chat window.

[1004] The user checks the displayed answer and feels reassured.

[1005] Example 2: A user's question expressing dissatisfaction with support.

[1006] The user types "Why is the support so slow?" and submits it.

[1007] The device acquires text data, and the emotion engine recognizes "anger."

[1008] The device sends text data and information about the emotion of anger to the server.

[1009] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[1010] The generative model generates the response, "We apologize for the inconvenience. We are currently working to resolve this as quickly as possible," and returns it to the server.

[1011] The server receives the generated response, formats it, and sends it to the terminal.

[1012] The device displays the answer in the user's chat window.

[1013] The user reviews the displayed answer and is somewhat convinced.

[1014] This system allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[1015] The following describes the processing flow.

[1016] Step 1:

[1017] The user types their question into the customer support chat window and presses the submit button.

[1018] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[1019] Step 2:

[1020] The terminal retrieves the user's input and formats that text data.

[1021] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[1022] Step 3:

[1023] The device sends formatted text data to the emotion engine, which then recognizes the user's emotions.

[1024] Specific operation: Text data is passed to the emotion engine, which then reads emotions such as "anxiety" from the text.

[1025] Step 4:

[1026] The emotion engine returns the recognized emotion information to the device.

[1027] Specific operation: The emotion engine performs text analysis and returns the identified emotion information to the device.

[1028] Step 5:

[1029] The device sends text data and sentiment information to the server.

[1030] Specific operation: Generate an HTTP POST request containing text data and sentiment information, and send it to the server's API endpoint.

[1031] Step 6:

[1032] The server receives text data and sentiment information from the terminal.

[1033] Specific operation: The server's API receives an HTTP POST request and extracts text data and sentiment information as input for analysis.

[1034] Step 7:

[1035] The server preprocesses the received text data and sentiment information.

[1036] Specific operations: Cleans text data (removes unnecessary characters and spaces), tokenizes (breaks down words and sentences into tokens), and converts the format of sentiment information.

[1037] Step 8:

[1038] The generative model receives pre-processed text data and sentiment information and generates an appropriate response.

[1039] Specific operation: A generative AI model (e.g., GPT-3) analyzes the input data and generates response text that takes sentiment into account.

[1040] Step 9:

[1041] The server receives the generated response and formats or corrects it as needed.

[1042] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[1043] Step 10:

[1044] The server sends the formatted response to the terminal.

[1045] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[1046] Step 11:

[1047] The terminal displays the response received from the server in the user's chat window.

[1048] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[1049] Step 12:

[1050] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[1051] Specific operation: The user reads the displayed answer text, enters a new question if necessary, and resubmits.

[1052] This allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[1053] (Example 2)

[1054] 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."

[1055] Traditional customer support systems struggled to accurately recognize and respond to user emotions, sometimes resulting in decreased user satisfaction and delays in problem resolution. Furthermore, efficiently providing 24 / 7 support placed a significant burden on human resources.

[1056] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1057] In this invention, the server includes means for receiving a question entered by a user, means for generating emotional information from the received question using an emotional analysis engine, means for inputting the question along with the generated emotional information into a generation model to generate an appropriate answer, means for formatting the generated answer in a way that takes the user's emotions into consideration, and means for sending the formatted answer to the user. This makes it possible to provide high-quality support that takes the user's emotions into consideration 24 hours a day.

[1058] "Means for receiving user-entered questions" refers to a mechanism for acquiring text data entered by users through customer support chat windows or other interfaces and sending it to the server.

[1059] "Means for generating emotional information using an emotion analysis engine" refers to a mechanism that analyzes text data entered by a user, recognizes the user's emotional state from its content, and generates appropriate emotional information.

[1060] "A means of inputting questions along with generated sentiment information into a generative model to generate appropriate answers" refers to a mechanism for inputting recognized sentiment information and user questions into a generative model-trained generative algorithm to generate appropriate answers that take the user's emotions into consideration.

[1061] "Methods for shaping generated responses to reflect user emotions" refers to a mechanism that processes responses generated by a generative model into a format and phrasing that aligns with the user's emotions, making them easy to understand and considerate for the user.

[1062] "Means for sending formatted responses to users" refers to a mechanism for sending formatted responses from a server to the user's terminal and displaying them in a chat window or other interface.

[1063] This system combines a generative AI model and an emotion analysis engine to generate responses that take user emotions into consideration, in order to provide effective customer support. The specific configuration and operation of the system are described below.

[1064] Hardware and software to use

[1065] Hardware:

[1066] Server: Cloud service (e.g., AWS EC2)

[1067] Devices: User devices such as PCs, smartphones, and tablets.

[1068] software:

[1069] Chat window platform: Customer support system (e.g., Intercom, Zendesk)

[1070] Sentiment analysis engine: Text analysis library (e.g., Affective Computing SDK)

[1071] Generative AI models: Natural language generation models (e.g., OpenAI GPT-3)

[1072] System response processing

[1073] The server first has a component that receives questions entered by the user. The user types their question in the chat window and presses the send button. For example, if the user types "Please tell me about the return policy for this product," this text data is captured by the terminal. The terminal then prepares to send this text data to the server.

[1074] The server also has a component that analyzes received questions using an emotion analysis engine and generates emotion information. The emotion analysis engine analyzes the text entered by the user and generates emotion information. For example, if it recognizes the emotion of "anxiety" from the user's statement, it will generate emotion information for "anxiety".

[1075] Next, the server inputs the question, along with the generated sentiment information, into the generating AI model to produce an appropriate answer. The generating model receives the pre-processed data, analyzes it, and generates the answer. This answer will have content and tone that takes the user's emotions into consideration. For example, it might generate an answer such as, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[1076] The generated response is then formatted to reflect the user's emotions and tone. This process makes the response more acceptable to the user. The formatted response is then sent back to the terminal by the server.

[1077] The device displays the received response in the chat window. The user can review the displayed response and determine whether the problem has been resolved.

[1078] Specific examples of operation

[1079] The following is a concrete example of actual system operation.

[1080] Example of a prompt:

[1081] 1. "Please tell me about the return policy for this product."

[1082] 2. "Why is the support so slow?"

[1083] For example, if a user types "Please tell me about the return policy for this product," the device receives this input, and its sentiment analysis engine recognizes "anxiety." The server then inputs this text data and sentiment information into a generating AI model, which generates and formats a response such as "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and sends it back to the device. The device then displays this response in the user's chat window, allowing the user to see it and feel reassured.

[1084] In this way, the system provides high-quality customer support that takes user emotions into consideration, improving user satisfaction and streamlining support operations. Furthermore, it enables 24 / 7 support, contributing to a reduction in the burden on human resources.

[1085] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1086] Step 1:

[1087] The user types their question into the customer support chat window and presses the send button. This is then sent to the device as text data.

[1088] Input: Text data entered by the user in the chat window

[1089] Output: Text data is sent to the terminal.

[1090] Specific action: The user types "Please tell me about the return policy for this product" into the chat window and presses the send button.

[1091] Step 2:

[1092] The terminal receives the input text data and stores it in memory. It also prepares the text data for transmission.

[1093] Input: Text data entered by the user

[1094] Output: Text data stored within the device

[1095] Specific operation: The device detects user input and saves that text data to temporary storage.

[1096] Step 3:

[1097] The device uses an emotion analysis engine to analyze text data. This analysis extracts user emotion information from the text data.

[1098] Input: Text data stored on the device

[1099] Output: Sentiment information generated based on text data

[1100] Specific operation: The device's emotion analysis engine analyzes the text data and recognizes that the user is feeling "anxious".

[1101] Step 4:

[1102] The device sends text data and sentiment information to the server.

[1103] Input: Text data and sentiment information

[1104] Output: Text data and sentiment information sent to the server

[1105] Specific operation: The terminal sends text data and sentiment information to the server in packet format.

[1106] Step 5:

[1107] The server receives text data and sentiment information sent from the terminal and performs preprocessing. Preprocessing includes text data cleansing and sentiment information format conversion.

[1108] Input: Text data and sentiment information received from the device.

[1109] Output: Preprocessed text data and sentiment information

[1110] Specific operation: The server cleanses the received data and converts it to the required format.

[1111] Step 6:

[1112] The server inputs pre-processed text data and sentiment information into a generating AI model to generate responses.

[1113] Input: Preprocessed text data and sentiment information

[1114] Output: Answers generated by the generative AI model

[1115] Specific operation: The server inputs data into the generated AI model and generates the response, "Rest assured, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[1116] Step 7:

[1117] The server receives the generated responses and performs formatting and necessary corrections. This formatting ensures the consistency of the responses.

[1118] Input: Answers generated by the generative AI model

[1119] Output: Formatted answer

[1120] Specific operation: The server inspects the response and corrects grammar and tone.

[1121] Step 8:

[1122] The server sends the formatted response to the terminal.

[1123] Input: Formatted answer

[1124] Output: Formatted response sent to the terminal

[1125] Specific operation: The server sends the formatted response to the terminal.

[1126] Step 9:

[1127] The device displays the received response in the user's chat window.

[1128] Input: Formatted response received from the server

[1129] Output: Answers displayed in the chat window

[1130] Specific operation: The device displays the received response in the chat window and notifies the user.

[1131] Step 10:

[1132] The user reviews the displayed answers and asks additional questions as needed.

[1133] Input: Answer displayed in the chat window

[1134] Output: Additional questions (if necessary)

[1135] Specific action: The user reads the displayed answer and enters whether they are satisfied or ask further questions.

[1136] (Application Example 2)

[1137] 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."

[1138] Traditional customer support systems often struggle to respond to user emotions and provide only uniform answers, which can lead to decreased customer satisfaction. Furthermore, in physical stores, situations may arise where staff are unavailable, resulting in customers not receiving prompt and appropriate support.

[1139] 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. In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generation model and generating an appropriate answer, means for sending the generated answer to the user, means for recognizing the user's emotions using an emotion engine, means for inputting the recognized emotion information into the generation model and generating an answer, and means for displaying or playing the generated answer in voice or text. This enables the provision of answers that correspond to the user's emotions, and allows for quick and appropriate support even in physical stores.

[1140] A "user" refers to a person who enters questions or information into a system.

[1141] "Means for receiving questions" refers to a part of the system that retrieves and processes questions entered by users.

[1142] A "generative model" refers to an algorithm used to generate appropriate answers based on received questions and data.

[1143] "Means for generating answers" refers to a function that uses a generative model to generate answers to received questions.

[1144] An "emotion engine" refers to an algorithm that analyzes and recognizes emotions from user input data.

[1145] "Means of recognizing emotions" refers to a function that uses an emotion engine to identify the user's emotions.

[1146] "Emotional information" refers to information about emotions recognized from user input data.

[1147] "Means of submitting responses" refers to a part of the system used to communicate generated responses to users.

[1148] "Means of displaying or playing" refers to a function that provides the generated response to the user visually or audibly in the form of audio or text.

[1149] This invention is a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions, enabling fast and emotionally responsive customer service in physical stores. The system components include a server, a terminal (smart device), an emotion engine, a generative AI model, and communication means.

[1150] System Configuration

[1151] 1. Voice input processing

[1152] Hardware: Smart glasses

[1153] Example: The general term is "smart glasses device".

[1154] Software: Speech Recognition API

[1155] Example: "Speech-to-text conversion service" as a general term.

[1156] Process: The user wears smart glasses and inputs a question by voice. The smart glasses convert the voice input into text data.

[1157] 2. Emotion Recognition Engine Processing

[1158] Software: Natural Language Processing Library

[1159] Example: "Natural language processing engine" as a general term

[1160] Processing: Text data converted from speech is input into the emotion recognition engine to identify the customer's emotion. For example, if a customer asks, "What are the features of this product?", the emotion engine identifies "interest." This allows the customer's emotion information to be input into the generative model.

[1161] 3. Data transmission process

[1162] Software: HTTP communication library

[1163] Example: "Communication Library" as a general term

[1164] Processing: Text data and sentiment information are sent to the server. The server receives this data and inputs it into a generative model to generate a response.

[1165] 4. Generative Model Processing

[1166] Hardware: Server environment

[1167] Example: "Cloud Server" as a general term

[1168] Software: Generative AI Model

[1169] Example: "Generative algorithm" as a general term

[1170] Processing: The server inputs the received text data and sentiment information into a generative model to generate an appropriate response. The generated response is expressed in a tone and content that is sensitive to the customer's emotions.

[1171] 5. Sending the response

[1172] Software: HTTP communication library

[1173] Example: "Communication Library" as a general term

[1174] Processing: The generated response is sent back to the smart glasses.

[1175] 6. Displaying the answer

[1176] Hardware: Smart glasses

[1177] Example: "Smart device" as a general term

[1178] Software: Display Management Library

[1179] Example: "Display management software" as a general term

[1180] Processing: The response is displayed as audio or text on the smart glasses worn by the customer, allowing the user to confirm it.

[1181] Specific example

[1182] In a physical store, if a customer asks a question to smart glasses by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the customer's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[1183] Example of a prompt

[1184] A customer asked the following question: "What are the features of this product?" Emotion: Interest

[1185] The generating AI model should consider the customer's interest in this question and respond as follows: "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design."

[1186] This form allows for a concrete understanding of how the invention is implemented and provides a detailed process that can be reproduced by anyone.

[1187] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1188] Step 1:

[1189] The user wears smart glasses and inputs questions by voice. The input voice data is captured through the microphone of the smart device.

[1190] Step 2:

[1191] The device sends the acquired audio data to a speech recognition API. The speech recognition API converts the audio data into text data. The input is audio data, and the output is text data.

[1192] Step 3:

[1193] The device inputs the converted text data into the emotion recognition engine. The emotion recognition engine analyzes the text data and identifies the user's emotions. The input is text data, and the output is emotion information.

[1194] Step 4:

[1195] The terminal sends text data and recognized sentiment information to the server via a communication library. The input is text data and sentiment information, and the output is the data sent to the server.

[1196] Step 5:

[1197] The server preprocesses the received text data and sentiment information. This includes data cleaning and format conversion to prepare the data for input into the generative model. The input is the received data, and the output is the preprocessed data.

[1198] Step 6:

[1199] The server inputs pre-processed data into a generative AI model, which then generates appropriate responses. The generative model analyzes the data and generates responses that take the user's emotions into consideration. The input is pre-processed data, and the output is the generated response.

[1200] Step 7:

[1201] The server receives the generated response and performs further formatting and editing as needed. The formatted response is then sent back to the terminal via the communication library. The input is the generated response, and the output is the formatted response.

[1202] Step 8:

[1203] The device displays the received response to the user using the smart glasses' display management library. Audio output may also be provided. The input is a formatted response, and the output is presented to the user as audio or text.

[1204] Specific example

[1205] For example, if a user asks the smart glasses a question by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the user's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[1206] 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.

[1207] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1208] 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.

[1209] [Fourth Embodiment]

[1210] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1211] 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.

[1212] 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).

[1213] 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.

[1214] 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.

[1215] 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).

[1216] 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.

[1217] 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.

[1218] 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.

[1219] 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.

[1220] 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.

[1221] 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.

[1222] 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".

[1223] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[1224] Program Processing Overview

[1225] 1. User input and submission of questions

[1226] The user types their question into the customer support chat window and presses the send button. This action causes the user's device to retrieve the entered text data and prepare to send it to the server.

[1227] 2. Server receives and analyzes the question.

[1228] The server receives text data sent from the terminal. The server then performs preprocessing on the received text data to input it into the generative model. This preprocessing includes text data cleansing and format conversion.

[1229] 3. Generating answers using generative models

[1230] A generative model (e.g., GPT-3) receives pre-processed text data, performs analysis, and generates an appropriate response. The generated response is then returned to the server.

[1231] 4. The server sends the response.

[1232] The server receives the generated response, reformats or corrects it if necessary, and then sends it to the terminal.

[1233] 5. Displaying responses via the device

[1234] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[1235] Specific example

[1236] Example 1: Questions about the product return policy

[1237] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[1238] The device retrieves this text data and sends it to the server.

[1239] The server receives text data, inputs it into a generative model, and performs analysis.

[1240] The generative model generates the response, "Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[1241] The server receives the generated response, formats it, and sends it to the terminal.

[1242] The device displays the answer in the user's chat window.

[1243] Users can review the displayed answers and ask further questions if necessary.

[1244] This system allows users to receive fast and consistent support 24 / 7. The use of generative AI models enables support across a wide range of topics, leading to consistent support quality and improved customer satisfaction. This, in turn, enables the efficient use of human resources in customer support.

[1245] The following describes the processing flow.

[1246] Step 1:

[1247] The user types their question into the customer support chat window and presses the submit button.

[1248] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[1249] Step 2:

[1250] The terminal retrieves the user's input and formats that text data.

[1251] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[1252] Step 3:

[1253] The terminal sends formatted text data to the server.

[1254] Specific operation: Generate an HTTP POST request and send the formatted text to the server's API endpoint.

[1255] Step 4:

[1256] The server receives text data from the terminal.

[1257] Specific operation: The server's API receives an HTTP POST request and extracts text data as input for analysis.

[1258] Step 5:

[1259] The server preprocesses the received text data before inputting it into the generative model.

[1260] Specific operations: Performs text data cleansing (removes unnecessary characters and spaces) and tokenization (breaks down words and sentences into tokens).

[1261] Step 6:

[1262] The generative model receives pre-processed text data and generates an appropriate response.

[1263] Specific operation: A generative AI model (e.g., GPT-3) analyzes the text data and generates the response text, "Our return policy allows returns within 30 days of receiving the product."

[1264] Step 7:

[1265] The server receives the generated response and formats or corrects it as needed.

[1266] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[1267] Step 8:

[1268] The server sends the formatted response to the terminal.

[1269] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[1270] Step 9:

[1271] The terminal displays the response received from the server in the user's chat window.

[1272] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[1273] Step 10:

[1274] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[1275] Specific actions: Read the response text, enter a new question if necessary, and submit again.

[1276] As a result, users will be able to receive customer support quickly and efficiently. This system is particularly capable of 24 / 7 support and provides consistently high-quality support. This leads to improved customer satisfaction and optimized human resources.

[1277] (Example 1)

[1278] 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".

[1279] Traditional customer support systems have a problem in that they struggle to provide quick and consistent answers to user inquiries. Furthermore, providing 24 / 7 support requires significant human resources, resulting in high operational costs. Additionally, the inconsistency in support quality due to different personnel handling inquiries is another challenge. It is necessary to improve this situation and achieve efficient and high-quality customer support.

[1280] 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.

[1281] In this invention, the server includes means for receiving questions entered by the user, means for cleansing the received text data and converting it into a prompt sentence, means for inputting the pre-processed text data into a generative model and generating an appropriate answer, and means for formatting the generated answer and sending it to the user. This makes it possible to provide rapid and consistent answers to user inquiries 24 hours a day. Furthermore, by using a generative AI model, it becomes possible to handle a wide range of topics, and it is expected that the uniformity of support quality and the improvement of customer satisfaction can be expected.

[1282] A "user" refers to anyone who uses this system to contact customer support.

[1283] A "question" is text data containing the content of a user's inquiry to customer support.

[1284] "Receiving" refers to the act of a server acquiring data sent from a terminal.

[1285] "Cleansing" refers to the process of removing unnecessary spaces and symbols from text data and standardizing the data format.

[1286] A "prompt statement" is a query statement generated to be input into a generative model in order to obtain a response.

[1287] A "generative model" refers to an artificial intelligence algorithm that analyzes input text data and automatically generates appropriate responses.

[1288] "Formatting" refers to the process of adding line breaks and symbols, or standardizing the writing style, to make the generated response text easier to read.

[1289] "Sending" refers to the act of the server sending back the response obtained from the generative model to the terminal.

[1290] This invention relates to a system for implementing a generative AI model for customer support. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[1291] The user enters a question into a customer support chat window and presses the submit button. At this point, the device retrieves the entered text data and prepares to send it to the server. Specifically, the device converts the entered data into JSON format and sends it to the server using an HTTP request. Hardware used includes the user's device (e.g., a PC or smartphone). Software used includes a web browser or mobile app that provides the chat window.

[1292] The server receives HTTP requests sent from terminals and extracts the text data portion. At this time, the server performs preprocessing on the received text data. This preprocessing includes text data cleansing, such as removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. The preprocessed data is then converted into prompt statements to be passed to the generating AI model. The server contains a program for receiving HTTP requests and processing text data.

[1293] The server sends the generated prompt to a generation AI model (for example, OpenAI's GPT-3). The generation model analyzes the prompt and generates an appropriate response. This generated response text is returned to the server. Specific examples of prompts include the following sentences:

[1294] "A user is asking, 'What is the return policy for this product?' Please provide an appropriate answer."

[1295] The server receives the response returned from the generative model and performs formatting if necessary, such as adding line breaks, correcting symbols, and unifying the writing style. The formatted response text is then converted back to JSON format and sent to the terminal as an HTTP response. The server includes a program to appropriately format and resend the output from the generative model.

[1296] The terminal parses the HTTP response received from the server and retrieves the response text. It can then display this retrieved response text in the user's chat window. The user can review the displayed response and ask additional questions as needed.

[1297] Specific examples are given below.

[1298] Example: Questions about the product return policy

[1299] 1. The user types "Please tell me about the return policy for this product" in the chat window.

[1300] 2. The terminal converts the entered text data into JSON format and sends an HTTP request to the server.

[1301] 3. The server receives the HTTP request, cleanses the text data, and generates a prompt message.

[1302] 4. The server sends the generated prompt message to the generation model.

[1303] 5. The generative model generates the response "Our return policy allows returns within 30 days of receiving the product" and returns it to the server.

[1304] 6. The server formats the generated response and sends it to the terminal.

[1305] 7. The device receives the response and displays it in the user's chat window.

[1306] 8. The user reviews the displayed answers. If necessary, they can enter additional questions and resubmit.

[1307] This embodiment improves the quality of customer support and enables the provision of fast and consistent service. Furthermore, by utilizing generative AI models, it becomes possible to address a wide range of topics, leading to improved customer satisfaction and more efficient use of human resources.

[1308] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1309] Step 1:

[1310] User input and submission of questions

[1311] The user enters a question into the customer support chat window and presses the submit button. The user enters the question, "Please tell me about the return policy for this product." The terminal receives this input text. Since the input is the user's text data, the terminal converts it to JSON format and sends it to the server in the format "{"question": "Please tell me about the return policy for this product"}". Specifically, it sends the data to the server using an HTTP request.

[1312] Step 2:

[1313] Server receives and analyzes questions.

[1314] The server receives an HTTP request sent from the terminal. The received data is input in JSON format, namely "{"question": "Please tell me about the return policy for this product"}". The server parses this data and extracts the text "Please tell me about the return policy for this product". After extraction, the server cleanses this text data. Specifically, this involves removing unnecessary spaces and symbols, converting full-width characters to half-width characters, and formatting using regular expressions. After processing, the pre-processed text data "Please tell me about the return policy for this product" is obtained.

[1315] Step 3:

[1316] Server-driven generation of prompt messages

[1317] Based on the pre-processed text data, "Please tell me about the return policy for this product," the server generates a prompt. This prompt is then input into the generative model. Specifically, it is generated in the format, "The user is asking, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." This prompt becomes the input for the next step.

[1318] Step 4:

[1319] Generating answers using generative models

[1320] The server sends the generated prompt to a generating AI model (e.g., GPT-3). The input is the generated prompt "The user asks, 'Please tell me about the return policy for this product.' Please provide an appropriate answer." The generating AI model analyzes this prompt and generates an appropriate answer: "Our return policy allows returns within 30 days of receiving the product." This generated answer is returned to the server, which then receives it.

[1321] Step 5:

[1322] Server formatting and sending of responses

[1323] The server receives the response "Our return policy allows returns within 30 days of receiving the product." from the generative model. The input is the response text from the generative model. If necessary, formatting is performed on the response, such as line breaks, symbol corrections, and consistency of writing style. The formatted response text is converted back into JSON format and sent to the terminal in the format "{"answer": "Our return policy allows returns within 30 days of receiving the product." The server sends the data to the terminal using an HTTP response.

[1324] Step 6:

[1325] Displaying responses via device

[1326] The terminal parses the HTTP response received from the server and obtains the response text: "Our return policy allows returns within 30 days of receiving the product." The input is the response data from the server. The terminal displays this response text in the user's chat window. The user can review the displayed response and ask additional questions if necessary. In this case, the process is repeated from step 1.

[1327] (Application Example 1)

[1328] 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".

[1329] In e-commerce sites, responding quickly and accurately to a wide variety of user inquiries is crucial for maintaining customer satisfaction. Traditionally, customer support has relied on human resources, making 24 / 7 support and immediate responses difficult. This increased the burden on departments and could lead to inconsistent support quality. This invention aims to solve these problems and achieve efficient and consistent support by using a generative AI model.

[1330] 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.

[1331] In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generative model and generating an appropriate answer, means for sending the generated answer to the user, means for cleansing the question as text data and performing format conversion, and means for formatting the generated answer and displaying it on the user's terminal. This enables a quick and accurate response to user questions, realizing 24-hour support and consistent response quality.

[1332] A "user" refers to someone who operates the system, inputs a question, and obtains an answer.

[1333] A "question" refers to text data entered by a user and analyzed by a generative model.

[1334] A "generative model" refers to AI technology that generates responses based on input text data.

[1335] "Cleansing" refers to the process of removing noise from text data and preparing it in a format that is easy to analyze.

[1336] "Format conversion" refers to the process of converting text data into a format that a generative model can understand.

[1337] "Formatting" refers to the process of converting the answers generated by a generative model into a format that is easy for the user to read.

[1338] A "terminal" refers to a device used by a user to access the system, enter questions, and check answers.

[1339] A "server" refers to a computer system that receives questions from users, generates answers using a generative model, and sends them back to the user.

[1340] A "chat interface" refers to the user interface of a system that allows users to input questions and view generated answers.

[1341] This invention provides a system for automatically answering user questions in customer support on e-commerce websites. The system works as follows: the user inputs a question, an AI model analyzes it to generate a response, and then sends the response back to the user. The detailed configuration and processing are described below.

[1342] The server includes means for receiving questions entered by the user, means for inputting the received questions into a generative model and generating appropriate answers, means for sending the generated answers to the user, means for cleansing the questions as text data and performing format conversion, and means for formatting the generated answers and displaying them on the user's terminal.

[1343] The system will utilize smartphones and servers as hardware. Users will operate the system via an application on their smartphones. The software used will include a generative AI model (OpenAI's GPT-3), a text analysis tool (NLTK Natural Language Processing Toolkit), and an application server (Node.js).

[1344] When the server receives text data sent by the user, it first cleanses the text data. Specifically, it uses NLTK to remove unnecessary spaces and symbols. Next, it formats the cleaned text data, converting it into a format that the generative AI model can understand. This is done by receiving the text data as plain text and generating appropriate prompt sentences. The generative model analyzes these prompt sentences and generates an appropriate response. The generated response is returned to the server and formatted as needed. The formatted response is sent to the user's smartphone and displayed in the chat window.

[1345] As a concrete example, a user enters the question "Please tell me about the return policy for this product" and sends it to the server from their smartphone. An example of the prompt text is shown below.

[1346] "Please provide the best answer to the following question: What is the return policy for this product?"

[1347] The server cleanses and formats this text data before inputting it into a generating AI model. This results in the output: "Our return policy allows returns within 30 days of receiving the product." This output is then formatted and displayed on the user's smartphone. This allows the user to receive quick and accurate responses and access 24 / 7 support.

[1348] In this way, the present invention makes it possible to improve the efficiency and quality of customer support on e-commerce sites.

[1349] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1350] Step 1:

[1351] The user enters a question.

[1352] The user opens a chat application on their smartphone, enters their question in text format, and presses the send button. The entered question is saved on the device as text data.

[1353] Step 2:

[1354] The terminal sends the question to the server.

[1355] The user's device sends the entered text data to the server as an HTTP request. The input data includes the user's question.

[1356] Step 3:

[1357] The server receives the question.

[1358] The server receives the HTTP request sent from the terminal and retrieves the question as text data. The received data is stored in a variable for subsequent processing.

[1359] Step 4:

[1360] The server cleanses the text data.

[1361] The server uses NLTK to cleanse the received text data, removing unnecessary spaces and symbols. The cleansed text data is then output.

[1362] Step 5:

[1363] The server converts the text data to a different format.

[1364] The cleansed text data is converted into a format that can be input into the AI ​​model. The server generates a prompt in plain text format. This prompt is as follows: "Provide the best answer to the following question: [cleansed text]".

[1365] Step 6:

[1366] The server inputs prompt messages into the AI ​​model and generates responses.

[1367] The server sends the generated prompt text as input to the generative AI model. The generative AI model (GPT-3) analyzes the prompt text and generates the appropriate response text. The generated response text is obtained as output.

[1368] Step 7:

[1369] The server formats the generated response.

[1370] The server formats the response text obtained from the generating AI model into a user-friendly format. It adds necessary line breaks and highlighting to match the specific format. The formatted text data is then output.

[1371] Step 8:

[1372] The server sends the formatted response to the user's device.

[1373] The server sends the formatted response text data as an HTTP response to the user's device. The device receives this data.

[1374] Step 9:

[1375] The device will display the answer.

[1376] The user's device displays the received response text data in the chat window. The user can review the displayed response and enter further questions if necessary.

[1377] 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.

[1378] This invention relates to a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions. The following outlines the system's program processing and provides specific examples of how this invention is implemented.

[1379] Program Processing Overview

[1380] 1. User input and submission of questions

[1381] The user types their question into the customer support chat window and presses the send button. At this point, the device retrieves the entered text data and prepares to send it to the server.

[1382] 2. Emotion recognition by the emotion engine

[1383] After the terminal receives the question entered by the user, the emotion engine analyzes the text data and recognizes the emotion. The recognized emotion information is sent to the server along with the text data.

[1384] 3. Receiving questions and sentiments from the server

[1385] The server receives text data and sentiment information sent from the terminal. The server then performs preprocessing on the received text data and sentiment information to input it into the generative model. This preprocessing includes formatting the sentiment information and cleaning the text data.

[1386] 4. Generating answers using generative models

[1387] The generative model receives pre-processed text data and sentiment information, performs analysis, and generates appropriate responses. The generated responses are expressed in a way that takes the user's emotions into consideration, both in content and tone.

[1388] 5. Server sends response

[1389] The server receives the generated response, formats and corrects it as needed, and then sends it to the terminal.

[1390] 6. Displaying responses via the device

[1391] The device displays the received response in the user's chat window. The user can review the displayed response and ask additional questions if necessary.

[1392] Specific example

[1393] Example 1: Questions about the product return policy

[1394] The user types "Please tell me about the return policy for this product" into the chat window and sends it.

[1395] The device retrieves this text data, and the emotion engine recognizes the emotion that "the user appears anxious."

[1396] The device sends text data and sentiment information to the server.

[1397] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[1398] The generative model generates the response, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and returns it to the server.

[1399] The server receives the generated response, formats it, and sends it to the terminal.

[1400] The device displays the answer in the user's chat window.

[1401] The user checks the displayed answer and feels reassured.

[1402] Example 2: A user's question expressing dissatisfaction with support.

[1403] The user types "Why is the support so slow?" and submits it.

[1404] The device acquires text data, and the emotion engine recognizes "anger."

[1405] The device sends text data and information about the emotion of anger to the server.

[1406] The server receives text data and sentiment information, inputs it into a generative model, and performs analysis.

[1407] The generative model generates the response, "We apologize for the inconvenience. We are currently working to resolve this as quickly as possible," and returns it to the server.

[1408] The server receives the generated response, formats it, and sends it to the terminal.

[1409] The device displays the answer in the user's chat window.

[1410] The user reviews the displayed answer and is somewhat convinced.

[1411] This system allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[1412] The following describes the processing flow.

[1413] Step 1:

[1414] The user types their question into the customer support chat window and presses the submit button.

[1415] Specific action: The user types "Please tell me about the return policy for this product" into the chat interface and sends the message.

[1416] Step 2:

[1417] The terminal retrieves the user's input and formats that text data.

[1418] Specific operation: The terminal's chat client catches the send event, extracts the entered text, and converts it to the appropriate format.

[1419] Step 3:

[1420] The device sends formatted text data to the emotion engine, which then recognizes the user's emotions.

[1421] Specific operation: Text data is passed to the emotion engine, which then reads emotions such as "anxiety" from the text.

[1422] Step 4:

[1423] The emotion engine returns the recognized emotion information to the device.

[1424] Specific operation: The emotion engine performs text analysis and returns the identified emotion information to the device.

[1425] Step 5:

[1426] The device sends text data and sentiment information to the server.

[1427] Specific operation: Generate an HTTP POST request containing text data and sentiment information, and send it to the server's API endpoint.

[1428] Step 6:

[1429] The server receives text data and sentiment information from the terminal.

[1430] Specific operation: The server's API receives an HTTP POST request and extracts text data and sentiment information as input for analysis.

[1431] Step 7:

[1432] The server preprocesses the received text data and sentiment information.

[1433] Specific operations: Cleans text data (removes unnecessary characters and spaces), tokenizes (breaks down words and sentences into tokens), and converts the format of sentiment information.

[1434] Step 8:

[1435] The generative model receives pre-processed text data and sentiment information and generates an appropriate response.

[1436] Specific operation: A generative AI model (e.g., GPT-3) analyzes the input data and generates response text that takes sentiment into account.

[1437] Step 9:

[1438] The server receives the generated response and formats or corrects it as needed.

[1439] Specific operation: Analyzes the output text from the generative AI model and corrects typos and grammatical errors, as well as formatting the writing style.

[1440] Step 10:

[1441] The server sends the formatted response to the terminal.

[1442] Specific operation: Generate an HTTP response containing the formatted response text and send it to the terminal.

[1443] Step 11:

[1444] The terminal displays the response received from the server in the user's chat window.

[1445] Specific operation: The device receives an HTTP response and displays the response text from the response body in the chat window.

[1446] Step 12:

[1447] The user reviews the displayed answer and decides whether they are satisfied or want to ask further questions.

[1448] Specific operation: The user reads the displayed answer text, enters a new question if necessary, and resubmits.

[1449] This allows users to receive 24 / 7 responses that take their emotions into consideration. By utilizing an emotion engine, the content and tone of responses are adjusted according to the user's emotional state, providing more helpful, consistent, and high-quality support. This leads to improved customer satisfaction and more efficient use of human resources.

[1450] (Example 2)

[1451] 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".

[1452] Traditional customer support systems struggled to accurately recognize and respond to user emotions, sometimes resulting in decreased user satisfaction and delays in problem resolution. Furthermore, efficiently providing 24 / 7 support placed a significant burden on human resources.

[1453] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1454] In this invention, the server includes means for receiving a question entered by a user, means for generating emotional information from the received question using an emotional analysis engine, means for inputting the question along with the generated emotional information into a generation model to generate an appropriate answer, means for formatting the generated answer in a way that takes the user's emotions into consideration, and means for sending the formatted answer to the user. This makes it possible to provide high-quality support that takes the user's emotions into consideration 24 hours a day.

[1455] "Means for receiving user-entered questions" refers to a mechanism for acquiring text data entered by users through customer support chat windows or other interfaces and sending it to the server.

[1456] "Means for generating emotional information using an emotion analysis engine" refers to a mechanism that analyzes text data entered by a user, recognizes the user's emotional state from its content, and generates appropriate emotional information.

[1457] "A means of inputting questions along with generated sentiment information into a generative model to generate appropriate answers" refers to a mechanism for inputting recognized sentiment information and user questions into a generative model-trained generative algorithm to generate appropriate answers that take the user's emotions into consideration.

[1458] "Methods for shaping generated responses to reflect user emotions" refers to a mechanism that processes responses generated by a generative model into a format and phrasing that aligns with the user's emotions, making them easy to understand and considerate for the user.

[1459] "Means for sending formatted responses to users" refers to a mechanism for sending formatted responses from a server to the user's terminal and displaying them in a chat window or other interface.

[1460] This system combines a generative AI model and an emotion analysis engine to generate responses that take user emotions into consideration, in order to provide effective customer support. The specific configuration and operation of the system are described below.

[1461] Hardware and software to use

[1462] Hardware:

[1463] Server: Cloud service (e.g., AWS EC2)

[1464] Devices: User devices such as PCs, smartphones, and tablets.

[1465] software:

[1466] Chat window platform: Customer support system (e.g., Intercom, Zendesk)

[1467] Sentiment analysis engine: Text analysis library (e.g., Affective Computing SDK)

[1468] Generative AI models: Natural language generation models (e.g., OpenAI GPT-3)

[1469] System response processing

[1470] The server first has a component that receives questions entered by the user. The user types their question in the chat window and presses the send button. For example, if the user types "Please tell me about the return policy for this product," this text data is captured by the terminal. The terminal then prepares to send this text data to the server.

[1471] The server also has a component that analyzes received questions using an emotion analysis engine and generates emotion information. The emotion analysis engine analyzes the text entered by the user and generates emotion information. For example, if it recognizes the emotion of "anxiety" from the user's statement, it will generate emotion information for "anxiety".

[1472] Next, the server inputs the question, along with the generated sentiment information, into the generating AI model to produce an appropriate answer. The generating model receives the pre-processed data, analyzes it, and generates the answer. This answer will have content and tone that takes the user's emotions into consideration. For example, it might generate an answer such as, "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[1473] The generated response is then formatted to reflect the user's emotions and tone. This process makes the response more acceptable to the user. The formatted response is then sent back to the terminal by the server.

[1474] The device displays the received response in the chat window. The user can review the displayed response and determine whether the problem has been resolved.

[1475] Specific examples of operation

[1476] The following is a concrete example of actual system operation.

[1477] Example of a prompt:

[1478] 1. "Please tell me about the return policy for this product."

[1479] 2. "Why is the support so slow?"

[1480] For example, if a user types "Please tell me about the return policy for this product," the device receives this input, and its sentiment analysis engine recognizes "anxiety." The server then inputs this text data and sentiment information into a generating AI model, which generates and formats a response such as "Don't worry, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product," and sends it back to the device. The device then displays this response in the user's chat window, allowing the user to see it and feel reassured.

[1481] In this way, the system provides high-quality customer support that takes user emotions into consideration, improving user satisfaction and streamlining support operations. Furthermore, it enables 24 / 7 support, contributing to a reduction in the burden on human resources.

[1482] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1483] Step 1:

[1484] The user types their question into the customer support chat window and presses the send button. This is then sent to the device as text data.

[1485] Input: Text data entered by the user in the chat window

[1486] Output: Text data is sent to the terminal.

[1487] Specific action: The user types "Please tell me about the return policy for this product" into the chat window and presses the send button.

[1488] Step 2:

[1489] The terminal receives the input text data and stores it in memory. It also prepares the text data for transmission.

[1490] Input: Text data entered by the user

[1491] Output: Text data stored within the device

[1492] Specific operation: The device detects user input and saves that text data to temporary storage.

[1493] Step 3:

[1494] The device uses an emotion analysis engine to analyze text data. This analysis extracts user emotion information from the text data.

[1495] Input: Text data stored on the device

[1496] Output: Sentiment information generated based on text data

[1497] Specific operation: The device's emotion analysis engine analyzes the text data and recognizes that the user is feeling "anxious".

[1498] Step 4:

[1499] The device sends text data and sentiment information to the server.

[1500] Input: Text data and sentiment information

[1501] Output: Text data and sentiment information sent to the server

[1502] Specific operation: The terminal sends text data and sentiment information to the server in packet format.

[1503] Step 5:

[1504] The server receives text data and sentiment information sent from the terminal and performs preprocessing. Preprocessing includes text data cleansing and sentiment information format conversion.

[1505] Input: Text data and sentiment information received from the device.

[1506] Output: Preprocessed text data and sentiment information

[1507] Specific operation: The server cleanses the received data and converts it to the required format.

[1508] Step 6:

[1509] The server inputs pre-processed text data and sentiment information into a generating AI model to generate responses.

[1510] Input: Preprocessed text data and sentiment information

[1511] Output: Answers generated by the generative AI model

[1512] Specific operation: The server inputs data into the generated AI model and generates the response, "Rest assured, we have a money-back guarantee. Our return policy allows returns within 30 days of receiving the product."

[1513] Step 7:

[1514] The server receives the generated responses and performs formatting and necessary corrections. This formatting ensures the consistency of the responses.

[1515] Input: Answers generated by the generative AI model

[1516] Output: Formatted answer

[1517] Specific operation: The server inspects the response and corrects grammar and tone.

[1518] Step 8:

[1519] The server sends the formatted response to the terminal.

[1520] Input: Formatted answer

[1521] Output: Formatted response sent to the terminal

[1522] Specific operation: The server sends the formatted response to the terminal.

[1523] Step 9:

[1524] The device displays the received response in the user's chat window.

[1525] Input: Formatted response received from the server

[1526] Output: Answers displayed in the chat window

[1527] Specific operation: The device displays the received response in the chat window and notifies the user.

[1528] Step 10:

[1529] The user reviews the displayed answers and asks additional questions as needed.

[1530] Input: Answer displayed in the chat window

[1531] Output: Additional questions (if necessary)

[1532] Specific action: The user reads the displayed answer and enters whether they are satisfied or ask further questions.

[1533] (Application Example 2)

[1534] 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".

[1535] Traditional customer support systems often struggle to respond to user emotions and provide only uniform answers, which can lead to decreased customer satisfaction. Furthermore, in physical stores, situations may arise where staff are unavailable, resulting in customers not receiving prompt and appropriate support.

[1536] 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. In this invention, the server includes means for receiving a question entered by the user, means for inputting the received question into a generation model and generating an appropriate answer, means for sending the generated answer to the user, means for recognizing the user's emotions using an emotion engine, means for inputting the recognized emotion information into the generation model and generating an answer, and means for displaying or playing the generated answer in voice or text. This enables the provision of answers that correspond to the user's emotions, and allows for quick and appropriate support even in physical stores.

[1537] A "user" refers to a person who enters questions or information into a system.

[1538] "Means for receiving questions" refers to a part of the system that retrieves and processes questions entered by users.

[1539] A "generative model" refers to an algorithm used to generate appropriate answers based on received questions and data.

[1540] "Means for generating answers" refers to a function that uses a generative model to generate answers to received questions.

[1541] An "emotion engine" refers to an algorithm that analyzes and recognizes emotions from user input data.

[1542] "Means of recognizing emotions" refers to a function that uses an emotion engine to identify the user's emotions.

[1543] "Emotional information" refers to information about emotions recognized from user input data.

[1544] "Means of submitting responses" refers to a part of the system used to communicate generated responses to users.

[1545] "Means of displaying or playing" refers to a function that provides the generated response to the user visually or audibly in the form of audio or text.

[1546] This invention is a system that combines a generative AI model for customer support with an emotion engine that recognizes user emotions, enabling fast and emotionally responsive customer service in physical stores. The system components include a server, a terminal (smart device), an emotion engine, a generative AI model, and communication means.

[1547] System Configuration

[1548] 1. Voice input processing

[1549] Hardware: Smart glasses

[1550] Example: The general term is "smart glasses device".

[1551] Software: Speech Recognition API

[1552] Example: "Speech-to-text conversion service" as a general term.

[1553] Process: The user wears smart glasses and inputs a question by voice. The smart glasses convert the voice input into text data.

[1554] 2. Emotion Recognition Engine Processing

[1555] Software: Natural Language Processing Library

[1556] Example: "Natural language processing engine" as a general term

[1557] Processing: Text data converted from speech is input into the emotion recognition engine to identify the customer's emotion. For example, if a customer asks, "What are the features of this product?", the emotion engine identifies "interest." This allows the customer's emotion information to be input into the generative model.

[1558] 3. Data transmission process

[1559] Software: HTTP communication library

[1560] Example: "Communication Library" as a general term

[1561] Processing: Text data and sentiment information are sent to the server. The server receives this data and inputs it into a generative model to generate a response.

[1562] 4. Generative Model Processing

[1563] Hardware: Server environment

[1564] Example: "Cloud Server" as a general term

[1565] Software: Generative AI Model

[1566] Example: "Generative algorithm" as a general term

[1567] Processing: The server inputs the received text data and sentiment information into a generative model to generate an appropriate response. The generated response is expressed in a tone and content that is sensitive to the customer's emotions.

[1568] 5. Sending the response

[1569] Software: HTTP communication library

[1570] Example: "Communication Library" as a general term

[1571] Processing: The generated response is sent back to the smart glasses.

[1572] 6. Displaying the answer

[1573] Hardware: Smart glasses

[1574] Example: "Smart device" as a general term

[1575] Software: Display Management Library

[1576] Example: "Display management software" as a general term

[1577] Processing: The response is displayed as audio or text on the smart glasses worn by the customer, allowing the user to confirm it.

[1578] Specific example

[1579] In a physical store, if a customer asks a question to smart glasses by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the customer's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[1580] Example of a prompt

[1581] A customer asked the following question: "What are the features of this product?" Emotion: Interest

[1582] The generating AI model should consider the customer's interest in this question and respond as follows: "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design."

[1583] This form allows for a concrete understanding of how the invention is implemented and provides a detailed process that can be reproduced by anyone.

[1584] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1585] Step 1:

[1586] The user wears smart glasses and inputs questions by voice. The input voice data is captured through the microphone of the smart device.

[1587] Step 2:

[1588] The device sends the acquired audio data to a speech recognition API. The speech recognition API converts the audio data into text data. The input is audio data, and the output is text data.

[1589] Step 3:

[1590] The device inputs the converted text data into the emotion recognition engine. The emotion recognition engine analyzes the text data and identifies the user's emotions. The input is text data, and the output is emotion information.

[1591] Step 4:

[1592] The terminal sends text data and recognized sentiment information to the server via a communication library. The input is text data and sentiment information, and the output is the data sent to the server.

[1593] Step 5:

[1594] The server preprocesses the received text data and sentiment information. This includes data cleaning and format conversion to prepare the data for input into the generative model. The input is the received data, and the output is the preprocessed data.

[1595] Step 6:

[1596] The server inputs pre-processed data into a generative AI model, which then generates appropriate responses. The generative model analyzes the data and generates responses that take the user's emotions into consideration. The input is pre-processed data, and the output is the generated response.

[1597] Step 7:

[1598] The server receives the generated response and performs further formatting and editing as needed. The formatted response is then sent back to the terminal via the communication library. The input is the generated response, and the output is the formatted response.

[1599] Step 8:

[1600] The device displays the received response to the user using the smart glasses' display management library. Audio output may also be provided. The input is a formatted response, and the output is presented to the user as audio or text.

[1601] Specific example

[1602] For example, if a user asks the smart glasses a question by voice, such as "What are the features of this product?", the voice recognition API converts the voice into text, and the emotion engine recognizes the user's "interest." The text data and emotion information are sent to the server via a communication library, and the generative model generates a response such as, "This product uses the latest technology and combines multi-functionality with ease of use. Furthermore, it features a long-lasting battery and a stylish design." The generated response is then sent back to the smart glasses and displayed on the screen.

[1603] 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.

[1604] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1605] 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.

[1606] 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.

[1607] 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.

[1608] 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.

[1609] 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.

[1610] 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.

[1611] 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."

[1612] 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.

[1613] 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.

[1614] 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.

[1615] 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.

[1616] 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.

[1617] 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.

[1618] 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.

[1619] 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.

[1620] 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.

[1621] 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.

[1622] 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.

[1623] 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.

[1624] The following is further disclosed regarding the embodiments described above.

[1625] (Claim 1)

[1626] A means of receiving questions entered by the user,

[1627] A means of inputting a received question into a generative model and generating an appropriate answer,

[1628] A means of sending the generated response to the user,

[1629] A system that includes this.

[1630] (Claim 2)

[1631] The system according to claim 1, which receives user questions 24 hours a day and generates answers.

[1632] (Claim 3)

[1633] The system according to claim 1, which displays the generated response in a chat interface.

[1634] "Example 1"

[1635] (Claim 1)

[1636] A means of receiving questions entered by the user,

[1637] A means of cleansing the received text data and converting it into a prompt message,

[1638] A means for inputting preprocessed text data into a generative model and generating an appropriate response,

[1639] A means of formatting the generated response and sending it to the user,

[1640] A system that includes this.

[1641] (Claim 2)

[1642] The system according to claim 1, which receives user questions 24 hours a day and generates answers.

[1643] (Claim 3)

[1644] The system according to claim 1, which displays the generated response in a chat interface.

[1645] "Application Example 1"

[1646] (Claim 1)

[1647] A means of receiving questions entered by the user,

[1648] A means of inputting a received question into a generative model and generating an appropriate answer,

[1649] A means of sending the generated response to the user,

[1650] A method for cleansing the questions as text data and converting the format,

[1651] A means of formatting the generated response and displaying it on the user's device,

[1652] A system that includes this.

[1653] (Claim 2)

[1654] The system according to claim 1, which receives user questions 24 hours a day and generates answers.

[1655] (Claim 3)

[1656] The system according to claim 1, which displays the generated response in a chat interface.

[1657] "Example 2 of combining an emotion engine"

[1658] (Claim 1)

[1659] A means of receiving questions entered by the user,

[1660] A means for generating emotional information from received questions using an emotional analysis engine,

[1661] A means of inputting a question along with generated sentiment information into a generative model to generate an appropriate answer,

[1662] A method for formatting the generated responses to reflect the user's emotions and tone,

[1663] A means of sending formatted answers to users,

[1664] A system that includes this.

[1665] (Claim 2)

[1666] The system according to claim 1, which receives user questions 24 hours a day and generates answers.

[1667] (Claim 3)

[1668] The system according to claim 1, which displays the generated response in a chat interface.

[1669] "Application example 2 when combining with an emotional engine"

[1670] (Claim 1)

[1671] A means of receiving questions entered by the user,

[1672] A means of inputting a received question into a generative model and generating an appropriate answer,

[1673] A means of sending the generated response to the user,

[1674] A means of recognizing the user's emotions using an emotion engine,

[1675] A means for inputting recognized emotion information into the generative model to generate a response,

[1676] A means for displaying or playing the generated response in audio or text,

[1677] A system that includes this.

[1678] (Claim 2)

[1679] The system according to claim 1, which receives questions from users 24 hours a day and generates answers based on the content of the questions and recognized sentiment information.

[1680] (Claim 3)

[1681] The system according to claim 1, which displays the generated response on a chat interface or the display of a smart device. [Explanation of Symbols]

[1682] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of receiving questions entered by the user, A means of inputting a received question into a generative model and generating an appropriate answer, A means of sending the generated response to the user, A system that includes this.

2. The system according to claim 1, which receives user questions 24 hours a day and generates answers.

3. The system according to claim 1, which displays the generated response in a chat interface.

Images