system

Abstract

We provide the system. [Solution] A means for receiving the input message, A means of analyzing received messages using natural language processing, A means of identifying the intent of a message based on the analysis results, A means of querying the database according to the identified intent, A means of generating an appropriate response using a generative AI, A means for sending the generated response to the user's terminal, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In current manned call center operations, high labor costs, delays in inquiry response, buck-passing, and long waiting times are major problems. This increases the risk of reducing customer satisfaction. Also, since the inquiry content is diverse and it is necessary to allocate it to an appropriate person in charge each time, the efficiency is poor. In order to solve such problems, automation and efficiency improvement of inquiry response are essential.

Means for Solving the Problems

[0005] The present invention solves the above problems by the following means.

[0006] A system that includes means for receiving input messages, means for analyzing received messages using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing queries on a database according to the identified intent, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal, makes it possible to respond to inquiries quickly and appropriately. This enables reductions in labor costs, faster response times to inquiries, and improved customer satisfaction.

[0007] "Means for receiving input messages" refers to a function that allows the system to receive the content of inquiries sent by the user through their terminal.

[0008] "Methods for analysis using natural language processing" refer to techniques for analyzing the text of received messages using tokenization, part-of-speech tagging, grammatical analysis, proposition extraction, etc., in order to understand their meaning.

[0009] "Means for identifying the intent of a message" refers to a function that identifies what a user's inquiry is intended to convey, based on the results of natural language processing.

[0010] "Means of querying a database" refers to the ability to perform searches and queries against a database to retrieve relevant information based on a specified intent.

[0011] "Means for generating appropriate responses using generative AI" refers to a function that automatically constructs appropriate answers to user inquiries using a generative AI model based on acquired data.

[0012] "Means for sending the generated response to the user's device" refers to a function that delivers the generated response to the device the user is using, enabling the user to receive the information quickly. [Brief explanation of the drawing]

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

[0014] An example of an embodiment of the system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

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

[0016] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0017] In the following embodiments, a numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

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

[0019] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0020] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0021] [First Embodiment]

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

[0023] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0024] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0025] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0026] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0027] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0028] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

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

[0030] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0031] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0032] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0033] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0034] The embodiments for carrying out the present invention are shown below. The present invention is an automated call center system that has the function of receiving inquiry messages from users, analyzing them, generating appropriate responses, and sending them back to the users. This system operates based on a server, a terminal, and a user interface.

[0035] System Overview

[0036] 1. Receive user inquiries.

[0037] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[0038] The device receives this message and generates an API request to send to the server.

[0039] 2. The server receives the message.

[0040] The server receives the user's message via the receiving API. This message is temporarily stored for analysis.

[0041] The server passes the received message to the natural language processing module.

[0042] 3. Natural language processing

[0043] The server uses a natural language processing (NLP) module to analyze the message content. It understands the meaning and intent of the message using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[0044] The server identifies the intent of the inquiry based on the analysis results. For example, it might determine that the inquiry is for "stock confirmation."

[0045] 4. Execute database queries

[0046] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[0047] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[0048] 5. Generate a response

[0049] The server uses generative AI to automatically construct appropriate responses to user inquiries. For example, if product A is in stock, it will generate a response such as "Product A is in stock."

[0050] Generate an API request to send the response message generated by the server to the user's terminal.

[0051] 6. Send a response back to the user.

[0052] The server sends the generated response to the user's terminal.

[0053] The terminal displays the response message it received to the user. For example, the message "Product A is in stock" is displayed on the user's screen.

[0054] Specific example

[0055] Example 1: Checking product inventory

[0056] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[0057] 2. The device sends a message to the server.

[0058] 3. The server receives the message and performs natural language processing. It identifies it as an "inquiry about stock availability."

[0059] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[0060] 5. The server receives the result "In stock".

[0061] 6. The server generates a response saying, "Smartphones are currently in stock. You can place an order," and sends it to the user's terminal.

[0062] 7. Display the response received by the terminal to the user.

[0063] Example 2: Technical support inquiry

[0064] 1. The user sends the message, "The app won't start. What should I do?"

[0065] 2. The server receives the message, performs natural language processing, and identifies it as a "technical support inquiry."

[0066] 3. The server queries the FAQ database for relevant information.

[0067] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0068] 5. The server generates a response saying, "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[0069] 6. Display the responses received by the device to the user.

[0070] This invention enables automated inquiry handling, reducing labor costs, shortening response times, and improving customer satisfaction.

[0071] The following describes the processing flow.

[0072] Step 1:

[0073] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[0074] Step 2:

[0075] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[0076] Step 3:

[0077] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[0078] Step 4:

[0079] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0080] Step 5:

[0081] The server identifies the intent of the message based on the NLP analysis results. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[0082] Step 6:

[0083] The server generates queries against the database according to the identified intent. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[0084] Step 7:

[0085] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[0086] Step 8:

[0087] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be generated.

[0088] Step 9:

[0089] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0090] Step 10:

[0091] The terminal receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be displayed.

[0092] (Example 1)

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

[0094] Traditional call center systems often involved manual responses to user inquiries, resulting in high labor costs and long response times. Furthermore, inconsistent quality of responses could lead to decreased customer satisfaction. Additionally, delays in handling complex inquiries could negatively impact the user experience.

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

[0096] In this invention, the server includes means for receiving a message entered by a user using a terminal, means for sending the received message to the server, means for the server to analyze the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing a query to a database according to the identified intent, means for generating an appropriate response using generative AI based on information obtained from the database, means for sending the generated response to the user's terminal, and means for displaying the response message received by the user's terminal. This enables automated, high-quality, and rapid inquiry handling.

[0097] A "user" refers to a person or entity that makes a request to the system.

[0098] A "device" refers to a device operated by a user, and includes smartphones, personal computers, tablets, and other similar devices.

[0099] A "server" is a central computing unit that receives user inquiries, performs analysis, and generates responses.

[0100] A "message" refers to the text or voice inquiry that a user enters through their device and sends to the server.

[0101] Natural Language Processing (NLP) refers to the technology used to analyze, interpret, and process human language using machines.

[0102] "Tokenization" refers to the process of dividing input text into smaller units such as words and phrases.

[0103] "Part-of-speech tagging" refers to the process of labeling tokenized words with their part of speech (e.g., noun, verb, adjective).

[0104] "Grammar analysis" refers to the process of analyzing the sentence structure of input text and understanding the relationships between each element.

[0105] "Propositional extraction" refers to the process of extracting important information and intentions from a text.

[0106] "Intent" refers to the purpose or requirements indicated in the user's inquiry message.

[0107] A "query" refers to a request sent to retrieve necessary information from a database.

[0108] A "database" refers to a system that stores and manages a collection of structured data.

[0109] "Generative AI" refers to artificial intelligence technology that automatically generates responses based on input data.

[0110] "Response" refers to the reply message that the server generates in response to a user's inquiry and sends to the terminal.

[0111] Modes for carrying out the invention

[0112] This invention provides an automated call center system. This system receives user inquiry messages, analyzes them, generates appropriate responses, and sends them back to the user, thereby achieving fast and high-quality service. Specific embodiments of this system are described below.

[0113] composition

[0114] This system operates based on users, terminals, and servers. Users use terminals to make inquiries, and the terminals send these inquiries to the server. The server analyzes the received messages, generates appropriate responses using a database and generative AI, and sends them back to the user via the terminal.

[0115] Hardware and software to be used

[0116] 1. Terminal: A device used by the user to enter inquiries, such as a smartphone or personal computer (PC).

[0117] 2. Server: A central computing unit that performs message analysis, response generation, and database query execution.

[0118] 3. Natural Language Processing (NLP) Module: Software that performs message tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[0119] 4. Generative AI: An artificial intelligence model that automatically generates appropriate responses based on information obtained from a database.

[0120] 5. Database system: A system for storing and managing structured data.

[0121] Data processing and data calculation

[0122] Message reception: The user's device enters the inquiry message, creates an HTTP POST request, and sends it to the server.

[0123] Message Analysis: The server analyzes received messages using a natural language processing (NLP) module. The NLP module performs tokenization, part-of-speech tagging, grammatical analysis, proposition extraction, and other operations to determine the intent of the message.

[0124] Database queries: The server generates and executes SQL queries against the database according to the specified intent. The retrieved data is temporarily stored.

[0125] Response generation: The server uses generative AI to automatically generate an appropriate response based on the acquired data.

[0126] Response transmission: The server sends the generated response to the user's terminal as an HTTP POST request, and the terminal displays it.

[0127] Specific example

[0128] The following shows a specific example of how this system works.

[0129] Example 1: Checking product inventory

[0130] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[0131] 2. The device sends the message to the server.

[0132] 3. The server analyzes the received message using a natural language processing module and identifies it as an "inquiry about inventory."

[0133] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[0134] 5. The server receives the result "In stock" and uses generative AI to generate the response "Smartphones are currently in stock. You can place an order."

[0135] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[0136] Example 2: Technical support inquiry

[0137] 1. The user sends the message, "The app won't start. What should I do?"

[0138] 2. The server analyzes the received message using a natural language processing module and identifies it as a "technical support inquiry."

[0139] 3. The server executes the following query on the FAQ database: SELECT solution FROM faq WHERE issue='App won't start';

[0140] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0141] 5. The server uses a generative AI to generate a response based on this information: "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again."

[0142] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[0143] Example of a prompt

[0144] The following are specific examples of prompt statements to be input into the generating AI model.

[0145] 1. A prompt to generate a response to the inquiry, "Do you have the latest smartphones in stock?"

[0146] Prompt: "Generate an appropriate response when a user asks, 'Do you have the latest smartphones in stock?'"

[0147] 2. A prompt to generate a response to a technical support inquiry: "The app won't start. What should I do?"

[0148] Prompt: "Provide a solution for when a user asks, 'The app won't start. What should I do?'"

[0149] This invention enables the automation of customer inquiry handling, which is expected to reduce labor costs, shorten response times, and improve customer satisfaction.

[0150] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0151] Step 1:

[0152] The user enters an inquiry.

[0153] Input: The user enters their inquiry into the input field on the terminal. For example, they might enter "Please tell me the stock status of product A."

[0154] Action: The user clicks the "Submit" button.

[0155] Output: The entered text is sent to the system, and the contents of the input field are cleared.

[0156] Step 2:

[0157] The device sends a message to the server.

[0158] Input: The inquiry message entered by the user.

[0159] Operation: The terminal generates an HTTP POST request and sends the message to the server. This request includes query text and metadata (e.g., timestamp, user ID).

[0160] Output: If the request is successful, the server receives a message.

[0161] Step 3:

[0162] The server receives the message

[0163] Input: HTTP POST request sent from the terminal.

[0164] Operation: The server receives requests via the receiving API and temporarily stores the message content. It also logs the request metadata.

[0165] Output: The temporarily stored message is passed to the natural language processing (NLP) module.

[0166] Step 4:

[0167] Perform Natural Language Processing (NLP)

[0168] Input: Query message temporarily stored on the server.

[0169] Operation: The server uses an NLP module to tokenize messages, tag them with parts of speech, perform grammatical analysis, extract propositions, and so on. In this process, it analyzes and identifies the intent of the message.

[0170] Output: Analyzed intent (e.g., "Check inventory") and message structure data.

[0171] Step 5:

[0172] Execute a database query

[0173] Input: Analyzed intent and database query format.

[0174] Operation: The server generates and executes appropriate SQL queries against the database based on a specific intent. For example, it might issue a query to check the inventory status of "product A".

[0175] Output: Results retrieved from the database (e.g., "In stock").

[0176] Step 6:

[0177] Generate a response

[0178] Input: Results retrieved from the database and prompts for the generated AI model.

[0179] Operation: The server uses generative AI to construct an appropriate response based on the retrieved database results. For example, it might generate the sentence, "Product A is in stock."

[0180] Output: The generated response message.

[0181] Step 7:

[0182] The server sends a response message.

[0183] Input: The generated response message and the user's terminal information.

[0184] Operation: The server generates an HTTP POST request and sends a response message to the user's terminal.

[0185] Output: The terminal receives a response message.

[0186] Step 8:

[0187] The device displays a response message.

[0188] Input: The response message sent from the server.

[0189] Action: The device updates the user interface (UI) and displays a response message on the screen. For example, it might display the message "Product A is in stock."

[0190] Output: A response message is displayed on the user's screen.

[0191] (Application Example 1)

[0192] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0193] The problem that this invention aims to solve is to establish a method for providing quick and accurate responses to user inquiries. In particular, it aims to provide a system that can automatically generate responses to a wide variety of user inquiries on e-commerce sites, such as product inventory checks and technical support, and provide quick answers. Conventional methods often resulted in delayed responses or required a large amount of human resources, which contributed to decreased customer satisfaction.

[0194] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0195] In this invention, the server includes means for receiving input messages, means for analyzing the received messages using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing queries to a database according to the identified intent, means for checking the inventory status of a specific product and responding immediately to inquiries from users regarding inventory status, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to provide quick and accurate responses to user inquiries and improve customer satisfaction.

[0196] "Means for receiving input messages" refers to the means by which a user inputs a message through a terminal and receives it within the system.

[0197] "Means for analyzing received messages using natural language processing" refers to methods for analyzing received messages using natural language processing techniques to understand their content and intent.

[0198] "Means for identifying the intent of a message based on analysis results" refers to methods for identifying what a message means and what the user is looking for, based on the results of analysis using natural language processing.

[0199] "Means of querying a database according to a specified intent" refers to means of querying a database to retrieve necessary information based on a specified intent.

[0200] "A means of checking the inventory status of a specific product and responding immediately to user inquiries about inventory status" refers to a means of immediately checking the inventory status of a specific product in order to respond to user inquiries about inventory status.

[0201] "Means for generating appropriate responses using generative AI" refers to methods for automatically generating appropriate responses to user inquiries using generative AI technology.

[0202] "Means for sending the generated response to the user's terminal" refers to means of sending the generated response message to the user's terminal and providing feedback to the user.

[0203] The embodiments for carrying out the present invention are shown below. The present invention provides an automated inquiry response system, specifically intended for use on e-commerce websites. This system receives inquiry messages from users, analyzes them using natural language processing, retrieves information from a database based on the identified intent, and generates an appropriate response using generative AI.

[0204] System Overview

[0205] 1. User input

[0206] The user uses their smartphone to open the dedicated application for the online shopping site. They enter their inquiry in text format into the chat box. For example, they might type, "Do you still have any XX brand dress shirts in stock?"

[0207] 2. Receiving and parsing messages

[0208] The terminal sends the entered message to the server. The server receives the message and analyzes it using natural language processing. Specifically, processes such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction are performed.

[0209] 3. Identifying Intent and Data Acquisition

[0210] The system analyzes the message to identify the user's intent. Based on the identified intent, it queries the database to check, for example, the inventory status of a specific product. The server then passes the retrieved inventory information to the generative AI.

[0211] 4. Generating and sending responses

[0212] Generative AI generates appropriate responses based on the data it receives. For example, it might generate a message like, "We currently have XX brand dress shirts in stock. You can place an order."

[0213] 5. Display of response

[0214] The generated response message is sent back to the user's device and displayed in the chat box. This allows the user to receive a quick and accurate response.

[0215] Hardware and software

[0216] This system uses the following hardware and software.

[0217] Hardware:

[0218] Server: High-performance server (e.g., AWS®, GCP)

[0219] User device: Smartphone (e.g., iPhone®, Android® devices)

[0220] software:

[0221] Server side: Flask, SQLite (database)

[0222] Natural Language Processing: spaCy and NLTK

[0223] Generating AI: OpenAI(registered trademark) GPT-3(registered trademark)

[0224] Frontend: Smartphone applications (e.g., React Native)

[0225] Specific example

[0226] Message Processing Example

[0227] A user enters the message "Do you still have XX brand dress shirts in stock?" and sends it to the server. This message is received by the server and analyzed using natural language processing. From the analysis results, it is determined that the user's intention is to check the stock availability. The server queries the database to check the stock status of XX brand dress shirts. Based on the results, a generative AI generates a response message: "We currently have XX brand dress shirts in stock. You can place an order." This message is sent to the user's device and displayed in the chat box.

[0228] Example of prompts for the generating AI

[0229] If the product is in stock, how should I respond?

[0230] This system allows users to obtain information quickly and accurately, enabling effective responses to inquiries. This technology will greatly contribute to reducing labor costs and improving customer satisfaction.

[0231] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0232] Step 1:

[0233] The user uses their smartphone to open the e-commerce site's application and enters their inquiry as text into the chat box. For example, they might type, "Do you still have XX brand dress shirts in stock?" This message is then sent to the server as input.

[0234] Step 2:

[0235] The device sends the user's message to the server. Specifically, the device generates an API request and sends the message to the server via the receiving API. The input here is the message entered by the user, and the output is the message sent to the server.

[0236] Step 3:

[0237] The server temporarily stores the received message and passes it to the natural language processing module. The natural language processing module performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to analyze the meaning of the message. The input here is the message received in the previous step, and the output is the analysis result.

[0238] Step 4:

[0239] The server identifies the intent of the user's inquiry based on the results of natural language processing. For example, it might determine that this message is an inquiry about stock availability. At this stage, the input is the result of natural language processing, and the output is the identification of the intent, "stock availability check."

[0240] Step 5:

[0241] Based on the identified intent, the server executes a query to retrieve relevant data from the database. For example, "Execute an SQL query to check the inventory status of XX brand dress shirts. The input for this step is the identified intent, and the output is the result of the database query."

[0242] Step 6:

[0243] The server retrieves information from the database and passes it to a generative AI to generate an appropriate response. The generative AI is given prompts based on the retrieved information, and generates a response message such as, "We currently have XX brand dress shirts in stock. You can place an order." The input here is the inventory information retrieved from the database, and the output is the generated response message.

[0244] Step 7:

[0245] The server generates an API request to send the generated response message to the user's terminal and sends it to the user's terminal. The user's terminal displays the received response message in the chat box. The input here is the generated response message, and the output is the response message displayed on the user's terminal.

[0246] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0247] The following describes embodiments for carrying out the present invention. The present invention is an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and an emotion engine, generates an appropriate response, and sends it back to the user. The system operates based on a server, terminal, and user interface, and focuses particularly on analyzing the user's emotions.

[0248] System Overview

[0249] 1. Receive user inquiries.

[0250] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[0251] The device receives this message and generates an API request to send to the server.

[0252] 2. The server receives the message.

[0253] The server receives the user's message at the receiving API endpoint. This message is temporarily stored for parsing.

[0254] The server passes the received message to the natural language processing module.

[0255] 3. Natural language processing

[0256] The server uses a natural language processing (NLP) module to analyze the message content. This analysis includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0257] 4. Perform emotion analysis

[0258] The server uses an emotion engine to analyze the user's emotions within a message. For example, if a message like "Do you have it in stock?" is determined to contain a sense of urgency, it will assign emotion tags such as positive, negative, or neutral.

[0259] 5. Identifying the Intent

[0260] The server identifies the intent of the message based on the NLP analysis results and the emotion engine results. For example, it might be identified as an "inquiry about stock availability."

[0261] 6. Executing database queries

[0262] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[0263] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[0264] 7. Response generation

[0265] The server uses generative AI to automatically construct appropriate responses to user inquiries. In doing so, it considers sentiment tags and adjusts the tone and content of the response accordingly. For example, if negative sentiment is detected, the response message will emphasize consideration and solutions.

[0266] The server generates a message saying, "Smartphone X is currently in stock. You can order it now." However, if the user is feeling anxious, a more considerate message such as, "Don't worry, Smartphone X is in stock. You can order it right away," will be generated.

[0267] 8. Sending a response to the user

[0268] The system generates and sends an API request to send the response message generated by the server to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0269] 9. Displaying user responses

[0270] The device receives a response message from the server and displays it to the user. For example, it might display, "Smartphone X is currently in stock. You can place an order." However, it can also display messages that are considerate of the user's feelings.

[0271] Specific example

[0272] Example 1: Checking product inventory

[0273] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[0274] 2. The device sends a message to the server.

[0275] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[0276] 4. The server executes a query "SELECT stock_status FROM products WHERE product_name='Smartphone X';" on the product inventory database.

[0277] 5. The server receives the result "In stock".

[0278] 6. The server generates a response "Smartphone X is currently in stock. Please rest assured and you can place an order immediately." and sends it to the user terminal.

[0279] 7. The terminal displays the received response to the user.

[0280] Example 2: Technical support inquiry

[0281] 1. The user sends a message "The app won't start. What should I do?"

[0282] 2. The server receives the message and performs natural language processing and sentiment analysis. The user's confusion and anxiety are detected from the message.

[0283] 3. The server queries the relevant information in the FAQ database.

[0284] 4. The server obtains the information "If the app won't start, please clear the cache or reinstall it."

[0285] 5. The server generates a considerate response "If the app won't start, please clear the cache or reinstall the app once. If it still doesn't solve the problem, please contact us again." and sends it to the user terminal.

[0286] 6. The terminal displays the received response to the user.

[0287] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[0288] The following describes the processing flow.

[0289] Step 1:

[0290] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[0291] Step 2:

[0292] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[0293] Step 3:

[0294] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[0295] Step 4:

[0296] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0297] Step 5:

[0298] The server identifies the intent of the message based on the results of natural language processing. For example, it is identified as an "inquiry about inventory check". Thereby, the system understands that the user is asking for inventory information.

[0299] Step 6:

[0300] The server calls the sentiment analysis module to analyze the user's sentiment in the message. A word sentiment dictionary or a machine learning model is used for sentiment analysis. For example, it is determined that a message such as "Is there any stock?" contains a sense of impatience.

[0301] Step 7:

[0302] Based on the sentiment analysis results, the server assigns a sentiment tag to the user message. For example, a sentiment tag of "impatience" is assigned to the message.

[0303] Step 8:

[0304] Based on the identified intent and sentiment tag, the server generates a query to retrieve relevant information from the database. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[0305] Step 9:

[0306] The server executes the generated SQL query on the database to retrieve relevant data. For example, a result such as "Smartphone X is in stock" is returned.

[0307] Step 10:

[0308] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. The response generation incorporates an algorithm that adjusts the tone and content considering emotion tags. For example, if a negative emotion is detected, the response message will emphasize consideration and solutions.

[0309] Step 11:

[0310] The server generates a message saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately." However, if the user is feeling anxious, a more considerate message such as "Please rest assured, Smartphone X is in stock. You can order it immediately." will be generated.

[0311] Step 12:

[0312] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0313] Step 13:

[0314] The device receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. Please rest assured, you can order it immediately." might be displayed.

[0315] (Example 2)

[0316] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0317] Traditional call center systems have struggled to provide efficient and effective responses to user inquiries, particularly in generating responses that take user emotions into consideration. Furthermore, their reliance on human resources has led to increased response times and rising labor costs. To address these challenges, a system is needed that accurately analyzes message content and automatically generates responses that consider user emotions.

[0318] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0319] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results and sentiment analysis, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to automatically generate a response that takes the user's emotions into consideration.

[0320] "Means for receiving input messages" refers to the communication methods and systems that allow a server to receive messages entered by a user using a terminal.

[0321] "Methods of analysis using natural language processing" refers to methods of analyzing human language using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction of input messages.

[0322] "Means of identifying the intent of a message" refers to methods of clarifying what the user is seeking in their inquiry, based on the results of natural language processing analysis and sentiment analysis.

[0323] "Means of querying a database" refers to a method of executing query commands to retrieve necessary information from a database based on a specified intent.

[0324] "Means of generating appropriate responses using generative AI" refers to a method of constructing responses to user inquiries based on information obtained from a database and sentiment tags, using generative AI models (e.g., GPT-3, BERT).

[0325] "Means of sending the generated response to the user's terminal" refers to the method of sending the response generated on the server to the user's terminal using a communication method or protocol (e.g., HTTP).

[0326] This invention relates to an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and sentiment analysis engines, generates appropriate responses, and sends them back to the users. This system operates based on interfaces from servers, terminals, and users, and places particular emphasis on analyzing user sentiment.

[0327] The server first receives a user inquiry message. The user enters the inquiry in text format using a device (e.g., smartphone, PC). For example, the user might enter the message, "Please tell me the stock status of product A." The device receives this message and generates an API request to send to the server. The server receives this message at the specified receiving API endpoint (e.g., / receive-message) and stores it temporarily.

[0328] Next, the server passes the received message to a natural language processing (NLP) module. The NLP module analyzes the message content using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Specifically, it uses an NLP library (e.g., spaCy or NLTK). For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0329] Furthermore, the server uses a sentiment analysis engine (e.g., IBM Watson®, Azure® Text Analytics) to analyze the user's emotions within the message. The server analyzes the user's message and, if negative emotions such as impatience are detected, adds corresponding sentiment tags. For example, if the message "Do you have this in stock?" is determined to contain impatience, it will be tagged with positive, negative, or neutral sentiment tags.

[0330] Next, the server identifies the intent of the message based on the NLP analysis results and sentiment analysis results. For example, it can be identified as an "inquiry about stock availability." Based on this intent, the server executes a query to a database (e.g., MySQL® or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query (e.g., SELECT stock_status FROM products WHERE product_name='Smartphone X';) and queries the database.

[0331] The server passes the acquired information to a generative AI model (e.g., GPT-3, BERT) to generate a response. The generative AI model generates an appropriate response based on the information and sentiment tags acquired from the database. In this process, it adjusts the tone and content of the response, taking sentiment tags into consideration. For example, based on the result "In stock," it generates the message, "Rest assured, smartphone X is in stock. You can order it immediately."

[0332] Finally, the server generates and sends an API request to send the generated response message to the user's device. This request is sent to the user's device using the HTTP protocol or similar. The device displays the received response message to the user, conveying a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[0333] Specific example

[0334] Example 1: Checking product inventory

[0335] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[0336] 2. The device sends a message to the server.

[0337] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[0338] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[0339] 5. The server receives the result "In stock".

[0340] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[0341] 7. Display the response received by the terminal to the user.

[0342] Example 2: Technical support inquiry

[0343] 1. The user sends the message, "The app won't start. What should I do?"

[0344] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[0345] 3. The server queries the FAQ database for relevant information.

[0346] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0347] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[0348] 6. Display the responses received by the device to the user.

[0349] Examples of input prompts for a generative AI model

[0350] "Generate a response based on the user's inquiry, 'Do you have the latest smartphone X in stock?' Assume the user is in a hurry."

[0351] "Based on the user inquiry, 'The app won't start. What should I do?', please generate an appropriate response for the confused user."

[0352] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[0353] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0354] Step 1: Receive user inquiry

[0355] The user enters their inquiry in text format using a device (e.g., smartphone, PC). An example of the entered message is "Please tell me the stock status of product A." The device receives this message and generates an API request (e.g., an HTTP POST request) to send to the server. The input is the user's message, and the output is the API request sent to the server.

[0356] Step 2: The server receives the message.

[0357] The server receives user messages at the specified receiving API endpoint (e.g., / receive-message). The server temporarily stores the received messages in an in-memory data store such as Redis and manages the order of processing. At this point, the input is the API request from the terminal, and the output is the temporarily stored messages.

[0358] Step 3: Process natural language

[0359] The server uses an NLP module (e.g., spaCy or NLTK) to tokenize the received message. The server then performs part-of-speech tagging and extracts propositions through grammatical analysis. Specifically, it might split a message like "Do you have the latest smartphone X in stock?" into tokens and assign a part-of-speech tag to each token. The input is the temporarily stored message, and the output is the tokenized data and part-of-speech tagging information as a result of the analysis.

[0360] Step 4: Perform sentiment analysis

[0361] The server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the sentiment of a message based on the NLP analysis results. The server analyzes the user's message, and if it detects negative emotions, such as anxiety, it adds the corresponding sentiment tag. The input is tokenized and part-of-speech tagged data, and the output is sentiment-tagged data.

[0362] Step 5: Identifying the Intent

[0363] The server integrates the NLP analysis results and sentiment analysis results to identify the intent of the message. For example, it might identify it as an "inquiry about stock availability." Specifically, the server prepares the data necessary for the next processing based on the identified intent. The input is analyzed data with sentiment tags, and the output is data containing the identified intent.

[0364] Step 6: Execute the database query

[0365] Based on the identified intent, the server issues queries to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query such as `SELECT stock_status FROM products WHERE product_name='smartphone X';` and queries the database. The input is data containing the identified intent, and the output is information retrieved from the database (e.g., "In stock").

[0366] Step 7: Generating the response

[0367] The server inputs data into a generative AI model (e.g., GPT-3, BERT) to generate appropriate responses to user inquiries. The tone and content of the responses are adjusted considering sentiment tags. For example, it might generate the message, "Rest assured, smartphone X is in stock. You can order it immediately." The input consists of information retrieved from a database and data with sentiment tags, while the output is the generated response.

[0368] Step 8: Sending a response to the user

[0369] This process generates and sends an API request to the user's terminal to send the response message generated by the server. This request is made using the HTTP protocol, etc. The input is the generated response message, and the output is the API request sent to the terminal.

[0370] Step 9: User response display

[0371] The terminal receives a response message sent from the server and displays it to the user. Specifically, it displays "Rest assured, smartphone X is in stock. You can order it immediately." The input is the response message sent from the server, and the output is the text displayed to the user.

[0372] These steps enable the system to provide timely and appropriate responses to user inquiries, while also considering their emotional impact.

[0373] (Application Example 2)

[0374] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0375] Traditional automated response systems often fail to consider user emotions, potentially leading to decreased user satisfaction. This is especially true for e-commerce sites, where prompt and appropriate responses to user inquiries are crucial; responses that disregard user feelings diminish credibility. Furthermore, traditional manual responses to a large volume of inquiries are inefficient and costly, necessitating automation.

[0376] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0377] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, means for analyzing the user's emotions, means for generating a response that takes the user's emotions into consideration based on the analysis results, and means for sending the generated response to the user's terminal. This makes it possible to provide an appropriate and prompt response that takes the user's emotions into consideration, thereby improving user satisfaction and reducing costs.

[0378] "Input messages" refer to text-based inquiries or requests that users send to the system through their devices.

[0379] "Means of receiving" refers to the function on the server side that receives messages sent by the user.

[0380] "Natural language processing" is a general term for technologies that analyze input messages and understand their content and structure.

[0381] "Means of analysis" refers to the function used to analyze received messages using natural language processing.

[0382] "Message intent" refers to the user's purpose or request, extracted from the content of the message as analyzed by natural language processing.

[0383] "Means of identification" refers to a function for identifying the intent of a message based on the analysis results.

[0384] "Means of querying a database" refers to means of retrieving necessary information from a database according to a specified intent.

[0385] "Generative AI" refers to artificial intelligence technology aimed at generation, specifically a system that automatically constructs responses to user inquiries based on information and analysis results obtained from a database.

[0386] "Means for generating appropriate responses" refers to a function that uses generative AI to generate appropriate responses to user inquiries.

[0387] "User sentiment" refers to the emotional nuances and tone that a user conveys when typing a message.

[0388] "Means of analysis (emotions)" refers to the function for analyzing the user's emotions within a message.

[0389] "Emotionally sensitive responses" refer to responses that are tailored to the user's emotions, based on the results of an emotional analysis of the user, and using appropriate tone and content.

[0390] "Means of generating (emotions)" refers to the function that generates responses that take into account the user's emotions, considering the results of emotion analysis.

[0391] "Means of sending to the user's terminal" refers to the function for sending the generated response from the server to the user's terminal.

[0392] System Overview

[0393] This invention is an automated call center system that automatically processes user inquiries and generates appropriate responses based on sentiment analysis. The system operates based on a server, terminals, and a user interface, and focuses particularly on analyzing the user's sentiment.

[0394] Hardware and software configuration

[0395] Hardware:

[0396] User devices: smartphones, PCs, etc.

[0397] Server: Equipped with a high-performance CPU, sufficient memory, and disk space.

[0398] software:

[0399] Natural language processing libraries: spaCy, NLTK, etc.

[0400] Emotion analysis tools: Emotion analysis models based on TENSORFLOW®, etc.

[0401] Server environment: Web framework such as Flask

[0402] Processing steps

[0403] 1. Receiving a message

[0404] Users submit inquiries in text format using devices such as smartphones and PCs. These messages are sent to the server via HTTP requests.

[0405] 2. Natural Language Processing and Sentiment Analysis

[0406] The server passes the received message to a natural language processing module for analysis such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Simultaneously, it uses an emotion engine to analyze the user's emotions. Sentiment analysis determines whether the message has multiple emotion tags, such as positive, negative, or anxiety.

[0407] 3. Identifying the Intent and Database Queries

[0408] Based on the analysis results, the server identifies the intent of the message. For example, it might categorize it as "inventory check" or "technical support." Based on the specific intent, the server queries the database to retrieve the necessary information.

[0409] 4. Generating a response

[0410] Based on the acquired information and sentiment tags, a generative AI is used to automatically generate a response message to the user. In this process, sentiment tags are taken into consideration to generate a considerate response.

[0411] 5. Sending and displaying responses

[0412] The server sends the generated response message to the user's terminal, and the terminal displays this message to the user.

[0413] Specific examples

[0414] Example of checking inventory:

[0415] User: "Do you have the latest smartphone X in stock?"

[0416] System response: "Rest assured, the Smartphone X is in stock. You can order it immediately."

[0417] Examples of technical support:

[0418] User: "The app won't start. What should I do?"

[0419] System response: "If the app won't launch, try clearing your cache or reinstalling the app. If the problem persists, please contact us again."

[0420] Example of a prompt

[0421] User query: "Do you have the latest smartphone X in stock?"

[0422] NLP analysis results: "Inventory check" "Smartphone X"

[0423] Emotion analysis result: "Anxiety"

[0424] Generate API prompt: Generate a response that is considerate of a user who is anxious about the availability of smartphone X.

[0425] Thus, the present invention makes it possible to achieve appropriate and prompt responses that take into account the user's feelings, thereby improving user satisfaction and reducing costs.

[0426] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0427] Step 1:

[0428] The user uses their device to enter their inquiry in text format. For example, they might enter the message, "Do you have the latest smartphone X in stock?" The entered message is received by the device's application and sent to the server in the form of an API request.

[0429] Input: User's text-based inquiry message

[0430] Output: Message in API request format

[0431] Step 2:

[0432] The server receives the message at the API endpoint. The received message is temporarily stored and then passed to the natural language processing (NLP) module.

[0433] Input: Message in API request format

[0434] Output: Message data to be passed to the NLP module

[0435] Step 3:

[0436] The server's natural language processing module analyzes the message. It performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to understand the message's content.

[0437] Input: Temporarily saved message data

[0438] Output: Analysis results (tokens, part-of-speech tagging, grammatical analysis results, proposition extraction results)

[0439] Step 4:

[0440] The server's emotion engine analyzes the user's emotions from the analyzed messages. For example, it assigns emotion tags such as positive, negative, and anxiety.

[0441] Input: NLP analysis results

[0442] Output: Sentiment analysis results (emotion tags)

[0443] Step 5:

[0444] The server identifies the intent of the message based on NLP analysis results and sentiment analysis results. For example, it might identify it as an "inquiry about stock availability."

[0445] Input: NLP analysis results and emotion analysis results

[0446] Output: Message intent (e.g., Inventory check)

[0447] Step 6:

[0448] Depending on the identified intent, the server executes SQL queries to retrieve relevant information from the database. For example, it might generate an SQL query to check the "stock status of product A" and then query the database.

[0449] Input: Message intent

[0450] Output: Database query results (e.g., inventory status)

[0451] Step 7:

[0452] The server uses generative AI to generate appropriate responses based on information and emotion tags retrieved from the database. For example, based on a database result such as "In stock" and the emotion tag "Anxiety," it generates a response such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[0453] Input: Database query results and sentiment tags

[0454] Output: Generated response message

[0455] Step 8:

[0456] Generate and send an API request to send the response message generated by the server to the user's terminal.

[0457] Input: Generated response message

[0458] Output: Response message in API request format

[0459] Step 9:

[0460] The terminal receives a response message from the server and displays it to the user. For example, it might display a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[0461] Input: Response message in API request format

[0462] Output: Response message displayed to the user

[0463] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0464] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0465] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0466] [Second Embodiment]

[0467] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0468] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0469] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0470] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0471] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0472] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0473] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0474] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0475] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0476] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0477] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0478] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0479] The embodiments for carrying out the present invention are shown below. The present invention is an automated call center system that has the function of receiving inquiry messages from users, analyzing them, generating appropriate responses, and sending them back to the users. This system operates based on a server, a terminal, and a user interface.

[0480] System Overview

[0481] 1. Receive user inquiries.

[0482] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[0483] The device receives this message and generates an API request to send to the server.

[0484] 2. The server receives the message.

[0485] The server receives the user's message via the receiving API. This message is temporarily stored for analysis.

[0486] The server passes the received message to the natural language processing module.

[0487] 3. Natural language processing

[0488] The server uses a natural language processing (NLP) module to analyze the message content. It understands the meaning and intent of the message using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[0489] The server identifies the intent of the inquiry based on the analysis results. For example, it might determine that the inquiry is for "stock confirmation."

[0490] 4. Execute database queries

[0491] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[0492] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[0493] 5. Generate a response

[0494] The server uses generative AI to automatically construct appropriate responses to user inquiries. For example, if product A is in stock, it will generate a response such as "Product A is in stock."

[0495] Generate an API request to send the response message generated by the server to the user's terminal.

[0496] 6. Send a response back to the user.

[0497] The server sends the generated response to the user's terminal.

[0498] The terminal displays the response message it received to the user. For example, the message "Product A is in stock" is displayed on the user's screen.

[0499] Specific example

[0500] Example 1: Checking product inventory

[0501] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[0502] 2. The device sends a message to the server.

[0503] 3. The server receives the message and performs natural language processing. It identifies it as an "inquiry about stock availability."

[0504] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[0505] 5. The server receives the result "In stock".

[0506] 6. The server generates a response saying, "Smartphones are currently in stock. You can place an order," and sends it to the user's terminal.

[0507] 7. Display the response received by the terminal to the user.

[0508] Example 2: Technical support inquiry

[0509] 1. The user sends the message, "The app won't start. What should I do?"

[0510] 2. The server receives the message, performs natural language processing, and identifies it as a "technical support inquiry."

[0511] 3. The server queries the FAQ database for relevant information.

[0512] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0513] 5. The server generates a response saying, "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[0514] 6. Display the responses received by the device to the user.

[0515] This invention enables automated inquiry handling, reducing labor costs, shortening response times, and improving customer satisfaction.

[0516] The following describes the processing flow.

[0517] Step 1:

[0518] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[0519] Step 2:

[0520] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[0521] Step 3:

[0522] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[0523] Step 4:

[0524] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0525] Step 5:

[0526] The server identifies the intent of the message based on the NLP analysis results. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[0527] Step 6:

[0528] The server generates queries against the database according to the identified intent. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[0529] Step 7:

[0530] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[0531] Step 8:

[0532] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be generated.

[0533] Step 9:

[0534] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0535] Step 10:

[0536] The terminal receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be displayed.

[0537] (Example 1)

[0538] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0539] Traditional call center systems often involved manual responses to user inquiries, resulting in high labor costs and long response times. Furthermore, inconsistent quality of responses could lead to decreased customer satisfaction. Additionally, delays in handling complex inquiries could negatively impact the user experience.

[0540] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0541] In this invention, the server includes means for receiving a message entered by a user using a terminal, means for sending the received message to the server, means for the server to analyze the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing a query to a database according to the identified intent, means for generating an appropriate response using generative AI based on information obtained from the database, means for sending the generated response to the user's terminal, and means for displaying the response message received by the user's terminal. This enables automated, high-quality, and rapid inquiry handling.

[0542] A "user" refers to a person or entity that makes a request to the system.

[0543] A "device" refers to a device operated by a user, and includes smartphones, personal computers, tablets, and other similar devices.

[0544] A "server" is a central computing unit that receives user inquiries, performs analysis, and generates responses.

[0545] A "message" refers to the text or voice inquiry that a user enters through their device and sends to the server.

[0546] Natural Language Processing (NLP) refers to the technology used to analyze, interpret, and process human language using machines.

[0547] "Tokenization" refers to the process of dividing input text into smaller units such as words and phrases.

[0548] "Part-of-speech tagging" refers to the process of labeling tokenized words with their part of speech (e.g., noun, verb, adjective).

[0549] "Grammar analysis" refers to the process of analyzing the sentence structure of input text and understanding the relationships between each element.

[0550] "Propositional extraction" refers to the process of extracting important information and intentions from a text.

[0551] "Intent" refers to the purpose or requirements indicated in the user's inquiry message.

[0552] A "query" refers to a request sent to retrieve necessary information from a database.

[0553] A "database" refers to a system that stores and manages a collection of structured data.

[0554] "Generative AI" refers to artificial intelligence technology that automatically generates responses based on input data.

[0555] "Response" refers to the reply message that the server generates in response to a user's inquiry and sends to the terminal.

[0556] Modes for carrying out the invention

[0557] This invention provides an automated call center system. This system receives user inquiry messages, analyzes them, generates appropriate responses, and sends them back to the user, thereby achieving fast and high-quality service. Specific embodiments of this system are described below.

[0558] composition

[0559] This system operates based on users, terminals, and servers. Users use terminals to make inquiries, and the terminals send these inquiries to the server. The server analyzes the received messages, generates appropriate responses using a database and generative AI, and sends them back to the user via the terminal.

[0560] Hardware and software to be used

[0561] 1. Terminal: A device used by the user to enter inquiries, such as a smartphone or personal computer (PC).

[0562] 2. Server: A central computing unit that performs message analysis, response generation, and database query execution.

[0563] 3. Natural Language Processing (NLP) Module: Software that performs message tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[0564] 4. Generative AI: An artificial intelligence model that automatically generates appropriate responses based on information obtained from a database.

[0565] 5. Database system: A system for storing and managing structured data.

[0566] Data processing and data calculation

[0567] Message reception: The user's device enters the inquiry message, creates an HTTP POST request, and sends it to the server.

[0568] Message Analysis: The server analyzes received messages using a natural language processing (NLP) module. The NLP module performs tokenization, part-of-speech tagging, grammatical analysis, proposition extraction, and other operations to determine the intent of the message.

[0569] Database queries: The server generates and executes SQL queries against the database according to the specified intent. The retrieved data is temporarily stored.

[0570] Response generation: The server uses generative AI to automatically generate an appropriate response based on the acquired data.

[0571] Response transmission: The server sends the generated response to the user's terminal as an HTTP POST request, and the terminal displays it.

[0572] Specific example

[0573] The following shows a specific example of how this system works.

[0574] Example 1: Checking product inventory

[0575] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[0576] 2. The device sends the message to the server.

[0577] 3. The server analyzes the received message using a natural language processing module and identifies it as an "inquiry about inventory."

[0578] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[0579] 5. The server receives the result "In stock" and uses generative AI to generate the response "Smartphones are currently in stock. You can place an order."

[0580] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[0581] Example 2: Technical support inquiry

[0582] 1. The user sends the message, "The app won't start. What should I do?"

[0583] 2. The server analyzes the received message using a natural language processing module and identifies it as a "technical support inquiry."

[0584] 3. The server executes the following query on the FAQ database: SELECT solution FROM faq WHERE issue='App won't start';

[0585] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0586] 5. The server uses a generative AI to generate a response based on this information: "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again."

[0587] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[0588] Example of a prompt

[0589] The following are specific examples of prompt statements to be input into the generating AI model.

[0590] 1. A prompt to generate a response to the inquiry, "Do you have the latest smartphones in stock?"

[0591] Prompt: "Generate an appropriate response when a user asks, 'Do you have the latest smartphones in stock?'"

[0592] 2. A prompt to generate a response to a technical support inquiry: "The app won't start. What should I do?"

[0593] Prompt: "Provide a solution for when a user asks, 'The app won't start. What should I do?'"

[0594] This invention enables the automation of customer inquiry handling, which is expected to reduce labor costs, shorten response times, and improve customer satisfaction.

[0595] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0596] Step 1:

[0597] The user enters an inquiry.

[0598] Input: The user enters their inquiry into the input field on the terminal. For example, they might enter "Please tell me the stock status of product A."

[0599] Action: The user clicks the "Submit" button.

[0600] Output: The entered text is sent to the system, and the contents of the input field are cleared.

[0601] Step 2:

[0602] The device sends a message to the server.

[0603] Input: The inquiry message entered by the user.

[0604] Operation: The terminal generates an HTTP POST request and sends the message to the server. This request includes query text and metadata (e.g., timestamp, user ID).

[0605] Output: If the request is successful, the server receives a message.

[0606] Step 3:

[0607] The server receives the message

[0608] Input: HTTP POST request sent from the terminal.

[0609] Operation: The server receives requests via the receiving API and temporarily stores the message content. It also logs the request metadata.

[0610] Output: The temporarily stored message is passed to the natural language processing (NLP) module.

[0611] Step 4:

[0612] Perform Natural Language Processing (NLP)

[0613] Input: Query message temporarily stored on the server.

[0614] Operation: The server uses an NLP module to tokenize messages, tag them with parts of speech, perform grammatical analysis, extract propositions, and so on. In this process, it analyzes and identifies the intent of the message.

[0615] Output: Analyzed intent (e.g., "Check inventory") and message structure data.

[0616] Step 5:

[0617] Execute a database query

[0618] Input: Analyzed intent and database query format.

[0619] Operation: The server generates and executes appropriate SQL queries against the database based on a specific intent. For example, it might issue a query to check the inventory status of "product A".

[0620] Output: Results retrieved from the database (e.g., "In stock").

[0621] Step 6:

[0622] Generate a response

[0623] Input: Results retrieved from the database and prompts for the generated AI model.

[0624] Operation: The server uses generative AI to construct an appropriate response based on the retrieved database results. For example, it might generate the sentence, "Product A is in stock."

[0625] Output: The generated response message.

[0626] Step 7:

[0627] The server sends a response message.

[0628] Input: The generated response message and the user's terminal information.

[0629] Operation: The server generates an HTTP POST request and sends a response message to the user's terminal.

[0630] Output: The terminal receives a response message.

[0631] Step 8:

[0632] The device displays a response message.

[0633] Input: The response message sent from the server.

[0634] Action: The device updates the user interface (UI) and displays a response message on the screen. For example, it might display the message "Product A is in stock."

[0635] Output: A response message is displayed on the user's screen.

[0636] (Application Example 1)

[0637] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0638] The problem that this invention aims to solve is to establish a method for providing quick and accurate responses to user inquiries. In particular, it aims to provide a system that can automatically generate responses to a wide variety of user inquiries on e-commerce sites, such as product inventory checks and technical support, and provide quick answers. Conventional methods often resulted in delayed responses or required a large amount of human resources, which contributed to decreased customer satisfaction.

[0639] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0640] In this invention, the server includes means for receiving input messages, means for analyzing the received messages using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing queries to a database according to the identified intent, means for checking the inventory status of a specific product and responding immediately to inquiries from users regarding inventory status, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to provide quick and accurate responses to user inquiries and improve customer satisfaction.

[0641] "Means for receiving input messages" refers to the means by which a user inputs a message through a terminal and receives it within the system.

[0642] "Means for analyzing received messages using natural language processing" refers to methods for analyzing received messages using natural language processing techniques to understand their content and intent.

[0643] "Means for identifying the intent of a message based on analysis results" refers to methods for identifying what a message means and what the user is looking for, based on the results of analysis using natural language processing.

[0644] "Means of querying a database according to a specified intent" refers to means of querying a database to retrieve necessary information based on a specified intent.

[0645] "A means of checking the inventory status of a specific product and responding immediately to user inquiries about inventory status" refers to a means of immediately checking the inventory status of a specific product in order to respond to user inquiries about inventory status.

[0646] "Means for generating appropriate responses using generative AI" refers to methods for automatically generating appropriate responses to user inquiries using generative AI technology.

[0647] "Means for sending the generated response to the user's terminal" refers to means of sending the generated response message to the user's terminal and providing feedback to the user.

[0648] The embodiments for carrying out the present invention are shown below. The present invention provides an automated inquiry response system, specifically intended for use on e-commerce websites. This system receives inquiry messages from users, analyzes them using natural language processing, retrieves information from a database based on the identified intent, and generates an appropriate response using generative AI.

[0649] System Overview

[0650] 1. User input

[0651] The user uses their smartphone to open the dedicated application for the online shopping site. They enter their inquiry in text format into the chat box. For example, they might type, "Do you still have any XX brand dress shirts in stock?"

[0652] 2. Receiving and parsing messages

[0653] The terminal sends the entered message to the server. The server receives the message and analyzes it using natural language processing. Specifically, processes such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction are performed.

[0654] 3. Identifying Intent and Data Acquisition

[0655] The system analyzes the message to identify the user's intent. Based on the identified intent, it queries the database to check, for example, the inventory status of a specific product. The server then passes the retrieved inventory information to the generative AI.

[0656] 4. Generating and sending responses

[0657] Generative AI generates appropriate responses based on the data it receives. For example, it might generate a message like, "We currently have XX brand dress shirts in stock. You can place an order."

[0658] 5. Display of response

[0659] The generated response message is sent back to the user's device and displayed in the chat box. This allows the user to receive a quick and accurate response.

[0660] Hardware and software

[0661] This system uses the following hardware and software.

[0662] Hardware:

[0663] Server: High-performance server (e.g., AWS, GCP)

[0664] User device: Smartphone (e.g., iPhone, Android device)

[0665] software:

[0666] Server side: Flask, SQLite (database)

[0667] Natural Language Processing: spaCy and NLTK

[0668] Generation AI:OpenAI GPT-3

[0669] Frontend: Smartphone applications (e.g., React Native)

[0670] Specific example

[0671] Message Processing Example

[0672] A user enters the message "Do you still have XX brand dress shirts in stock?" and sends it to the server. This message is received by the server and analyzed using natural language processing. From the analysis results, it is determined that the user's intention is to check the stock availability. The server queries the database to check the stock status of XX brand dress shirts. Based on the results, a generative AI generates a response message: "We currently have XX brand dress shirts in stock. You can place an order." This message is sent to the user's device and displayed in the chat box.

[0673] Example of prompts for the generating AI

[0674] If the product is in stock, how should I respond?

[0675] This system allows users to obtain information quickly and accurately, enabling effective responses to inquiries. This technology will greatly contribute to reducing labor costs and improving customer satisfaction.

[0676] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0677] Step 1:

[0678] The user uses their smartphone to open the e-commerce site's application and enters their inquiry as text into the chat box. For example, they might type, "Do you still have XX brand dress shirts in stock?" This message is then sent to the server as input.

[0679] Step 2:

[0680] The device sends the user's message to the server. Specifically, the device generates an API request and sends the message to the server via the receiving API. The input here is the message entered by the user, and the output is the message sent to the server.

[0681] Step 3:

[0682] The server temporarily stores the received message and passes it to the natural language processing module. The natural language processing module performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to analyze the meaning of the message. The input here is the message received in the previous step, and the output is the analysis result.

[0683] Step 4:

[0684] The server identifies the intent of the user's inquiry based on the results of natural language processing. For example, it might determine that this message is an inquiry about stock availability. At this stage, the input is the result of natural language processing, and the output is the identification of the intent, "stock availability check."

[0685] Step 5:

[0686] Based on the identified intent, the server executes a query to retrieve relevant data from the database. For example, "Execute an SQL query to check the inventory status of XX brand dress shirts. The input for this step is the identified intent, and the output is the result of the database query."

[0687] Step 6:

[0688] The server retrieves information from the database and passes it to a generative AI to generate an appropriate response. The generative AI is given prompts based on the retrieved information, and generates a response message such as, "We currently have XX brand dress shirts in stock. You can place an order." The input here is the inventory information retrieved from the database, and the output is the generated response message.

[0689] Step 7:

[0690] The server generates an API request to send the generated response message to the user's terminal and sends it to the user's terminal. The user's terminal displays the received response message in the chat box. The input here is the generated response message, and the output is the response message displayed on the user's terminal.

[0691] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0692] The following describes embodiments for carrying out the present invention. The present invention is an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and an emotion engine, generates an appropriate response, and sends it back to the user. The system operates based on a server, terminal, and user interface, and focuses particularly on analyzing the user's emotions.

[0693] System Overview

[0694] 1. Receive user inquiries.

[0695] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[0696] The device receives this message and generates an API request to send to the server.

[0697] 2. The server receives the message.

[0698] The server receives the user's message at the receiving API endpoint. This message is temporarily stored for parsing.

[0699] The server passes the received message to the natural language processing module.

[0700] 3. Natural language processing

[0701] The server uses a natural language processing (NLP) module to analyze the message content. This analysis includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0702] 4. Perform emotion analysis

[0703] The server uses an emotion engine to analyze the user's emotions within a message. For example, if a message like "Do you have it in stock?" is determined to contain a sense of urgency, it will assign emotion tags such as positive, negative, or neutral.

[0704] 5. Identifying the Intent

[0705] The server identifies the intent of the message based on the NLP analysis results and the emotion engine results. For example, it might be identified as an "inquiry about stock availability."

[0706] 6. Executing database queries

[0707] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[0708] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[0709] 7. Response generation

[0710] The server uses generative AI to automatically construct appropriate responses to user inquiries. In doing so, it considers sentiment tags and adjusts the tone and content of the response accordingly. For example, if negative sentiment is detected, the response message will emphasize consideration and solutions.

[0711] The server generates a message saying, "Smartphone X is currently in stock. You can order it now." However, if the user is feeling anxious, a more considerate message such as, "Don't worry, Smartphone X is in stock. You can order it right away," will be generated.

[0712] 8. Sending a response to the user

[0713] The system generates and sends an API request to send the response message generated by the server to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0714] 9. Displaying user responses

[0715] The device receives a response message from the server and displays it to the user. For example, it might display, "Smartphone X is currently in stock. You can place an order." However, it can also display messages that are considerate of the user's feelings.

[0716] Specific example

[0717] Example 1: Checking product inventory

[0718] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[0719] 2. The device sends a message to the server.

[0720] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[0721] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[0722] 5. The server receives the result "In stock".

[0723] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[0724] 7. Display the response received by the terminal to the user.

[0725] Example 2: Technical support inquiry

[0726] 1. The user sends the message, "The app won't start. What should I do?"

[0727] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[0728] 3. The server queries the FAQ database for relevant information.

[0729] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0730] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[0731] 6. Display the responses received by the device to the user.

[0732] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[0733] The following describes the processing flow.

[0734] Step 1:

[0735] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[0736] Step 2:

[0737] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[0738] Step 3:

[0739] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[0740] Step 4:

[0741] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0742] Step 5:

[0743] The server identifies the intent of the message based on the results of natural language processing. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[0744] Step 6:

[0745] The server calls an emotion analysis module to analyze the user's emotions within a message. Emotion analysis uses word-emotion dictionaries and machine learning models. For example, the message "Do you have this in stock?" might be judged to contain a sense of urgency.

[0746] Step 7:

[0747] The server assigns emotion tags to user messages based on the emotion analysis results. For example, the emotion tag "anxiety" might be assigned to a message.

[0748] Step 8:

[0749] Based on the identified intent and sentiment tags, the server generates queries to retrieve relevant information from the database. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[0750] Step 9:

[0751] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[0752] Step 10:

[0753] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. The response generation incorporates an algorithm that adjusts the tone and content considering emotion tags. For example, if a negative emotion is detected, the response message will emphasize consideration and solutions.

[0754] Step 11:

[0755] The server generates a message saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately." However, if the user is feeling anxious, a more considerate message such as "Please rest assured, Smartphone X is in stock. You can order it immediately." will be generated.

[0756] Step 12:

[0757] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0758] Step 13:

[0759] The device receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. Please rest assured, you can order it immediately." might be displayed.

[0760] (Example 2)

[0761] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0762] Traditional call center systems have struggled to provide efficient and effective responses to user inquiries, particularly in generating responses that take user emotions into consideration. Furthermore, their reliance on human resources has led to increased response times and rising labor costs. To address these challenges, a system is needed that accurately analyzes message content and automatically generates responses that consider user emotions.

[0763] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0764] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results and sentiment analysis, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to automatically generate a response that takes the user's emotions into consideration.

[0765] "Means for receiving input messages" refers to the communication methods and systems that allow a server to receive messages entered by a user using a terminal.

[0766] "Methods of analysis using natural language processing" refers to methods of analyzing human language using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction of input messages.

[0767] "Means of identifying the intent of a message" refers to methods of clarifying what the user is seeking in their inquiry, based on the results of natural language processing analysis and sentiment analysis.

[0768] "Means of querying a database" refers to a method of executing query commands to retrieve necessary information from a database based on a specified intent.

[0769] "Means of generating appropriate responses using generative AI" refers to a method of constructing responses to user inquiries based on information obtained from a database and sentiment tags, using generative AI models (e.g., GPT-3, BERT).

[0770] "Means of sending the generated response to the user's terminal" refers to the method of sending the response generated on the server to the user's terminal using a communication method or protocol (e.g., HTTP).

[0771] This invention relates to an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and sentiment analysis engines, generates appropriate responses, and sends them back to the users. This system operates based on interfaces from servers, terminals, and users, and places particular emphasis on analyzing user sentiment.

[0772] The server first receives a user inquiry message. The user enters the inquiry in text format using a device (e.g., smartphone, PC). For example, the user might enter the message, "Please tell me the stock status of product A." The device receives this message and generates an API request to send to the server. The server receives this message at the specified receiving API endpoint (e.g., / receive-message) and stores it temporarily.

[0773] Next, the server passes the received message to a natural language processing (NLP) module. The NLP module analyzes the message content using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Specifically, it uses an NLP library (e.g., spaCy or NLTK). For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0774] Furthermore, the server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the user's emotions within the message. The server analyzes the user's message and, if it detects negative emotions such as impatience, adds corresponding sentiment tags. For example, if the message "Do you have this in stock?" is determined to contain impatience, it will be tagged with positive, negative, or neutral sentiment tags.

[0775] Next, the server identifies the intent of the message based on the NLP analysis results and sentiment analysis results. For example, it can be identified as an "inquiry about stock availability." Based on this intent, the server executes a query to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query (e.g., SELECT stock_status FROM products WHERE product_name='Smartphone X';) and queries the database.

[0776] The server passes the acquired information to a generative AI model (e.g., GPT-3, BERT) to generate a response. The generative AI model generates an appropriate response based on the information and sentiment tags acquired from the database. In this process, it adjusts the tone and content of the response, taking sentiment tags into consideration. For example, based on the result "In stock," it generates the message, "Rest assured, smartphone X is in stock. You can order it immediately."

[0777] Finally, the server generates and sends an API request to send the generated response message to the user's device. This request is sent to the user's device using the HTTP protocol or similar. The device displays the received response message to the user, conveying a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[0778] Specific example

[0779] Example 1: Checking product inventory

[0780] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[0781] 2. The device sends a message to the server.

[0782] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[0783] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[0784] 5. The server receives the result "In stock".

[0785] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[0786] 7. Display the response received by the terminal to the user.

[0787] Example 2: Technical support inquiry

[0788] 1. The user sends the message, "The app won't start. What should I do?"

[0789] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[0790] 3. The server queries the FAQ database for relevant information.

[0791] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0792] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[0793] 6. Display the responses received by the device to the user.

[0794] Examples of input prompts for a generative AI model

[0795] "Generate a response based on the user's inquiry, 'Do you have the latest smartphone X in stock?' Assume the user is in a hurry."

[0796] "Based on the user inquiry, 'The app won't start. What should I do?', please generate an appropriate response for the confused user."

[0797] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[0798] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0799] Step 1: Receive user inquiry

[0800] The user enters their inquiry in text format using a device (e.g., smartphone, PC). An example of the entered message is "Please tell me the stock status of product A." The device receives this message and generates an API request (e.g., an HTTP POST request) to send to the server. The input is the user's message, and the output is the API request sent to the server.

[0801] Step 2: The server receives the message.

[0802] The server receives user messages at the specified receiving API endpoint (e.g., / receive-message). The server temporarily stores the received messages in an in-memory data store such as Redis and manages the order of processing. At this point, the input is the API request from the terminal, and the output is the temporarily stored messages.

[0803] Step 3: Process natural language

[0804] The server uses an NLP module (e.g., spaCy or NLTK) to tokenize the received message. The server then performs part-of-speech tagging and extracts propositions through grammatical analysis. Specifically, it might split a message like "Do you have the latest smartphone X in stock?" into tokens and assign a part-of-speech tag to each token. The input is the temporarily stored message, and the output is the tokenized data and part-of-speech tagging information as a result of the analysis.

[0805] Step 4: Perform sentiment analysis

[0806] The server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the sentiment of a message based on the NLP analysis results. The server analyzes the user's message, and if it detects negative emotions, such as anxiety, it adds the corresponding sentiment tag. The input is tokenized and part-of-speech tagged data, and the output is sentiment-tagged data.

[0807] Step 5: Identifying the Intent

[0808] The server integrates the NLP analysis results and sentiment analysis results to identify the intent of the message. For example, it might identify it as an "inquiry about stock availability." Specifically, the server prepares the data necessary for the next processing based on the identified intent. The input is analyzed data with sentiment tags, and the output is data containing the identified intent.

[0809] Step 6: Execute the database query

[0810] Based on the identified intent, the server issues queries to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query such as `SELECT stock_status FROM products WHERE product_name='smartphone X';` and queries the database. The input is data containing the identified intent, and the output is information retrieved from the database (e.g., "In stock").

[0811] Step 7: Generating the response

[0812] The server inputs data into a generative AI model (e.g., GPT-3, BERT) to generate appropriate responses to user inquiries. The tone and content of the responses are adjusted considering sentiment tags. For example, it might generate the message, "Rest assured, smartphone X is in stock. You can order it immediately." The input consists of information retrieved from a database and data with sentiment tags, while the output is the generated response.

[0813] Step 8: Sending a response to the user

[0814] This process generates and sends an API request to the user's terminal to send the response message generated by the server. This request is made using the HTTP protocol, etc. The input is the generated response message, and the output is the API request sent to the terminal.

[0815] Step 9: User response display

[0816] The terminal receives a response message sent from the server and displays it to the user. Specifically, it displays "Rest assured, smartphone X is in stock. You can order it immediately." The input is the response message sent from the server, and the output is the text displayed to the user.

[0817] These steps enable the system to provide timely and appropriate responses to user inquiries, while also considering their emotional impact.

[0818] (Application Example 2)

[0819] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0820] Traditional automated response systems often fail to consider user emotions, potentially leading to decreased user satisfaction. This is especially true for e-commerce sites, where prompt and appropriate responses to user inquiries are crucial; responses that disregard user feelings diminish credibility. Furthermore, traditional manual responses to a large volume of inquiries are inefficient and costly, necessitating automation.

[0821] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0822] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, means for analyzing the user's emotions, means for generating a response that takes the user's emotions into consideration based on the analysis results, and means for sending the generated response to the user's terminal. This makes it possible to provide an appropriate and prompt response that takes the user's emotions into consideration, thereby improving user satisfaction and reducing costs.

[0823] "Input messages" refer to text-based inquiries or requests that users send to the system through their devices.

[0824] "Means of receiving" refers to the function on the server side that receives messages sent by the user.

[0825] "Natural language processing" is a general term for technologies that analyze input messages and understand their content and structure.

[0826] "Means of analysis" refers to the function used to analyze received messages using natural language processing.

[0827] "Message intent" refers to the user's purpose or request, extracted from the content of the message as analyzed by natural language processing.

[0828] "Means of identification" refers to a function for identifying the intent of a message based on the analysis results.

[0829] "Means of querying a database" refers to means of retrieving necessary information from a database according to a specified intent.

[0830] "Generative AI" refers to artificial intelligence technology aimed at generation, specifically a system that automatically constructs responses to user inquiries based on information and analysis results obtained from a database.

[0831] "Means for generating appropriate responses" refers to a function that uses generative AI to generate appropriate responses to user inquiries.

[0832] "User sentiment" refers to the emotional nuances and tone that a user conveys when typing a message.

[0833] "Means of analysis (emotions)" refers to the function for analyzing the user's emotions within a message.

[0834] "Emotionally sensitive responses" refer to responses that are tailored to the user's emotions, based on the results of an emotional analysis of the user, and using appropriate tone and content.

[0835] "Means of generating (emotions)" refers to the function that generates responses that take into account the user's emotions, considering the results of emotion analysis.

[0836] "Means of sending to the user's terminal" refers to the function for sending the generated response from the server to the user's terminal.

[0837] System Overview

[0838] This invention is an automated call center system that automatically processes user inquiries and generates appropriate responses based on sentiment analysis. The system operates based on a server, terminals, and a user interface, and focuses particularly on analyzing the user's sentiment.

[0839] Hardware and software configuration

[0840] Hardware:

[0841] User devices: smartphones, PCs, etc.

[0842] Server: Equipped with a high-performance CPU, sufficient memory, and disk space.

[0843] software:

[0844] Natural language processing libraries: spaCy, NLTK, etc.

[0845] Sentiment analysis tools: TensorFlow-based sentiment analysis models, etc.

[0846] Server environment: Web framework such as Flask

[0847] Processing steps

[0848] 1. Receiving a message

[0849] Users submit inquiries in text format using devices such as smartphones and PCs. These messages are sent to the server via HTTP requests.

[0850] 2. Natural Language Processing and Sentiment Analysis

[0851] The server passes the received message to a natural language processing module for analysis such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Simultaneously, it uses an emotion engine to analyze the user's emotions. Sentiment analysis determines whether the message has multiple emotion tags, such as positive, negative, or anxiety.

[0852] 3. Identifying the Intent and Database Queries

[0853] Based on the analysis results, the server identifies the intent of the message. For example, it might categorize it as "inventory check" or "technical support." Based on the specific intent, the server queries the database to retrieve the necessary information.

[0854] 4. Generating a response

[0855] Based on the acquired information and sentiment tags, a generative AI is used to automatically generate a response message to the user. In this process, sentiment tags are taken into consideration to generate a considerate response.

[0856] 5. Sending and displaying responses

[0857] The server sends the generated response message to the user's terminal, and the terminal displays this message to the user.

[0858] Specific examples

[0859] Example of checking inventory:

[0860] User: "Do you have the latest smartphone X in stock?"

[0861] System response: "Rest assured, the Smartphone X is in stock. You can order it immediately."

[0862] Examples of technical support:

[0863] User: "The app won't start. What should I do?"

[0864] System response: "If the app won't launch, try clearing your cache or reinstalling the app. If the problem persists, please contact us again."

[0865] Example of a prompt

[0866] User query: "Do you have the latest smartphone X in stock?"

[0867] NLP analysis results: "Inventory check" "Smartphone X"

[0868] Emotion analysis result: "Anxiety"

[0869] Generate API prompt: Generate a response that is considerate of a user who is anxious about the availability of smartphone X.

[0870] Thus, the present invention makes it possible to achieve appropriate and prompt responses that take into account the user's feelings, thereby improving user satisfaction and reducing costs.

[0871] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0872] Step 1:

[0873] The user uses their device to enter their inquiry in text format. For example, they might enter the message, "Do you have the latest smartphone X in stock?" The entered message is received by the device's application and sent to the server in the form of an API request.

[0874] Input: User's text-based inquiry message

[0875] Output: Message in API request format

[0876] Step 2:

[0877] The server receives the message at the API endpoint. The received message is temporarily stored and then passed to the natural language processing (NLP) module.

[0878] Input: Message in API request format

[0879] Output: Message data to be passed to the NLP module

[0880] Step 3:

[0881] The server's natural language processing module analyzes the message. It performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to understand the message's content.

[0882] Input: Temporarily saved message data

[0883] Output: Analysis results (tokens, part-of-speech tagging, grammatical analysis results, proposition extraction results)

[0884] Step 4:

[0885] The server's emotion engine analyzes the user's emotions from the analyzed messages. For example, it assigns emotion tags such as positive, negative, and anxiety.

[0886] Input: NLP analysis results

[0887] Output: Sentiment analysis results (emotion tags)

[0888] Step 5:

[0889] The server identifies the intent of the message based on NLP analysis results and sentiment analysis results. For example, it might identify it as an "inquiry about stock availability."

[0890] Input: NLP analysis results and emotion analysis results

[0891] Output: Message intent (e.g., Inventory check)

[0892] Step 6:

[0893] Depending on the identified intent, the server executes SQL queries to retrieve relevant information from the database. For example, it might generate an SQL query to check the "stock status of product A" and then query the database.

[0894] Input: Message intent

[0895] Output: Database query results (e.g., inventory status)

[0896] Step 7:

[0897] The server uses generative AI to generate appropriate responses based on information and emotion tags retrieved from the database. For example, based on a database result such as "In stock" and the emotion tag "Anxiety," it generates a response such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[0898] Input: Database query results and sentiment tags

[0899] Output: Generated response message

[0900] Step 8:

[0901] Generate and send an API request to send the response message generated by the server to the user's terminal.

[0902] Input: Generated response message

[0903] Output: Response message in API request format

[0904] Step 9:

[0905] The terminal receives a response message from the server and displays it to the user. For example, it might display a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[0906] Input: Response message in API request format

[0907] Output: Response message displayed to the user

[0908] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0909] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0910] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0911] [Third Embodiment]

[0912] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0913] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0914] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0915] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0916] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0917] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0918] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0919] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0920] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0921] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0922] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0923] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0924] The embodiments for carrying out the present invention are shown below. The present invention is an automated call center system that has the function of receiving inquiry messages from users, analyzing them, generating appropriate responses, and sending them back to the users. This system operates based on a server, a terminal, and a user interface.

[0925] System Overview

[0926] 1. Receive user inquiries.

[0927] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[0928] The device receives this message and generates an API request to send to the server.

[0929] 2. The server receives the message.

[0930] The server receives the user's message via the receiving API. This message is temporarily stored for analysis.

[0931] The server passes the received message to the natural language processing module.

[0932] 3. Natural language processing

[0933] The server uses a natural language processing (NLP) module to analyze the message content. It understands the meaning and intent of the message using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[0934] The server identifies the intent of the inquiry based on the analysis results. For example, it might determine that the inquiry is for "stock confirmation."

[0935] 4. Execute database queries

[0936] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[0937] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[0938] 5. Generate a response

[0939] The server uses generative AI to automatically construct appropriate responses to user inquiries. For example, if product A is in stock, it will generate a response such as "Product A is in stock."

[0940] Generate an API request to send the response message generated by the server to the user's terminal.

[0941] 6. Send a response back to the user.

[0942] The server sends the generated response to the user's terminal.

[0943] The terminal displays the response message it received to the user. For example, the message "Product A is in stock" is displayed on the user's screen.

[0944] Specific example

[0945] Example 1: Checking product inventory

[0946] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[0947] 2. The device sends a message to the server.

[0948] 3. The server receives the message and performs natural language processing. It identifies it as an "inquiry about stock availability."

[0949] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[0950] 5. The server receives the result "In stock".

[0951] 6. The server generates a response saying, "Smartphones are currently in stock. You can place an order," and sends it to the user's terminal.

[0952] 7. Display the response received by the terminal to the user.

[0953] Example 2: Technical support inquiry

[0954] 1. The user sends the message, "The app won't start. What should I do?"

[0955] 2. The server receives the message, performs natural language processing, and identifies it as a "technical support inquiry."

[0956] 3. The server queries the FAQ database for relevant information.

[0957] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[0958] 5. The server generates a response saying, "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[0959] 6. Display the responses received by the device to the user.

[0960] This invention enables automated inquiry handling, reducing labor costs, shortening response times, and improving customer satisfaction.

[0961] The following describes the processing flow.

[0962] Step 1:

[0963] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[0964] Step 2:

[0965] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[0966] Step 3:

[0967] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[0968] Step 4:

[0969] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[0970] Step 5:

[0971] The server identifies the intent of the message based on the NLP analysis results. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[0972] Step 6:

[0973] The server generates queries against the database according to the identified intent. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[0974] Step 7:

[0975] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[0976] Step 8:

[0977] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be generated.

[0978] Step 9:

[0979] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[0980] Step 10:

[0981] The terminal receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be displayed.

[0982] (Example 1)

[0983] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0984] Traditional call center systems often involved manual responses to user inquiries, resulting in high labor costs and long response times. Furthermore, inconsistent quality of responses could lead to decreased customer satisfaction. Additionally, delays in handling complex inquiries could negatively impact the user experience.

[0985] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0986] In this invention, the server includes means for receiving a message entered by a user using a terminal, means for sending the received message to the server, means for the server to analyze the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing a query to a database according to the identified intent, means for generating an appropriate response using generative AI based on information obtained from the database, means for sending the generated response to the user's terminal, and means for displaying the response message received by the user's terminal. This enables automated, high-quality, and rapid inquiry handling.

[0987] A "user" refers to a person or entity that makes a request to the system.

[0988] A "device" refers to a device operated by a user, and includes smartphones, personal computers, tablets, and other similar devices.

[0989] A "server" is a central computing unit that receives user inquiries, performs analysis, and generates responses.

[0990] A "message" refers to the text or voice inquiry that a user enters through their device and sends to the server.

[0991] Natural Language Processing (NLP) refers to the technology used to analyze, interpret, and process human language using machines.

[0992] "Tokenization" refers to the process of dividing input text into smaller units such as words and phrases.

[0993] "Part-of-speech tagging" refers to the process of labeling tokenized words with their part of speech (e.g., noun, verb, adjective).

[0994] "Grammar analysis" refers to the process of analyzing the sentence structure of input text and understanding the relationships between each element.

[0995] "Propositional extraction" refers to the process of extracting important information and intentions from a text.

[0996] "Intent" refers to the purpose or requirements indicated in the user's inquiry message.

[0997] A "query" refers to a request sent to retrieve necessary information from a database.

[0998] A "database" refers to a system that stores and manages a collection of structured data.

[0999] "Generative AI" refers to artificial intelligence technology that automatically generates responses based on input data.

[1000] "Response" refers to the reply message that the server generates in response to a user's inquiry and sends to the terminal.

[1001] Modes for carrying out the invention

[1002] This invention provides an automated call center system. This system receives user inquiry messages, analyzes them, generates appropriate responses, and sends them back to the user, thereby achieving fast and high-quality service. Specific embodiments of this system are described below.

[1003] composition

[1004] This system operates based on users, terminals, and servers. Users use terminals to make inquiries, and the terminals send these inquiries to the server. The server analyzes the received messages, generates appropriate responses using a database and generative AI, and sends them back to the user via the terminal.

[1005] Hardware and software to be used

[1006] 1. Terminal: A device used by the user to enter inquiries, such as a smartphone or personal computer (PC).

[1007] 2. Server: A central computing unit that performs message analysis, response generation, and database query execution.

[1008] 3. Natural Language Processing (NLP) Module: Software that performs message tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1009] 4. Generative AI: An artificial intelligence model that automatically generates appropriate responses based on information obtained from a database.

[1010] 5. Database system: A system for storing and managing structured data.

[1011] Data processing and data calculation

[1012] Message reception: The user's device enters the inquiry message, creates an HTTP POST request, and sends it to the server.

[1013] Message Analysis: The server analyzes received messages using a natural language processing (NLP) module. The NLP module performs tokenization, part-of-speech tagging, grammatical analysis, proposition extraction, and other operations to determine the intent of the message.

[1014] Database queries: The server generates and executes SQL queries against the database according to the specified intent. The retrieved data is temporarily stored.

[1015] Response generation: The server uses generative AI to automatically generate an appropriate response based on the acquired data.

[1016] Response transmission: The server sends the generated response to the user's terminal as an HTTP POST request, and the terminal displays it.

[1017] Specific example

[1018] The following shows a specific example of how this system works.

[1019] Example 1: Checking product inventory

[1020] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[1021] 2. The device sends the message to the server.

[1022] 3. The server analyzes the received message using a natural language processing module and identifies it as an "inquiry about inventory."

[1023] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[1024] 5. The server receives the result "In stock" and uses generative AI to generate the response "Smartphones are currently in stock. You can place an order."

[1025] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[1026] Example 2: Technical support inquiry

[1027] 1. The user sends the message, "The app won't start. What should I do?"

[1028] 2. The server analyzes the received message using a natural language processing module and identifies it as a "technical support inquiry."

[1029] 3. The server executes the following query on the FAQ database: SELECT solution FROM faq WHERE issue='App won't start';

[1030] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1031] 5. The server uses a generative AI to generate a response based on this information: "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again."

[1032] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[1033] Example of a prompt

[1034] The following are specific examples of prompt statements to be input into the generating AI model.

[1035] 1. A prompt to generate a response to the inquiry, "Do you have the latest smartphones in stock?"

[1036] Prompt: "Generate an appropriate response when a user asks, 'Do you have the latest smartphones in stock?'"

[1037] 2. A prompt to generate a response to a technical support inquiry: "The app won't start. What should I do?"

[1038] Prompt: "Provide a solution for when a user asks, 'The app won't start. What should I do?'"

[1039] This invention enables the automation of customer inquiry handling, which is expected to reduce labor costs, shorten response times, and improve customer satisfaction.

[1040] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1041] Step 1:

[1042] The user enters an inquiry.

[1043] Input: The user enters their inquiry into the input field on the terminal. For example, they might enter "Please tell me the stock status of product A."

[1044] Action: The user clicks the "Submit" button.

[1045] Output: The entered text is sent to the system, and the contents of the input field are cleared.

[1046] Step 2:

[1047] The device sends a message to the server.

[1048] Input: The inquiry message entered by the user.

[1049] Operation: The terminal generates an HTTP POST request and sends the message to the server. This request includes query text and metadata (e.g., timestamp, user ID).

[1050] Output: If the request is successful, the server receives a message.

[1051] Step 3:

[1052] The server receives the message

[1053] Input: HTTP POST request sent from the terminal.

[1054] Operation: The server receives requests via the receiving API and temporarily stores the message content. It also logs the request metadata.

[1055] Output: The temporarily stored message is passed to the natural language processing (NLP) module.

[1056] Step 4:

[1057] Perform Natural Language Processing (NLP)

[1058] Input: Query message temporarily stored on the server.

[1059] Operation: The server uses an NLP module to tokenize messages, tag them with parts of speech, perform grammatical analysis, extract propositions, and so on. In this process, it analyzes and identifies the intent of the message.

[1060] Output: Analyzed intent (e.g., "Check inventory") and message structure data.

[1061] Step 5:

[1062] Execute a database query

[1063] Input: Analyzed intent and database query format.

[1064] Operation: The server generates and executes appropriate SQL queries against the database based on a specific intent. For example, it might issue a query to check the inventory status of "product A".

[1065] Output: Results retrieved from the database (e.g., "In stock").

[1066] Step 6:

[1067] Generate a response

[1068] Input: Results retrieved from the database and prompts for the generated AI model.

[1069] Operation: The server uses generative AI to construct an appropriate response based on the retrieved database results. For example, it might generate the sentence, "Product A is in stock."

[1070] Output: The generated response message.

[1071] Step 7:

[1072] The server sends a response message.

[1073] Input: The generated response message and the user's terminal information.

[1074] Operation: The server generates an HTTP POST request and sends a response message to the user's terminal.

[1075] Output: The terminal receives a response message.

[1076] Step 8:

[1077] The device displays a response message.

[1078] Input: The response message sent from the server.

[1079] Action: The device updates the user interface (UI) and displays a response message on the screen. For example, it might display the message "Product A is in stock."

[1080] Output: A response message is displayed on the user's screen.

[1081] (Application Example 1)

[1082] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1083] The problem that this invention aims to solve is to establish a method for providing quick and accurate responses to user inquiries. In particular, it aims to provide a system that can automatically generate responses to a wide variety of user inquiries on e-commerce sites, such as product inventory checks and technical support, and provide quick answers. Conventional methods often resulted in delayed responses or required a large amount of human resources, which contributed to decreased customer satisfaction.

[1084] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1085] In this invention, the server includes means for receiving input messages, means for analyzing the received messages using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing queries to a database according to the identified intent, means for checking the inventory status of a specific product and responding immediately to inquiries from users regarding inventory status, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to provide quick and accurate responses to user inquiries and improve customer satisfaction.

[1086] "Means for receiving input messages" refers to the means by which a user inputs a message through a terminal and receives it within the system.

[1087] "Means for analyzing received messages using natural language processing" refers to methods for analyzing received messages using natural language processing techniques to understand their content and intent.

[1088] "Means for identifying the intent of a message based on analysis results" refers to methods for identifying what a message means and what the user is looking for, based on the results of analysis using natural language processing.

[1089] "Means of querying a database according to a specified intent" refers to means of querying a database to retrieve necessary information based on a specified intent.

[1090] "A means of checking the inventory status of a specific product and responding immediately to user inquiries about inventory status" refers to a means of immediately checking the inventory status of a specific product in order to respond to user inquiries about inventory status.

[1091] "Means for generating appropriate responses using generative AI" refers to methods for automatically generating appropriate responses to user inquiries using generative AI technology.

[1092] "Means for sending the generated response to the user's terminal" refers to means of sending the generated response message to the user's terminal and providing feedback to the user.

[1093] The embodiments for carrying out the present invention are shown below. The present invention provides an automated inquiry response system, specifically intended for use on e-commerce websites. This system receives inquiry messages from users, analyzes them using natural language processing, retrieves information from a database based on the identified intent, and generates an appropriate response using generative AI.

[1094] System Overview

[1095] 1. User input

[1096] The user uses their smartphone to open the dedicated application for the online shopping site. They enter their inquiry in text format into the chat box. For example, they might type, "Do you still have any XX brand dress shirts in stock?"

[1097] 2. Receiving and parsing messages

[1098] The terminal sends the entered message to the server. The server receives the message and analyzes it using natural language processing. Specifically, processes such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction are performed.

[1099] 3. Identifying Intent and Data Acquisition

[1100] The system analyzes the message to identify the user's intent. Based on the identified intent, it queries the database to check, for example, the inventory status of a specific product. The server then passes the retrieved inventory information to the generative AI.

[1101] 4. Generating and sending responses

[1102] Generative AI generates appropriate responses based on the data it receives. For example, it might generate a message like, "We currently have XX brand dress shirts in stock. You can place an order."

[1103] 5. Display of response

[1104] The generated response message is sent back to the user's device and displayed in the chat box. This allows the user to receive a quick and accurate response.

[1105] Hardware and software

[1106] This system uses the following hardware and software.

[1107] Hardware:

[1108] Server: High-performance server (e.g., AWS, GCP)

[1109] User device: Smartphone (e.g., iPhone, Android device)

[1110] software:

[1111] Server side: Flask, SQLite (database)

[1112] Natural Language Processing: spaCy and NLTK

[1113] Generation AI:OpenAI GPT-3

[1114] Frontend: Smartphone applications (e.g., React Native)

[1115] Specific example

[1116] Message Processing Example

[1117] A user enters the message "Do you still have XX brand dress shirts in stock?" and sends it to the server. This message is received by the server and analyzed using natural language processing. From the analysis results, it is determined that the user's intention is to check the stock availability. The server queries the database to check the stock status of XX brand dress shirts. Based on the results, a generative AI generates a response message: "We currently have XX brand dress shirts in stock. You can place an order." This message is sent to the user's device and displayed in the chat box.

[1118] Example of prompts for the generating AI

[1119] If the product is in stock, how should I respond?

[1120] This system allows users to obtain information quickly and accurately, enabling effective responses to inquiries. This technology will greatly contribute to reducing labor costs and improving customer satisfaction.

[1121] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1122] Step 1:

[1123] The user uses their smartphone to open the e-commerce site's application and enters their inquiry as text into the chat box. For example, they might type, "Do you still have XX brand dress shirts in stock?" This message is then sent to the server as input.

[1124] Step 2:

[1125] The device sends the user's message to the server. Specifically, the device generates an API request and sends the message to the server via the receiving API. The input here is the message entered by the user, and the output is the message sent to the server.

[1126] Step 3:

[1127] The server temporarily stores the received message and passes it to the natural language processing module. The natural language processing module performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to analyze the meaning of the message. The input here is the message received in the previous step, and the output is the analysis result.

[1128] Step 4:

[1129] The server identifies the intent of the user's inquiry based on the results of natural language processing. For example, it might determine that this message is an inquiry about stock availability. At this stage, the input is the result of natural language processing, and the output is the identification of the intent, "stock availability check."

[1130] Step 5:

[1131] Based on the identified intent, the server executes a query to retrieve relevant data from the database. For example, "Execute an SQL query to check the inventory status of XX brand dress shirts. The input for this step is the identified intent, and the output is the result of the database query."

[1132] Step 6:

[1133] The server retrieves information from the database and passes it to a generative AI to generate an appropriate response. The generative AI is given prompts based on the retrieved information, and generates a response message such as, "We currently have XX brand dress shirts in stock. You can place an order." The input here is the inventory information retrieved from the database, and the output is the generated response message.

[1134] Step 7:

[1135] The server generates an API request to send the generated response message to the user's terminal and sends it to the user's terminal. The user's terminal displays the received response message in the chat box. The input here is the generated response message, and the output is the response message displayed on the user's terminal.

[1136] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1137] The following describes embodiments for carrying out the present invention. The present invention is an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and an emotion engine, generates an appropriate response, and sends it back to the user. The system operates based on a server, terminal, and user interface, and focuses particularly on analyzing the user's emotions.

[1138] System Overview

[1139] 1. Receive user inquiries.

[1140] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[1141] The device receives this message and generates an API request to send to the server.

[1142] 2. The server receives the message.

[1143] The server receives the user's message at the receiving API endpoint. This message is temporarily stored for parsing.

[1144] The server passes the received message to the natural language processing module.

[1145] 3. Natural language processing

[1146] The server uses a natural language processing (NLP) module to analyze the message content. This analysis includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1147] 4. Perform emotion analysis

[1148] The server uses an emotion engine to analyze the user's emotions within a message. For example, if a message like "Do you have it in stock?" is determined to contain a sense of urgency, it will assign emotion tags such as positive, negative, or neutral.

[1149] 5. Identifying the Intent

[1150] The server identifies the intent of the message based on the NLP analysis results and the emotion engine results. For example, it might be identified as an "inquiry about stock availability."

[1151] 6. Executing database queries

[1152] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[1153] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[1154] 7. Response generation

[1155] The server uses generative AI to automatically construct appropriate responses to user inquiries. In doing so, it considers sentiment tags and adjusts the tone and content of the response accordingly. For example, if negative sentiment is detected, the response message will emphasize consideration and solutions.

[1156] The server generates a message saying, "Smartphone X is currently in stock. You can order it now." However, if the user is feeling anxious, a more considerate message such as, "Don't worry, Smartphone X is in stock. You can order it right away," will be generated.

[1157] 8. Sending a response to the user

[1158] The system generates and sends an API request to send the response message generated by the server to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[1159] 9. Displaying user responses

[1160] The device receives a response message from the server and displays it to the user. For example, it might display, "Smartphone X is currently in stock. You can place an order." However, it can also display messages that are considerate of the user's feelings.

[1161] Specific example

[1162] Example 1: Checking product inventory

[1163] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[1164] 2. The device sends a message to the server.

[1165] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[1166] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[1167] 5. The server receives the result "In stock".

[1168] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[1169] 7. Display the response received by the terminal to the user.

[1170] Example 2: Technical support inquiry

[1171] 1. The user sends the message, "The app won't start. What should I do?"

[1172] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[1173] 3. The server queries the FAQ database for relevant information.

[1174] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1175] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[1176] 6. Display the responses received by the device to the user.

[1177] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[1178] The following describes the processing flow.

[1179] Step 1:

[1180] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[1181] Step 2:

[1182] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[1183] Step 3:

[1184] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[1185] Step 4:

[1186] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1187] Step 5:

[1188] The server identifies the intent of the message based on the results of natural language processing. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[1189] Step 6:

[1190] The server calls an emotion analysis module to analyze the user's emotions within a message. Emotion analysis uses word-emotion dictionaries and machine learning models. For example, the message "Do you have this in stock?" might be judged to contain a sense of urgency.

[1191] Step 7:

[1192] The server assigns emotion tags to user messages based on the emotion analysis results. For example, the emotion tag "anxiety" might be assigned to a message.

[1193] Step 8:

[1194] Based on the identified intent and sentiment tags, the server generates queries to retrieve relevant information from the database. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[1195] Step 9:

[1196] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[1197] Step 10:

[1198] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. The response generation incorporates an algorithm that adjusts the tone and content considering emotion tags. For example, if a negative emotion is detected, the response message will emphasize consideration and solutions.

[1199] Step 11:

[1200] The server generates a message saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately." However, if the user is feeling anxious, a more considerate message such as "Please rest assured, Smartphone X is in stock. You can order it immediately." will be generated.

[1201] Step 12:

[1202] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[1203] Step 13:

[1204] The device receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. Please rest assured, you can order it immediately." might be displayed.

[1205] (Example 2)

[1206] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1207] Traditional call center systems have struggled to provide efficient and effective responses to user inquiries, particularly in generating responses that take user emotions into consideration. Furthermore, their reliance on human resources has led to increased response times and rising labor costs. To address these challenges, a system is needed that accurately analyzes message content and automatically generates responses that consider user emotions.

[1208] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1209] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results and sentiment analysis, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to automatically generate a response that takes the user's emotions into consideration.

[1210] "Means for receiving input messages" refers to the communication methods and systems that allow a server to receive messages entered by a user using a terminal.

[1211] "Methods of analysis using natural language processing" refers to methods of analyzing human language using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction of input messages.

[1212] "Means of identifying the intent of a message" refers to methods of clarifying what the user is seeking in their inquiry, based on the results of natural language processing analysis and sentiment analysis.

[1213] "Means of querying a database" refers to a method of executing query commands to retrieve necessary information from a database based on a specified intent.

[1214] "Means of generating appropriate responses using generative AI" refers to a method of constructing responses to user inquiries based on information obtained from a database and sentiment tags, using generative AI models (e.g., GPT-3, BERT).

[1215] "Means of sending the generated response to the user's terminal" refers to the method of sending the response generated on the server to the user's terminal using a communication method or protocol (e.g., HTTP).

[1216] This invention relates to an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and sentiment analysis engines, generates appropriate responses, and sends them back to the users. This system operates based on interfaces from servers, terminals, and users, and places particular emphasis on analyzing user sentiment.

[1217] The server first receives a user inquiry message. The user enters the inquiry in text format using a device (e.g., smartphone, PC). For example, the user might enter the message, "Please tell me the stock status of product A." The device receives this message and generates an API request to send to the server. The server receives this message at the specified receiving API endpoint (e.g., / receive-message) and stores it temporarily.

[1218] Next, the server passes the received message to a natural language processing (NLP) module. The NLP module analyzes the message content using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Specifically, it uses an NLP library (e.g., spaCy or NLTK). For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1219] Furthermore, the server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the user's emotions within the message. The server analyzes the user's message and, if it detects negative emotions such as impatience, adds corresponding sentiment tags. For example, if the message "Do you have this in stock?" is determined to contain impatience, it will be tagged with positive, negative, or neutral sentiment tags.

[1220] Next, the server identifies the intent of the message based on the NLP analysis results and sentiment analysis results. For example, it can be identified as an "inquiry about stock availability." Based on this intent, the server executes a query to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query (e.g., SELECT stock_status FROM products WHERE product_name='Smartphone X';) and queries the database.

[1221] The server passes the acquired information to a generative AI model (e.g., GPT-3, BERT) to generate a response. The generative AI model generates an appropriate response based on the information and sentiment tags acquired from the database. In this process, it adjusts the tone and content of the response, taking sentiment tags into consideration. For example, based on the result "In stock," it generates the message, "Rest assured, smartphone X is in stock. You can order it immediately."

[1222] Finally, the server generates and sends an API request to send the generated response message to the user's device. This request is sent to the user's device using the HTTP protocol or similar. The device displays the received response message to the user, conveying a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[1223] Specific example

[1224] Example 1: Checking product inventory

[1225] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[1226] 2. The device sends a message to the server.

[1227] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[1228] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[1229] 5. The server receives the result "In stock".

[1230] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[1231] 7. Display the response received by the terminal to the user.

[1232] Example 2: Technical support inquiry

[1233] 1. The user sends the message, "The app won't start. What should I do?"

[1234] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[1235] 3. The server queries the FAQ database for relevant information.

[1236] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1237] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[1238] 6. Display the responses received by the device to the user.

[1239] Examples of input prompts for a generative AI model

[1240] "Generate a response based on the user's inquiry, 'Do you have the latest smartphone X in stock?' Assume the user is in a hurry."

[1241] "Based on the user inquiry, 'The app won't start. What should I do?', please generate an appropriate response for the confused user."

[1242] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[1243] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1244] Step 1: Receive user inquiry

[1245] The user enters their inquiry in text format using a device (e.g., smartphone, PC). An example of the entered message is "Please tell me the stock status of product A." The device receives this message and generates an API request (e.g., an HTTP POST request) to send to the server. The input is the user's message, and the output is the API request sent to the server.

[1246] Step 2: The server receives the message.

[1247] The server receives user messages at the specified receiving API endpoint (e.g., / receive-message). The server temporarily stores the received messages in an in-memory data store such as Redis and manages the order of processing. At this point, the input is the API request from the terminal, and the output is the temporarily stored messages.

[1248] Step 3: Process natural language

[1249] The server uses an NLP module (e.g., spaCy or NLTK) to tokenize the received message. The server then performs part-of-speech tagging and extracts propositions through grammatical analysis. Specifically, it might split a message like "Do you have the latest smartphone X in stock?" into tokens and assign a part-of-speech tag to each token. The input is the temporarily stored message, and the output is the tokenized data and part-of-speech tagging information as a result of the analysis.

[1250] Step 4: Perform sentiment analysis

[1251] The server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the sentiment of a message based on the NLP analysis results. The server analyzes the user's message, and if it detects negative emotions, such as anxiety, it adds the corresponding sentiment tag. The input is tokenized and part-of-speech tagged data, and the output is sentiment-tagged data.

[1252] Step 5: Identifying the Intent

[1253] The server integrates the NLP analysis results and sentiment analysis results to identify the intent of the message. For example, it might identify it as an "inquiry about stock availability." Specifically, the server prepares the data necessary for the next processing based on the identified intent. The input is analyzed data with sentiment tags, and the output is data containing the identified intent.

[1254] Step 6: Execute the database query

[1255] Based on the identified intent, the server issues queries to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query such as `SELECT stock_status FROM products WHERE product_name='smartphone X';` and queries the database. The input is data containing the identified intent, and the output is information retrieved from the database (e.g., "In stock").

[1256] Step 7: Generating the response

[1257] The server inputs data into a generative AI model (e.g., GPT-3, BERT) to generate appropriate responses to user inquiries. The tone and content of the responses are adjusted considering sentiment tags. For example, it might generate the message, "Rest assured, smartphone X is in stock. You can order it immediately." The input consists of information retrieved from a database and data with sentiment tags, while the output is the generated response.

[1258] Step 8: Sending a response to the user

[1259] This process generates and sends an API request to the user's terminal to send the response message generated by the server. This request is made using the HTTP protocol, etc. The input is the generated response message, and the output is the API request sent to the terminal.

[1260] Step 9: User response display

[1261] The terminal receives a response message sent from the server and displays it to the user. Specifically, it displays "Rest assured, smartphone X is in stock. You can order it immediately." The input is the response message sent from the server, and the output is the text displayed to the user.

[1262] These steps enable the system to provide timely and appropriate responses to user inquiries, while also considering their emotional impact.

[1263] (Application Example 2)

[1264] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1265] Traditional automated response systems often fail to consider user emotions, potentially leading to decreased user satisfaction. This is especially true for e-commerce sites, where prompt and appropriate responses to user inquiries are crucial; responses that disregard user feelings diminish credibility. Furthermore, traditional manual responses to a large volume of inquiries are inefficient and costly, necessitating automation.

[1266] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1267] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, means for analyzing the user's emotions, means for generating a response that takes the user's emotions into consideration based on the analysis results, and means for sending the generated response to the user's terminal. This makes it possible to provide an appropriate and prompt response that takes the user's emotions into consideration, thereby improving user satisfaction and reducing costs.

[1268] "Input messages" refer to text-based inquiries or requests that users send to the system through their devices.

[1269] "Means of receiving" refers to the function on the server side that receives messages sent by the user.

[1270] "Natural language processing" is a general term for technologies that analyze input messages and understand their content and structure.

[1271] "Means of analysis" refers to the function used to analyze received messages using natural language processing.

[1272] "Message intent" refers to the user's purpose or request, extracted from the content of the message as analyzed by natural language processing.

[1273] "Means of identification" refers to a function for identifying the intent of a message based on the analysis results.

[1274] "Means of querying a database" refers to means of retrieving necessary information from a database according to a specified intent.

[1275] "Generative AI" refers to artificial intelligence technology aimed at generation, specifically a system that automatically constructs responses to user inquiries based on information and analysis results obtained from a database.

[1276] "Means for generating appropriate responses" refers to a function that uses generative AI to generate appropriate responses to user inquiries.

[1277] "User sentiment" refers to the emotional nuances and tone that a user conveys when typing a message.

[1278] "Means of analysis (emotions)" refers to the function for analyzing the user's emotions within a message.

[1279] "Emotionally sensitive responses" refer to responses that are tailored to the user's emotions, based on the results of an emotional analysis of the user, and using appropriate tone and content.

[1280] "Means of generating (emotions)" refers to the function that generates responses that take into account the user's emotions, considering the results of emotion analysis.

[1281] "Means of sending to the user's terminal" refers to the function for sending the generated response from the server to the user's terminal.

[1282] System Overview

[1283] This invention is an automated call center system that automatically processes user inquiries and generates appropriate responses based on sentiment analysis. The system operates based on a server, terminals, and a user interface, and focuses particularly on analyzing the user's sentiment.

[1284] Hardware and software configuration

[1285] Hardware:

[1286] User devices: smartphones, PCs, etc.

[1287] Server: Equipped with a high-performance CPU, sufficient memory, and disk space.

[1288] software:

[1289] Natural language processing libraries: spaCy, NLTK, etc.

[1290] Sentiment analysis tools: TensorFlow-based sentiment analysis models, etc.

[1291] Server environment: Web framework such as Flask

[1292] Processing steps

[1293] 1. Receiving a message

[1294] Users submit inquiries in text format using devices such as smartphones and PCs. These messages are sent to the server via HTTP requests.

[1295] 2. Natural Language Processing and Sentiment Analysis

[1296] The server passes the received message to a natural language processing module for analysis such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Simultaneously, it uses an emotion engine to analyze the user's emotions. Sentiment analysis determines whether the message has multiple emotion tags, such as positive, negative, or anxiety.

[1297] 3. Identifying the Intent and Database Queries

[1298] Based on the analysis results, the server identifies the intent of the message. For example, it might categorize it as "inventory check" or "technical support." Based on the specific intent, the server queries the database to retrieve the necessary information.

[1299] 4. Generating a response

[1300] Based on the acquired information and sentiment tags, a generative AI is used to automatically generate a response message to the user. In this process, sentiment tags are taken into consideration to generate a considerate response.

[1301] 5. Sending and displaying responses

[1302] The server sends the generated response message to the user's terminal, and the terminal displays this message to the user.

[1303] Specific examples

[1304] Example of checking inventory:

[1305] User: "Do you have the latest smartphone X in stock?"

[1306] System response: "Rest assured, the Smartphone X is in stock. You can order it immediately."

[1307] Examples of technical support:

[1308] User: "The app won't start. What should I do?"

[1309] System response: "If the app won't launch, try clearing your cache or reinstalling the app. If the problem persists, please contact us again."

[1310] Example of a prompt

[1311] User query: "Do you have the latest smartphone X in stock?"

[1312] NLP analysis results: "Inventory check" "Smartphone X"

[1313] Emotion analysis result: "Anxiety"

[1314] Generate API prompt: Generate a response that is considerate of a user who is anxious about the availability of smartphone X.

[1315] Thus, the present invention makes it possible to achieve appropriate and prompt responses that take into account the user's feelings, thereby improving user satisfaction and reducing costs.

[1316] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1317] Step 1:

[1318] The user uses their device to enter their inquiry in text format. For example, they might enter the message, "Do you have the latest smartphone X in stock?" The entered message is received by the device's application and sent to the server in the form of an API request.

[1319] Input: User's text-based inquiry message

[1320] Output: Message in API request format

[1321] Step 2:

[1322] The server receives the message at the API endpoint. The received message is temporarily stored and then passed to the natural language processing (NLP) module.

[1323] Input: Message in API request format

[1324] Output: Message data to be passed to the NLP module

[1325] Step 3:

[1326] The server's natural language processing module analyzes the message. It performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to understand the message's content.

[1327] Input: Temporarily saved message data

[1328] Output: Analysis results (tokens, part-of-speech tagging, grammatical analysis results, proposition extraction results)

[1329] Step 4:

[1330] The server's emotion engine analyzes the user's emotions from the analyzed messages. For example, it assigns emotion tags such as positive, negative, and anxiety.

[1331] Input: NLP analysis results

[1332] Output: Sentiment analysis results (emotion tags)

[1333] Step 5:

[1334] The server identifies the intent of the message based on NLP analysis results and sentiment analysis results. For example, it might identify it as an "inquiry about stock availability."

[1335] Input: NLP analysis results and emotion analysis results

[1336] Output: Message intent (e.g., Inventory check)

[1337] Step 6:

[1338] Depending on the identified intent, the server executes SQL queries to retrieve relevant information from the database. For example, it might generate an SQL query to check the "stock status of product A" and then query the database.

[1339] Input: Message intent

[1340] Output: Database query results (e.g., inventory status)

[1341] Step 7:

[1342] The server uses generative AI to generate appropriate responses based on information and emotion tags retrieved from the database. For example, based on a database result such as "In stock" and the emotion tag "Anxiety," it generates a response such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[1343] Input: Database query results and sentiment tags

[1344] Output: Generated response message

[1345] Step 8:

[1346] Generate and send an API request to send the response message generated by the server to the user's terminal.

[1347] Input: Generated response message

[1348] Output: Response message in API request format

[1349] Step 9:

[1350] The terminal receives a response message from the server and displays it to the user. For example, it might display a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[1351] Input: Response message in API request format

[1352] Output: Response message displayed to the user

[1353] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1354] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1355] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1356] [Fourth Embodiment]

[1357] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1358] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1359] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1360] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1361] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1362] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1363] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1364] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1365] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1366] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1367] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1368] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1369] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1370] The embodiments for carrying out the present invention are shown below. The present invention is an automated call center system that has the function of receiving inquiry messages from users, analyzing them, generating appropriate responses, and sending them back to the users. This system operates based on a server, a terminal, and a user interface.

[1371] System Overview

[1372] 1. Receive user inquiries.

[1373] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[1374] The device receives this message and generates an API request to send to the server.

[1375] 2. The server receives the message.

[1376] The server receives the user's message via the receiving API. This message is temporarily stored for analysis.

[1377] The server passes the received message to the natural language processing module.

[1378] 3. Natural language processing

[1379] The server uses a natural language processing (NLP) module to analyze the message content. It understands the meaning and intent of the message using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1380] The server identifies the intent of the inquiry based on the analysis results. For example, it might determine that the inquiry is for "stock confirmation."

[1381] 4. Execute database queries

[1382] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[1383] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[1384] 5. Generate a response

[1385] The server uses generative AI to automatically construct appropriate responses to user inquiries. For example, if product A is in stock, it will generate a response such as "Product A is in stock."

[1386] Generate an API request to send the response message generated by the server to the user's terminal.

[1387] 6. Send a response back to the user.

[1388] The server sends the generated response to the user's terminal.

[1389] The terminal displays the response message it received to the user. For example, the message "Product A is in stock" is displayed on the user's screen.

[1390] Specific example

[1391] Example 1: Checking product inventory

[1392] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[1393] 2. The device sends a message to the server.

[1394] 3. The server receives the message and performs natural language processing. It identifies it as an "inquiry about stock availability."

[1395] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[1396] 5. The server receives the result "In stock".

[1397] 6. The server generates a response saying, "Smartphones are currently in stock. You can place an order," and sends it to the user's terminal.

[1398] 7. Display the response received by the terminal to the user.

[1399] Example 2: Technical support inquiry

[1400] 1. The user sends the message, "The app won't start. What should I do?"

[1401] 2. The server receives the message, performs natural language processing, and identifies it as a "technical support inquiry."

[1402] 3. The server queries the FAQ database for relevant information.

[1403] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1404] 5. The server generates a response saying, "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[1405] 6. Display the responses received by the device to the user.

[1406] This invention enables automated inquiry handling, reducing labor costs, shortening response times, and improving customer satisfaction.

[1407] The following describes the processing flow.

[1408] Step 1:

[1409] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[1410] Step 2:

[1411] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[1412] Step 3:

[1413] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[1414] Step 4:

[1415] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1416] Step 5:

[1417] The server identifies the intent of the message based on the NLP analysis results. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[1418] Step 6:

[1419] The server generates queries against the database according to the identified intent. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[1420] Step 7:

[1421] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[1422] Step 8:

[1423] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be generated.

[1424] Step 9:

[1425] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[1426] Step 10:

[1427] The terminal receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. You can place an order." might be displayed.

[1428] (Example 1)

[1429] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1430] Traditional call center systems often involved manual responses to user inquiries, resulting in high labor costs and long response times. Furthermore, inconsistent quality of responses could lead to decreased customer satisfaction. Additionally, delays in handling complex inquiries could negatively impact the user experience.

[1431] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1432] In this invention, the server includes means for receiving a message entered by a user using a terminal, means for sending the received message to the server, means for the server to analyze the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing a query to a database according to the identified intent, means for generating an appropriate response using generative AI based on information obtained from the database, means for sending the generated response to the user's terminal, and means for displaying the response message received by the user's terminal. This enables automated, high-quality, and rapid inquiry handling.

[1433] A "user" refers to a person or entity that makes a request to the system.

[1434] A "device" refers to a device operated by a user, and includes smartphones, personal computers, tablets, and other similar devices.

[1435] A "server" is a central computing unit that receives user inquiries, performs analysis, and generates responses.

[1436] A "message" refers to the text or voice inquiry that a user enters through their device and sends to the server.

[1437] Natural Language Processing (NLP) refers to the technology used to analyze, interpret, and process human language using machines.

[1438] "Tokenization" refers to the process of dividing input text into smaller units such as words and phrases.

[1439] "Part-of-speech tagging" refers to the process of labeling tokenized words with their part of speech (e.g., noun, verb, adjective).

[1440] "Grammar analysis" refers to the process of analyzing the sentence structure of input text and understanding the relationships between each element.

[1441] "Propositional extraction" refers to the process of extracting important information and intentions from a text.

[1442] "Intent" refers to the purpose or requirements indicated in the user's inquiry message.

[1443] A "query" refers to a request sent to retrieve necessary information from a database.

[1444] A "database" refers to a system that stores and manages a collection of structured data.

[1445] "Generative AI" refers to artificial intelligence technology that automatically generates responses based on input data.

[1446] "Response" refers to the reply message that the server generates in response to a user's inquiry and sends to the terminal.

[1447] Modes for carrying out the invention

[1448] This invention provides an automated call center system. This system receives user inquiry messages, analyzes them, generates appropriate responses, and sends them back to the user, thereby achieving fast and high-quality service. Specific embodiments of this system are described below.

[1449] composition

[1450] This system operates based on users, terminals, and servers. Users use terminals to make inquiries, and the terminals send these inquiries to the server. The server analyzes the received messages, generates appropriate responses using a database and generative AI, and sends them back to the user via the terminal.

[1451] Hardware and software to be used

[1452] 1. Terminal: A device used by the user to enter inquiries, such as a smartphone or personal computer (PC).

[1453] 2. Server: A central computing unit that performs message analysis, response generation, and database query execution.

[1454] 3. Natural Language Processing (NLP) Module: Software that performs message tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1455] 4. Generative AI: An artificial intelligence model that automatically generates appropriate responses based on information obtained from a database.

[1456] 5. Database system: A system for storing and managing structured data.

[1457] Data processing and data calculation

[1458] Message reception: The user's device enters the inquiry message, creates an HTTP POST request, and sends it to the server.

[1459] Message Analysis: The server analyzes received messages using a natural language processing (NLP) module. The NLP module performs tokenization, part-of-speech tagging, grammatical analysis, proposition extraction, and other operations to determine the intent of the message.

[1460] Database queries: The server generates and executes SQL queries against the database according to the specified intent. The retrieved data is temporarily stored.

[1461] Response generation: The server uses generative AI to automatically generate an appropriate response based on the acquired data.

[1462] Response transmission: The server sends the generated response to the user's terminal as an HTTP POST request, and the terminal displays it.

[1463] Specific example

[1464] The following shows a specific example of how this system works.

[1465] Example 1: Checking product inventory

[1466] 1. The user enters "Do you have the latest smartphones in stock?" into the device.

[1467] 2. The device sends the message to the server.

[1468] 3. The server analyzes the received message using a natural language processing module and identifies it as an "inquiry about inventory."

[1469] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='smartphone';

[1470] 5. The server receives the result "In stock" and uses generative AI to generate the response "Smartphones are currently in stock. You can place an order."

[1471] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[1472] Example 2: Technical support inquiry

[1473] 1. The user sends the message, "The app won't start. What should I do?"

[1474] 2. The server analyzes the received message using a natural language processing module and identifies it as a "technical support inquiry."

[1475] 3. The server executes the following query on the FAQ database: SELECT solution FROM faq WHERE issue='App won't start';

[1476] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1477] 5. The server uses a generative AI to generate a response based on this information: "If the app does not start, please try clearing the cache or reinstalling the app. If the problem persists, please contact us again."

[1478] 6. The server sends the generated response to the user's terminal, which then displays it to the user.

[1479] Example of a prompt

[1480] The following are specific examples of prompt statements to be input into the generating AI model.

[1481] 1. A prompt to generate a response to the inquiry, "Do you have the latest smartphones in stock?"

[1482] Prompt: "Generate an appropriate response when a user asks, 'Do you have the latest smartphones in stock?'"

[1483] 2. A prompt to generate a response to a technical support inquiry: "The app won't start. What should I do?"

[1484] Prompt: "Provide a solution for when a user asks, 'The app won't start. What should I do?'"

[1485] This invention enables the automation of customer inquiry handling, which is expected to reduce labor costs, shorten response times, and improve customer satisfaction.

[1486] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1487] Step 1:

[1488] The user enters an inquiry.

[1489] Input: The user enters their inquiry into the input field on the terminal. For example, they might enter "Please tell me the stock status of product A."

[1490] Action: The user clicks the "Submit" button.

[1491] Output: The entered text is sent to the system, and the contents of the input field are cleared.

[1492] Step 2:

[1493] The device sends a message to the server.

[1494] Input: The inquiry message entered by the user.

[1495] Operation: The terminal generates an HTTP POST request and sends the message to the server. This request includes query text and metadata (e.g., timestamp, user ID).

[1496] Output: If the request is successful, the server receives a message.

[1497] Step 3:

[1498] The server receives the message

[1499] Input: HTTP POST request sent from the terminal.

[1500] Operation: The server receives requests via the receiving API and temporarily stores the message content. It also logs the request metadata.

[1501] Output: The temporarily stored message is passed to the natural language processing (NLP) module.

[1502] Step 4:

[1503] Perform Natural Language Processing (NLP)

[1504] Input: Query message temporarily stored on the server.

[1505] Operation: The server uses an NLP module to tokenize messages, tag them with parts of speech, perform grammatical analysis, extract propositions, and so on. In this process, it analyzes and identifies the intent of the message.

[1506] Output: Analyzed intent (e.g., "Check inventory") and message structure data.

[1507] Step 5:

[1508] Execute a database query

[1509] Input: Analyzed intent and database query format.

[1510] Operation: The server generates and executes appropriate SQL queries against the database based on a specific intent. For example, it might issue a query to check the inventory status of "product A".

[1511] Output: Results retrieved from the database (e.g., "In stock").

[1512] Step 6:

[1513] Generate a response

[1514] Input: Results retrieved from the database and prompts for the generated AI model.

[1515] Operation: The server uses generative AI to construct an appropriate response based on the retrieved database results. For example, it might generate the sentence, "Product A is in stock."

[1516] Output: The generated response message.

[1517] Step 7:

[1518] The server sends a response message.

[1519] Input: The generated response message and the user's terminal information.

[1520] Operation: The server generates an HTTP POST request and sends a response message to the user's terminal.

[1521] Output: The terminal receives a response message.

[1522] Step 8:

[1523] The device displays a response message.

[1524] Input: The response message sent from the server.

[1525] Action: The device updates the user interface (UI) and displays a response message on the screen. For example, it might display the message "Product A is in stock."

[1526] Output: A response message is displayed on the user's screen.

[1527] (Application Example 1)

[1528] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1529] The problem that this invention aims to solve is to establish a method for providing quick and accurate responses to user inquiries. In particular, it aims to provide a system that can automatically generate responses to a wide variety of user inquiries on e-commerce sites, such as product inventory checks and technical support, and provide quick answers. Conventional methods often resulted in delayed responses or required a large amount of human resources, which contributed to decreased customer satisfaction.

[1530] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1531] In this invention, the server includes means for receiving input messages, means for analyzing the received messages using natural language processing, means for identifying the intent of the message based on the analysis results, means for executing queries to a database according to the identified intent, means for checking the inventory status of a specific product and responding immediately to inquiries from users regarding inventory status, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to provide quick and accurate responses to user inquiries and improve customer satisfaction.

[1532] "Means for receiving input messages" refers to the means by which a user inputs a message through a terminal and receives it within the system.

[1533] "Means for analyzing received messages using natural language processing" refers to methods for analyzing received messages using natural language processing techniques to understand their content and intent.

[1534] "Means for identifying the intent of a message based on analysis results" refers to methods for identifying what a message means and what the user is looking for, based on the results of analysis using natural language processing.

[1535] "Means of querying a database according to a specified intent" refers to means of querying a database to retrieve necessary information based on a specified intent.

[1536] "A means of checking the inventory status of a specific product and responding immediately to user inquiries about inventory status" refers to a means of immediately checking the inventory status of a specific product in order to respond to user inquiries about inventory status.

[1537] "Means for generating appropriate responses using generative AI" refers to methods for automatically generating appropriate responses to user inquiries using generative AI technology.

[1538] "Means for sending the generated response to the user's terminal" refers to means of sending the generated response message to the user's terminal and providing feedback to the user.

[1539] The embodiments for carrying out the present invention are shown below. The present invention provides an automated inquiry response system, specifically intended for use on e-commerce websites. This system receives inquiry messages from users, analyzes them using natural language processing, retrieves information from a database based on the identified intent, and generates an appropriate response using generative AI.

[1540] System Overview

[1541] 1. User input

[1542] The user uses their smartphone to open the dedicated application for the online shopping site. They enter their inquiry in text format into the chat box. For example, they might type, "Do you still have any XX brand dress shirts in stock?"

[1543] 2. Receiving and parsing messages

[1544] The terminal sends the entered message to the server. The server receives the message and analyzes it using natural language processing. Specifically, processes such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction are performed.

[1545] 3. Identifying Intent and Data Acquisition

[1546] The system analyzes the message to identify the user's intent. Based on the identified intent, it queries the database to check, for example, the inventory status of a specific product. The server then passes the retrieved inventory information to the generative AI.

[1547] 4. Generating and sending responses

[1548] Generative AI generates appropriate responses based on the data it receives. For example, it might generate a message like, "We currently have XX brand dress shirts in stock. You can place an order."

[1549] 5. Display of response

[1550] The generated response message is sent back to the user's device and displayed in the chat box. This allows the user to receive a quick and accurate response.

[1551] Hardware and software

[1552] This system uses the following hardware and software.

[1553] Hardware:

[1554] Server: High-performance server (e.g., AWS, GCP)

[1555] User device: Smartphone (e.g., iPhone, Android device)

[1556] software:

[1557] Server side: Flask, SQLite (database)

[1558] Natural Language Processing: spaCy and NLTK

[1559] Generation AI:OpenAI GPT-3

[1560] Frontend: Smartphone applications (e.g., React Native)

[1561] Specific example

[1562] Message Processing Example

[1563] A user enters the message "Do you still have XX brand dress shirts in stock?" and sends it to the server. This message is received by the server and analyzed using natural language processing. From the analysis results, it is determined that the user's intention is to check the stock availability. The server queries the database to check the stock status of XX brand dress shirts. Based on the results, a generative AI generates a response message: "We currently have XX brand dress shirts in stock. You can place an order." This message is sent to the user's device and displayed in the chat box.

[1564] Example of prompts for the generating AI

[1565] If the product is in stock, how should I respond?

[1566] This system allows users to obtain information quickly and accurately, enabling effective responses to inquiries. This technology will greatly contribute to reducing labor costs and improving customer satisfaction.

[1567] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1568] Step 1:

[1569] The user uses their smartphone to open the e-commerce site's application and enters their inquiry as text into the chat box. For example, they might type, "Do you still have XX brand dress shirts in stock?" This message is then sent to the server as input.

[1570] Step 2:

[1571] The device sends the user's message to the server. Specifically, the device generates an API request and sends the message to the server via the receiving API. The input here is the message entered by the user, and the output is the message sent to the server.

[1572] Step 3:

[1573] The server temporarily stores the received message and passes it to the natural language processing module. The natural language processing module performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to analyze the meaning of the message. The input here is the message received in the previous step, and the output is the analysis result.

[1574] Step 4:

[1575] The server identifies the intent of the user's inquiry based on the results of natural language processing. For example, it might determine that this message is an inquiry about stock availability. At this stage, the input is the result of natural language processing, and the output is the identification of the intent, "stock availability check."

[1576] Step 5:

[1577] Based on the identified intent, the server executes a query to retrieve relevant data from the database. For example, "Execute an SQL query to check the inventory status of XX brand dress shirts. The input for this step is the identified intent, and the output is the result of the database query."

[1578] Step 6:

[1579] The server retrieves information from the database and passes it to a generative AI to generate an appropriate response. The generative AI is given prompts based on the retrieved information, and generates a response message such as, "We currently have XX brand dress shirts in stock. You can place an order." The input here is the inventory information retrieved from the database, and the output is the generated response message.

[1580] Step 7:

[1581] The server generates an API request to send the generated response message to the user's terminal and sends it to the user's terminal. The user's terminal displays the received response message in the chat box. The input here is the generated response message, and the output is the response message displayed on the user's terminal.

[1582] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1583] The following describes embodiments for carrying out the present invention. The present invention is an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and an emotion engine, generates an appropriate response, and sends it back to the user. The system operates based on a server, terminal, and user interface, and focuses particularly on analyzing the user's emotions.

[1584] System Overview

[1585] 1. Receive user inquiries.

[1586] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter a message such as "Please tell me the stock status of product A."

[1587] The device receives this message and generates an API request to send to the server.

[1588] 2. The server receives the message.

[1589] The server receives the user's message at the receiving API endpoint. This message is temporarily stored for parsing.

[1590] The server passes the received message to the natural language processing module.

[1591] 3. Natural language processing

[1592] The server uses a natural language processing (NLP) module to analyze the message content. This analysis includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1593] 4. Perform emotion analysis

[1594] The server uses an emotion engine to analyze the user's emotions within a message. For example, if a message like "Do you have it in stock?" is determined to contain a sense of urgency, it will assign emotion tags such as positive, negative, or neutral.

[1595] 5. Identifying the Intent

[1596] The server identifies the intent of the message based on the NLP analysis results and the emotion engine results. For example, it might be identified as an "inquiry about stock availability."

[1597] 6. Executing database queries

[1598] The server executes queries to retrieve relevant information from the database based on a specific intention. For example, it generates an SQL query to check the "stock status of product A" and queries the database.

[1599] The server passes the data to a generative AI to generate a response based on the information it has acquired.

[1600] 7. Response generation

[1601] The server uses generative AI to automatically construct appropriate responses to user inquiries. In doing so, it considers sentiment tags and adjusts the tone and content of the response accordingly. For example, if negative sentiment is detected, the response message will emphasize consideration and solutions.

[1602] The server generates a message saying, "Smartphone X is currently in stock. You can order it now." However, if the user is feeling anxious, a more considerate message such as, "Don't worry, Smartphone X is in stock. You can order it right away," will be generated.

[1603] 8. Sending a response to the user

[1604] The system generates and sends an API request to send the response message generated by the server to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[1605] 9. Displaying user responses

[1606] The device receives a response message from the server and displays it to the user. For example, it might display, "Smartphone X is currently in stock. You can place an order." However, it can also display messages that are considerate of the user's feelings.

[1607] Specific example

[1608] Example 1: Checking product inventory

[1609] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[1610] 2. The device sends a message to the server.

[1611] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[1612] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[1613] 5. The server receives the result "In stock".

[1614] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[1615] 7. Display the response received by the terminal to the user.

[1616] Example 2: Technical support inquiry

[1617] 1. The user sends the message, "The app won't start. What should I do?"

[1618] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[1619] 3. The server queries the FAQ database for relevant information.

[1620] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1621] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[1622] 6. Display the responses received by the device to the user.

[1623] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[1624] The following describes the processing flow.

[1625] Step 1:

[1626] The user uses a device (e.g., smartphone, PC) to enter their inquiry in text format. For example, they might enter, "Do you have the latest smartphone X in stock?"

[1627] Step 2:

[1628] The terminal receives the user's input message and generates an API request to send that message to the server. The generated request is sent to the server using the HTTP protocol or similar.

[1629] Step 3:

[1630] The server receives user messages at the receiving API endpoint. The received messages are temporarily stored in the database or memory and used for subsequent processing.

[1631] Step 4:

[1632] The server invokes a natural language processing (NLP) module to parse the received message. This parsing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1633] Step 5:

[1634] The server identifies the intent of the message based on the results of natural language processing. For example, it might be identified as an "inventory check inquiry." This allows the system to understand that the user is seeking inventory information.

[1635] Step 6:

[1636] The server calls an emotion analysis module to analyze the user's emotions within a message. Emotion analysis uses word-emotion dictionaries and machine learning models. For example, the message "Do you have this in stock?" might be judged to contain a sense of urgency.

[1637] Step 7:

[1638] The server assigns emotion tags to user messages based on the emotion analysis results. For example, the emotion tag "anxiety" might be assigned to a message.

[1639] Step 8:

[1640] Based on the identified intent and sentiment tags, the server generates queries to retrieve relevant information from the database. For example, it generates an SQL query such as "SELECT stock_status FROM products WHERE product_name='Smartphone X';".

[1641] Step 9:

[1642] The server executes the generated SQL query on the database and retrieves the relevant data. For example, it might return a result such as "Smartphone X is in stock."

[1643] Step 10:

[1644] Based on the data acquired by the server, a generative AI is used to generate an appropriate response. The response generation incorporates an algorithm that adjusts the tone and content considering emotion tags. For example, if a negative emotion is detected, the response message will emphasize consideration and solutions.

[1645] Step 11:

[1646] The server generates a message saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately." However, if the user is feeling anxious, a more considerate message such as "Please rest assured, Smartphone X is in stock. You can order it immediately." will be generated.

[1647] Step 12:

[1648] The server generates and sends an API request to send the generated response message to the user's terminal. This request is sent to the user's terminal using the HTTP protocol or similar.

[1649] Step 13:

[1650] The device receives a response message from the server and displays it to the user. For example, a message such as "Smartphone X is currently in stock. Please rest assured, you can order it immediately." might be displayed.

[1651] (Example 2)

[1652] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1653] Traditional call center systems have struggled to provide efficient and effective responses to user inquiries, particularly in generating responses that take user emotions into consideration. Furthermore, their reliance on human resources has led to increased response times and rising labor costs. To address these challenges, a system is needed that accurately analyzes message content and automatically generates responses that consider user emotions.

[1654] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1655] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results and sentiment analysis, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, and means for sending the generated response to the user's terminal. This makes it possible to automatically generate a response that takes the user's emotions into consideration.

[1656] "Means for receiving input messages" refers to the communication methods and systems that allow a server to receive messages entered by a user using a terminal.

[1657] "Methods of analysis using natural language processing" refers to methods of analyzing human language using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction of input messages.

[1658] "Means of identifying the intent of a message" refers to methods of clarifying what the user is seeking in their inquiry, based on the results of natural language processing analysis and sentiment analysis.

[1659] "Means of querying a database" refers to a method of executing query commands to retrieve necessary information from a database based on a specified intent.

[1660] "Means of generating appropriate responses using generative AI" refers to a method of constructing responses to user inquiries based on information obtained from a database and sentiment tags, using generative AI models (e.g., GPT-3, BERT).

[1661] "Means of sending the generated response to the user's terminal" refers to the method of sending the response generated on the server to the user's terminal using a communication method or protocol (e.g., HTTP).

[1662] This invention relates to an automated call center system that receives inquiry messages from users, analyzes them using natural language processing and sentiment analysis engines, generates appropriate responses, and sends them back to the users. This system operates based on interfaces from servers, terminals, and users, and places particular emphasis on analyzing user sentiment.

[1663] The server first receives a user inquiry message. The user enters the inquiry in text format using a device (e.g., smartphone, PC). For example, the user might enter the message, "Please tell me the stock status of product A." The device receives this message and generates an API request to send to the server. The server receives this message at the specified receiving API endpoint (e.g., / receive-message) and stores it temporarily.

[1664] Next, the server passes the received message to a natural language processing (NLP) module. The NLP module analyzes the message content using techniques such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Specifically, it uses an NLP library (e.g., spaCy or NLTK). For example, if the message is "Do you have the latest smartphone X in stock?", part-of-speech tagging would recognize "latest" as an adjective, "smartphone X" as a noun, "stock" as a noun, and "do you have" as a verb.

[1665] Furthermore, the server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the user's emotions within the message. The server analyzes the user's message and, if it detects negative emotions such as impatience, adds corresponding sentiment tags. For example, if the message "Do you have this in stock?" is determined to contain impatience, it will be tagged with positive, negative, or neutral sentiment tags.

[1666] Next, the server identifies the intent of the message based on the NLP analysis results and sentiment analysis results. For example, it can be identified as an "inquiry about stock availability." Based on this intent, the server executes a query to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query (e.g., SELECT stock_status FROM products WHERE product_name='Smartphone X';) and queries the database.

[1667] The server passes the acquired information to a generative AI model (e.g., GPT-3, BERT) to generate a response. The generative AI model generates an appropriate response based on the information and sentiment tags acquired from the database. In this process, it adjusts the tone and content of the response, taking sentiment tags into consideration. For example, based on the result "In stock," it generates the message, "Rest assured, smartphone X is in stock. You can order it immediately."

[1668] Finally, the server generates and sends an API request to send the generated response message to the user's device. This request is sent to the user's device using the HTTP protocol or similar. The device displays the received response message to the user, conveying a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[1669] Specific example

[1670] Example 1: Checking product inventory

[1671] 1. The user enters "Do you have the latest smartphone X in stock?" into the device.

[1672] 2. The device sends a message to the server.

[1673] 3. The server receives the message and performs natural language processing and sentiment analysis. As a result, it determines that the user is in a hurry.

[1674] 4. The server executes the following query on the product inventory database: SELECT stock_status FROM products WHERE product_name='Smartphone X';

[1675] 5. The server receives the result "In stock".

[1676] 6. The server generates a response saying, "Smartphone X is currently in stock. Please rest assured, you can order it immediately," and sends it to the user's terminal.

[1677] 7. Display the response received by the terminal to the user.

[1678] Example 2: Technical support inquiry

[1679] 1. The user sends the message, "The app won't start. What should I do?"

[1680] 2. The server receives the message and performs natural language processing and sentiment analysis. User confusion and anxiety are detected from the message.

[1681] 3. The server queries the FAQ database for relevant information.

[1682] 4. The server receives the message, "If the app does not start, please clear the cache or reinstall the app."

[1683] 5. The server generates a thoughtful response, such as "If the app doesn't start, please try clearing your cache or reinstalling the app. If the problem persists, please contact us again," and sends it to the user's device.

[1684] 6. Display the responses received by the device to the user.

[1685] Examples of input prompts for a generative AI model

[1686] "Generate a response based on the user's inquiry, 'Do you have the latest smartphone X in stock?' Assume the user is in a hurry."

[1687] "Based on the user inquiry, 'The app won't start. What should I do?', please generate an appropriate response for the confused user."

[1688] This invention enables automated inquiry handling that combines natural dialogue with consideration for emotions, resulting in reduced labor costs, shorter response times, and improved customer satisfaction.

[1689] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1690] Step 1: Receive user inquiry

[1691] The user enters their inquiry in text format using a device (e.g., smartphone, PC). An example of the entered message is "Please tell me the stock status of product A." The device receives this message and generates an API request (e.g., an HTTP POST request) to send to the server. The input is the user's message, and the output is the API request sent to the server.

[1692] Step 2: The server receives the message.

[1693] The server receives user messages at the specified receiving API endpoint (e.g., / receive-message). The server temporarily stores the received messages in an in-memory data store such as Redis and manages the order of processing. At this point, the input is the API request from the terminal, and the output is the temporarily stored messages.

[1694] Step 3: Process natural language

[1695] The server uses an NLP module (e.g., spaCy or NLTK) to tokenize the received message. The server then performs part-of-speech tagging and extracts propositions through grammatical analysis. Specifically, it might split a message like "Do you have the latest smartphone X in stock?" into tokens and assign a part-of-speech tag to each token. The input is the temporarily stored message, and the output is the tokenized data and part-of-speech tagging information as a result of the analysis.

[1696] Step 4: Perform sentiment analysis

[1697] The server uses a sentiment analysis engine (e.g., IBM Watson, Azure Text Analytics) to analyze the sentiment of a message based on the NLP analysis results. The server analyzes the user's message, and if it detects negative emotions, such as anxiety, it adds the corresponding sentiment tag. The input is tokenized and part-of-speech tagged data, and the output is sentiment-tagged data.

[1698] Step 5: Identifying the Intent

[1699] The server integrates the NLP analysis results and sentiment analysis results to identify the intent of the message. For example, it might identify it as an "inquiry about stock availability." Specifically, the server prepares the data necessary for the next processing based on the identified intent. The input is analyzed data with sentiment tags, and the output is data containing the identified intent.

[1700] Step 6: Execute the database query

[1701] Based on the identified intent, the server issues queries to a database (e.g., MySQL or PostgreSQL) to retrieve the necessary information. Specifically, it generates an SQL query such as `SELECT stock_status FROM products WHERE product_name='smartphone X';` and queries the database. The input is data containing the identified intent, and the output is information retrieved from the database (e.g., "In stock").

[1702] Step 7: Generating the response

[1703] The server inputs data into a generative AI model (e.g., GPT-3, BERT) to generate appropriate responses to user inquiries. The tone and content of the responses are adjusted considering sentiment tags. For example, it might generate the message, "Rest assured, smartphone X is in stock. You can order it immediately." The input consists of information retrieved from a database and data with sentiment tags, while the output is the generated response.

[1704] Step 8: Sending a response to the user

[1705] This process generates and sends an API request to the user's terminal to send the response message generated by the server. This request is made using the HTTP protocol, etc. The input is the generated response message, and the output is the API request sent to the terminal.

[1706] Step 9: User response display

[1707] The terminal receives a response message sent from the server and displays it to the user. Specifically, it displays "Rest assured, smartphone X is in stock. You can order it immediately." The input is the response message sent from the server, and the output is the text displayed to the user.

[1708] These steps enable the system to provide timely and appropriate responses to user inquiries, while also considering their emotional impact.

[1709] (Application Example 2)

[1710] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1711] Traditional automated response systems often fail to consider user emotions, potentially leading to decreased user satisfaction. This is especially true for e-commerce sites, where prompt and appropriate responses to user inquiries are crucial; responses that disregard user feelings diminish credibility. Furthermore, traditional manual responses to a large volume of inquiries are inefficient and costly, necessitating automation.

[1712] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1713] In this invention, the server includes means for receiving an input message, means for analyzing the received message using natural language processing, means for identifying the intent of the message based on the analysis results, means for querying a database according to the identified intent, means for generating an appropriate response using generative AI, means for analyzing the user's emotions, means for generating a response that takes the user's emotions into consideration based on the analysis results, and means for sending the generated response to the user's terminal. This makes it possible to provide an appropriate and prompt response that takes the user's emotions into consideration, thereby improving user satisfaction and reducing costs.

[1714] "Input messages" refer to text-based inquiries or requests that users send to the system through their devices.

[1715] "Means of receiving" refers to the function on the server side that receives messages sent by the user.

[1716] "Natural language processing" is a general term for technologies that analyze input messages and understand their content and structure.

[1717] "Means of analysis" refers to the function used to analyze received messages using natural language processing.

[1718] "Message intent" refers to the user's purpose or request, extracted from the content of the message as analyzed by natural language processing.

[1719] "Means of identification" refers to a function for identifying the intent of a message based on the analysis results.

[1720] "Means of querying a database" refers to means of retrieving necessary information from a database according to a specified intent.

[1721] "Generative AI" refers to artificial intelligence technology aimed at generation, specifically a system that automatically constructs responses to user inquiries based on information and analysis results obtained from a database.

[1722] "Means for generating appropriate responses" refers to a function that uses generative AI to generate appropriate responses to user inquiries.

[1723] "User sentiment" refers to the emotional nuances and tone that a user conveys when typing a message.

[1724] "Means of analysis (emotions)" refers to the function for analyzing the user's emotions within a message.

[1725] "Emotionally sensitive responses" refer to responses that are tailored to the user's emotions, based on the results of an emotional analysis of the user, and using appropriate tone and content.

[1726] "Means of generating (emotions)" refers to the function that generates responses that take into account the user's emotions, considering the results of emotion analysis.

[1727] "Means of sending to the user's terminal" refers to the function for sending the generated response from the server to the user's terminal.

[1728] System Overview

[1729] This invention is an automated call center system that automatically processes user inquiries and generates appropriate responses based on sentiment analysis. The system operates based on a server, terminals, and a user interface, and focuses particularly on analyzing the user's sentiment.

[1730] Hardware and software configuration

[1731] Hardware:

[1732] User devices: smartphones, PCs, etc.

[1733] Server: Equipped with a high-performance CPU, sufficient memory, and disk space.

[1734] software:

[1735] Natural language processing libraries: spaCy, NLTK, etc.

[1736] Sentiment analysis tools: TensorFlow-based sentiment analysis models, etc.

[1737] Server environment: Web framework such as Flask

[1738] Processing steps

[1739] 1. Receiving a message

[1740] Users submit inquiries in text format using devices such as smartphones and PCs. These messages are sent to the server via HTTP requests.

[1741] 2. Natural Language Processing and Sentiment Analysis

[1742] The server passes the received message to a natural language processing module for analysis such as tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction. Simultaneously, it uses an emotion engine to analyze the user's emotions. Sentiment analysis determines whether the message has multiple emotion tags, such as positive, negative, or anxiety.

[1743] 3. Identifying the Intent and Database Queries

[1744] Based on the analysis results, the server identifies the intent of the message. For example, it might categorize it as "inventory check" or "technical support." Based on the specific intent, the server queries the database to retrieve the necessary information.

[1745] 4. Generating a response

[1746] Based on the acquired information and sentiment tags, a generative AI is used to automatically generate a response message to the user. In this process, sentiment tags are taken into consideration to generate a considerate response.

[1747] 5. Sending and displaying responses

[1748] The server sends the generated response message to the user's terminal, and the terminal displays this message to the user.

[1749] Specific examples

[1750] Example of checking inventory:

[1751] User: "Do you have the latest smartphone X in stock?"

[1752] System response: "Rest assured, the Smartphone X is in stock. You can order it immediately."

[1753] Examples of technical support:

[1754] User: "The app won't start. What should I do?"

[1755] System response: "If the app won't launch, try clearing your cache or reinstalling the app. If the problem persists, please contact us again."

[1756] Example of a prompt

[1757] User query: "Do you have the latest smartphone X in stock?"

[1758] NLP analysis results: "Inventory check" "Smartphone X"

[1759] Emotion analysis result: "Anxiety"

[1760] Generate API prompt: Generate a response that is considerate of a user who is anxious about the availability of smartphone X.

[1761] Thus, the present invention makes it possible to achieve appropriate and prompt responses that take into account the user's feelings, thereby improving user satisfaction and reducing costs.

[1762] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1763] Step 1:

[1764] The user uses their device to enter their inquiry in text format. For example, they might enter the message, "Do you have the latest smartphone X in stock?" The entered message is received by the device's application and sent to the server in the form of an API request.

[1765] Input: User's text-based inquiry message

[1766] Output: Message in API request format

[1767] Step 2:

[1768] The server receives the message at the API endpoint. The received message is temporarily stored and then passed to the natural language processing (NLP) module.

[1769] Input: Message in API request format

[1770] Output: Message data to be passed to the NLP module

[1771] Step 3:

[1772] The server's natural language processing module analyzes the message. It performs tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction to understand the message's content.

[1773] Input: Temporarily saved message data

[1774] Output: Analysis results (tokens, part-of-speech tagging, grammatical analysis results, proposition extraction results)

[1775] Step 4:

[1776] The server's emotion engine analyzes the user's emotions from the analyzed messages. For example, it assigns emotion tags such as positive, negative, and anxiety.

[1777] Input: NLP analysis results

[1778] Output: Sentiment analysis results (emotion tags)

[1779] Step 5:

[1780] The server identifies the intent of the message based on NLP analysis results and sentiment analysis results. For example, it might identify it as an "inquiry about stock availability."

[1781] Input: NLP analysis results and emotion analysis results

[1782] Output: Message intent (e.g., Inventory check)

[1783] Step 6:

[1784] Depending on the identified intent, the server executes SQL queries to retrieve relevant information from the database. For example, it might generate an SQL query to check the "stock status of product A" and then query the database.

[1785] Input: Message intent

[1786] Output: Database query results (e.g., inventory status)

[1787] Step 7:

[1788] The server uses generative AI to generate appropriate responses based on information and emotion tags retrieved from the database. For example, based on a database result such as "In stock" and the emotion tag "Anxiety," it generates a response such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[1789] Input: Database query results and sentiment tags

[1790] Output: Generated response message

[1791] Step 8:

[1792] Generate and send an API request to send the response message generated by the server to the user's terminal.

[1793] Input: Generated response message

[1794] Output: Response message in API request format

[1795] Step 9:

[1796] The terminal receives a response message from the server and displays it to the user. For example, it might display a message such as, "Rest assured, smartphone X is in stock. You can order it immediately."

[1797] Input: Response message in API request format

[1798] Output: Response message displayed to the user

[1799] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1800] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1801] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1802] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1803] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1804] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1805] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1806] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1807] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1808] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1809] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1810] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1811] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1812] 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.

[1813] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1814] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1815] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1816] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1817] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1818] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1819] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[1820] The following is further disclosed regarding the embodiments described above.

[1821] (Claim 1)

[1822] A means for receiving input messages,

[1823] A means of analyzing received messages using natural language processing,

[1824] A means of identifying the intent of a message based on the analysis results,

[1825] A means of querying the database according to the identified intent,

[1826] A means of generating an appropriate response using a generative AI,

[1827] A means for sending the generated response to the user's terminal,

[1828] A system that includes this.

[1829] (Claim 2)

[1830] The system according to claim 1, wherein the means for natural language processing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1831] (Claim 3)

[1832] The system according to claim 1, wherein the generation AI generates a response based on information obtained from a database.

[1833] "Example 1"

[1834] (Claim 1)

[1835] A means of receiving messages entered by a user using a terminal,

[1836] A means of sending received messages to the server,

[1837] A means for analyzing messages received by a server using natural language processing,

[1838] A means of identifying the intent of a message based on the analysis results,

[1839] A means of querying the database according to the identified intent,

[1840] A means of generating an appropriate response using generative AI based on information obtained from a database,

[1841] A means for sending the generated response to the user's terminal,

[1842] A means for displaying the response message received by the user's terminal,

[1843] A system that includes this.

[1844] (Claim 2)

[1845] The system according to claim 1, wherein the means for natural language processing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1846] (Claim 3)

[1847] The system according to claim 1, wherein the generation AI generates a response based on information obtained from a database.

[1848] "Application Example 1"

[1849] (Claim 1)

[1850] A means for receiving input messages,

[1851] A means of analyzing received messages using natural language processing,

[1852] A means of identifying the intent of a message based on the analysis results,

[1853] A means of querying the database according to the identified intent,

[1854] A means to check the inventory status of specific products and respond immediately to inquiries from users regarding inventory availability,

[1855] A means of generating an appropriate response using a generative AI,

[1856] A means for sending the generated response to the user's terminal,

[1857] A system that includes this.

[1858] (Claim 2)

[1859] The system according to claim 1, wherein the means for natural language processing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1860] (Claim 3)

[1861] The system according to claim 1, wherein the generation AI generates a response based on information obtained from a database.

[1862] "Example 2 of combining an emotion engine"

[1863] (Claim 1)

[1864] A means for receiving input messages,

[1865] A means of analyzing received messages using natural language processing,

[1866] A means of identifying the intent of a message based on analysis results and sentiment analysis,

[1867] A means of querying the database according to the identified intent,

[1868] A means of generating an appropriate response using a generative AI,

[1869] A means for sending the generated response to the user's terminal,

[1870] A system that includes this.

[1871] (Claim 2)

[1872] The system according to claim 1, wherein the natural language processing means includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction, and performs sentiment tagging based on the analysis results.

[1873] (Claim 3)

[1874] The system according to claim 1, wherein the generative AI generates a response based on information and emotion tags obtained from a database.

[1875] "Application example 2 when combining with an emotional engine"

[1876] (Claim 1)

[1877] A means for receiving input messages,

[1878] A means of analyzing received messages using natural language processing,

[1879] A means of identifying the intent of a message based on the analysis results,

[1880] A means of querying the database according to the identified intent,

[1881] A means of generating an appropriate response using a generative AI,

[1882] A means of analyzing user emotions,

[1883] A means for generating a response that takes into account the user's emotions based on the analysis results,

[1884] A means for sending the generated response to the user's terminal,

[1885] A system that includes this.

[1886] (Claim 2)

[1887] The system according to claim 1, wherein the means for natural language processing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

[1888] (Claim 3)

[1889] The system according to claim 1, wherein the generative AI generates a response based on information obtained from a database and the result of sentiment analysis. [Explanation of Symbols]

[1890] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for receiving input messages, A means of analyzing received messages using natural language processing, A means of identifying the intent of a message based on the analysis results, A means of querying the database according to the identified intent, A means of generating an appropriate response using a generative AI, A means for sending the generated response to the user's terminal, A system that includes this.

2. The system according to claim 1, wherein the means for natural language processing includes tokenization, part-of-speech tagging, grammatical analysis, and proposition extraction.

3. The system according to claim 1, wherein the generation AI generates a response based on information obtained from a database.

Images