System

Abstract

A system is provided.SOLUTION: A system for performing a simulated conversation of customer serving using generative artificial intelligence, comprising: means for storing a serving procedure and dummy customer information; means for generating a question based on the dummy customer information using the generative artificial intelligence; means for matching a response input by a trainee with the serving procedure and performing an evaluation; and means for providing feedback to the trainee based on the evaluation.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In the past, it was difficult to provide efficient, high-quality customer service training to a large number of trainees. Especially during busy periods, the number of inquiries increases several times over normal times, often resulting in a lack of resources for training trainees. Furthermore, training new operators requires a lot of time and effort, and there is a significant individual difference in customer service quality. This invention aims to solve these problems and provide efficient, high-quality customer service training. [Means for solving the problem]

[0005] The present invention provides a system for conducting simulated customer service conversations using generative artificial intelligence. The system includes a means for storing response procedures and dummy customer information, a means for generating questions based on the dummy customer information using generative artificial intelligence, a means for comparing responses entered by trainees with the response procedures and evaluating them, and a means for providing feedback to the trainees based on the evaluation. This allows trainees to efficiently learn appropriate response methods and prevents declines in response quality due to resource shortages during busy periods or individual differences. Furthermore, the system supports trainees' skill improvement by including a means for pointing out the correct response method and suggesting corrections using a display means in the event of an inappropriate response. The system also includes a means for generating dummy customer information and scenario data and transmitting them to the trainee's device, enabling training in an environment similar to actual customer service.

[0006] "Generative AI" is an AI technology that generates natural language based on input data and engages in dialogue and responses.

[0007] "Customer service" is a business process in which a company or organization responds to inquiries and requests from customers.

[0008] A "mock conversation" is a simulated conversation that imitates actual customer interactions, and is used for training and evaluation purposes.

[0009] "Procedures" are a set of standard operating procedures and guidelines to be followed when responding to customer inquiries and requests.

[0010] "Dummy customer information" is information about fictitious customers used for simulated conversations, including names, addresses, and past transaction history.

[0011] "Scenario data" is data that defines the flow of a series of questions and responses used in a simulated conversation.

[0012] The "means of evaluation" refers to a device or software that has the function of analyzing the trainee's responses and comparing them with response procedures and FAQ data to determine whether they are correct or incorrect.

[0013] A "means for providing feedback" is a device or software that has the function of presenting comments and areas for improvement to trainees based on the evaluation results.

[0014] A "display means" is a device such as a monitor, display, or projector used to visually display information.

[0015] The "means for suggesting corrections" refers to a device or software that has the function of providing a correction suggestion when an incorrect response is made and presenting the trainee with an appropriate response method.

[0016] "Devices" are electronic devices used by trainees, such as computers, tablets, and smartphones. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0018] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0019] First, the terms used in the following description will be explained.

[0020] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0021] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0022] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0023] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0024] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0025] [First embodiment]

[0026] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0027] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0028] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0029] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0030] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0031] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0032] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0033] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0034] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0035] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0036] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0037] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0038] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide efficient and high-quality training. This system is composed of a server, terminals, and users.

[0039] The server hosts the data and applications necessary for the training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, etc. The server is equipped with a generative artificial intelligence that plays the role of a customer and generates mock conversations.

[0040] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also displays feedback and evaluation results sent from the server to the trainee.

[0041] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[0042] As a specific example, the flow of a user logging in to a system will be described.

[0043] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0044] 2. The terminal sends the entered authentication information to the server.

[0045] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[0046] Next, the flow of role-playing is shown below.

[0047] 1. The user selects "Start Role-Playing" from the main menu.

[0048] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[0049] 3. The terminal displays the interactive screen.

[0050] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[0051] 5. The user enters and submits a response to the question.

[0052] 6. The terminal sends the entered response to the server,

[0053] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[0054] 7. The server generates the evaluation results and generates appropriate feedback messages.

[0055] 8. The terminal displays the feedback message on the interactive screen.

[0056] By using this system, trainees can learn in an environment that is similar to actual customer service, allowing them to efficiently master appropriate customer service techniques. In addition, if a trainee makes a mistake, the system will immediately suggest the correct way to respond, improving the quality of the training.

[0057] Furthermore, the system has the ability to generate dummy customer information and scenario data and send them to trainees' devices, allowing for practical learning that can handle a variety of scenarios. This system helps prevent a decline in customer service quality due to resource shortages during busy periods and individual differences.

[0058] The above is an embodiment of the present invention.

[0059] The processing flow will be explained below.

[0060] Login Process

[0061] Step 1:

[0062] The user displays the login screen for the system on the terminal and enters their ID and password.

[0063] Step 2:

[0064] The terminal transmits the entered authentication information to the server.

[0065] Step 3:

[0066] The server compares the received authentication information with a database and generates an authentication result.

[0067] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[0068] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[0069] Step 4:

[0070] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[0071] Role-playing begins

[0072] Step 1:

[0073] The user clicks the "Start Role-Playing" button from the main menu.

[0074] Step 2:

[0075] The device notifies the server of the click event.

[0076] Step 3:

[0077] The server generates dummy customer information and scenario data and transmits them to the terminal.

[0078] The dummy customer information includes names, addresses, and past donation details.

[0079] The scenario data includes specific conversation flows and questions.

[0080] Step 4:

[0081] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[0082] Mock conversation progression

[0083] Step 1:

[0084] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[0085] Step 2:

[0086] The terminal displays the generated question on an interactive screen.

[0087] Step 3:

[0088] The user enters a response to the displayed question and clicks the submit button.

[0089] Step 4:

[0090] The terminal sends the entered response to the server.

[0091] Step 5:

[0092] The server analyzes the received response and matches it with procedures and FAQ data.

[0093] Step 6:

[0094] The server generates an evaluation result based on the response content and generates an appropriate feedback message.

[0095] Step 7:

[0096] The terminal displays the feedback message on the interactive screen.

[0097] Providing feedback

[0098] Step 1:

[0099] The terminal receives and displays the feedback message sent from the server.

[0100] The feedback includes a pointer to what went wrong and the correct way to respond.

[0101] Step 2:

[0102] The user checks the feedback and inputs a revised response.

[0103] Step 3:

[0104] The terminal retransmits the corrected response to the server.

[0105] Step 4:

[0106] The server again analyzes the response and evaluates whether it is an appropriate response.

[0107] If appropriate, the following scenarios proceed:

[0108] If it's not appropriate, provide feedback again.

[0109] Role-playing ends

[0110] Step 1:

[0111] The user can complete all scenarios or click the "Exit" button midway through.

[0112] Step 2:

[0113] The terminal notifies the server of the termination operation.

[0114] Step 3:

[0115] The server records the user's progress data and evaluation results in a database and ends the session.

[0116] Step 4:

[0117] The terminal displays a logout screen to notify the user that the session has ended.

[0118] Example 1

[0119] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0120] Conventional customer service training systems have struggled to provide trainees with realistic mock conversations to help them acquire practical customer service skills. Additionally, feedback is often delayed, and it's difficult to suggest appropriate corrections in real time. As a result, the quality of training declines, and trainees may lack customer service skills in their actual work.

[0121] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0122] In this invention, the server includes means for storing training data and virtual customer data, means for generating dialogue based on the virtual customer data using generative artificial intelligence, means for comparing responses input by the trainee with training procedures and evaluating them, and means for providing feedback to the trainee based on the evaluation. This enables the trainee to acquire practical conversation skills through realistic simulated conversations, and by providing appropriate, real-time feedback and suggestions for correction, the quality of the training can be improved.

[0123] - "Training purpose" refers to the purpose for trainees to acquire response skills and knowledge.

[0124] "Generative AI" is a type of AI that performs natural language processing and dialogue generation.

[0125] "Simulating a dialogue" means recreating an actual conversation environment and conducting training through virtual conversations.

[0126] A "system" is a device that integrates multiple elements and functions and is designed to achieve a specific purpose.

[0127] "Training Data" is information regarding guidelines and procedures used during training.

[0128] "Virtual customer data" refers to information about fictitious customers used in simulated conversations.

[0129] A "means" is a method or device used to achieve a particular purpose.

[0130] A "Trainer" is an individual who uses the system to improve their customer service skills.

[0131] "Training procedures" are procedures or guidelines to be followed when dealing with customers.

[0132] "Evaluating" means judging the appropriateness of the trainee's response.

[0133] "Feedback" refers to information about the trainee's evaluation of their response and areas for improvement.

[0134] "Dialogue based on virtual customer data" is a conversation generated based on fictitious customer information.

[0135] "Suggesting corrections" means suggesting procedures or methods for improvement when a response is inappropriate.

[0136] A "terminal" is a device used to access the system.

[0137] This invention is a training system that uses a generative AI model to conduct simulated customer service conversations, and aims to provide efficient and high-quality training. This system is composed of a server, a terminal, and a user.

[0138] The server hosts the data and applications required for training. Specifically, training data, virtual customer data, an FAQ database, and other items are stored on the server. The server is equipped with a generative AI model, which acts as the virtual customer and generates simulated conversations. Specific examples of servers include high-performance server machines (e.g., highly reliable server equipment, cloud computing platforms). Examples of software include the Linux OS, a generative AI model (e.g., an AI model that performs natural language processing), and a database management system (e.g., MySQL).

[0139] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also has the role of displaying feedback and evaluation results sent from the server to the trainee. A browser (e.g., a standard web browser) or a communication application (e.g., a communication app using the WebSocket protocol) is used.

[0140] The users are trainees and system administrators. Trainees log in to the system using a terminal and receive training in customer service through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[0141] As a concrete example, the flow of a user logging in to the system is shown below.

[0142] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0143] 2. The terminal sends the entered authentication information to the server.

[0144] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[0145] Next, the flow of role-playing is shown below.

[0146] 1. The user selects "Start Role-Playing" from the main menu.

[0147] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[0148] 3. The terminal displays the interactive screen.

[0149] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[0150] 5. The user enters and submits a response to the question.

[0151] 6. The terminal sends the entered response to the server.

[0152] 7. The server analyzes the response and compares it with pre-defined training procedures and FAQ data.

[0153] 8. The server generates the evaluation results and generates appropriate feedback messages.

[0154] 9. The terminal displays the feedback message on the interactive screen.

[0155] As a concrete example, the prompt sentence is shown below.

[0156] Prompt statement:

[0157] "My phone bill seems unusually high lately. How does it compare to my last bill?"

[0158] Using this system, trainees can learn in an environment that closely resembles actual customer interactions, allowing them to efficiently master customer interaction methods. In addition, if an incorrect interaction occurs, the correct interaction method is immediately displayed, improving the quality of training. Furthermore, the system generates virtual customer data and scenario data, allowing for practical learning that corresponds to a variety of scenarios. This system can prevent a decline in interaction quality due to resource shortages during busy periods or individual differences.

[0159] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0160] Explain the processing flow of the system program by dividing it into processing steps

[0161] Step 1: User Login

[0162] Input: User ID and password

[0163] Output: Authentication result (success / failure)

[0164] What happens:

[0165] 1. The user uses the terminal to display the system login screen and enters their ID and password.

[0166] Specific operation: Open the login page in your browser and enter the ID "user01" and password "password123" in the form.

[0167] 2. The terminal sends the entered authentication information to the server.

[0168] What it does: Sends authentication information in an HTTP POST request.

[0169] 3. The server checks the received authentication information against its database.

[0170] Specific behavior: Performs a database query to verify the ID and password.

[0171] 4. The server generates an authentication result and sends it to the terminal.

[0172] Specific operation: If authentication is successful, generate and send a token; if it fails, send an error message "Authentication failed."

[0173] Step 2: Prepare to start role-playing

[0174] Input: Request to start roleplaying

[0175] Output: Virtual customer profile and scenario data

[0176] What happens:

[0177] 1. The user selects "Start Role-Playing" from the main menu.

[0178] Specific action: Click a button on the menu.

[0179] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[0180] Specific operation: Generate a virtual customer name "Customer A" and a scenario "Regarding bill payment" and send them to the terminal in JSON format.

[0181] 3. The terminal displays the interactive screen.

[0182] Specific operation: Display the received profile and scenario data on the screen.

[0183] Step 3: Role-playing dialogue

[0184] Input: User response

[0185] Output: The next question generated by the generative AI

[0186] What happens:

[0187] 1. The generation AI (server) generates an initial question based on the scenario and sends it to the device.

[0188] Specific behavior: Generate and send the question "Hello, this is Customer A. I'm having trouble paying my bill."

[0189] 2. The user enters and submits a response to the question.

[0190] Specific action: Enter the response "Please tell me about the specific problem" and click the send button.

[0191] 3. The terminal sends the entered response to the server.

[0192] Specific operation: The response content is sent to the server via an HTTP POST request.

[0193] 4. The server analyzes the response and compares it with the configured training procedures and FAQ data.

[0194] Specific behavior: Performs text analysis and searches for relevant entries in the FAQ database.

[0195] Step 4: Feedback and Rating

[0196] Input: User responses and analysis results

[0197] Output: Feedback message

[0198] What happens:

[0199] 1. The server generates the evaluation results and generates a feedback message.

[0200] What it does: Evaluate the response and generate feedback like "Your response was unclear. Please check the amount next time."

[0201] 2. The terminal displays the feedback message on the interactive screen.

[0202] What it does: Displays the feedback received on the screen.

[0203] Step 5: Logout

[0204] Input: Logout request

[0205] Output: Logout result, login screen displayed

[0206] What happens:

[0207] 1. The user selects logout from the menu.

[0208] Specific action: Click the "Logout" button on the main menu.

[0209] 2. The device sends a logout request to the server.

[0210] Specific behavior: Sends a logout request as an HTTP POST request.

[0211] 3. The server invalidates the session and redirects to the login screen.

[0212] Specific actions: End the session and send the URL of the login screen to the terminal.

[0213] The above are the specific processing steps of this system.

[0214] (Application example 1)

[0215] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0216] Current customer service training programs struggle to provide a practical training environment that closely resembles actual customer interactions, making it difficult for trainees to efficiently acquire the skills to deal with the diverse situations they will encounter in the field. Furthermore, current systems often lack real-time feedback and evaluation of trainees' customer service quality, limiting the effectiveness of training. In particular, there are few systems that support training while on the move or in a brick-and-mortar store, making it difficult to provide realistic training.

[0217] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0218] In this invention, the server includes means for generating questions based on the dummy customer information using a generative AI, means for comparing responses entered by the trainee with the service procedures and evaluating them, means for providing feedback to the trainee based on the evaluation, means for providing a simulation environment for the trainee using smart glasses, means for analyzing voice input from the smart glasses and transmitting it to the generative AI model, and means for displaying feedback received from the server on the display of the smart glasses. This enables the trainee to effectively acquire customer service skills while receiving evaluation and feedback in real time in an immersive environment close to reality.

[0219] "Generative AI" refers to AI technology that uses natural language processing technology to generate human-like conversations and content.

[0220] "Customer service" refers to the business process of responding to inquiries and requests from customers.

[0221] "Simulated conversation" refers to fictitious dialogue generated to simulate real conversations.

[0222] "Service procedures" refer to standard procedures and guidelines to be followed when dealing with customers.

[0223] "Dummy customer information" refers to fictitious customer profiles and scenario information used in training.

[0224] "Evaluation" refers to the process of judging the quality and appropriateness of the trainee's response and giving them a score and feedback.

[0225] "Feedback" refers to information provided to trainees, including areas for improvement and appropriate guidance.

[0226] "Smart glasses" are glasses with an integrated display that are devices capable of displaying information and inputting and outputting voice.

[0227] "Simulation environment" refers to a virtual environment in which trainees can practice in situations similar to their actual work.

[0228] "Generative AI model" refers to an artificial intelligence model that has been pre-trained to perform generative tasks.

[0229] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide real-time feedback to trainees in brick-and-mortar stores using smart glasses.

[0230] Hardware used

[0231] Smart Glasses

[0232] These are glasses with an integrated display that can display information and input and output voice. Specific examples include Google Glass and Vuzix Blade.

[0233] server

[0234] Hosting generative artificial intelligence and databases for real-time data processing. Specific examples include AWS EC2 instances.

[0235] Software used

[0236] Generative Artificial Intelligence

[0237] It uses natural language processing technology to generate simulated conversations in real time, a specific example being OpenAI GPT-4.

[0238] Database

[0239] Dummy customer information and scenario data are stored and managed. A specific example is AWS RDS (MySQL).

[0240] Real-time communication

[0241] A technology for data communication between a server and smart glasses. A specific example is WebSocket communication.

[0242] System configuration and data processing

[0243] server

[0244] 1. Generate dummy customer information

[0245] The server generates dummy customer information and scenario data using the generative AI model, using the following prompt sentences for this process:

[0246] Sample prompt 1: Generate a customer profile

[0247] Generate a customer profile containing the following information:

[0248] name

[0249] age

[0250] sex

[0251] Purchase Intent

[0252] Customer questions and concerns

[0253] 2. Mock conversation scenario generation

[0254] The server generates questions based on the scenario data and uses the following prompts to make the simulation feel more realistic for the trainee:

[0255] Prompt example 2: Training scenario generation

[0256] Train your customer service representatives using the following scenarios:

[0257] A customer asks for more information about a product.

[0258] A customer asks about returns or exchanges.

[0259] Customers want to know about specific promotions and offers.

[0260] 3. Ratings and Feedback

[0261] The server analyzes the trainee's responses, compares them with the response procedures, and evaluates them. The evaluation results and feedback are generated in real time and displayed on the smart glasses.

[0262] Smart Glasses

[0263] 1. Real-time interaction

[0264] The smart glasses receive scenario data from the server and display it on the screen. They also analyze the trainee's voice input and send it to the server.

[0265] 2. Feedback display

[0266] Feedback received from the server is displayed in real time, allowing trainees to quickly learn how to respond appropriately.

[0267] A concrete example of the entire system

[0268] 1. Part of a training scenario

[0269] For example, a dialogue about a "question about product description" is simulated. The following flow is an example.

[0270] Dummy customer question: "What is the battery life of this product?"

[0271] Trainee response: "About 10 hours."

[0272] Generative AI evaluation: "That's a good answer, but it depends on the specific use case, so it would be good to explain that it varies depending on the use case."

[0273] Using this system, trainees can effectively acquire customer service skills in a realistic environment, receiving real-time evaluations and feedback.

[0274] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0275] Step 1:

[0276] The user puts on the smart glasses and logs into the system.

[0277] Input: User ID and password

[0278] Data processing / calculation: The authentication information is sent to the server and checked against a database.

[0279] Output: The authentication result is sent from the server to the smart glasses.

[0280] Specific operation: If authentication is successful, the server displays the main menu on the smart glasses display. If authentication is unsuccessful, it displays an error message.

[0281] Step 2:

[0282] The user selects "Start Training" from the main menu.

[0283] Input: Select Start Training on the main menu

[0284] Data processing / calculation: The server generates dummy customer information and scenario data using the generative AI model.

[0285] Output: The generated dummy customer information and scenario data are sent to the smart glasses.

[0286] Specific operation: The server generates data using the prompt sentence and sends it to the smart glasses.

[0287] Step 3:

[0288] An interactive screen is displayed on the smart glasses.

[0289] Input: Dummy customer information and scenario data sent from the server

[0290] Data processing / calculation: Dummy customer questions are displayed on the smart glasses display.

[0291] Output: Trainee reviews the question and prepares a response.

[0292] Specific operation: The smart glasses analyze the information received from the server and display it on the screen.

[0293] Step 4:

[0294] The user responds to the dummy customer's questions by voice input.

[0295] Input: Trainee's voice response

[0296] Data processing / calculation: The smart glasses convert voice input into text and send it to the server.

[0297] Output: Voice input is sent to the server as text data.

[0298] How it works: The smart glasses use a built-in microphone to convert voice data into text using voice recognition technology.

[0299] Step 5:

[0300] The server analyzes and evaluates the response it receives.

[0301] Input: Transcribed trainee responses

[0302] Data processing / calculation: The server compares the response with existing procedures and evaluates it using a generative AI model.

[0303] Output: Generates evaluation results and feedback messages.

[0304] Specific Actions: The server uses an analysis algorithm to assess relevance and generate a feedback message.

[0305] Step 6:

[0306] The feedback message received from the server is displayed on the smart glasses.

[0307] Input: Feedback message sent by the server

[0308] Data processing / calculation: Display feedback on the smart glasses display.

[0309] Output: Trainee reviews the feedback and understands areas for improvement.

[0310] How it works: The smart glasses provide real-time feedback and show trainees how to respond appropriately.

[0311] Step 7:

[0312] The user selects the next training scenario or ends the training.

[0313] Input: Trainee selection (next scenario or exit)

[0314] Data processing / calculation: The server receives commands to generate new scenario data or to shut down the system.

[0315] Output: New scenario data is sent to the smart glasses or a completion message is displayed.

[0316] Specific behavior: The server generates appropriate data depending on the situation and provides the next action based on the trainee's selection.

[0317] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0318] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide more advanced and adaptable training by combining it with an emotion engine that recognizes the user's emotions. This system is composed of a server, a terminal, an emotion engine, and a user.

[0319] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative artificial intelligence. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[0320] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[0321] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[0322] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[0323] As a concrete example, we will explain the flow of a user logging in to the system and performing role-playing with an emotion engine built in.

[0324] Login Process

[0325] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0326] 2. The terminal sends the entered authentication information to the server.

[0327] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[0328] Role-playing begins

[0329] 1. The user selects "Start Role-Playing" from the main menu.

[0330] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[0331] 3. The terminal displays the interactive screen.

[0332] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[0333] Mock conversation progression

[0334] 1. The user enters and submits a response to a question.

[0335] 2. The terminal sends the entered response to the server.

[0336] 3. The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[0337] 4. The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[0338] 5. The server generates evaluation results and feedback messages based on the recognized emotions.

[0339] 6. The terminal displays the feedback message on the interactive screen.

[0340] For example, if a user has a strong emotion (anger or anxiety), the emotion engine will recognize this and the server can change the response and tone of the feedback to teach the trainee how to respond appropriately.

[0341] Providing feedback

[0342] 1. The device receives and displays the feedback message sent from the server. The feedback includes the error indication and the correct response.

[0343] 2. The user reviews the feedback and re-enters the corrected response.

[0344] 3. The terminal resends the corrected response to the server.

[0345] 4. The server again analyzes the response and evaluates whether it is an appropriate response.

[0346] If appropriate, the following scenarios proceed:

[0347] If it's not appropriate, provide feedback again.

[0348] Role-playing ends

[0349] 1. The user completes all scenarios or clicks the "Exit" button midway through.

[0350] 2. The terminal notifies the server of the termination operation.

[0351] 3. The server records the user's progress and evaluation results in a database and ends the session.

[0352] 4. The terminal displays a logout screen to notify the user that the session has ended.

[0353] By using this system, trainees can learn in an environment that is close to actual customer interactions, allowing them to efficiently master appropriate ways of interacting with customers. In addition, the introduction of an emotion engine provides appropriate feedback based on the user's emotions, enabling more practical training.

[0354] The processing flow will be explained below.

[0355] Login Process

[0356] Step 1:

[0357] The user displays the login screen for the system on the terminal and enters their ID and password.

[0358] Step 2:

[0359] The terminal transmits the entered authentication information to the server.

[0360] Step 3:

[0361] The server compares the received authentication information with a database and generates an authentication result.

[0362] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[0363] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[0364] Step 4:

[0365] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[0366] Role-playing begins

[0367] Step 1:

[0368] The user clicks the "Start Role-Playing" button from the main menu.

[0369] Step 2:

[0370] The device notifies the server of the click event.

[0371] Step 3:

[0372] The server generates dummy customer information and scenario data and transmits them to the terminal.

[0373] The dummy customer information includes names, addresses, and past donation details.

[0374] The scenario data includes specific conversation flows and questions.

[0375] Step 4:

[0376] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[0377] Mock conversation progression

[0378] Step 1:

[0379] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[0380] Step 2:

[0381] The terminal displays the generated question on an interactive screen.

[0382] Step 3:

[0383] The user enters a response to the displayed question and clicks the submit button.

[0384] Step 4:

[0385] The terminal sends the entered response to the server.

[0386] Step 5:

[0387] The server analyzes the response and compares it with the response procedures and FAQ data.

[0388] Step 6:

[0389] The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[0390] Step 7:

[0391] The server generates an evaluation result and a feedback message based on the recognized emotion.

[0392] Step 8:

[0393] The terminal displays the feedback message on the interactive screen.

[0394] Providing feedback

[0395] Step 1:

[0396] The terminal receives and displays the feedback message sent from the server.

[0397] The feedback includes a pointer to what went wrong and the correct way to respond.

[0398] Step 2:

[0399] The user checks the feedback and re-enters the corrected response.

[0400] Step 3:

[0401] The terminal retransmits the corrected response to the server.

[0402] Step 4:

[0403] The server again analyzes the response and evaluates whether it is an appropriate response.

[0404] If appropriate, the following scenarios proceed:

[0405] If it's not appropriate, provide feedback again.

[0406] Role-playing ends

[0407] Step 1:

[0408] The user can complete all scenarios or click the "Exit" button midway through.

[0409] Step 2:

[0410] The terminal notifies the server of the termination operation.

[0411] Step 3:

[0412] The server records the user's progress data and evaluation results in a database and ends the session.

[0413] Step 4:

[0414] The terminal displays a logout screen to notify the user that the session has ended.

[0415] Example 2

[0416] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0417] Conventional customer service training systems have made it difficult for trainees to learn in an environment similar to actual customer service. Furthermore, general feedback systems are unable to provide appropriate feedback that takes into account the trainee's emotional state, limiting the effectiveness of the training. Therefore, it has been a challenge for trainees to efficiently acquire the appropriate skills for actual customer service situations.

[0418] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for storing response procedures and virtual customer information, means for generating questions based on the virtual customer information using the generative AI model, and means for comparing responses entered by the trainee with the response procedures and evaluating them. This enables the trainee to efficiently train in a simulated conversation environment that is similar to actual customer service while recognizing user emotions using an emotion analysis engine.

[0419] A "generative artificial intelligence model" is an artificial intelligence technology that generates natural language based on training data, and is used to generate questions and responses from virtual customers.

[0420] An "emotion analysis engine" is a technology that recognizes emotions from text or voice data entered by a user and provides feedback based on those emotions.

[0421] "Service procedures" define the standard procedures and responses to be used when dealing with customers.

[0422] "Virtual customer information" refers to fictitious customer profiles and background information generated based on simulated conversation scenarios.

[0423] "Trainees" refer to learners who intend to use this system to acquire customer service skills.

[0424] The "means for generating questions" refers to techniques or algorithms for generating questions based on virtual customer information using a generative artificial intelligence model.

[0425] "Means for verification and evaluation" refers to techniques or methods for comparing responses entered by trainees with pre-defined response procedures and evaluating their accuracy and appropriateness.

[0426] "Means for providing feedback" refers to technologies and algorithms that provide trainees with appropriate advice and suggestions for correction based on the results of evaluation and sentiment analysis.

[0427] "Scenario data" refers to data that includes a series of hypothetical interactions and situations used to structure the progress of a simulated conversation.

[0428] This invention is a simulated conversation system for customer service that combines a generative artificial intelligence model and an emotion analysis engine, and aims to enable trainees to efficiently train in an environment that closely resembles actual customer service. The system is composed of a server, a terminal, an emotion analysis engine, and a user.

[0429] Hardware and software used

[0430] server

[0431] The server hosts the generative AI model, sentiment analysis engine, response procedures, virtual customer information, and FAQ database. Specifically, it is implemented with the following configuration.

[0432] Hardware: powerful processor, sufficient memory, and large storage capacity for database storage

[0433] Software: Linux-based operating systems such as Ubuntu or CentOS, database management systems such as MySQL or PostgreSQL, and machine learning libraries such as TensorFlow

[0434] Terminal

[0435] Terminals are devices used by trainees, including PCs, tablets, smartphones, etc. Terminals communicate with the server and perform the following processes:

[0436] Hardware: Any device that supports a web browser

[0437] Software: A modern web browser (Google Chrome, Mozilla Firefox, etc.), JavaScript, and a front-end framework such as React

[0438] Sentiment Analysis Engine

[0439] The sentiment analysis engine is used to analyze text and voice data to recognize the emotions of the user.

[0440] Software: Natural language processing (NLP) libraries (NLTK, spaCy, etc.), speech recognition libraries (Google Speech to Text API, etc.)

[0441] Explanation of program processing

[0442] Generating training scenarios

[0443] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0444] 2. The device sends this information to the server.

[0445] 3. The server checks the user information against the database, and if authentication is successful, displays the main menu on the terminal.

[0446] 4. When the user selects "Start role-playing," the server generates virtual customer information and scenario data and sends them to the terminal.

[0447] 5. The device displays a dialogue screen and asks questions from the generating AI.

[0448] Mock conversation progression

[0449] 1. The user enters a response to the question posed by the generating AI and submits it.

[0450] 2. The terminal sends the response data to the server.

[0451] 3. The server analyzes the received data and compares it with the response procedures and FAQ data.

[0452] 4. The emotion analysis engine (server) analyzes the user's input text and assigns emotion labels.

[0453] 5. Based on this information, the server generates evaluation results and feedback messages and sends them to the terminal.

[0454] Providing feedback

[0455] 1. The device receives and displays a feedback message, which includes a description of the error and the correct response.

[0456] 2. The user reviews the feedback, re-enters the revised response, and submits it.

[0457] 3. The terminal retransmits the retyped response to the server.

[0458] 4. The server analyzes again and decides whether to proceed to the next scenario or provide feedback again.

[0459] End of role-playing

[0460] 1. The user ends the role-playing by completing all scenarios or by clicking the "Exit" button.

[0461] 2. The device notifies the server of this information, and the server records the user's progress data and evaluation results in a database.

[0462] 3. The terminal displays a logout screen to notify the user that the session has ended.

[0463] Specific examples

[0464] As a concrete example, the following mock conversation scenario can be considered.

[0465] 1. User: "I have a question about a product."

[0466] 2. Generative AI (server): "Which feature of which product are you asking about?"

[0467] 3. User: "What is the return process?"

[0468] 4. The emotion analysis engine (server) recognizes "anxiety" from the user's message and softens the tone of the response.

[0469] 5. Server: "For more information on the return process, please see this guide."

[0470] Prompt Sentence Examples

[0471] "Generate the following role-playing scenario: A user has a question about a new product. Initiate a dialogue that describes the product's features."

[0472] "Generate an appropriate response based on the user's input: 'I would like to return this item.'"

[0473] The specific processes and procedures used in implementing the present invention have been described above. This system enables trainees to efficiently receive practical training that takes into account emotions in an environment that closely resembles actual customer interactions.

[0474] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0475] Step 1:

[0476] The user displays the system login screen on the terminal and enters their ID and password.

[0477] Input: ID, password

[0478] Output: None

[0479] Specifically, the user opens a browser, accesses the specified URL (login page), enters the ID and password in the login form, and clicks the submit button.

[0480] Step 2:

[0481] The device sends the entered authentication information to the server, encrypting it and using the HTTPS protocol.

[0482] Input: ID, password (encrypted)

[0483] Output: Authentication request data

[0484] Specifically, it uses JavaScript to capture form data and sends the data to the server using an AJAX request.

[0485] Step 3:

[0486] The server compares the received authentication information with a database and generates an authentication result.

[0487] Input: Authentication request data

[0488] Output: Authentication result (success / failure)

[0489] Specifically, the server executes an SQL query to match the corresponding user information from the database, and if authentication is successful, generates a token to start a session.

[0490] Step 4:

[0491] The server sends the authentication result to the terminal. If the authentication is successful, it sends an HTML page to display the main menu to the terminal. If not, it sends a page containing an error message.

[0492] Input: Authentication result

[0493] Output: HTML page (main menu or error messages)

[0494] Specifically, the server selects an appropriate HTML template based on the authentication result and sends it to the client.

[0495] Step 5:

[0496] The user selects "Start Role-Playing" from the main menu.

[0497] Input: User clicks

[0498] Output: Request data

[0499] As a specific operation, the user clicks the "Start Role Playing" button on the main menu, and request data is generated.

[0500] Step 6:

[0501] The device sends a "start role-playing" request to the server.

[0502] Input: Request data

[0503] Output: Start request

[0504] As a specific operation, the terminal sends a start request to the server using an AJAX request.

[0505] Step 7:

[0506] The server generates virtual customer information and scenario data and transmits them to the terminal. The server generates a scenario using a generative artificial intelligence model.

[0507] Input: Start Request

[0508] Output: Virtual customer information, scenario data

[0509] Specifically, the server calls the generation AI, sends a scenario generation prompt, receives a response, and then sends the generated scenario data and virtual customer information to the client.

[0510] Step 8:

[0511] The device displays a dialogue screen and asks the first question from the generating AI.

[0512] Input: Virtual customer information, scenario data

[0513] Output: Showing the first question

[0514] Specifically, the device dynamically renders an interactive screen using HTML and JavaScript based on the data it receives.

[0515] Step 9:

[0516] The user enters and submits responses to questions posed by the generating AI.

[0517] Input: User response

[0518] Output: Response data

[0519] As a specific operation, the user enters information into the text box on the interactive screen and clicks the send button.

[0520] Step 10:

[0521] The terminal captures the entered responses and sends them to the server.

[0522] Input: Response data

[0523] Output: Response data (sent)

[0524] As a specific operation, the terminal uses an AJAX request to send response data to the server.

[0525] Step 11:

[0526] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[0527] Input: Response data

[0528] Output: Evaluation results

[0529] Specifically, the server uses natural language processing technology to analyze the response data and compare it with response procedures and FAQ data.

[0530] Step 12:

[0531] The emotion analysis engine (server) recognizes emotions from the user's input and sends the results to the server.

[0532] Input: Response data

[0533] Output: Emotion label

[0534] Specifically, the sentiment analysis engine uses NLP techniques to recognize the emotional state of text and generate sentiment labels.

[0535] Step 13:

[0536] The server generates a feedback message based on the evaluation results and emotion labels and sends it to the device.

[0537] Input: Evaluation result, emotion label

[0538] Output: Feedback message

[0539] Specifically, the server uses a rating algorithm to generate appropriate feedback content.

[0540] Step 14:

[0541] The terminal displays the feedback message on the interactive screen.

[0542] Input: Feedback message

[0543] Output: Show feedback

[0544] As a specific operation, the terminal renders the received feedback message at a specified location on the interactive screen.

[0545] Step 15:

[0546] The user checks the feedback, re-enters the corrected response, and submits it.

[0547] Input: Corrected response

[0548] Output: Corrected response data

[0549] Specifically, the user checks the feedback, then enters a revised response in the response text box, and clicks the send button again.

[0550] Step 16:

[0551] The terminal resends the modified response to the server.

[0552] Input: Corrected response data

[0553] Output: Corrected response data (sent)

[0554] As a specific operation, the terminal again uses an AJAX request to send the modified response data to the server.

[0555] Step 17:

[0556] The server again analyzes the response and evaluates whether it is an appropriate response.

[0557] Input: Corrected response data

[0558] Output: Reevaluation results

[0559] Specifically, the server performs the same analysis and evaluation, and based on the results, decides whether to proceed to the next scenario or provide feedback again.

[0560] Step 18:

[0561] The user can complete all scenarios or click the "Exit" button midway through.

[0562] Input: End operation

[0563] Output: Termination request data

[0564] Specifically, when the user clicks the quit button, a quit confirmation dialog box is displayed.

[0565] Step 19:

[0566] The terminal notifies the server of the termination operation.

[0567] Input: Termination request data

[0568] Output: Completion notice

[0569] As a specific operation, the terminal uses an AJAX request to send termination request data to the server.

[0570] Step 20:

[0571] The server records the user's progress data and evaluation results in a database and ends the session.

[0572] Input: Completion notice, progress data, evaluation results

[0573] Output: Recording success notification

[0574] Specifically, the server uses an SQL insert or update statement to record the necessary information in the database.

[0575] Step 21:

[0576] The terminal displays a logout screen to notify the user that the session has ended.

[0577] Input: Recording success notification

[0578] Output: Logout screen

[0579] Specifically, the terminal displays an HTML page that includes a message such as "You have been logged out."

[0580] (Application example 2)

[0581] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0582] Existing customer service training systems have the following problems. First, the mock conversation scenarios are fixed, making it difficult to reproduce the diverse situations that occur in actual customer service interactions. Second, they are unable to respond to fluctuations in customer emotions, limiting opportunities to acquire realistic customer service skills. Third, they lack a mechanism for accurately assessing trainees' emotions and responses and providing appropriate feedback. There is a need for a system that can solve these problems and provide more comprehensive and practical customer service training.

[0583] The identification process by the identification 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 storing response procedures and dummy customer information, means for generating questions based on the dummy customer information using the generative artificial intelligence, means for recognizing emotions from responses entered by trainees, means for adjusting feedback and response content based on the emotion recognition results, and means for providing feedback to trainees based on the evaluation. This enables practical customer response training that can be adapted to a variety of situations.

[0584] "Generative AI" is an AI system that automatically generates appropriate responses and scenarios based on user input and the situation.

[0585] "Service procedures" refer to specific steps or protocols to be followed when serving customers.

[0586] "Dummy customer information" refers to fictitious customer data created based on actual customer information and used in training.

[0587] "Means for generating questions" refers to the function of artificial intelligence to automatically create questions based on dummy customer information and scenario data.

[0588] "Means for recognizing emotions" refers to technology for analyzing and understanding a user's emotional state from their responses and behavior.

[0589] The "means for adjusting feedback and response content" is a mechanism for appropriately changing or optimizing the feedback provided or the next response based on the emotion recognition results.

[0590] The "means for providing evaluation" is a function for judging the accuracy and appropriateness of the user's response and notifying the trainee of the result.

[0591] This invention is a system that uses a generative AI model and an emotion engine to conduct simulated customer service conversations. The system is composed of a server, a terminal, an emotion engine, and a user.

[0592] server:

[0593] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative AI model. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[0594] Device:

[0595] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[0596] Emotion Engine:

[0597] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[0598] User:

[0599] The participants are mainly trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through mock conversations. System administrators monitor the trainees' progress and evaluations, and provide appropriate feedback and guidance.

[0600] Program processing:

[0601] The server hosts the application and generates mock conversation scenarios for trainees to perform. A generative AI model runs on the server and generates appropriate questions and scenarios based on dummy customer information and an FAQ database. The trainee's response data is sent to the server and compared with the response procedures. An emotion engine recognizes the user's emotions, and feedback and response content are adjusted based on the results.

[0602] Examples:

[0603] For example, consider a scenario in which a new employee at a brick-and-mortar store is using their smartphone to train their customer service skills with this system. If the following prompt is used:

[0604] plaintext

[0605] "We'll learn what to do if a customer is in a hurry and wants to know where an item is. Let's simulate how you would respond when they ask, 'I'm in a hurry, where's the shampoo?'"

[0606] Based on this prompt, the server generates a scenario and displays it on the user's (trainee's) smartphone. When the trainee enters a response, it is sent to the server and compared with the customer service procedures and FAQ database. At the same time, the emotion engine recognizes the emotion from the response, and the server generates appropriate feedback based on that and sends it back to the trainee. For example, if a customer expresses that they are "in a hurry," and the emotion engine recognizes tension or confusion from the user's response, the server will provide feedback to soften the tone, helping to ensure a smoother response next time.

[0607] In this way, trainees can acquire skills to handle a variety of scenarios in a realistic environment, which is expected to improve their performance when dealing with actual customers.

[0608] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0609] Step 1:

[0610] The user displays the system login screen on the terminal and enters their ID and password. The input data is sent to the server, which compares the authentication information with the database and generates an authentication result. If the authentication is successful, the terminal displays the main menu; if it fails, it displays an error message.

[0611] Step 2:

[0612] The user selects "Start Role-Playing" from the main menu. The server generates a dummy customer profile and scenario data and sends them to the device. The device receives this and displays a dialogue screen. The generative AI model generates the first question based on the scenario and sends it to the device.

[0613] Step 3:

[0614] The user inputs a response to the question and sends it to the terminal. The terminal then sends the input data to the server, which then analyzes the response. The analysis includes checking the response against the response procedure and FAQ database.

[0615] Step 4:

[0616] Based on the analysis results, the server uses an emotion engine to recognize emotions from the user's responses. The emotion engine analyzes text and voice data to recognize the user's emotional state. The recognized emotion data is then sent to the server.

[0617] Step 5:

[0618] The server adjusts the feedback and response content based on the emotion recognition results, optimizing the tone and content of the feedback message depending on the recognized emotion, and generating the next question or scenario as needed and sending it to the device.

[0619] Step 6:

[0620] As a substep, the terminal displays a feedback message on the interactive screen and provides appropriate feedback to the user, who can then confirm the feedback and re-enter a corrected response if necessary.

[0621] Step 7:

[0622] When the user completes all scenarios or clicks the "Exit" button at any point, the terminal notifies the server of the end operation. The server records the user's progress data and evaluation results in a database and ends the session. A logout screen is displayed on the terminal to notify the user of the end of the session.

[0623] Through the above processing steps, this system can provide realistic customer service training and support users in improving their skills.

[0624] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0625] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0626] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0627] [Second embodiment]

[0628] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0629] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0630] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0631] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0632] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0633] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0634] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0635] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0636] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0637] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0638] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0639] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0640] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide efficient and high-quality training. This system is composed of a server, terminals, and users.

[0641] The server hosts the data and applications necessary for the training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, etc. The server is equipped with a generative artificial intelligence that plays the role of a customer and generates mock conversations.

[0642] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also displays feedback and evaluation results sent from the server to the trainee.

[0643] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[0644] As a specific example, the flow of a user logging in to a system will be described.

[0645] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0646] 2. The terminal sends the entered authentication information to the server.

[0647] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[0648] Next, the flow of role-playing is shown below.

[0649] 1. The user selects "Start Role-Playing" from the main menu.

[0650] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[0651] 3. The terminal displays the interactive screen.

[0652] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[0653] 5. The user enters and submits a response to the question.

[0654] 6. The terminal sends the entered response to the server,

[0655] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[0656] 7. The server generates the evaluation results and generates appropriate feedback messages.

[0657] 8. The terminal displays the feedback message on the interactive screen.

[0658] By using this system, trainees can learn in an environment that is similar to actual customer service, allowing them to efficiently master appropriate customer service techniques. In addition, if a trainee makes a mistake, the system will immediately suggest the correct way to respond, improving the quality of the training.

[0659] Furthermore, the system has the ability to generate dummy customer information and scenario data and send them to trainees' devices, allowing for practical learning that can handle a variety of scenarios. This system helps prevent a decline in customer service quality due to resource shortages during busy periods and individual differences.

[0660] The above is an embodiment of the present invention.

[0661] The processing flow will be explained below.

[0662] Login Process

[0663] Step 1:

[0664] The user displays the login screen for the system on the terminal and enters their ID and password.

[0665] Step 2:

[0666] The terminal transmits the entered authentication information to the server.

[0667] Step 3:

[0668] The server compares the received authentication information with a database and generates an authentication result.

[0669] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[0670] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[0671] Step 4:

[0672] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[0673] Role-playing begins

[0674] Step 1:

[0675] The user clicks the "Start Role-Playing" button from the main menu.

[0676] Step 2:

[0677] The device notifies the server of the click event.

[0678] Step 3:

[0679] The server generates dummy customer information and scenario data and transmits them to the terminal.

[0680] The dummy customer information includes names, addresses, and past donation details.

[0681] The scenario data includes specific conversation flows and questions.

[0682] Step 4:

[0683] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[0684] Mock conversation progression

[0685] Step 1:

[0686] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[0687] Step 2:

[0688] The terminal displays the generated question on an interactive screen.

[0689] Step 3:

[0690] The user enters a response to the displayed question and clicks the submit button.

[0691] Step 4:

[0692] The terminal sends the entered response to the server.

[0693] Step 5:

[0694] The server analyzes the received response and matches it with procedures and FAQ data.

[0695] Step 6:

[0696] The server generates an evaluation result based on the response content and generates an appropriate feedback message.

[0697] Step 7:

[0698] The terminal displays the feedback message on the interactive screen.

[0699] Providing feedback

[0700] Step 1:

[0701] The terminal receives and displays the feedback message sent from the server.

[0702] The feedback includes a pointer to what went wrong and the correct way to respond.

[0703] Step 2:

[0704] The user checks the feedback and inputs a revised response.

[0705] Step 3:

[0706] The terminal retransmits the corrected response to the server.

[0707] Step 4:

[0708] The server again analyzes the response and evaluates whether it is an appropriate response.

[0709] If appropriate, the following scenarios proceed:

[0710] If it's not appropriate, provide feedback again.

[0711] Role-playing ends

[0712] Step 1:

[0713] The user can complete all scenarios or click the "Exit" button midway through.

[0714] Step 2:

[0715] The terminal notifies the server of the termination operation.

[0716] Step 3:

[0717] The server records the user's progress data and evaluation results in a database and ends the session.

[0718] Step 4:

[0719] The terminal displays a logout screen to notify the user that the session has ended.

[0720] Example 1

[0721] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0722] Conventional customer service training systems have struggled to provide trainees with realistic mock conversations to help them acquire practical customer service skills. Additionally, feedback is often delayed, and it's difficult to suggest appropriate corrections in real time. As a result, the quality of training declines, and trainees may lack customer service skills in their actual work.

[0723] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0724] In this invention, the server includes means for storing training data and virtual customer data, means for generating dialogue based on the virtual customer data using generative artificial intelligence, means for comparing responses input by the trainee with training procedures and evaluating them, and means for providing feedback to the trainee based on the evaluation. This enables the trainee to acquire practical conversation skills through realistic simulated conversations, and by providing appropriate, real-time feedback and suggestions for correction, the quality of the training can be improved.

[0725] - "Training purpose" refers to the purpose for trainees to acquire response skills and knowledge.

[0726] "Generative AI" is a type of AI that performs natural language processing and dialogue generation.

[0727] "Simulating a dialogue" means recreating an actual conversation environment and conducting training through virtual conversations.

[0728] A "system" is a device that integrates multiple elements and functions and is designed to achieve a specific purpose.

[0729] "Training Data" is information regarding guidelines and procedures used during training.

[0730] "Virtual customer data" refers to information about fictitious customers used in simulated conversations.

[0731] A "means" is a method or device used to achieve a particular purpose.

[0732] A "Trainer" is an individual who uses the system to improve their customer service skills.

[0733] "Training procedures" are procedures or guidelines to be followed when dealing with customers.

[0734] "Evaluating" means judging the appropriateness of the trainee's response.

[0735] "Feedback" refers to information about the trainee's evaluation of their response and areas for improvement.

[0736] "Dialogue based on virtual customer data" is a conversation generated based on fictitious customer information.

[0737] "Suggesting corrections" means suggesting procedures or methods for improvement when a response is inappropriate.

[0738] A "terminal" is a device used to access the system.

[0739] This invention is a training system that uses a generative AI model to conduct simulated customer service conversations, and aims to provide efficient and high-quality training. This system is composed of a server, a terminal, and a user.

[0740] The server hosts the data and applications required for training. Specifically, training data, virtual customer data, an FAQ database, and other items are stored on the server. The server is equipped with a generative AI model, which acts as the virtual customer and generates simulated conversations. Specific examples of servers include high-performance server machines (e.g., highly reliable server equipment, cloud computing platforms). Examples of software include the Linux OS, a generative AI model (e.g., an AI model that performs natural language processing), and a database management system (e.g., MySQL).

[0741] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also has the role of displaying feedback and evaluation results sent from the server to the trainee. A browser (e.g., a standard web browser) or a communication application (e.g., a communication app using the WebSocket protocol) is used.

[0742] The users are trainees and system administrators. Trainees log in to the system using a terminal and receive training in customer service through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[0743] As a concrete example, the flow of a user logging in to the system is shown below.

[0744] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0745] 2. The terminal sends the entered authentication information to the server.

[0746] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[0747] Next, the flow of role-playing is shown below.

[0748] 1. The user selects "Start Role-Playing" from the main menu.

[0749] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[0750] 3. The terminal displays the interactive screen.

[0751] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[0752] 5. The user enters and submits a response to the question.

[0753] 6. The terminal sends the entered response to the server.

[0754] 7. The server analyzes the response and compares it with pre-defined training procedures and FAQ data.

[0755] 8. The server generates the evaluation results and generates appropriate feedback messages.

[0756] 9. The terminal displays the feedback message on the interactive screen.

[0757] As a concrete example, the prompt sentence is shown below.

[0758] Prompt statement:

[0759] "My phone bill seems unusually high lately. How does it compare to my last bill?"

[0760] Using this system, trainees can learn in an environment that closely resembles actual customer interactions, allowing them to efficiently master customer interaction methods. In addition, if an incorrect interaction occurs, the correct interaction method is immediately displayed, improving the quality of training. Furthermore, the system generates virtual customer data and scenario data, allowing for practical learning that corresponds to a variety of scenarios. This system can prevent a decline in interaction quality due to resource shortages during busy periods or individual differences.

[0761] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0762] Explain the processing flow of the system program by dividing it into processing steps

[0763] Step 1: User Login

[0764] Input: User ID and password

[0765] Output: Authentication result (success / failure)

[0766] What happens:

[0767] 1. The user uses the terminal to display the system login screen and enters their ID and password.

[0768] Specific operation: Open the login page in your browser and enter the ID "user01" and password "password123" in the form.

[0769] 2. The terminal sends the entered authentication information to the server.

[0770] What it does: Sends authentication information in an HTTP POST request.

[0771] 3. The server checks the received authentication information against its database.

[0772] Specific behavior: Performs a database query to verify the ID and password.

[0773] 4. The server generates an authentication result and sends it to the terminal.

[0774] Specific operation: If authentication is successful, generate and send a token; if it fails, send an error message "Authentication failed."

[0775] Step 2: Prepare to start role-playing

[0776] Input: Request to start roleplaying

[0777] Output: Virtual customer profile and scenario data

[0778] What happens:

[0779] 1. The user selects "Start Role-Playing" from the main menu.

[0780] Specific action: Click a button on the menu.

[0781] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[0782] Specific operation: Generate a virtual customer name "Customer A" and a scenario "Regarding bill payment" and send them to the terminal in JSON format.

[0783] 3. The terminal displays the interactive screen.

[0784] Specific operation: Display the received profile and scenario data on the screen.

[0785] Step 3: Role-playing dialogue

[0786] Input: User response

[0787] Output: The next question generated by the generative AI

[0788] What happens:

[0789] 1. The generation AI (server) generates an initial question based on the scenario and sends it to the device.

[0790] Specific behavior: Generate and send the question "Hello, this is Customer A. I'm having trouble paying my bill."

[0791] 2. The user enters and submits a response to the question.

[0792] Specific action: Enter the response "Please tell me about the specific problem" and click the send button.

[0793] 3. The terminal sends the entered response to the server.

[0794] Specific operation: The response content is sent to the server via an HTTP POST request.

[0795] 4. The server analyzes the response and compares it with the configured training procedures and FAQ data.

[0796] Specific behavior: Performs text analysis and searches for relevant entries in the FAQ database.

[0797] Step 4: Feedback and Rating

[0798] Input: User responses and analysis results

[0799] Output: Feedback message

[0800] What happens:

[0801] 1. The server generates the evaluation results and generates a feedback message.

[0802] What it does: Evaluate the response and generate feedback like "Your response was unclear. Please check the amount next time."

[0803] 2. The terminal displays the feedback message on the interactive screen.

[0804] What it does: Displays the feedback received on the screen.

[0805] Step 5: Logout

[0806] Input: Logout request

[0807] Output: Logout result, login screen displayed

[0808] What happens:

[0809] 1. The user selects logout from the menu.

[0810] Specific action: Click the "Logout" button on the main menu.

[0811] 2. The device sends a logout request to the server.

[0812] Specific behavior: Sends a logout request as an HTTP POST request.

[0813] 3. The server invalidates the session and redirects to the login screen.

[0814] Specific actions: End the session and send the URL of the login screen to the terminal.

[0815] The above are the specific processing steps of this system.

[0816] (Application example 1)

[0817] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0818] Current customer service training programs struggle to provide a practical training environment that closely resembles actual customer interactions, making it difficult for trainees to efficiently acquire the skills to deal with the diverse situations they will encounter in the field. Furthermore, current systems often lack real-time feedback and evaluation of trainees' customer service quality, limiting the effectiveness of training. In particular, there are few systems that support training while on the move or in a brick-and-mortar store, making it difficult to provide realistic training.

[0819] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0820] In this invention, the server includes means for generating questions based on the dummy customer information using a generative AI, means for comparing responses entered by the trainee with the service procedures and evaluating them, means for providing feedback to the trainee based on the evaluation, means for providing a simulation environment for the trainee using smart glasses, means for analyzing voice input from the smart glasses and transmitting it to the generative AI model, and means for displaying feedback received from the server on the display of the smart glasses. This enables the trainee to effectively acquire customer service skills while receiving evaluation and feedback in real time in an immersive environment close to reality.

[0821] "Generative AI" refers to AI technology that uses natural language processing technology to generate human-like conversations and content.

[0822] "Customer service" refers to the business process of responding to inquiries and requests from customers.

[0823] "Simulated conversation" refers to fictitious dialogue generated to simulate real conversations.

[0824] "Service procedures" refer to standard procedures and guidelines to be followed when dealing with customers.

[0825] "Dummy customer information" refers to fictitious customer profiles and scenario information used in training.

[0826] "Evaluation" refers to the process of judging the quality and appropriateness of the trainee's response and giving them a score and feedback.

[0827] "Feedback" refers to information provided to trainees, including areas for improvement and appropriate guidance.

[0828] "Smart glasses" are glasses with an integrated display that are devices capable of displaying information and inputting and outputting voice.

[0829] "Simulation environment" refers to a virtual environment in which trainees can practice in situations similar to their actual work.

[0830] "Generative AI model" refers to an artificial intelligence model that has been pre-trained to perform generative tasks.

[0831] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide real-time feedback to trainees in brick-and-mortar stores using smart glasses.

[0832] Hardware used

[0833] Smart Glasses

[0834] These are glasses with an integrated display that can display information and input and output voice. Specific examples include Google Glass and Vuzix Blade.

[0835] server

[0836] Hosting generative artificial intelligence and databases for real-time data processing. Specific examples include AWS EC2 instances.

[0837] Software used

[0838] Generative Artificial Intelligence

[0839] It uses natural language processing technology to generate simulated conversations in real time, a specific example being OpenAI GPT-4.

[0840] Database

[0841] Dummy customer information and scenario data are stored and managed. A specific example is AWS RDS (MySQL).

[0842] Real-time communication

[0843] A technology for data communication between a server and smart glasses. A specific example is WebSocket communication.

[0844] System configuration and data processing

[0845] server

[0846] 1. Generate dummy customer information

[0847] The server generates dummy customer information and scenario data using the generative AI model, using the following prompt sentences for this process:

[0848] Sample prompt 1: Generate a customer profile

[0849] Generate a customer profile containing the following information:

[0850] name

[0851] age

[0852] sex

[0853] Purchase Intent

[0854] Customer questions and concerns

[0855] 2. Mock conversation scenario generation

[0856] The server generates questions based on the scenario data and uses the following prompts to make the simulation feel more realistic for the trainee:

[0857] Prompt example 2: Training scenario generation

[0858] Train your customer service representatives using the following scenarios:

[0859] A customer asks for more information about a product.

[0860] A customer asks about returns or exchanges.

[0861] Customers want to know about specific promotions and offers.

[0862] 3. Ratings and Feedback

[0863] The server analyzes the trainee's responses, compares them with the response procedures, and evaluates them. The evaluation results and feedback are generated in real time and displayed on the smart glasses.

[0864] Smart Glasses

[0865] 1. Real-time interaction

[0866] The smart glasses receive scenario data from the server and display it on the screen. They also analyze the trainee's voice input and send it to the server.

[0867] 2. Feedback display

[0868] Feedback received from the server is displayed in real time, allowing trainees to quickly learn how to respond appropriately.

[0869] A concrete example of the entire system

[0870] 1. Part of a training scenario

[0871] For example, a dialogue about a "question about product description" is simulated. The following flow is an example.

[0872] Dummy customer question: "What is the battery life of this product?"

[0873] Trainee response: "About 10 hours."

[0874] Generative AI evaluation: "That's a good answer, but it depends on the specific use case, so it would be good to explain that it varies depending on the use case."

[0875] Using this system, trainees can effectively acquire customer service skills in a realistic environment, receiving real-time evaluations and feedback.

[0876] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0877] Step 1:

[0878] The user puts on the smart glasses and logs into the system.

[0879] Input: User ID and password

[0880] Data processing / calculation: The authentication information is sent to the server and checked against a database.

[0881] Output: The authentication result is sent from the server to the smart glasses.

[0882] Specific operation: If authentication is successful, the server displays the main menu on the smart glasses display. If authentication is unsuccessful, it displays an error message.

[0883] Step 2:

[0884] The user selects "Start Training" from the main menu.

[0885] Input: Select Start Training on the main menu

[0886] Data processing / calculation: The server generates dummy customer information and scenario data using the generative AI model.

[0887] Output: The generated dummy customer information and scenario data are sent to the smart glasses.

[0888] Specific operation: The server generates data using the prompt sentence and sends it to the smart glasses.

[0889] Step 3:

[0890] An interactive screen is displayed on the smart glasses.

[0891] Input: Dummy customer information and scenario data sent from the server

[0892] Data processing / calculation: Dummy customer questions are displayed on the smart glasses display.

[0893] Output: Trainee reviews the question and prepares a response.

[0894] Specific operation: The smart glasses analyze the information received from the server and display it on the screen.

[0895] Step 4:

[0896] The user responds to the dummy customer's questions by voice input.

[0897] Input: Trainee's voice response

[0898] Data processing / calculation: The smart glasses convert voice input into text and send it to the server.

[0899] Output: Voice input is sent to the server as text data.

[0900] How it works: The smart glasses use a built-in microphone to convert voice data into text using voice recognition technology.

[0901] Step 5:

[0902] The server analyzes and evaluates the response it receives.

[0903] Input: Transcribed trainee responses

[0904] Data processing / calculation: The server compares the response with existing procedures and evaluates it using a generative AI model.

[0905] Output: Generates evaluation results and feedback messages.

[0906] Specific Actions: The server uses an analysis algorithm to assess relevance and generate a feedback message.

[0907] Step 6:

[0908] The feedback message received from the server is displayed on the smart glasses.

[0909] Input: Feedback message sent by the server

[0910] Data processing / calculation: Display feedback on the smart glasses display.

[0911] Output: Trainee reviews the feedback and understands areas for improvement.

[0912] How it works: The smart glasses provide real-time feedback and show trainees how to respond appropriately.

[0913] Step 7:

[0914] The user selects the next training scenario or ends the training.

[0915] Input: Trainee selection (next scenario or exit)

[0916] Data processing / calculation: The server receives commands to generate new scenario data or to shut down the system.

[0917] Output: New scenario data is sent to the smart glasses or a completion message is displayed.

[0918] Specific behavior: The server generates appropriate data depending on the situation and provides the next action based on the trainee's selection.

[0919] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0920] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide more advanced and adaptable training by combining it with an emotion engine that recognizes the user's emotions. This system is composed of a server, a terminal, an emotion engine, and a user.

[0921] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative artificial intelligence. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[0922] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[0923] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[0924] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[0925] As a concrete example, we will explain the flow of a user logging in to the system and performing role-playing with an emotion engine built in.

[0926] Login Process

[0927] 1. The user displays the system login screen on the terminal and enters their ID and password.

[0928] 2. The terminal sends the entered authentication information to the server.

[0929] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[0930] Role-playing begins

[0931] 1. The user selects "Start Role-Playing" from the main menu.

[0932] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[0933] 3. The terminal displays the interactive screen.

[0934] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[0935] Mock conversation progression

[0936] 1. The user enters and submits a response to a question.

[0937] 2. The terminal sends the entered response to the server.

[0938] 3. The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[0939] 4. The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[0940] 5. The server generates evaluation results and feedback messages based on the recognized emotions.

[0941] 6. The terminal displays the feedback message on the interactive screen.

[0942] For example, if a user has a strong emotion (anger or anxiety), the emotion engine will recognize this and the server can change the response and tone of the feedback to teach the trainee how to respond appropriately.

[0943] Providing feedback

[0944] 1. The device receives and displays the feedback message sent from the server. The feedback includes the error indication and the correct response.

[0945] 2. The user reviews the feedback and re-enters the corrected response.

[0946] 3. The terminal resends the corrected response to the server.

[0947] 4. The server again analyzes the response and evaluates whether it is an appropriate response.

[0948] If appropriate, the following scenarios proceed:

[0949] If it's not appropriate, provide feedback again.

[0950] Role-playing ends

[0951] 1. The user completes all scenarios or clicks the "Exit" button midway through.

[0952] 2. The terminal notifies the server of the termination operation.

[0953] 3. The server records the user's progress and evaluation results in a database and ends the session.

[0954] 4. The terminal displays a logout screen to notify the user that the session has ended.

[0955] By using this system, trainees can learn in an environment that is close to actual customer interactions, allowing them to efficiently master appropriate ways of interacting with customers. In addition, the introduction of an emotion engine provides appropriate feedback based on the user's emotions, enabling more practical training.

[0956] The processing flow will be explained below.

[0957] Login Process

[0958] Step 1:

[0959] The user displays the login screen for the system on the terminal and enters their ID and password.

[0960] Step 2:

[0961] The terminal transmits the entered authentication information to the server.

[0962] Step 3:

[0963] The server compares the received authentication information with a database and generates an authentication result.

[0964] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[0965] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[0966] Step 4:

[0967] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[0968] Role-playing begins

[0969] Step 1:

[0970] The user clicks the "Start Role-Playing" button from the main menu.

[0971] Step 2:

[0972] The device notifies the server of the click event.

[0973] Step 3:

[0974] The server generates dummy customer information and scenario data and transmits them to the terminal.

[0975] The dummy customer information includes names, addresses, and past donation details.

[0976] The scenario data includes specific conversation flows and questions.

[0977] Step 4:

[0978] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[0979] Mock conversation progression

[0980] Step 1:

[0981] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[0982] Step 2:

[0983] The terminal displays the generated question on an interactive screen.

[0984] Step 3:

[0985] The user enters a response to the displayed question and clicks the submit button.

[0986] Step 4:

[0987] The terminal sends the entered response to the server.

[0988] Step 5:

[0989] The server analyzes the response and compares it with the response procedures and FAQ data.

[0990] Step 6:

[0991] The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[0992] Step 7:

[0993] The server generates an evaluation result and a feedback message based on the recognized emotion.

[0994] Step 8:

[0995] The terminal displays the feedback message on the interactive screen.

[0996] Providing feedback

[0997] Step 1:

[0998] The terminal receives and displays the feedback message sent from the server.

[0999] The feedback includes a pointer to what went wrong and the correct way to respond.

[1000] Step 2:

[1001] The user checks the feedback and re-enters the corrected response.

[1002] Step 3:

[1003] The terminal retransmits the corrected response to the server.

[1004] Step 4:

[1005] The server again analyzes the response and evaluates whether it is an appropriate response.

[1006] If appropriate, the following scenarios proceed:

[1007] If it's not appropriate, provide feedback again.

[1008] Role-playing ends

[1009] Step 1:

[1010] The user can complete all scenarios or click the "Exit" button midway through.

[1011] Step 2:

[1012] The terminal notifies the server of the termination operation.

[1013] Step 3:

[1014] The server records the user's progress data and evaluation results in a database and ends the session.

[1015] Step 4:

[1016] The terminal displays a logout screen to notify the user that the session has ended.

[1017] Example 2

[1018] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1019] Conventional customer service training systems have made it difficult for trainees to learn in an environment similar to actual customer service. Furthermore, general feedback systems are unable to provide appropriate feedback that takes into account the trainee's emotional state, limiting the effectiveness of the training. Therefore, it has been a challenge for trainees to efficiently acquire the appropriate skills for actual customer service situations.

[1020] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for storing response procedures and virtual customer information, means for generating questions based on the virtual customer information using the generative AI model, and means for comparing responses entered by the trainee with the response procedures and evaluating them. This enables the trainee to efficiently train in a simulated conversation environment that is similar to actual customer service while recognizing user emotions using an emotion analysis engine.

[1021] A "generative artificial intelligence model" is an artificial intelligence technology that generates natural language based on training data, and is used to generate questions and responses from virtual customers.

[1022] An "emotion analysis engine" is a technology that recognizes emotions from text or voice data entered by a user and provides feedback based on those emotions.

[1023] "Service procedures" define the standard procedures and responses to be used when dealing with customers.

[1024] "Virtual customer information" refers to fictitious customer profiles and background information generated based on simulated conversation scenarios.

[1025] "Trainees" refer to learners who intend to use this system to acquire customer service skills.

[1026] The "means for generating questions" refers to techniques or algorithms for generating questions based on virtual customer information using a generative artificial intelligence model.

[1027] "Means for verification and evaluation" refers to techniques or methods for comparing responses entered by trainees with pre-defined response procedures and evaluating their accuracy and appropriateness.

[1028] "Means for providing feedback" refers to technologies and algorithms that provide trainees with appropriate advice and suggestions for correction based on the results of evaluation and sentiment analysis.

[1029] "Scenario data" refers to data that includes a series of hypothetical interactions and situations used to structure the progress of a simulated conversation.

[1030] This invention is a simulated conversation system for customer service that combines a generative artificial intelligence model and an emotion analysis engine, and aims to enable trainees to efficiently train in an environment that closely resembles actual customer service. The system is composed of a server, a terminal, an emotion analysis engine, and a user.

[1031] Hardware and software used

[1032] server

[1033] The server hosts the generative AI model, sentiment analysis engine, response procedures, virtual customer information, and FAQ database. Specifically, it is implemented with the following configuration.

[1034] Hardware: powerful processor, sufficient memory, and large storage capacity for database storage

[1035] Software: Linux-based operating systems such as Ubuntu or CentOS, database management systems such as MySQL or PostgreSQL, and machine learning libraries such as TensorFlow

[1036] Terminal

[1037] Terminals are devices used by trainees, including PCs, tablets, smartphones, etc. Terminals communicate with the server and perform the following processes:

[1038] Hardware: Any device that supports a web browser

[1039] Software: A modern web browser (Google Chrome, Mozilla Firefox, etc.), JavaScript, and a front-end framework such as React

[1040] Sentiment Analysis Engine

[1041] The sentiment analysis engine is used to analyze text and voice data to recognize the emotions of the user.

[1042] Software: Natural language processing (NLP) libraries (NLTK, spaCy, etc.), speech recognition libraries (Google Speech to Text API, etc.)

[1043] Explanation of program processing

[1044] Generating training scenarios

[1045] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1046] 2. The device sends this information to the server.

[1047] 3. The server checks the user information against the database, and if authentication is successful, displays the main menu on the terminal.

[1048] 4. When the user selects "Start role-playing," the server generates virtual customer information and scenario data and sends them to the terminal.

[1049] 5. The device displays a dialogue screen and asks questions from the generating AI.

[1050] Mock conversation progression

[1051] 1. The user enters a response to the question posed by the generating AI and submits it.

[1052] 2. The terminal sends the response data to the server.

[1053] 3. The server analyzes the received data and compares it with the response procedures and FAQ data.

[1054] 4. The emotion analysis engine (server) analyzes the user's input text and assigns emotion labels.

[1055] 5. Based on this information, the server generates evaluation results and feedback messages and sends them to the terminal.

[1056] Providing feedback

[1057] 1. The device receives and displays a feedback message, which includes a description of the error and the correct response.

[1058] 2. The user reviews the feedback, re-enters the revised response, and submits it.

[1059] 3. The terminal retransmits the retyped response to the server.

[1060] 4. The server analyzes again and decides whether to proceed to the next scenario or provide feedback again.

[1061] End of role-playing

[1062] 1. The user ends the role-playing by completing all scenarios or by clicking the "Exit" button.

[1063] 2. The device notifies the server of this information, and the server records the user's progress data and evaluation results in a database.

[1064] 3. The terminal displays a logout screen to notify the user that the session has ended.

[1065] Specific examples

[1066] As a concrete example, the following mock conversation scenario can be considered.

[1067] 1. User: "I have a question about a product."

[1068] 2. Generative AI (server): "Which feature of which product are you asking about?"

[1069] 3. User: "What is the return process?"

[1070] 4. The emotion analysis engine (server) recognizes "anxiety" from the user's message and softens the tone of the response.

[1071] 5. Server: "For more information on the return process, please see this guide."

[1072] Prompt Sentence Examples

[1073] "Generate the following role-playing scenario: A user has a question about a new product. Initiate a dialogue that describes the product's features."

[1074] "Generate an appropriate response based on the user's input: 'I would like to return this item.'"

[1075] The specific processes and procedures used in implementing the present invention have been described above. This system enables trainees to efficiently receive practical training that takes into account emotions in an environment that closely resembles actual customer interactions.

[1076] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1077] Step 1:

[1078] The user displays the system login screen on the terminal and enters their ID and password.

[1079] Input: ID, password

[1080] Output: None

[1081] Specifically, the user opens a browser, accesses the specified URL (login page), enters the ID and password in the login form, and clicks the submit button.

[1082] Step 2:

[1083] The device sends the entered authentication information to the server, encrypting it and using the HTTPS protocol.

[1084] Input: ID, password (encrypted)

[1085] Output: Authentication request data

[1086] Specifically, it uses JavaScript to capture form data and sends the data to the server using an AJAX request.

[1087] Step 3:

[1088] The server compares the received authentication information with a database and generates an authentication result.

[1089] Input: Authentication request data

[1090] Output: Authentication result (success / failure)

[1091] Specifically, the server executes an SQL query to match the corresponding user information from the database, and if authentication is successful, generates a token to start a session.

[1092] Step 4:

[1093] The server sends the authentication result to the terminal. If the authentication is successful, it sends an HTML page to display the main menu to the terminal. If not, it sends a page containing an error message.

[1094] Input: Authentication result

[1095] Output: HTML page (main menu or error messages)

[1096] Specifically, the server selects an appropriate HTML template based on the authentication result and sends it to the client.

[1097] Step 5:

[1098] The user selects "Start Role-Playing" from the main menu.

[1099] Input: User clicks

[1100] Output: Request data

[1101] As a specific operation, the user clicks the "Start Role Playing" button on the main menu, and request data is generated.

[1102] Step 6:

[1103] The device sends a "start role-playing" request to the server.

[1104] Input: Request data

[1105] Output: Start request

[1106] As a specific operation, the terminal sends a start request to the server using an AJAX request.

[1107] Step 7:

[1108] The server generates virtual customer information and scenario data and transmits them to the terminal. The server generates a scenario using a generative artificial intelligence model.

[1109] Input: Start Request

[1110] Output: Virtual customer information, scenario data

[1111] Specifically, the server calls the generation AI, sends a scenario generation prompt, receives a response, and then sends the generated scenario data and virtual customer information to the client.

[1112] Step 8:

[1113] The device displays a dialogue screen and asks the first question from the generating AI.

[1114] Input: Virtual customer information, scenario data

[1115] Output: Showing the first question

[1116] Specifically, the device dynamically renders an interactive screen using HTML and JavaScript based on the data it receives.

[1117] Step 9:

[1118] The user enters and submits responses to questions posed by the generating AI.

[1119] Input: User response

[1120] Output: Response data

[1121] As a specific operation, the user enters information into the text box on the interactive screen and clicks the send button.

[1122] Step 10:

[1123] The terminal captures the entered responses and sends them to the server.

[1124] Input: Response data

[1125] Output: Response data (sent)

[1126] As a specific operation, the terminal uses an AJAX request to send response data to the server.

[1127] Step 11:

[1128] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[1129] Input: Response data

[1130] Output: Evaluation results

[1131] Specifically, the server uses natural language processing technology to analyze the response data and compare it with response procedures and FAQ data.

[1132] Step 12:

[1133] The emotion analysis engine (server) recognizes emotions from the user's input and sends the results to the server.

[1134] Input: Response data

[1135] Output: Emotion label

[1136] Specifically, the sentiment analysis engine uses NLP techniques to recognize the emotional state of text and generate sentiment labels.

[1137] Step 13:

[1138] The server generates a feedback message based on the evaluation results and emotion labels and sends it to the device.

[1139] Input: Evaluation result, emotion label

[1140] Output: Feedback message

[1141] Specifically, the server uses a rating algorithm to generate appropriate feedback content.

[1142] Step 14:

[1143] The terminal displays the feedback message on the interactive screen.

[1144] Input: Feedback message

[1145] Output: Show feedback

[1146] As a specific operation, the terminal renders the received feedback message at a specified location on the interactive screen.

[1147] Step 15:

[1148] The user checks the feedback, re-enters the corrected response, and submits it.

[1149] Input: Corrected response

[1150] Output: Corrected response data

[1151] Specifically, the user checks the feedback, then enters a revised response in the response text box, and clicks the send button again.

[1152] Step 16:

[1153] The terminal resends the modified response to the server.

[1154] Input: Corrected response data

[1155] Output: Corrected response data (sent)

[1156] As a specific operation, the terminal again uses an AJAX request to send the modified response data to the server.

[1157] Step 17:

[1158] The server again analyzes the response and evaluates whether it is an appropriate response.

[1159] Input: Corrected response data

[1160] Output: Reevaluation results

[1161] Specifically, the server performs the same analysis and evaluation, and based on the results, decides whether to proceed to the next scenario or provide feedback again.

[1162] Step 18:

[1163] The user can complete all scenarios or click the "Exit" button midway through.

[1164] Input: End operation

[1165] Output: Termination request data

[1166] Specifically, when the user clicks the quit button, a quit confirmation dialog box is displayed.

[1167] Step 19:

[1168] The terminal notifies the server of the termination operation.

[1169] Input: Termination request data

[1170] Output: Completion notice

[1171] As a specific operation, the terminal uses an AJAX request to send termination request data to the server.

[1172] Step 20:

[1173] The server records the user's progress data and evaluation results in a database and ends the session.

[1174] Input: Completion notice, progress data, evaluation results

[1175] Output: Recording success notification

[1176] Specifically, the server uses an SQL insert or update statement to record the necessary information in the database.

[1177] Step 21:

[1178] The terminal displays a logout screen to notify the user that the session has ended.

[1179] Input: Recording success notification

[1180] Output: Logout screen

[1181] Specifically, the terminal displays an HTML page that includes a message such as "You have been logged out."

[1182] (Application example 2)

[1183] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1184] Existing customer service training systems have the following problems. First, the mock conversation scenarios are fixed, making it difficult to reproduce the diverse situations that occur in actual customer service interactions. Second, they are unable to respond to fluctuations in customer emotions, limiting opportunities to acquire realistic customer service skills. Third, they lack a mechanism for accurately assessing trainees' emotions and responses and providing appropriate feedback. There is a need for a system that can solve these problems and provide more comprehensive and practical customer service training.

[1185] The identification process by the identification 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 storing response procedures and dummy customer information, means for generating questions based on the dummy customer information using the generative artificial intelligence, means for recognizing emotions from responses entered by trainees, means for adjusting feedback and response content based on the emotion recognition results, and means for providing feedback to trainees based on the evaluation. This enables practical customer response training that can be adapted to a variety of situations.

[1186] "Generative AI" is an AI system that automatically generates appropriate responses and scenarios based on user input and the situation.

[1187] "Service procedures" refer to specific steps or protocols to be followed when serving customers.

[1188] "Dummy customer information" refers to fictitious customer data created based on actual customer information and used in training.

[1189] "Means for generating questions" refers to the function of artificial intelligence to automatically create questions based on dummy customer information and scenario data.

[1190] "Means for recognizing emotions" refers to technology for analyzing and understanding a user's emotional state from their responses and behavior.

[1191] The "means for adjusting feedback and response content" is a mechanism for appropriately changing or optimizing the feedback provided or the next response based on the emotion recognition results.

[1192] The "means for providing evaluation" is a function for judging the accuracy and appropriateness of the user's response and notifying the trainee of the result.

[1193] This invention is a system that uses a generative AI model and an emotion engine to conduct simulated customer service conversations. The system is composed of a server, a terminal, an emotion engine, and a user.

[1194] server:

[1195] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative AI model. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[1196] Device:

[1197] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[1198] Emotion Engine:

[1199] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[1200] User:

[1201] The participants are mainly trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through mock conversations. System administrators monitor the trainees' progress and evaluations, and provide appropriate feedback and guidance.

[1202] Program processing:

[1203] The server hosts the application and generates mock conversation scenarios for trainees to perform. A generative AI model runs on the server and generates appropriate questions and scenarios based on dummy customer information and an FAQ database. The trainee's response data is sent to the server and compared with the response procedures. An emotion engine recognizes the user's emotions, and feedback and response content are adjusted based on the results.

[1204] Examples:

[1205] For example, consider a scenario in which a new employee at a brick-and-mortar store is using their smartphone to train their customer service skills with this system. If the following prompt is used:

[1206] plaintext

[1207] "We'll learn what to do if a customer is in a hurry and wants to know where an item is. Let's simulate how you would respond when they ask, 'I'm in a hurry, where's the shampoo?'"

[1208] Based on this prompt, the server generates a scenario and displays it on the user's (trainee's) smartphone. When the trainee enters a response, it is sent to the server and compared with the customer service procedures and FAQ database. At the same time, the emotion engine recognizes the emotion from the response, and the server generates appropriate feedback based on that and sends it back to the trainee. For example, if a customer expresses that they are "in a hurry," and the emotion engine recognizes tension or confusion from the user's response, the server will provide feedback to soften the tone, helping to ensure a smoother response next time.

[1209] In this way, trainees can acquire skills to handle a variety of scenarios in a realistic environment, which is expected to improve their performance when dealing with actual customers.

[1210] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1211] Step 1:

[1212] The user displays the system login screen on the terminal and enters their ID and password. The input data is sent to the server, which compares the authentication information with the database and generates an authentication result. If the authentication is successful, the terminal displays the main menu; if it fails, it displays an error message.

[1213] Step 2:

[1214] The user selects "Start Role-Playing" from the main menu. The server generates a dummy customer profile and scenario data and sends them to the device. The device receives this and displays a dialogue screen. The generative AI model generates the first question based on the scenario and sends it to the device.

[1215] Step 3:

[1216] The user inputs a response to the question and sends it to the terminal. The terminal then sends the input data to the server, which then analyzes the response. The analysis includes checking the response against the response procedure and FAQ database.

[1217] Step 4:

[1218] Based on the analysis results, the server uses an emotion engine to recognize emotions from the user's responses. The emotion engine analyzes text and voice data to recognize the user's emotional state. The recognized emotion data is then sent to the server.

[1219] Step 5:

[1220] The server adjusts the feedback and response content based on the emotion recognition results, optimizing the tone and content of the feedback message depending on the recognized emotion, and generating the next question or scenario as needed and sending it to the device.

[1221] Step 6:

[1222] As a substep, the terminal displays a feedback message on the interactive screen and provides appropriate feedback to the user, who can then confirm the feedback and re-enter a corrected response if necessary.

[1223] Step 7:

[1224] When the user completes all scenarios or clicks the "Exit" button at any point, the terminal notifies the server of the end operation. The server records the user's progress data and evaluation results in a database and ends the session. A logout screen is displayed on the terminal to notify the user of the end of the session.

[1225] Through the above processing steps, this system can provide realistic customer service training and support users in improving their skills.

[1226] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1227] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1228] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1229] [Third embodiment]

[1230] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1231] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1232] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1233] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1234] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1235] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1236] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1237] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1238] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1239] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1240] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1241] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1242] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide efficient and high-quality training. This system is composed of a server, terminals, and users.

[1243] The server hosts the data and applications necessary for the training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, etc. The server is equipped with a generative artificial intelligence that plays the role of a customer and generates mock conversations.

[1244] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also displays feedback and evaluation results sent from the server to the trainee.

[1245] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[1246] As a specific example, the flow of a user logging in to a system will be described.

[1247] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1248] 2. The terminal sends the entered authentication information to the server.

[1249] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[1250] Next, the flow of role-playing is shown below.

[1251] 1. The user selects "Start Role-Playing" from the main menu.

[1252] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[1253] 3. The terminal displays the interactive screen.

[1254] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[1255] 5. The user enters and submits a response to the question.

[1256] 6. The terminal sends the entered response to the server,

[1257] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[1258] 7. The server generates the evaluation results and generates appropriate feedback messages.

[1259] 8. The terminal displays the feedback message on the interactive screen.

[1260] By using this system, trainees can learn in an environment that is similar to actual customer service, allowing them to efficiently master appropriate customer service techniques. In addition, if a trainee makes a mistake, the system will immediately suggest the correct way to respond, improving the quality of the training.

[1261] Furthermore, the system has the ability to generate dummy customer information and scenario data and send them to trainees' devices, allowing for practical learning that can handle a variety of scenarios. This system helps prevent a decline in customer service quality due to resource shortages during busy periods and individual differences.

[1262] The above is an embodiment of the present invention.

[1263] The processing flow will be explained below.

[1264] Login Process

[1265] Step 1:

[1266] The user displays the login screen for the system on the terminal and enters their ID and password.

[1267] Step 2:

[1268] The terminal transmits the entered authentication information to the server.

[1269] Step 3:

[1270] The server compares the received authentication information with a database and generates an authentication result.

[1271] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[1272] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[1273] Step 4:

[1274] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[1275] Role-playing begins

[1276] Step 1:

[1277] The user clicks the "Start Role-Playing" button from the main menu.

[1278] Step 2:

[1279] The device notifies the server of the click event.

[1280] Step 3:

[1281] The server generates dummy customer information and scenario data and transmits them to the terminal.

[1282] The dummy customer information includes names, addresses, and past donation details.

[1283] The scenario data includes specific conversation flows and questions.

[1284] Step 4:

[1285] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[1286] Mock conversation progression

[1287] Step 1:

[1288] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[1289] Step 2:

[1290] The terminal displays the generated question on an interactive screen.

[1291] Step 3:

[1292] The user enters a response to the displayed question and clicks the submit button.

[1293] Step 4:

[1294] The terminal sends the entered response to the server.

[1295] Step 5:

[1296] The server analyzes the received response and matches it with procedures and FAQ data.

[1297] Step 6:

[1298] The server generates an evaluation result based on the response content and generates an appropriate feedback message.

[1299] Step 7:

[1300] The terminal displays the feedback message on the interactive screen.

[1301] Providing feedback

[1302] Step 1:

[1303] The terminal receives and displays the feedback message sent from the server.

[1304] The feedback includes a pointer to what went wrong and the correct way to respond.

[1305] Step 2:

[1306] The user checks the feedback and inputs a revised response.

[1307] Step 3:

[1308] The terminal retransmits the corrected response to the server.

[1309] Step 4:

[1310] The server again analyzes the response and evaluates whether it is an appropriate response.

[1311] If appropriate, the following scenarios proceed:

[1312] If it's not appropriate, provide feedback again.

[1313] Role-playing ends

[1314] Step 1:

[1315] The user can complete all scenarios or click the "Exit" button midway through.

[1316] Step 2:

[1317] The terminal notifies the server of the termination operation.

[1318] Step 3:

[1319] The server records the user's progress data and evaluation results in a database and ends the session.

[1320] Step 4:

[1321] The terminal displays a logout screen to notify the user that the session has ended.

[1322] Example 1

[1323] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1324] Conventional customer service training systems have struggled to provide trainees with realistic mock conversations to help them acquire practical customer service skills. Additionally, feedback is often delayed, and it's difficult to suggest appropriate corrections in real time. As a result, the quality of training declines, and trainees may lack customer service skills in their actual work.

[1325] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1326] In this invention, the server includes means for storing training data and virtual customer data, means for generating dialogue based on the virtual customer data using generative artificial intelligence, means for comparing responses input by the trainee with training procedures and evaluating them, and means for providing feedback to the trainee based on the evaluation. This enables the trainee to acquire practical conversation skills through realistic simulated conversations, and by providing appropriate, real-time feedback and suggestions for correction, the quality of the training can be improved.

[1327] - "Training purpose" refers to the purpose for trainees to acquire response skills and knowledge.

[1328] "Generative AI" is a type of AI that performs natural language processing and dialogue generation.

[1329] "Simulating a dialogue" means recreating an actual conversation environment and conducting training through virtual conversations.

[1330] A "system" is a device that integrates multiple elements and functions and is designed to achieve a specific purpose.

[1331] "Training Data" is information regarding guidelines and procedures used during training.

[1332] "Virtual customer data" refers to information about fictitious customers used in simulated conversations.

[1333] A "means" is a method or device used to achieve a particular purpose.

[1334] A "Trainer" is an individual who uses the system to improve their customer service skills.

[1335] "Training procedures" are procedures or guidelines to be followed when dealing with customers.

[1336] "Evaluating" means judging the appropriateness of the trainee's response.

[1337] "Feedback" refers to information about the trainee's evaluation of their response and areas for improvement.

[1338] "Dialogue based on virtual customer data" is a conversation generated based on fictitious customer information.

[1339] "Suggesting corrections" means suggesting procedures or methods for improvement when a response is inappropriate.

[1340] A "terminal" is a device used to access the system.

[1341] This invention is a training system that uses a generative AI model to conduct simulated customer service conversations, and aims to provide efficient and high-quality training. This system is composed of a server, a terminal, and a user.

[1342] The server hosts the data and applications required for training. Specifically, training data, virtual customer data, an FAQ database, and other items are stored on the server. The server is equipped with a generative AI model, which acts as the virtual customer and generates simulated conversations. Specific examples of servers include high-performance server machines (e.g., highly reliable server equipment, cloud computing platforms). Examples of software include the Linux OS, a generative AI model (e.g., an AI model that performs natural language processing), and a database management system (e.g., MySQL).

[1343] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also has the role of displaying feedback and evaluation results sent from the server to the trainee. A browser (e.g., a standard web browser) or a communication application (e.g., a communication app using the WebSocket protocol) is used.

[1344] The users are trainees and system administrators. Trainees log in to the system using a terminal and receive training in customer service through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[1345] As a concrete example, the flow of a user logging in to the system is shown below.

[1346] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1347] 2. The terminal sends the entered authentication information to the server.

[1348] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[1349] Next, the flow of role-playing is shown below.

[1350] 1. The user selects "Start Role-Playing" from the main menu.

[1351] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[1352] 3. The terminal displays the interactive screen.

[1353] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[1354] 5. The user enters and submits a response to the question.

[1355] 6. The terminal sends the entered response to the server.

[1356] 7. The server analyzes the response and compares it with pre-defined training procedures and FAQ data.

[1357] 8. The server generates the evaluation results and generates appropriate feedback messages.

[1358] 9. The terminal displays the feedback message on the interactive screen.

[1359] As a concrete example, the prompt sentence is shown below.

[1360] Prompt statement:

[1361] "My phone bill seems unusually high lately. How does it compare to my last bill?"

[1362] Using this system, trainees can learn in an environment that closely resembles actual customer interactions, allowing them to efficiently master customer interaction methods. In addition, if an incorrect interaction occurs, the correct interaction method is immediately displayed, improving the quality of training. Furthermore, the system generates virtual customer data and scenario data, allowing for practical learning that corresponds to a variety of scenarios. This system can prevent a decline in interaction quality due to resource shortages during busy periods or individual differences.

[1363] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1364] Explain the processing flow of the system program by dividing it into processing steps

[1365] Step 1: User Login

[1366] Input: User ID and password

[1367] Output: Authentication result (success / failure)

[1368] What happens:

[1369] 1. The user uses the terminal to display the system login screen and enters their ID and password.

[1370] Specific operation: Open the login page in your browser and enter the ID "user01" and password "password123" in the form.

[1371] 2. The terminal sends the entered authentication information to the server.

[1372] What it does: Sends authentication information in an HTTP POST request.

[1373] 3. The server checks the received authentication information against its database.

[1374] Specific behavior: Performs a database query to verify the ID and password.

[1375] 4. The server generates an authentication result and sends it to the terminal.

[1376] Specific operation: If authentication is successful, generate and send a token; if it fails, send an error message "Authentication failed."

[1377] Step 2: Prepare to start role-playing

[1378] Input: Request to start roleplaying

[1379] Output: Virtual customer profile and scenario data

[1380] What happens:

[1381] 1. The user selects "Start Role-Playing" from the main menu.

[1382] Specific action: Click a button on the menu.

[1383] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[1384] Specific operation: Generate a virtual customer name "Customer A" and a scenario "Regarding bill payment" and send them to the terminal in JSON format.

[1385] 3. The terminal displays the interactive screen.

[1386] Specific operation: Display the received profile and scenario data on the screen.

[1387] Step 3: Role-playing dialogue

[1388] Input: User response

[1389] Output: The next question generated by the generative AI

[1390] What happens:

[1391] 1. The generation AI (server) generates an initial question based on the scenario and sends it to the device.

[1392] Specific behavior: Generate and send the question "Hello, this is Customer A. I'm having trouble paying my bill."

[1393] 2. The user enters and submits a response to the question.

[1394] Specific action: Enter the response "Please tell me about the specific problem" and click the send button.

[1395] 3. The terminal sends the entered response to the server.

[1396] Specific operation: The response content is sent to the server via an HTTP POST request.

[1397] 4. The server analyzes the response and compares it with the configured training procedures and FAQ data.

[1398] Specific behavior: Performs text analysis and searches for relevant entries in the FAQ database.

[1399] Step 4: Feedback and Rating

[1400] Input: User responses and analysis results

[1401] Output: Feedback message

[1402] What happens:

[1403] 1. The server generates the evaluation results and generates a feedback message.

[1404] What it does: Evaluate the response and generate feedback like "Your response was unclear. Please check the amount next time."

[1405] 2. The terminal displays the feedback message on the interactive screen.

[1406] What it does: Displays the feedback received on the screen.

[1407] Step 5: Logout

[1408] Input: Logout request

[1409] Output: Logout result, login screen displayed

[1410] What happens:

[1411] 1. The user selects logout from the menu.

[1412] Specific action: Click the "Logout" button on the main menu.

[1413] 2. The device sends a logout request to the server.

[1414] Specific behavior: Sends a logout request as an HTTP POST request.

[1415] 3. The server invalidates the session and redirects to the login screen.

[1416] Specific actions: End the session and send the URL of the login screen to the terminal.

[1417] The above are the specific processing steps of this system.

[1418] (Application example 1)

[1419] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1420] Current customer service training programs struggle to provide a practical training environment that closely resembles actual customer interactions, making it difficult for trainees to efficiently acquire the skills to deal with the diverse situations they will encounter in the field. Furthermore, current systems often lack real-time feedback and evaluation of trainees' customer service quality, limiting the effectiveness of training. In particular, there are few systems that support training while on the move or in a brick-and-mortar store, making it difficult to provide realistic training.

[1421] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1422] In this invention, the server includes means for generating questions based on the dummy customer information using a generative AI, means for comparing responses entered by the trainee with the service procedures and evaluating them, means for providing feedback to the trainee based on the evaluation, means for providing a simulation environment for the trainee using smart glasses, means for analyzing voice input from the smart glasses and transmitting it to the generative AI model, and means for displaying feedback received from the server on the display of the smart glasses. This enables the trainee to effectively acquire customer service skills while receiving evaluation and feedback in real time in an immersive environment close to reality.

[1423] "Generative AI" refers to AI technology that uses natural language processing technology to generate human-like conversations and content.

[1424] "Customer service" refers to the business process of responding to inquiries and requests from customers.

[1425] "Simulated conversation" refers to fictitious dialogue generated to simulate real conversations.

[1426] "Service procedures" refer to standard procedures and guidelines to be followed when dealing with customers.

[1427] "Dummy customer information" refers to fictitious customer profiles and scenario information used in training.

[1428] "Evaluation" refers to the process of judging the quality and appropriateness of the trainee's response and giving them a score and feedback.

[1429] "Feedback" refers to information provided to trainees, including areas for improvement and appropriate guidance.

[1430] "Smart glasses" are glasses with an integrated display that are devices capable of displaying information and inputting and outputting voice.

[1431] "Simulation environment" refers to a virtual environment in which trainees can practice in situations similar to their actual work.

[1432] "Generative AI model" refers to an artificial intelligence model that has been pre-trained to perform generative tasks.

[1433] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide real-time feedback to trainees in brick-and-mortar stores using smart glasses.

[1434] Hardware used

[1435] Smart Glasses

[1436] These are glasses with an integrated display that can display information and input and output voice. Specific examples include Google Glass and Vuzix Blade.

[1437] server

[1438] Hosting generative artificial intelligence and databases for real-time data processing. Specific examples include AWS EC2 instances.

[1439] Software used

[1440] Generative Artificial Intelligence

[1441] It uses natural language processing technology to generate simulated conversations in real time, a specific example being OpenAI GPT-4.

[1442] Database

[1443] Dummy customer information and scenario data are stored and managed. A specific example is AWS RDS (MySQL).

[1444] Real-time communication

[1445] A technology for data communication between a server and smart glasses. A specific example is WebSocket communication.

[1446] System configuration and data processing

[1447] server

[1448] 1. Generate dummy customer information

[1449] The server generates dummy customer information and scenario data using the generative AI model, using the following prompt sentences for this process:

[1450] Sample prompt 1: Generate a customer profile

[1451] Generate a customer profile containing the following information:

[1452] name

[1453] age

[1454] sex

[1455] Purchase Intent

[1456] Customer questions and concerns

[1457] 2. Mock conversation scenario generation

[1458] The server generates questions based on the scenario data and uses the following prompts to make the simulation feel more realistic for the trainee:

[1459] Prompt example 2: Training scenario generation

[1460] Train your customer service representatives using the following scenarios:

[1461] A customer asks for more information about a product.

[1462] A customer asks about returns or exchanges.

[1463] Customers want to know about specific promotions and offers.

[1464] 3. Ratings and Feedback

[1465] The server analyzes the trainee's responses, compares them with the response procedures, and evaluates them. The evaluation results and feedback are generated in real time and displayed on the smart glasses.

[1466] Smart Glasses

[1467] 1. Real-time interaction

[1468] The smart glasses receive scenario data from the server and display it on the screen. They also analyze the trainee's voice input and send it to the server.

[1469] 2. Feedback display

[1470] Feedback received from the server is displayed in real time, allowing trainees to quickly learn how to respond appropriately.

[1471] A concrete example of the entire system

[1472] 1. Part of a training scenario

[1473] For example, a dialogue about a "question about product description" is simulated. The following flow is an example.

[1474] Dummy customer question: "What is the battery life of this product?"

[1475] Trainee response: "About 10 hours."

[1476] Generative AI evaluation: "That's a good answer, but it depends on the specific use case, so it would be good to explain that it varies depending on the use case."

[1477] Using this system, trainees can effectively acquire customer service skills in a realistic environment, receiving real-time evaluations and feedback.

[1478] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1479] Step 1:

[1480] The user puts on the smart glasses and logs into the system.

[1481] Input: User ID and password

[1482] Data processing / calculation: The authentication information is sent to the server and checked against a database.

[1483] Output: The authentication result is sent from the server to the smart glasses.

[1484] Specific operation: If authentication is successful, the server displays the main menu on the smart glasses display. If authentication is unsuccessful, it displays an error message.

[1485] Step 2:

[1486] The user selects "Start Training" from the main menu.

[1487] Input: Select Start Training on the main menu

[1488] Data processing / calculation: The server generates dummy customer information and scenario data using the generative AI model.

[1489] Output: The generated dummy customer information and scenario data are sent to the smart glasses.

[1490] Specific operation: The server generates data using the prompt sentence and sends it to the smart glasses.

[1491] Step 3:

[1492] An interactive screen is displayed on the smart glasses.

[1493] Input: Dummy customer information and scenario data sent from the server

[1494] Data processing / calculation: Dummy customer questions are displayed on the smart glasses display.

[1495] Output: Trainee reviews the question and prepares a response.

[1496] Specific operation: The smart glasses analyze the information received from the server and display it on the screen.

[1497] Step 4:

[1498] The user responds to the dummy customer's questions by voice input.

[1499] Input: Trainee's voice response

[1500] Data processing / calculation: The smart glasses convert voice input into text and send it to the server.

[1501] Output: Voice input is sent to the server as text data.

[1502] How it works: The smart glasses use a built-in microphone to convert voice data into text using voice recognition technology.

[1503] Step 5:

[1504] The server analyzes and evaluates the response it receives.

[1505] Input: Transcribed trainee responses

[1506] Data processing / calculation: The server compares the response with existing procedures and evaluates it using a generative AI model.

[1507] Output: Generates evaluation results and feedback messages.

[1508] Specific Actions: The server uses an analysis algorithm to assess relevance and generate a feedback message.

[1509] Step 6:

[1510] The feedback message received from the server is displayed on the smart glasses.

[1511] Input: Feedback message sent by the server

[1512] Data processing / calculation: Display feedback on the smart glasses display.

[1513] Output: Trainee reviews the feedback and understands areas for improvement.

[1514] How it works: The smart glasses provide real-time feedback and show trainees how to respond appropriately.

[1515] Step 7:

[1516] The user selects the next training scenario or ends the training.

[1517] Input: Trainee selection (next scenario or exit)

[1518] Data processing / calculation: The server receives commands to generate new scenario data or to shut down the system.

[1519] Output: New scenario data is sent to the smart glasses or a completion message is displayed.

[1520] Specific behavior: The server generates appropriate data depending on the situation and provides the next action based on the trainee's selection.

[1521] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1522] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide more advanced and adaptable training by combining it with an emotion engine that recognizes the user's emotions. This system is composed of a server, a terminal, an emotion engine, and a user.

[1523] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative artificial intelligence. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[1524] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[1525] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[1526] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[1527] As a concrete example, we will explain the flow of a user logging in to the system and performing role-playing with an emotion engine built in.

[1528] Login Process

[1529] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1530] 2. The terminal sends the entered authentication information to the server.

[1531] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[1532] Role-playing begins

[1533] 1. The user selects "Start Role-Playing" from the main menu.

[1534] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[1535] 3. The terminal displays the interactive screen.

[1536] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[1537] Mock conversation progression

[1538] 1. The user enters and submits a response to a question.

[1539] 2. The terminal sends the entered response to the server.

[1540] 3. The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[1541] 4. The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[1542] 5. The server generates evaluation results and feedback messages based on the recognized emotions.

[1543] 6. The terminal displays the feedback message on the interactive screen.

[1544] For example, if a user has a strong emotion (anger or anxiety), the emotion engine will recognize this and the server can change the response and tone of the feedback to teach the trainee how to respond appropriately.

[1545] Providing feedback

[1546] 1. The device receives and displays the feedback message sent from the server. The feedback includes the error indication and the correct response.

[1547] 2. The user reviews the feedback and re-enters the corrected response.

[1548] 3. The terminal resends the corrected response to the server.

[1549] 4. The server again analyzes the response and evaluates whether it is an appropriate response.

[1550] If appropriate, the following scenarios proceed:

[1551] If it's not appropriate, provide feedback again.

[1552] Role-playing ends

[1553] 1. The user completes all scenarios or clicks the "Exit" button midway through.

[1554] 2. The terminal notifies the server of the termination operation.

[1555] 3. The server records the user's progress and evaluation results in a database and ends the session.

[1556] 4. The terminal displays a logout screen to notify the user that the session has ended.

[1557] By using this system, trainees can learn in an environment that is close to actual customer interactions, allowing them to efficiently master appropriate ways of interacting with customers. In addition, the introduction of an emotion engine provides appropriate feedback based on the user's emotions, enabling more practical training.

[1558] The processing flow will be explained below.

[1559] Login Process

[1560] Step 1:

[1561] The user displays the login screen for the system on the terminal and enters their ID and password.

[1562] Step 2:

[1563] The terminal transmits the entered authentication information to the server.

[1564] Step 3:

[1565] The server compares the received authentication information with a database and generates an authentication result.

[1566] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[1567] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[1568] Step 4:

[1569] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[1570] Role-playing begins

[1571] Step 1:

[1572] The user clicks the "Start Role-Playing" button from the main menu.

[1573] Step 2:

[1574] The device notifies the server of the click event.

[1575] Step 3:

[1576] The server generates dummy customer information and scenario data and transmits them to the terminal.

[1577] The dummy customer information includes names, addresses, and past donation details.

[1578] The scenario data includes specific conversation flows and questions.

[1579] Step 4:

[1580] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[1581] Mock conversation progression

[1582] Step 1:

[1583] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[1584] Step 2:

[1585] The terminal displays the generated question on an interactive screen.

[1586] Step 3:

[1587] The user enters a response to the displayed question and clicks the submit button.

[1588] Step 4:

[1589] The terminal sends the entered response to the server.

[1590] Step 5:

[1591] The server analyzes the response and compares it with the response procedures and FAQ data.

[1592] Step 6:

[1593] The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[1594] Step 7:

[1595] The server generates an evaluation result and a feedback message based on the recognized emotion.

[1596] Step 8:

[1597] The terminal displays the feedback message on the interactive screen.

[1598] Providing feedback

[1599] Step 1:

[1600] The terminal receives and displays the feedback message sent from the server.

[1601] The feedback includes a pointer to what went wrong and the correct way to respond.

[1602] Step 2:

[1603] The user checks the feedback and re-enters the corrected response.

[1604] Step 3:

[1605] The terminal retransmits the corrected response to the server.

[1606] Step 4:

[1607] The server again analyzes the response and evaluates whether it is an appropriate response.

[1608] If appropriate, the following scenarios proceed:

[1609] If it's not appropriate, provide feedback again.

[1610] Role-playing ends

[1611] Step 1:

[1612] The user can complete all scenarios or click the "Exit" button midway through.

[1613] Step 2:

[1614] The terminal notifies the server of the termination operation.

[1615] Step 3:

[1616] The server records the user's progress data and evaluation results in a database and ends the session.

[1617] Step 4:

[1618] The terminal displays a logout screen to notify the user that the session has ended.

[1619] Example 2

[1620] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1621] Conventional customer service training systems have made it difficult for trainees to learn in an environment similar to actual customer service. Furthermore, general feedback systems are unable to provide appropriate feedback that takes into account the trainee's emotional state, limiting the effectiveness of the training. Therefore, it has been a challenge for trainees to efficiently acquire the appropriate skills for actual customer service situations.

[1622] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for storing response procedures and virtual customer information, means for generating questions based on the virtual customer information using the generative AI model, and means for comparing responses entered by the trainee with the response procedures and evaluating them. This enables the trainee to efficiently train in a simulated conversation environment that is similar to actual customer service while recognizing user emotions using an emotion analysis engine.

[1623] A "generative artificial intelligence model" is an artificial intelligence technology that generates natural language based on training data, and is used to generate questions and responses from virtual customers.

[1624] An "emotion analysis engine" is a technology that recognizes emotions from text or voice data entered by a user and provides feedback based on those emotions.

[1625] "Service procedures" define the standard procedures and responses to be used when dealing with customers.

[1626] "Virtual customer information" refers to fictitious customer profiles and background information generated based on simulated conversation scenarios.

[1627] "Trainees" refer to learners who intend to use this system to acquire customer service skills.

[1628] The "means for generating questions" refers to techniques or algorithms for generating questions based on virtual customer information using a generative artificial intelligence model.

[1629] "Means for verification and evaluation" refers to techniques or methods for comparing responses entered by trainees with pre-defined response procedures and evaluating their accuracy and appropriateness.

[1630] "Means for providing feedback" refers to technologies and algorithms that provide trainees with appropriate advice and suggestions for correction based on the results of evaluation and sentiment analysis.

[1631] "Scenario data" refers to data that includes a series of hypothetical interactions and situations used to structure the progress of a simulated conversation.

[1632] This invention is a simulated conversation system for customer service that combines a generative artificial intelligence model and an emotion analysis engine, and aims to enable trainees to efficiently train in an environment that closely resembles actual customer service. The system is composed of a server, a terminal, an emotion analysis engine, and a user.

[1633] Hardware and software used

[1634] server

[1635] The server hosts the generative AI model, sentiment analysis engine, response procedures, virtual customer information, and FAQ database. Specifically, it is implemented with the following configuration.

[1636] Hardware: powerful processor, sufficient memory, and large storage capacity for database storage

[1637] Software: Linux-based operating systems such as Ubuntu or CentOS, database management systems such as MySQL or PostgreSQL, and machine learning libraries such as TensorFlow

[1638] Terminal

[1639] Terminals are devices used by trainees, including PCs, tablets, smartphones, etc. Terminals communicate with the server and perform the following processes:

[1640] Hardware: Any device that supports a web browser

[1641] Software: A modern web browser (Google Chrome, Mozilla Firefox, etc.), JavaScript, and a front-end framework such as React

[1642] Sentiment Analysis Engine

[1643] The sentiment analysis engine is used to analyze text and voice data to recognize the emotions of the user.

[1644] Software: Natural language processing (NLP) libraries (NLTK, spaCy, etc.), speech recognition libraries (Google Speech to Text API, etc.)

[1645] Explanation of program processing

[1646] Generating training scenarios

[1647] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1648] 2. The device sends this information to the server.

[1649] 3. The server checks the user information against the database, and if authentication is successful, displays the main menu on the terminal.

[1650] 4. When the user selects "Start role-playing," the server generates virtual customer information and scenario data and sends them to the terminal.

[1651] 5. The device displays a dialogue screen and asks questions from the generating AI.

[1652] Mock conversation progression

[1653] 1. The user enters a response to the question posed by the generating AI and submits it.

[1654] 2. The terminal sends the response data to the server.

[1655] 3. The server analyzes the received data and compares it with the response procedures and FAQ data.

[1656] 4. The emotion analysis engine (server) analyzes the user's input text and assigns emotion labels.

[1657] 5. Based on this information, the server generates evaluation results and feedback messages and sends them to the terminal.

[1658] Providing feedback

[1659] 1. The device receives and displays a feedback message, which includes a description of the error and the correct response.

[1660] 2. The user reviews the feedback, re-enters the revised response, and submits it.

[1661] 3. The terminal retransmits the retyped response to the server.

[1662] 4. The server analyzes again and decides whether to proceed to the next scenario or provide feedback again.

[1663] End of role-playing

[1664] 1. The user ends the role-playing by completing all scenarios or by clicking the "Exit" button.

[1665] 2. The device notifies the server of this information, and the server records the user's progress data and evaluation results in a database.

[1666] 3. The terminal displays a logout screen to notify the user that the session has ended.

[1667] Specific examples

[1668] As a concrete example, the following mock conversation scenario can be considered.

[1669] 1. User: "I have a question about a product."

[1670] 2. Generative AI (server): "Which feature of which product are you asking about?"

[1671] 3. User: "What is the return process?"

[1672] 4. The emotion analysis engine (server) recognizes "anxiety" from the user's message and softens the tone of the response.

[1673] 5. Server: "For more information on the return process, please see this guide."

[1674] Prompt Sentence Examples

[1675] "Generate the following role-playing scenario: A user has a question about a new product. Initiate a dialogue that describes the product's features."

[1676] "Generate an appropriate response based on the user's input: 'I would like to return this item.'"

[1677] The specific processes and procedures used in implementing the present invention have been described above. This system enables trainees to efficiently receive practical training that takes into account emotions in an environment that closely resembles actual customer interactions.

[1678] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1679] Step 1:

[1680] The user displays the system login screen on the terminal and enters their ID and password.

[1681] Input: ID, password

[1682] Output: None

[1683] Specifically, the user opens a browser, accesses the specified URL (login page), enters the ID and password in the login form, and clicks the submit button.

[1684] Step 2:

[1685] The device sends the entered authentication information to the server, encrypting it and using the HTTPS protocol.

[1686] Input: ID, password (encrypted)

[1687] Output: Authentication request data

[1688] Specifically, it uses JavaScript to capture form data and sends the data to the server using an AJAX request.

[1689] Step 3:

[1690] The server compares the received authentication information with a database and generates an authentication result.

[1691] Input: Authentication request data

[1692] Output: Authentication result (success / failure)

[1693] Specifically, the server executes an SQL query to match the corresponding user information from the database, and if authentication is successful, generates a token to start a session.

[1694] Step 4:

[1695] The server sends the authentication result to the terminal. If the authentication is successful, it sends an HTML page to display the main menu to the terminal. If not, it sends a page containing an error message.

[1696] Input: Authentication result

[1697] Output: HTML page (main menu or error messages)

[1698] Specifically, the server selects an appropriate HTML template based on the authentication result and sends it to the client.

[1699] Step 5:

[1700] The user selects "Start Role-Playing" from the main menu.

[1701] Input: User clicks

[1702] Output: Request data

[1703] As a specific operation, the user clicks the "Start Role Playing" button on the main menu, and request data is generated.

[1704] Step 6:

[1705] The device sends a "start role-playing" request to the server.

[1706] Input: Request data

[1707] Output: Start request

[1708] As a specific operation, the terminal sends a start request to the server using an AJAX request.

[1709] Step 7:

[1710] The server generates virtual customer information and scenario data and transmits them to the terminal. The server generates a scenario using a generative artificial intelligence model.

[1711] Input: Start Request

[1712] Output: Virtual customer information, scenario data

[1713] Specifically, the server calls the generation AI, sends a scenario generation prompt, receives a response, and then sends the generated scenario data and virtual customer information to the client.

[1714] Step 8:

[1715] The device displays a dialogue screen and asks the first question from the generating AI.

[1716] Input: Virtual customer information, scenario data

[1717] Output: Showing the first question

[1718] Specifically, the device dynamically renders an interactive screen using HTML and JavaScript based on the data it receives.

[1719] Step 9:

[1720] The user enters and submits responses to questions posed by the generating AI.

[1721] Input: User response

[1722] Output: Response data

[1723] As a specific operation, the user enters information into the text box on the interactive screen and clicks the send button.

[1724] Step 10:

[1725] The terminal captures the entered responses and sends them to the server.

[1726] Input: Response data

[1727] Output: Response data (sent)

[1728] As a specific operation, the terminal uses an AJAX request to send response data to the server.

[1729] Step 11:

[1730] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[1731] Input: Response data

[1732] Output: Evaluation results

[1733] Specifically, the server uses natural language processing technology to analyze the response data and compare it with response procedures and FAQ data.

[1734] Step 12:

[1735] The emotion analysis engine (server) recognizes emotions from the user's input and sends the results to the server.

[1736] Input: Response data

[1737] Output: Emotion label

[1738] Specifically, the sentiment analysis engine uses NLP techniques to recognize the emotional state of text and generate sentiment labels.

[1739] Step 13:

[1740] The server generates a feedback message based on the evaluation results and emotion labels and sends it to the device.

[1741] Input: Evaluation result, emotion label

[1742] Output: Feedback message

[1743] Specifically, the server uses a rating algorithm to generate appropriate feedback content.

[1744] Step 14:

[1745] The terminal displays the feedback message on the interactive screen.

[1746] Input: Feedback message

[1747] Output: Show feedback

[1748] As a specific operation, the terminal renders the received feedback message at a specified location on the interactive screen.

[1749] Step 15:

[1750] The user checks the feedback, re-enters the corrected response, and submits it.

[1751] Input: Corrected response

[1752] Output: Corrected response data

[1753] Specifically, the user checks the feedback, then enters a revised response in the response text box, and clicks the send button again.

[1754] Step 16:

[1755] The terminal resends the modified response to the server.

[1756] Input: Corrected response data

[1757] Output: Corrected response data (sent)

[1758] As a specific operation, the terminal again uses an AJAX request to send the modified response data to the server.

[1759] Step 17:

[1760] The server again analyzes the response and evaluates whether it is an appropriate response.

[1761] Input: Corrected response data

[1762] Output: Reevaluation results

[1763] Specifically, the server performs the same analysis and evaluation, and based on the results, decides whether to proceed to the next scenario or provide feedback again.

[1764] Step 18:

[1765] The user can complete all scenarios or click the "Exit" button midway through.

[1766] Input: End operation

[1767] Output: Termination request data

[1768] Specifically, when the user clicks the quit button, a quit confirmation dialog box is displayed.

[1769] Step 19:

[1770] The terminal notifies the server of the termination operation.

[1771] Input: Termination request data

[1772] Output: Completion notice

[1773] As a specific operation, the terminal uses an AJAX request to send termination request data to the server.

[1774] Step 20:

[1775] The server records the user's progress data and evaluation results in a database and ends the session.

[1776] Input: Completion notice, progress data, evaluation results

[1777] Output: Recording success notification

[1778] Specifically, the server uses an SQL insert or update statement to record the necessary information in the database.

[1779] Step 21:

[1780] The terminal displays a logout screen to notify the user that the session has ended.

[1781] Input: Recording success notification

[1782] Output: Logout screen

[1783] Specifically, the terminal displays an HTML page that includes a message such as "You have been logged out."

[1784] (Application example 2)

[1785] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1786] Existing customer service training systems have the following problems. First, the mock conversation scenarios are fixed, making it difficult to reproduce the diverse situations that occur in actual customer service interactions. Second, they are unable to respond to fluctuations in customer emotions, limiting opportunities to acquire realistic customer service skills. Third, they lack a mechanism for accurately assessing trainees' emotions and responses and providing appropriate feedback. There is a need for a system that can solve these problems and provide more comprehensive and practical customer service training.

[1787] The identification process by the identification 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 storing response procedures and dummy customer information, means for generating questions based on the dummy customer information using the generative artificial intelligence, means for recognizing emotions from responses entered by trainees, means for adjusting feedback and response content based on the emotion recognition results, and means for providing feedback to trainees based on the evaluation. This enables practical customer response training that can be adapted to a variety of situations.

[1788] "Generative AI" is an AI system that automatically generates appropriate responses and scenarios based on user input and the situation.

[1789] "Service procedures" refer to specific steps or protocols to be followed when serving customers.

[1790] "Dummy customer information" refers to fictitious customer data created based on actual customer information and used in training.

[1791] "Means for generating questions" refers to the function of artificial intelligence to automatically create questions based on dummy customer information and scenario data.

[1792] "Means for recognizing emotions" refers to technology for analyzing and understanding a user's emotional state from their responses and behavior.

[1793] The "means for adjusting feedback and response content" is a mechanism for appropriately changing or optimizing the feedback provided or the next response based on the emotion recognition results.

[1794] The "means for providing evaluation" is a function for judging the accuracy and appropriateness of the user's response and notifying the trainee of the result.

[1795] This invention is a system that uses a generative AI model and an emotion engine to conduct simulated customer service conversations. The system is composed of a server, a terminal, an emotion engine, and a user.

[1796] server:

[1797] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative AI model. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[1798] Device:

[1799] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[1800] Emotion Engine:

[1801] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[1802] User:

[1803] The participants are mainly trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through mock conversations. System administrators monitor the trainees' progress and evaluations, and provide appropriate feedback and guidance.

[1804] Program processing:

[1805] The server hosts the application and generates mock conversation scenarios for trainees to perform. A generative AI model runs on the server and generates appropriate questions and scenarios based on dummy customer information and an FAQ database. The trainee's response data is sent to the server and compared with the response procedures. An emotion engine recognizes the user's emotions, and feedback and response content are adjusted based on the results.

[1806] Examples:

[1807] For example, consider a scenario in which a new employee at a brick-and-mortar store is using their smartphone to train their customer service skills with this system. If the following prompt is used:

[1808] plaintext

[1809] "We'll learn what to do if a customer is in a hurry and wants to know where an item is. Let's simulate how you would respond when they ask, 'I'm in a hurry, where's the shampoo?'"

[1810] Based on this prompt, the server generates a scenario and displays it on the user's (trainee's) smartphone. When the trainee enters a response, it is sent to the server and compared with the customer service procedures and FAQ database. At the same time, the emotion engine recognizes the emotion from the response, and the server generates appropriate feedback based on that and sends it back to the trainee. For example, if a customer expresses that they are "in a hurry," and the emotion engine recognizes tension or confusion from the user's response, the server will provide feedback to soften the tone, helping to ensure a smoother response next time.

[1811] In this way, trainees can acquire skills to handle a variety of scenarios in a realistic environment, which is expected to improve their performance when dealing with actual customers.

[1812] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1813] Step 1:

[1814] The user displays the system login screen on the terminal and enters their ID and password. The input data is sent to the server, which compares the authentication information with the database and generates an authentication result. If the authentication is successful, the terminal displays the main menu; if it fails, it displays an error message.

[1815] Step 2:

[1816] The user selects "Start Role-Playing" from the main menu. The server generates a dummy customer profile and scenario data and sends them to the device. The device receives this and displays a dialogue screen. The generative AI model generates the first question based on the scenario and sends it to the device.

[1817] Step 3:

[1818] The user inputs a response to the question and sends it to the terminal. The terminal then sends the input data to the server, which then analyzes the response. The analysis includes checking the response against the response procedure and FAQ database.

[1819] Step 4:

[1820] Based on the analysis results, the server uses an emotion engine to recognize emotions from the user's responses. The emotion engine analyzes text and voice data to recognize the user's emotional state. The recognized emotion data is then sent to the server.

[1821] Step 5:

[1822] The server adjusts the feedback and response content based on the emotion recognition results, optimizing the tone and content of the feedback message depending on the recognized emotion, and generating the next question or scenario as needed and sending it to the device.

[1823] Step 6:

[1824] As a substep, the terminal displays a feedback message on the interactive screen and provides appropriate feedback to the user, who can then confirm the feedback and re-enter a corrected response if necessary.

[1825] Step 7:

[1826] When the user completes all scenarios or clicks the "Exit" button at any point, the terminal notifies the server of the end operation. The server records the user's progress data and evaluation results in a database and ends the session. A logout screen is displayed on the terminal to notify the user of the end of the session.

[1827] Through the above processing steps, this system can provide realistic customer service training and support users in improving their skills.

[1828] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1829] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1830] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1831] [Fourth embodiment]

[1832] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1833] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1834] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1835] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1836] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1837] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1838] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1839] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1840] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1841] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1842] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1843] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1844] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1845] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide efficient and high-quality training. This system is composed of a server, terminals, and users.

[1846] The server hosts the data and applications necessary for the training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, etc. The server is equipped with a generative artificial intelligence that plays the role of a customer and generates mock conversations.

[1847] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also displays feedback and evaluation results sent from the server to the trainee.

[1848] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[1849] As a specific example, the flow of a user logging in to a system will be described.

[1850] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1851] 2. The terminal sends the entered authentication information to the server.

[1852] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[1853] Next, the flow of role-playing is shown below.

[1854] 1. The user selects "Start Role-Playing" from the main menu.

[1855] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[1856] 3. The terminal displays the interactive screen.

[1857] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[1858] 5. The user enters and submits a response to the question.

[1859] 6. The terminal sends the entered response to the server,

[1860] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[1861] 7. The server generates the evaluation results and generates appropriate feedback messages.

[1862] 8. The terminal displays the feedback message on the interactive screen.

[1863] By using this system, trainees can learn in an environment that is similar to actual customer service, allowing them to efficiently master appropriate customer service techniques. In addition, if a trainee makes a mistake, the system will immediately suggest the correct way to respond, improving the quality of the training.

[1864] Furthermore, the system has the ability to generate dummy customer information and scenario data and send them to trainees' devices, allowing for practical learning that can handle a variety of scenarios. This system helps prevent a decline in customer service quality due to resource shortages during busy periods and individual differences.

[1865] The above is an embodiment of the present invention.

[1866] The processing flow will be explained below.

[1867] Login Process

[1868] Step 1:

[1869] The user displays the login screen for the system on the terminal and enters their ID and password.

[1870] Step 2:

[1871] The terminal transmits the entered authentication information to the server.

[1872] Step 3:

[1873] The server compares the received authentication information with a database and generates an authentication result.

[1874] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[1875] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[1876] Step 4:

[1877] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[1878] Role-playing begins

[1879] Step 1:

[1880] The user clicks the "Start Role-Playing" button from the main menu.

[1881] Step 2:

[1882] The device notifies the server of the click event.

[1883] Step 3:

[1884] The server generates dummy customer information and scenario data and transmits them to the terminal.

[1885] The dummy customer information includes names, addresses, and past donation details.

[1886] The scenario data includes specific conversation flows and questions.

[1887] Step 4:

[1888] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[1889] Mock conversation progression

[1890] Step 1:

[1891] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[1892] Step 2:

[1893] The terminal displays the generated question on an interactive screen.

[1894] Step 3:

[1895] The user enters a response to the displayed question and clicks the submit button.

[1896] Step 4:

[1897] The terminal sends the entered response to the server.

[1898] Step 5:

[1899] The server analyzes the received response and matches it with procedures and FAQ data.

[1900] Step 6:

[1901] The server generates an evaluation result based on the response content and generates an appropriate feedback message.

[1902] Step 7:

[1903] The terminal displays the feedback message on the interactive screen.

[1904] Providing feedback

[1905] Step 1:

[1906] The terminal receives and displays the feedback message sent from the server.

[1907] The feedback includes a pointer to what went wrong and the correct way to respond.

[1908] Step 2:

[1909] The user checks the feedback and inputs a revised response.

[1910] Step 3:

[1911] The terminal retransmits the corrected response to the server.

[1912] Step 4:

[1913] The server again analyzes the response and evaluates whether it is an appropriate response.

[1914] If appropriate, the following scenarios proceed:

[1915] If it's not appropriate, provide feedback again.

[1916] Role-playing ends

[1917] Step 1:

[1918] The user can complete all scenarios or click the "Exit" button midway through.

[1919] Step 2:

[1920] The terminal notifies the server of the termination operation.

[1921] Step 3:

[1922] The server records the user's progress data and evaluation results in a database and ends the session.

[1923] Step 4:

[1924] The terminal displays a logout screen to notify the user that the session has ended.

[1925] Example 1

[1926] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1927] Conventional customer service training systems have struggled to provide trainees with realistic mock conversations to help them acquire practical customer service skills. Additionally, feedback is often delayed, and it's difficult to suggest appropriate corrections in real time. As a result, the quality of training declines, and trainees may lack customer service skills in their actual work.

[1928] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1929] In this invention, the server includes means for storing training data and virtual customer data, means for generating dialogue based on the virtual customer data using generative artificial intelligence, means for comparing responses input by the trainee with training procedures and evaluating them, and means for providing feedback to the trainee based on the evaluation. This enables the trainee to acquire practical conversation skills through realistic simulated conversations, and by providing appropriate, real-time feedback and suggestions for correction, the quality of the training can be improved.

[1930] - "Training purpose" refers to the purpose for trainees to acquire response skills and knowledge.

[1931] "Generative AI" is a type of AI that performs natural language processing and dialogue generation.

[1932] "Simulating a dialogue" means recreating an actual conversation environment and conducting training through virtual conversations.

[1933] A "system" is a device that integrates multiple elements and functions and is designed to achieve a specific purpose.

[1934] "Training Data" is information regarding guidelines and procedures used during training.

[1935] "Virtual customer data" refers to information about fictitious customers used in simulated conversations.

[1936] A "means" is a method or device used to achieve a particular purpose.

[1937] A "Trainer" is an individual who uses the system to improve their customer service skills.

[1938] "Training procedures" are procedures or guidelines to be followed when dealing with customers.

[1939] "Evaluating" means judging the appropriateness of the trainee's response.

[1940] "Feedback" refers to information about the trainee's evaluation of their response and areas for improvement.

[1941] "Dialogue based on virtual customer data" is a conversation generated based on fictitious customer information.

[1942] "Suggesting corrections" means suggesting procedures or methods for improvement when a response is inappropriate.

[1943] A "terminal" is a device used to access the system.

[1944] This invention is a training system that uses a generative AI model to conduct simulated customer service conversations, and aims to provide efficient and high-quality training. This system is composed of a server, a terminal, and a user.

[1945] The server hosts the data and applications required for training. Specifically, training data, virtual customer data, an FAQ database, and other items are stored on the server. The server is equipped with a generative AI model, which acts as the virtual customer and generates simulated conversations. Specific examples of servers include high-performance server machines (e.g., highly reliable server equipment, cloud computing platforms). Examples of software include the Linux OS, a generative AI model (e.g., an AI model that performs natural language processing), and a database management system (e.g., MySQL).

[1946] The terminal is the device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and sends responses entered by the trainee to the server. It also has the role of displaying feedback and evaluation results sent from the server to the trainee. A browser (e.g., a standard web browser) or a communication application (e.g., a communication app using the WebSocket protocol) is used.

[1947] The users are trainees and system administrators. Trainees log in to the system using a terminal and receive training in customer service through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[1948] As a concrete example, the flow of a user logging in to the system is shown below.

[1949] 1. The user displays the system login screen on the terminal and enters their ID and password.

[1950] 2. The terminal sends the entered authentication information to the server.

[1951] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[1952] Next, the flow of role-playing is shown below.

[1953] 1. The user selects "Start Role-Playing" from the main menu.

[1954] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[1955] 3. The terminal displays the interactive screen.

[1956] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[1957] 5. The user enters and submits a response to the question.

[1958] 6. The terminal sends the entered response to the server.

[1959] 7. The server analyzes the response and compares it with pre-defined training procedures and FAQ data.

[1960] 8. The server generates the evaluation results and generates appropriate feedback messages.

[1961] 9. The terminal displays the feedback message on the interactive screen.

[1962] As a concrete example, the prompt sentence is shown below.

[1963] Prompt statement:

[1964] "My phone bill seems unusually high lately. How does it compare to my last bill?"

[1965] Using this system, trainees can learn in an environment that closely resembles actual customer interactions, allowing them to efficiently master customer interaction methods. In addition, if an incorrect interaction occurs, the correct interaction method is immediately displayed, improving the quality of training. Furthermore, the system generates virtual customer data and scenario data, allowing for practical learning that corresponds to a variety of scenarios. This system can prevent a decline in interaction quality due to resource shortages during busy periods or individual differences.

[1966] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1967] Explain the processing flow of the system program by dividing it into processing steps

[1968] Step 1: User Login

[1969] Input: User ID and password

[1970] Output: Authentication result (success / failure)

[1971] What happens:

[1972] 1. The user uses the terminal to display the system login screen and enters their ID and password.

[1973] Specific operation: Open the login page in your browser and enter the ID "user01" and password "password123" in the form.

[1974] 2. The terminal sends the entered authentication information to the server.

[1975] What it does: Sends authentication information in an HTTP POST request.

[1976] 3. The server checks the received authentication information against its database.

[1977] Specific behavior: Performs a database query to verify the ID and password.

[1978] 4. The server generates an authentication result and sends it to the terminal.

[1979] Specific operation: If authentication is successful, generate and send a token; if it fails, send an error message "Authentication failed."

[1980] Step 2: Prepare to start role-playing

[1981] Input: Request to start roleplaying

[1982] Output: Virtual customer profile and scenario data

[1983] What happens:

[1984] 1. The user selects "Start Role-Playing" from the main menu.

[1985] Specific action: Click a button on the menu.

[1986] 2. The server generates a virtual customer profile and scenario data and sends them to the terminal.

[1987] Specific operation: Generate a virtual customer name "Customer A" and a scenario "Regarding bill payment" and send them to the terminal in JSON format.

[1988] 3. The terminal displays the interactive screen.

[1989] Specific operation: Display the received profile and scenario data on the screen.

[1990] Step 3: Role-playing dialogue

[1991] Input: User response

[1992] Output: The next question generated by the generative AI

[1993] What happens:

[1994] 1. The generation AI (server) generates an initial question based on the scenario and sends it to the device.

[1995] Specific behavior: Generate and send the question "Hello, this is Customer A. I'm having trouble paying my bill."

[1996] 2. The user enters and submits a response to the question.

[1997] Specific action: Enter the response "Please tell me about the specific problem" and click the send button.

[1998] 3. The terminal sends the entered response to the server.

[1999] Specific operation: The response content is sent to the server via an HTTP POST request.

[2000] 4. The server analyzes the response and compares it with the configured training procedures and FAQ data.

[2001] Specific behavior: Performs text analysis and searches for relevant entries in the FAQ database.

[2002] Step 4: Feedback and Rating

[2003] Input: User responses and analysis results

[2004] Output: Feedback message

[2005] What happens:

[2006] 1. The server generates the evaluation results and generates a feedback message.

[2007] What it does: Evaluate the response and generate feedback like "Your response was unclear. Please check the amount next time."

[2008] 2. The terminal displays the feedback message on the interactive screen.

[2009] What it does: Displays the feedback received on the screen.

[2010] Step 5: Logout

[2011] Input: Logout request

[2012] Output: Logout result, login screen displayed

[2013] What happens:

[2014] 1. The user selects logout from the menu.

[2015] Specific action: Click the "Logout" button on the main menu.

[2016] 2. The device sends a logout request to the server.

[2017] Specific behavior: Sends a logout request as an HTTP POST request.

[2018] 3. The server invalidates the session and redirects to the login screen.

[2019] Specific actions: End the session and send the URL of the login screen to the terminal.

[2020] The above are the specific processing steps of this system.

[2021] (Application example 1)

[2022] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2023] Current customer service training programs struggle to provide a practical training environment that closely resembles actual customer interactions, making it difficult for trainees to efficiently acquire the skills to deal with the diverse situations they will encounter in the field. Furthermore, current systems often lack real-time feedback and evaluation of trainees' customer service quality, limiting the effectiveness of training. In particular, there are few systems that support training while on the move or in a brick-and-mortar store, making it difficult to provide realistic training.

[2024] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[2025] In this invention, the server includes means for generating questions based on the dummy customer information using a generative AI, means for comparing responses entered by the trainee with the service procedures and evaluating them, means for providing feedback to the trainee based on the evaluation, means for providing a simulation environment for the trainee using smart glasses, means for analyzing voice input from the smart glasses and transmitting it to the generative AI model, and means for displaying feedback received from the server on the display of the smart glasses. This enables the trainee to effectively acquire customer service skills while receiving evaluation and feedback in real time in an immersive environment close to reality.

[2026] "Generative AI" refers to AI technology that uses natural language processing technology to generate human-like conversations and content.

[2027] "Customer service" refers to the business process of responding to inquiries and requests from customers.

[2028] "Simulated conversation" refers to fictitious dialogue generated to simulate real conversations.

[2029] "Service procedures" refer to standard procedures and guidelines to be followed when dealing with customers.

[2030] "Dummy customer information" refers to fictitious customer profiles and scenario information used in training.

[2031] "Evaluation" refers to the process of judging the quality and appropriateness of the trainee's response and giving them a score and feedback.

[2032] "Feedback" refers to information provided to trainees, including areas for improvement and appropriate guidance.

[2033] "Smart glasses" are glasses with an integrated display that are devices capable of displaying information and inputting and outputting voice.

[2034] "Simulation environment" refers to a virtual environment in which trainees can practice in situations similar to their actual work.

[2035] "Generative AI model" refers to an artificial intelligence model that has been pre-trained to perform generative tasks.

[2036] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide real-time feedback to trainees in brick-and-mortar stores using smart glasses.

[2037] Hardware used

[2038] Smart Glasses

[2039] These are glasses with an integrated display that can display information and input and output voice. Specific examples include Google Glass and Vuzix Blade.

[2040] server

[2041] Hosting generative artificial intelligence and databases for real-time data processing. Specific examples include AWS EC2 instances.

[2042] Software used

[2043] Generative Artificial Intelligence

[2044] It uses natural language processing technology to generate simulated conversations in real time, a specific example being OpenAI GPT-4.

[2045] Database

[2046] Dummy customer information and scenario data are stored and managed. A specific example is AWS RDS (MySQL).

[2047] Real-time communication

[2048] A technology for data communication between a server and smart glasses. A specific example is WebSocket communication.

[2049] System configuration and data processing

[2050] server

[2051] 1. Generate dummy customer information

[2052] The server generates dummy customer information and scenario data using the generative AI model, using the following prompt sentences for this process:

[2053] Sample prompt 1: Generate a customer profile

[2054] Generate a customer profile containing the following information:

[2055] name

[2056] age

[2057] sex

[2058] Purchase Intent

[2059] Customer questions and concerns

[2060] 2. Mock conversation scenario generation

[2061] The server generates questions based on the scenario data and uses the following prompts to make the simulation feel more realistic for the trainee:

[2062] Prompt example 2: Training scenario generation

[2063] Train your customer service representatives using the following scenarios:

[2064] A customer asks for more information about a product.

[2065] A customer asks about returns or exchanges.

[2066] Customers want to know about specific promotions and offers.

[2067] 3. Ratings and Feedback

[2068] The server analyzes the trainee's responses, compares them with the response procedures, and evaluates them. The evaluation results and feedback are generated in real time and displayed on the smart glasses.

[2069] Smart Glasses

[2070] 1. Real-time interaction

[2071] The smart glasses receive scenario data from the server and display it on the screen. They also analyze the trainee's voice input and send it to the server.

[2072] 2. Feedback display

[2073] Feedback received from the server is displayed in real time, allowing trainees to quickly learn how to respond appropriately.

[2074] A concrete example of the entire system

[2075] 1. Part of a training scenario

[2076] For example, a dialogue about a "question about product description" is simulated. The following flow is an example.

[2077] Dummy customer question: "What is the battery life of this product?"

[2078] Trainee response: "About 10 hours."

[2079] Generative AI evaluation: "That's a good answer, but it depends on the specific use case, so it would be good to explain that it varies depending on the use case."

[2080] Using this system, trainees can effectively acquire customer service skills in a realistic environment, receiving real-time evaluations and feedback.

[2081] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[2082] Step 1:

[2083] The user puts on the smart glasses and logs into the system.

[2084] Input: User ID and password

[2085] Data processing / calculation: The authentication information is sent to the server and checked against a database.

[2086] Output: The authentication result is sent from the server to the smart glasses.

[2087] Specific operation: If authentication is successful, the server displays the main menu on the smart glasses display. If authentication is unsuccessful, it displays an error message.

[2088] Step 2:

[2089] The user selects "Start Training" from the main menu.

[2090] Input: Select Start Training on the main menu

[2091] Data processing / calculation: The server generates dummy customer information and scenario data using the generative AI model.

[2092] Output: The generated dummy customer information and scenario data are sent to the smart glasses.

[2093] Specific operation: The server generates data using the prompt sentence and sends it to the smart glasses.

[2094] Step 3:

[2095] An interactive screen is displayed on the smart glasses.

[2096] Input: Dummy customer information and scenario data sent from the server

[2097] Data processing / calculation: Dummy customer questions are displayed on the smart glasses display.

[2098] Output: Trainee reviews the question and prepares a response.

[2099] Specific operation: The smart glasses analyze the information received from the server and display it on the screen.

[2100] Step 4:

[2101] The user responds to the dummy customer's questions by voice input.

[2102] Input: Trainee's voice response

[2103] Data processing / calculation: The smart glasses convert voice input into text and send it to the server.

[2104] Output: Voice input is sent to the server as text data.

[2105] How it works: The smart glasses use a built-in microphone to convert voice data into text using voice recognition technology.

[2106] Step 5:

[2107] The server analyzes and evaluates the response it receives.

[2108] Input: Transcribed trainee responses

[2109] Data processing / calculation: The server compares the response with existing procedures and evaluates it using a generative AI model.

[2110] Output: Generates evaluation results and feedback messages.

[2111] Specific Actions: The server uses an analysis algorithm to assess relevance and generate a feedback message.

[2112] Step 6:

[2113] The feedback message received from the server is displayed on the smart glasses.

[2114] Input: Feedback message sent by the server

[2115] Data processing / calculation: Display feedback on the smart glasses display.

[2116] Output: Trainee reviews the feedback and understands areas for improvement.

[2117] How it works: The smart glasses provide real-time feedback and show trainees how to respond appropriately.

[2118] Step 7:

[2119] The user selects the next training scenario or ends the training.

[2120] Input: Trainee selection (next scenario or exit)

[2121] Data processing / calculation: The server receives commands to generate new scenario data or to shut down the system.

[2122] Output: New scenario data is sent to the smart glasses or a completion message is displayed.

[2123] Specific behavior: The server generates appropriate data depending on the situation and provides the next action based on the trainee's selection.

[2124] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[2125] This invention is a system for conducting simulated customer service conversations using generative artificial intelligence, and aims to provide more advanced and adaptable training by combining it with an emotion engine that recognizes the user's emotions. This system is composed of a server, a terminal, an emotion engine, and a user.

[2126] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative artificial intelligence. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[2127] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[2128] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[2129] The users are trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through simulated conversations. The system administrator monitors the trainees' progress and evaluations, and provides appropriate feedback and guidance.

[2130] As a concrete example, we will explain the flow of a user logging in to the system and performing role-playing with an emotion engine built in.

[2131] Login Process

[2132] 1. The user displays the system login screen on the terminal and enters their ID and password.

[2133] 2. The terminal sends the entered authentication information to the server.

[2134] 3. The server checks the received authentication information against the database and generates an authentication result. If successful, it displays the main menu, otherwise it displays an error message.

[2135] Role-playing begins

[2136] 1. The user selects "Start Role-Playing" from the main menu.

[2137] 2. The server generates a profile and scenario data for the dummy customer and sends it to the terminal.

[2138] 3. The terminal displays the interactive screen.

[2139] 4. The generation AI (server) generates the first question based on the scenario and sends it to the device.

[2140] Mock conversation progression

[2141] 1. The user enters and submits a response to a question.

[2142] 2. The terminal sends the entered response to the server.

[2143] 3. The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[2144] 4. The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[2145] 5. The server generates evaluation results and feedback messages based on the recognized emotions.

[2146] 6. The terminal displays the feedback message on the interactive screen.

[2147] For example, if a user has a strong emotion (anger or anxiety), the emotion engine will recognize this and the server can change the response and tone of the feedback to teach the trainee how to respond appropriately.

[2148] Providing feedback

[2149] 1. The device receives and displays the feedback message sent from the server. The feedback includes the error indication and the correct response.

[2150] 2. The user reviews the feedback and re-enters the corrected response.

[2151] 3. The terminal resends the corrected response to the server.

[2152] 4. The server again analyzes the response and evaluates whether it is an appropriate response.

[2153] If appropriate, the following scenarios proceed:

[2154] If it's not appropriate, provide feedback again.

[2155] Role-playing ends

[2156] 1. The user completes all scenarios or clicks the "Exit" button midway through.

[2157] 2. The terminal notifies the server of the termination operation.

[2158] 3. The server records the user's progress and evaluation results in a database and ends the session.

[2159] 4. The terminal displays a logout screen to notify the user that the session has ended.

[2160] By using this system, trainees can learn in an environment that is close to actual customer interactions, allowing them to efficiently master appropriate ways of interacting with customers. In addition, the introduction of an emotion engine provides appropriate feedback based on the user's emotions, enabling more practical training.

[2161] The processing flow will be explained below.

[2162] Login Process

[2163] Step 1:

[2164] The user displays the login screen for the system on the terminal and enters their ID and password.

[2165] Step 2:

[2166] The terminal transmits the entered authentication information to the server.

[2167] Step 3:

[2168] The server compares the received authentication information with a database and generates an authentication result.

[2169] If successful: User session information is generated and the main menu screen data is returned to the terminal.

[2170] If unsuccessful: Authentication failure data including an error message is returned to the terminal.

[2171] Step 4:

[2172] The terminal displays the authentication result on the screen, and if successful, transitions to the main menu screen.

[2173] Role-playing begins

[2174] Step 1:

[2175] The user clicks the "Start Role-Playing" button from the main menu.

[2176] Step 2:

[2177] The device notifies the server of the click event.

[2178] Step 3:

[2179] The server generates dummy customer information and scenario data and transmits them to the terminal.

[2180] The dummy customer information includes names, addresses, and past donation details.

[2181] The scenario data includes specific conversation flows and questions.

[2182] Step 4:

[2183] The terminal creates and displays an interactive screen based on the received dummy customer information and scenario data.

[2184] Mock conversation progression

[2185] Step 1:

[2186] The generation AI (server) generates the first question based on the scenario and sends it to the terminal.

[2187] Step 2:

[2188] The terminal displays the generated question on an interactive screen.

[2189] Step 3:

[2190] The user enters a response to the displayed question and clicks the submit button.

[2191] Step 4:

[2192] The terminal sends the entered response to the server.

[2193] Step 5:

[2194] The server analyzes the response and compares it with the response procedures and FAQ data.

[2195] Step 6:

[2196] The emotion engine (server) recognizes emotions from the user's input and sends the results to the server.

[2197] Step 7:

[2198] The server generates an evaluation result and a feedback message based on the recognized emotion.

[2199] Step 8:

[2200] The terminal displays the feedback message on the interactive screen.

[2201] Providing feedback

[2202] Step 1:

[2203] The terminal receives and displays the feedback message sent from the server.

[2204] The feedback includes a pointer to what went wrong and the correct way to respond.

[2205] Step 2:

[2206] The user checks the feedback and re-enters the corrected response.

[2207] Step 3:

[2208] The terminal retransmits the corrected response to the server.

[2209] Step 4:

[2210] The server again analyzes the response and evaluates whether it is an appropriate response.

[2211] If appropriate, the following scenarios proceed:

[2212] If it's not appropriate, provide feedback again.

[2213] Role-playing ends

[2214] Step 1:

[2215] The user can complete all scenarios or click the "Exit" button midway through.

[2216] Step 2:

[2217] The terminal notifies the server of the termination operation.

[2218] Step 3:

[2219] The server records the user's progress data and evaluation results in a database and ends the session.

[2220] Step 4:

[2221] The terminal displays a logout screen to notify the user that the session has ended.

[2222] Example 2

[2223] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2224] Conventional customer service training systems have made it difficult for trainees to learn in an environment similar to actual customer service. Furthermore, general feedback systems are unable to provide appropriate feedback that takes into account the trainee's emotional state, limiting the effectiveness of the training. Therefore, it has been a challenge for trainees to efficiently acquire the appropriate skills for actual customer service situations.

[2225] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for storing response procedures and virtual customer information, means for generating questions based on the virtual customer information using the generative AI model, and means for comparing responses entered by the trainee with the response procedures and evaluating them. This enables the trainee to efficiently train in a simulated conversation environment that is similar to actual customer service while recognizing user emotions using an emotion analysis engine.

[2226] A "generative artificial intelligence model" is an artificial intelligence technology that generates natural language based on training data, and is used to generate questions and responses from virtual customers.

[2227] An "emotion analysis engine" is a technology that recognizes emotions from text or voice data entered by a user and provides feedback based on those emotions.

[2228] "Service procedures" define the standard procedures and responses to be used when dealing with customers.

[2229] "Virtual customer information" refers to fictitious customer profiles and background information generated based on simulated conversation scenarios.

[2230] "Trainees" refer to learners who intend to use this system to acquire customer service skills.

[2231] The "means for generating questions" refers to techniques or algorithms for generating questions based on virtual customer information using a generative artificial intelligence model.

[2232] "Means for verification and evaluation" refers to techniques or methods for comparing responses entered by trainees with pre-defined response procedures and evaluating their accuracy and appropriateness.

[2233] "Means for providing feedback" refers to technologies and algorithms that provide trainees with appropriate advice and suggestions for correction based on the results of evaluation and sentiment analysis.

[2234] "Scenario data" refers to data that includes a series of hypothetical interactions and situations used to structure the progress of a simulated conversation.

[2235] This invention is a simulated conversation system for customer service that combines a generative artificial intelligence model and an emotion analysis engine, and aims to enable trainees to efficiently train in an environment that closely resembles actual customer service. The system is composed of a server, a terminal, an emotion analysis engine, and a user.

[2236] Hardware and software used

[2237] server

[2238] The server hosts the generative AI model, sentiment analysis engine, response procedures, virtual customer information, and FAQ database. Specifically, it is implemented with the following configuration.

[2239] Hardware: powerful processor, sufficient memory, and large storage capacity for database storage

[2240] Software: Linux-based operating systems such as Ubuntu or CentOS, database management systems such as MySQL or PostgreSQL, and machine learning libraries such as TensorFlow

[2241] Terminal

[2242] Terminals are devices used by trainees, including PCs, tablets, smartphones, etc. Terminals communicate with the server and perform the following processes:

[2243] Hardware: Any device that supports a web browser

[2244] Software: A modern web browser (Google Chrome, Mozilla Firefox, etc.), JavaScript, and a front-end framework such as React

[2245] Sentiment Analysis Engine

[2246] The sentiment analysis engine is used to analyze text and voice data to recognize the emotions of the user.

[2247] Software: Natural language processing (NLP) libraries (NLTK, spaCy, etc.), speech recognition libraries (Google Speech to Text API, etc.)

[2248] Explanation of program processing

[2249] Generating training scenarios

[2250] 1. The user displays the system login screen on the terminal and enters their ID and password.

[2251] 2. The device sends this information to the server.

[2252] 3. The server checks the user information against the database, and if authentication is successful, displays the main menu on the terminal.

[2253] 4. When the user selects "Start role-playing," the server generates virtual customer information and scenario data and sends them to the terminal.

[2254] 5. The device displays a dialogue screen and asks questions from the generating AI.

[2255] Mock conversation progression

[2256] 1. The user enters a response to the question posed by the generating AI and submits it.

[2257] 2. The terminal sends the response data to the server.

[2258] 3. The server analyzes the received data and compares it with the response procedures and FAQ data.

[2259] 4. The emotion analysis engine (server) analyzes the user's input text and assigns emotion labels.

[2260] 5. Based on this information, the server generates evaluation results and feedback messages and sends them to the terminal.

[2261] Providing feedback

[2262] 1. The device receives and displays a feedback message, which includes a description of the error and the correct response.

[2263] 2. The user reviews the feedback, re-enters the revised response, and submits it.

[2264] 3. The terminal retransmits the retyped response to the server.

[2265] 4. The server analyzes again and decides whether to proceed to the next scenario or provide feedback again.

[2266] End of role-playing

[2267] 1. The user ends the role-playing by completing all scenarios or by clicking the "Exit" button.

[2268] 2. The device notifies the server of this information, and the server records the user's progress data and evaluation results in a database.

[2269] 3. The terminal displays a logout screen to notify the user that the session has ended.

[2270] Specific examples

[2271] As a concrete example, the following mock conversation scenario can be considered.

[2272] 1. User: "I have a question about a product."

[2273] 2. Generative AI (server): "Which feature of which product are you asking about?"

[2274] 3. User: "What is the return process?"

[2275] 4. The emotion analysis engine (server) recognizes "anxiety" from the user's message and softens the tone of the response.

[2276] 5. Server: "For more information on the return process, please see this guide."

[2277] Prompt Sentence Examples

[2278] "Generate the following role-playing scenario: A user has a question about a new product. Initiate a dialogue that describes the product's features."

[2279] "Generate an appropriate response based on the user's input: 'I would like to return this item.'"

[2280] The specific processes and procedures used in implementing the present invention have been described above. This system enables trainees to efficiently receive practical training that takes into account emotions in an environment that closely resembles actual customer interactions.

[2281] The flow of the identification process in the second embodiment will be described with reference to FIG.

[2282] Step 1:

[2283] The user displays the system login screen on the terminal and enters their ID and password.

[2284] Input: ID, password

[2285] Output: None

[2286] Specifically, the user opens a browser, accesses the specified URL (login page), enters the ID and password in the login form, and clicks the submit button.

[2287] Step 2:

[2288] The device sends the entered authentication information to the server, encrypting it and using the HTTPS protocol.

[2289] Input: ID, password (encrypted)

[2290] Output: Authentication request data

[2291] Specifically, it uses JavaScript to capture form data and sends the data to the server using an AJAX request.

[2292] Step 3:

[2293] The server compares the received authentication information with a database and generates an authentication result.

[2294] Input: Authentication request data

[2295] Output: Authentication result (success / failure)

[2296] Specifically, the server executes an SQL query to match the corresponding user information from the database, and if authentication is successful, generates a token to start a session.

[2297] Step 4:

[2298] The server sends the authentication result to the terminal. If the authentication is successful, it sends an HTML page to display the main menu to the terminal. If not, it sends a page containing an error message.

[2299] Input: Authentication result

[2300] Output: HTML page (main menu or error messages)

[2301] Specifically, the server selects an appropriate HTML template based on the authentication result and sends it to the client.

[2302] Step 5:

[2303] The user selects "Start Role-Playing" from the main menu.

[2304] Input: User clicks

[2305] Output: Request data

[2306] As a specific operation, the user clicks the "Start Role Playing" button on the main menu, and request data is generated.

[2307] Step 6:

[2308] The device sends a "start role-playing" request to the server.

[2309] Input: Request data

[2310] Output: Start request

[2311] As a specific operation, the terminal sends a start request to the server using an AJAX request.

[2312] Step 7:

[2313] The server generates virtual customer information and scenario data and transmits them to the terminal. The server generates a scenario using a generative artificial intelligence model.

[2314] Input: Start Request

[2315] Output: Virtual customer information, scenario data

[2316] Specifically, the server calls the generation AI, sends a scenario generation prompt, receives a response, and then sends the generated scenario data and virtual customer information to the client.

[2317] Step 8:

[2318] The device displays a dialogue screen and asks the first question from the generating AI.

[2319] Input: Virtual customer information, scenario data

[2320] Output: Showing the first question

[2321] Specifically, the device dynamically renders an interactive screen using HTML and JavaScript based on the data it receives.

[2322] Step 9:

[2323] The user enters and submits responses to questions posed by the generating AI.

[2324] Input: User response

[2325] Output: Response data

[2326] As a specific operation, the user enters information into the text box on the interactive screen and clicks the send button.

[2327] Step 10:

[2328] The terminal captures the entered responses and sends them to the server.

[2329] Input: Response data

[2330] Output: Response data (sent)

[2331] As a specific operation, the terminal uses an AJAX request to send response data to the server.

[2332] Step 11:

[2333] The server analyzes the response and compares it with pre-defined response procedures and FAQ data.

[2334] Input: Response data

[2335] Output: Evaluation results

[2336] Specifically, the server uses natural language processing technology to analyze the response data and compare it with response procedures and FAQ data.

[2337] Step 12:

[2338] The emotion analysis engine (server) recognizes emotions from the user's input and sends the results to the server.

[2339] Input: Response data

[2340] Output: Emotion label

[2341] Specifically, the sentiment analysis engine uses NLP techniques to recognize the emotional state of text and generate sentiment labels.

[2342] Step 13:

[2343] The server generates a feedback message based on the evaluation results and emotion labels and sends it to the device.

[2344] Input: Evaluation result, emotion label

[2345] Output: Feedback message

[2346] Specifically, the server uses a rating algorithm to generate appropriate feedback content.

[2347] Step 14:

[2348] The terminal displays the feedback message on the interactive screen.

[2349] Input: Feedback message

[2350] Output: Show feedback

[2351] As a specific operation, the terminal renders the received feedback message at a specified location on the interactive screen.

[2352] Step 15:

[2353] The user checks the feedback, re-enters the corrected response, and submits it.

[2354] Input: Corrected response

[2355] Output: Corrected response data

[2356] Specifically, the user checks the feedback, then enters a revised response in the response text box, and clicks the send button again.

[2357] Step 16:

[2358] The terminal resends the modified response to the server.

[2359] Input: Corrected response data

[2360] Output: Corrected response data (sent)

[2361] As a specific operation, the terminal again uses an AJAX request to send the modified response data to the server.

[2362] Step 17:

[2363] The server again analyzes the response and evaluates whether it is an appropriate response.

[2364] Input: Corrected response data

[2365] Output: Reevaluation results

[2366] Specifically, the server performs the same analysis and evaluation, and based on the results, decides whether to proceed to the next scenario or provide feedback again.

[2367] Step 18:

[2368] The user can complete all scenarios or click the "Exit" button midway through.

[2369] Input: End operation

[2370] Output: Termination request data

[2371] Specifically, when the user clicks the quit button, a quit confirmation dialog box is displayed.

[2372] Step 19:

[2373] The terminal notifies the server of the termination operation.

[2374] Input: Termination request data

[2375] Output: Completion notice

[2376] As a specific operation, the terminal uses an AJAX request to send termination request data to the server.

[2377] Step 20:

[2378] The server records the user's progress data and evaluation results in a database and ends the session.

[2379] Input: Completion notice, progress data, evaluation results

[2380] Output: Recording success notification

[2381] Specifically, the server uses an SQL insert or update statement to record the necessary information in the database.

[2382] Step 21:

[2383] The terminal displays a logout screen to notify the user that the session has ended.

[2384] Input: Recording success notification

[2385] Output: Logout screen

[2386] Specifically, the terminal displays an HTML page that includes a message such as "You have been logged out."

[2387] (Application example 2)

[2388] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2389] Existing customer service training systems have the following problems. First, the mock conversation scenarios are fixed, making it difficult to reproduce the diverse situations that occur in actual customer service interactions. Second, they are unable to respond to fluctuations in customer emotions, limiting opportunities to acquire realistic customer service skills. Third, they lack a mechanism for accurately assessing trainees' emotions and responses and providing appropriate feedback. There is a need for a system that can solve these problems and provide more comprehensive and practical customer service training.

[2390] The identification process by the identification 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 storing response procedures and dummy customer information, means for generating questions based on the dummy customer information using the generative artificial intelligence, means for recognizing emotions from responses entered by trainees, means for adjusting feedback and response content based on the emotion recognition results, and means for providing feedback to trainees based on the evaluation. This enables practical customer response training that can be adapted to a variety of situations.

[2391] "Generative AI" is an AI system that automatically generates appropriate responses and scenarios based on user input and the situation.

[2392] "Service procedures" refer to specific steps or protocols to be followed when serving customers.

[2393] "Dummy customer information" refers to fictitious customer data created based on actual customer information and used in training.

[2394] "Means for generating questions" refers to the function of artificial intelligence to automatically create questions based on dummy customer information and scenario data.

[2395] "Means for recognizing emotions" refers to technology for analyzing and understanding a user's emotional state from their responses and behavior.

[2396] The "means for adjusting feedback and response content" is a mechanism for appropriately changing or optimizing the feedback provided or the next response based on the emotion recognition results.

[2397] The "means for providing evaluation" is a function for judging the accuracy and appropriateness of the user's response and notifying the trainee of the result.

[2398] This invention is a system that uses a generative AI model and an emotion engine to conduct simulated customer service conversations. The system is composed of a server, a terminal, an emotion engine, and a user.

[2399] server:

[2400] The server hosts the data and applications necessary for training. Specifically, the server stores response procedures, dummy customer information, an FAQ database, and a generative AI model. Based on this, the server generates questions based on the dummy customer information and conducts simulated conversations. It also analyzes responses entered by users, evaluates whether they are correct, and provides feedback.

[2401] Device:

[2402] The terminal is a device used by the trainee, such as a PC, tablet, or smartphone. The terminal communicates with the server and displays the responses entered by the trainee, as well as feedback and evaluation results sent from the server.

[2403] Emotion Engine:

[2404] The emotion engine is a component that recognizes emotions based on user input. The emotion engine analyzes the user's text and voice data to recognize their emotional state. The recognized emotions are used to adjust the content of the feedback and responses generated by the server.

[2405] User:

[2406] The participants are mainly trainees and system administrators. Trainees log in to the system using terminals and receive customer service training through mock conversations. System administrators monitor the trainees' progress and evaluations, and provide appropriate feedback and guidance.

[2407] Program processing:

[2408] The server hosts the application and generates mock conversation scenarios for trainees to perform. A generative AI model runs on the server and generates appropriate questions and scenarios based on dummy customer information and an FAQ database. The trainee's response data is sent to the server and compared with the response procedures. An emotion engine recognizes the user's emotions, and feedback and response content are adjusted based on the results.

[2409] Examples:

[2410] For example, consider a scenario in which a new employee at a brick-and-mortar store is using their smartphone to train their customer service skills with this system. If the following prompt is used:

[2411] plaintext

[2412] "We'll learn what to do if a customer is in a hurry and wants to know where an item is. Let's simulate how you would respond when they ask, 'I'm in a hurry, where's the shampoo?'"

[2413] Based on this prompt, the server generates a scenario and displays it on the user's (trainee's) smartphone. When the trainee enters a response, it is sent to the server and compared with the customer service procedures and FAQ database. At the same time, the emotion engine recognizes the emotion from the response, and the server generates appropriate feedback based on that and sends it back to the trainee. For example, if a customer expresses that they are "in a hurry," and the emotion engine recognizes tension or confusion from the user's response, the server will provide feedback to soften the tone, helping to ensure a smoother response next time.

[2414] In this way, trainees can acquire skills to handle a variety of scenarios in a realistic environment, which is expected to improve their performance when dealing with actual customers.

[2415] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2416] Step 1:

[2417] The user displays the system login screen on the terminal and enters their ID and password. The input data is sent to the server, which compares the authentication information with the database and generates an authentication result. If the authentication is successful, the terminal displays the main menu; if it fails, it displays an error message.

[2418] Step 2:

[2419] The user selects "Start Role-Playing" from the main menu. The server generates a dummy customer profile and scenario data and sends them to the device. The device receives this and displays a dialogue screen. The generative AI model generates the first question based on the scenario and sends it to the device.

[2420] Step 3:

[2421] The user inputs a response to the question and sends it to the terminal. The terminal then sends the input data to the server, which then analyzes the response. The analysis includes checking the response against the response procedure and FAQ database.

[2422] Step 4:

[2423] Based on the analysis results, the server uses an emotion engine to recognize emotions from the user's responses. The emotion engine analyzes text and voice data to recognize the user's emotional state. The recognized emotion data is then sent to the server.

[2424] Step 5:

[2425] The server adjusts the feedback and response content based on the emotion recognition results, optimizing the tone and content of the feedback message depending on the recognized emotion, and generating the next question or scenario as needed and sending it to the device.

[2426] Step 6:

[2427] As a substep, the terminal displays a feedback message on the interactive screen and provides appropriate feedback to the user, who can then confirm the feedback and re-enter a corrected response if necessary.

[2428] Step 7:

[2429] When the user completes all scenarios or clicks the "Exit" button at any point, the terminal notifies the server of the end operation. The server records the user's progress data and evaluation results in a database and ends the session. A logout screen is displayed on the terminal to notify the user of the end of the session.

[2430] Through the above processing steps, this system can provide realistic customer service training and support users in improving their skills.

[2431] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[2432] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[2433] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[2434] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[2435] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[2436] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[2437] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[2438] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[2439] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[2440] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[2441] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[2442] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[2443] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[2444] 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.

[2445] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[2446] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[2447] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[2448] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[2449] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[2450] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[2451] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[2452] The following is further disclosed regarding the above embodiment.

[2453] (Claim 1)

[2454] A system for conducting simulated conversations with customers using generative artificial intelligence,

[2455] a means for storing a response procedure and dummy customer information;

[2456] means for generating questions based on the dummy customer information using the generation artificial intelligence;

[2457] A means for comparing a response input by a trainee with the response procedure and performing an evaluation;

[2458] means for providing feedback to the trainee based on said evaluation;

[2459] A system including:

[2460] (Claim 2)

[2461] 10. The system of claim 1, further comprising means for suggesting correction using corrective methods and display means if the evaluation is inappropriate.

[2462] (Claim 3)

[2463] The system according to claim 1, further comprising means for generating the dummy customer information and scenario data and transmitting the dummy customer information and scenario data to a trainee's terminal.

[2464] "Example 1"

[2465] (Claim 1)

[2466] 1. A system for simulating dialogue using generative artificial intelligence for training purposes, comprising:

[2467] means for storing training data and virtual customer data;

[2468] means for generating a dialogue based on the virtual customer data using the generating artificial intelligence;

[2469] a means for comparing and evaluating responses entered by the trainee against training procedures;

[2470] means for providing feedback to the trainee based on said evaluation;

[2471] A system including:

[2472] (Claim 2)

[2473] 10. The system of claim 1, further comprising means for suggesting corrections using corrective procedures and display devices if the evaluation is inappropriate.

[2474] (Claim 3)

[2475] 2. The system according to claim 1, further comprising means for generating the virtual customer data and scenario information and transmitting the generated data and scenario information to a trainee's terminal.

[2476] "Application Example 1"

[2477] (Claim 1)

[2478] A system for conducting simulated conversations with customers using generative artificial intelligence,

[2479] a means for storing a response procedure and dummy customer information;

[2480] means for generating questions based on the dummy customer information using the generation artificial intelligence;

[2481] A means for comparing a response input by a trainee with the response procedure and performing an evaluation;

[2482] means for providing feedback to the trainee based on said evaluation;

[2483] a means for providing a simulated environment to a trainee using smart glasses;

[2484] A means for analyzing voice input from the smart glasses and transmitting it to a generative AI model;

[2485] means for displaying the feedback received from the server on a display of the smart glasses;

[2486] A system including:

[2487] (Claim 2)

[2488] 10. The system of claim 1, further comprising means for suggesting correction using corrective methods and display means if the evaluation is inappropriate.

[2489] (Claim 3)

[2490] The system according to claim 1, further comprising means for generating the dummy customer information and scenario data and transmitting the dummy customer information and scenario data to a trainee's terminal.

[2491] "Example 2: Combining Emotion Engines"

[2492] (Claim 1)

[2493] A system for conducting simulated conversations with customers using generative artificial intelligence,

[2494] means for storing procedures and virtual customer information;

[2495] means for generating questions based on the virtual customer information using the generative artificial intelligence model;

[2496] a means for comparing a response input by a trainee with the response procedure and performing an evaluation;

[2497] a means for providing feedback to the trainee based on the evaluation, the means including an emotion analysis engine for recognizing the evaluation and the user's emotion;

[2498] A system including:

[2499] (Claim 2)

[2500] 10. The system of claim 1, further comprising means for suggesting correction using corrective methods and display means if the evaluation is inappropriate.

[2501] (Claim 3)

[2502] 2. The system according to claim 1, further comprising means for generating the virtual customer information and scenario data and transmitting the generated information and scenario data to a trainee's terminal.

[2503] "Application example 2 when combining emotion engines"

[2504] (Claim 1)

[2505] A system for conducting simulated conversations with customers using generative artificial intelligence,

[2506] a means for storing a response procedure and dummy customer information;

[2507] means for generating questions based on the dummy customer information using the generation artificial intelligence;

[2508] A means for comparing a response input by a trainee with the response procedure and performing an evaluation;

[2509] a means for recognizing emotions from responses entered by the trainee;

[2510] a means for adjusting feedback and response content based on the emotion recognition result;

[2511] means for providing feedback to the trainee based on said evaluation;

[2512] A system including:

[2513] (Claim 2)

[2514] 10. The system of claim 1, further comprising means for suggesting correction using corrective methods and display means if the evaluation is inappropriate.

[2515] (Claim 3)

[2516] The system according to claim 1, further comprising means for generating the dummy customer information and scenario data and transmitting the dummy customer information and scenario data to a trainee's terminal. [Explanation of symbols]

[2517] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A system for conducting simulated conversations with customers using generative artificial intelligence, a means for storing a response procedure and dummy customer information; means for generating questions based on the dummy customer information using the generation artificial intelligence; A means for comparing a response input by a trainee with the response procedure and performing an evaluation; means for providing feedback to the trainee based on said evaluation; A system including:

2. The system of claim 1 , further comprising means for suggesting corrections using corrective methods and display means if the evaluation is inappropriate.

3. The system according to claim 1 , further comprising means for generating the dummy customer information and scenario data and transmitting the dummy customer information and scenario data to a trainee's terminal.

Images