system

Abstract

We provide the system. [Solution] Means for receiving inquiry data, A natural language processing means for analyzing the content of the aforementioned query data, A means for classifying the content of inquiries based on the aforementioned analysis results, A means of generating responses to inquiries using generative artificial intelligence, Means for transmitting the generated response, A means of transferring the inquiry to a human operator based on the aforementioned analysis results, 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 in response to the user utterance.

Prior Art Document

Patent Document

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a conventional customer support system, it is difficult to respond quickly and accurately to a large number of inquiries, and particularly, a great deal of time and resources are required for responding to common questions and claims. As a result, there have been problems of a decrease in customer satisfaction and an increase in operation costs. The present invention aims to realize automatic classification and response of inquiries by utilizing generative artificial intelligence and improve the efficiency of customer support operations.

Means for Solving the Problems

[0005] The present invention is a system that includes means for receiving inquiry data, means for natural language processing to analyze the content of the inquiry data, means for classifying the content of the inquiry based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, and means for transferring the inquiry to a human operator based on the analysis results. Furthermore, the system further includes means for preprocessing the inquiry data, means for performing natural language processing using the preprocessed data, and means for storing the inquiry data and the generated response in a database, thereby further improving the efficiency of customer support.

[0006] "Inquiry data" refers to information, including questions and requests, that users send to customer support.

[0007] "Means of receiving" refers to the functions and devices that receive inquiry data via email, web forms, or chat systems.

[0008] "Natural language processing means" refers to technologies and algorithms for analyzing the content of query data and understanding its meaning and intent.

[0009] "Analysis results" refer to information such as the intent of the inquiry and important keywords obtained through natural language processing.

[0010] "Means for classifying inquiry content" refers to techniques and algorithms for preprocessing inquiry data and identifying the type of inquiry based on the analysis results.

[0011] "Generative artificial intelligence" refers to an artificial intelligence model that automatically generates appropriate answers to user inquiries.

[0012] "Means for generating answers" refers to functions and devices that use generative artificial intelligence to create appropriate answers.

[0013] "Means of sending responses" refers to functions and devices that send generated responses to users via email, chat, or other means.

[0014] "Means of handing over to human operators" refers to functions and devices that appropriately assign inquiries that are difficult for generative artificial intelligence to handle to human operators.

[0015] "Preprocessing means" refers to functions and devices that remove unnecessary information from query data and convert it into a format that is easy to analyze.

[0016] "Means of saving to a database" refers to a storage device and its operation method for maintaining a history of queries and generated responses. [Brief explanation of the drawing]

[0017] [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]Shows an emotion map to which a plurality of emotions are mapped. [Figure 10] Shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Mode for Carrying Out the Invention

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

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

[0020] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be one arithmetic unit or a combination of a plurality of arithmetic units. Also, the processor may be one type of arithmetic unit or a combination of a plurality of 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.

[0021] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

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

[0023] 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).

[0024] 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."

[0025] [First Embodiment]

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

[0027] 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.

[0028] 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).

[0029] 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.

[0030] 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.

[0031] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form 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.

[0032] 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.

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

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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".

[0038] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence. The system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The necessary embodiments for carrying out this invention are described in detail below.

[0039] 1. Receiving inquiry data

[0040] When a user submits an inquiry via a web form, email, or chat system,

[0041] The server receives this.

[0042] The data received includes inquiry details and user contact information.

[0043] 2. Analysis and classification of inquiry content

[0044] The server preprocesses the query data it receives.

[0045] During preprocessing, unnecessary information (such as HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0046] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[0047] NLP models are implemented using libraries such as spaCy or transformers.

[0048] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[0049] The server classifies the query content based on the analysis results.

[0050] For example, these can be categorized into general questions, requests for technical support, and emotional complaints.

[0051] 3. Automated responses using generative AI

[0052] Extract queries that the server can automatically respond to.

[0053] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4®) and has it generate a response.

[0054] The generative artificial intelligence generates the answer and returns it to the server.

[0055] The server sends the generated response to the user.

[0056] For example, in response to an inquiry such as "Please tell me more about the data plans," the response would be in the format of "The following types of data plans are currently available..."

[0057] The server saves the query content and the generated response to the database.

[0058] 4. Handover to a human operator

[0059] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[0060] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0061] In this process, the server provides the operator with the necessary inquiry history and related information.

[0062] We will provide appropriate responses to inquiries that have been handed over to human operators.

[0063] As an example, let's consider the processing flow when a user inquires, "Please tell me about the latest mobile phone plans."

[0064] 1. The user submits an inquiry via a web form.

[0065] 2. The server receives this and performs preprocessing.

[0066] 3. The server analyzes the inquiry using an NLP model and classifies it as "capable of automated response."

[0067] 4. The server sends the question to the generative AI and instructs it to generate an answer.

[0068] 5. The generative artificial intelligence generates the answer and returns it to the server.

[0069] 6. The server sends the response to the user.

[0070] 7. The server saves all interactions to the database.

[0071] This allows for quick and accurate responses to user inquiries, as well as the ability to flexibly transfer them to human operators when necessary. This system is expected to improve the efficiency of customer support operations and enhance customer satisfaction.

[0072] The following describes the processing flow.

[0073] Step 1:

[0074] A user submits an inquiry. Inquiries may be submitted via a web form, email, or chat system.

[0075] Step 2:

[0076] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information (email address and phone number).

[0077] Step 3:

[0078] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0079] Step 4:

[0080] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[0081] Step 5:

[0082] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[0083] Step 6:

[0084] The server classifies the inquiry based on the analysis results. Classification categories include general questions, requests for technical support, and emotional complaints.

[0085] Step 7:

[0086] The server extracts queries that it has classified as "capable of automatic response."

[0087] Step 8:

[0088] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[0089] Step 9:

[0090] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[0091] Step 10:

[0092] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[0093] Step 11:

[0094] The server saves all interactions to a database. This database includes query details, generated responses, user information, and timestamps.

[0095] Step 12:

[0096] The server extracts inquiries that it has classified as requiring human intervention.

[0097] Step 13:

[0098] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0099] Step 14:

[0100] The server provides the operator with the inquiry history and related information.

[0101] Step 15:

[0102] For inquiries taken over by human operators, specific actions will be taken.

[0103] These steps streamline customer support operations by automatically classifying and processing inquiries, and only handing them over to human operators when necessary.

[0104] (Example 1)

[0105] 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."

[0106] Conventional customer support systems have often failed to adequately streamline inquiry processing, leading to decreased customer satisfaction. Furthermore, automated response systems struggled to handle complex inquiries, placing a burden on human operators. This invention aims to solve these problems by providing an efficient inquiry processing system utilizing generative artificial intelligence.

[0107] 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.

[0108] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for natural language processing to analyze the content of the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for storing the inquiry data and the generated response in a database, means for transferring the inquiry to a human operator based on the analysis results, means for sending a notification to an internal operation system, and means for removing unnecessary information in order to convert the inquiry content into an analyzable format. This enables increased efficiency in customer support operations and improved customer satisfaction.

[0109] "Inquiry data" refers to information that users send to the customer support system, including questions, requests, and complaints.

[0110] "Preprocessing" is the process of removing unnecessary information in order to convert query data into a parseable format.

[0111] "Natural language processing (NLP) techniques" refer to technologies for analyzing received query data and understanding its content, and include text analysis and keyword extraction.

[0112] A "classification method" is a process that classifies query content into a specific category based on the results of analysis using natural language processing methods.

[0113] "Generative artificial intelligence (AI)" is an artificial intelligence technology that automatically generates responses based on the content of inquiries.

[0114] "Methods for generating answers" refers to the process of generating appropriate answers to inquiries using generative artificial intelligence.

[0115] "Means of submitting responses" refers to the process of sending the generated responses to the user.

[0116] A "database" is a digital storage system for storing query data and generated responses.

[0117] "Methods for transferring inquiries" refers to the process of transferring inquiries that are difficult for generative artificial intelligence to handle to human operators.

[0118] "Methods for sending notifications" refers to the process of notifying human operators of important inquiries or inquiries that are difficult for generative artificial intelligence to handle.

[0119] "Converting to a parsable format" means transforming query data into a form that is easy to process, and this includes removing unnecessary information.

[0120] This invention is a system aimed at improving the efficiency of inquiry processing and customer satisfaction in customer support systems. Specific embodiments are described in detail below.

[0121] System Configuration

[0122] This system is primarily composed of the following hardware and software.

[0123] Server: Receives, processes, analyzes, generates responses, stores data, and sends notifications.

[0124] Device: The device that the user uses to submit an inquiry (e.g., a personal computer or smartphone).

[0125] software:

[0126] Natural Language Processing (NLP) libraries (e.g., spaCy, transformers)

[0127] Generative artificial intelligence (AI) models (e.g., GPT-4)

[0128] Database management systems (e.g., MySQL (registered trademark), PostgreSQL)

[0129] Description of specific embodiments

[0130] Receiving inquiry data

[0131] The user submits an inquiry using a device via a web form, email, or chat system. During this process, data containing the inquiry details and the user's contact information is generated. The server receives this data.

[0132] Pre-processing of inquiry content

[0133] The query data received by the server is preprocessed. Specifically, the software's regular expression library and string manipulation functions are used to remove unnecessary information (e.g., HTML tags, special characters, unnecessary line breaks and spaces) and convert it into a parseable format.

[0134] Analysis and classification of inquiry content

[0135] The server passes pre-processed query data to a natural language processing (NLP) model (e.g., spaCy, transformers) for analysis. The NLP model extracts the intent and key keywords of the query. Based on this analysis, the server classifies the query into a specific category (e.g., general question, request for technical support, emotional complaint, etc.).

[0136] Generation of automated responses using generative AI

[0137] The server extracts queries that can be answered automatically. If an automatic response is applicable, the server inputs the query details into a generative artificial intelligence (e.g., GPT-4) to create an appropriate prompt. For example, it might use a prompt like this:

[0138] "Please tell me about the latest mobile phone plans."

[0139] The server inputs this prompt into a generative AI model, which then generates a response. The generative artificial intelligence generates an appropriate response and returns it to the server.

[0140] Submitting responses and saving data

[0141] The server sends the generated response to the user. An example of the response might be, "The following data plans are currently available to you..." The server also saves the inquiry and the generated response to a database. The information saved includes the inquiry, analysis results, generated response, user contact information, date and time, etc.

[0142] Handover to human operator

[0143] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends a notification to the internal operation system to transfer these inquiries to human operators. The server provides the operators with the necessary inquiry history and related information. Human operators then take appropriate action on the inquiries they receive.

[0144] The above describes a specific embodiment of the present invention. This system enables increased efficiency in customer support operations and improved customer satisfaction.

[0145] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0146] Step 1:

[0147] The user submits an inquiry. The user enters and submits their inquiry and contact information using a web form, email, or chat system.

[0148] Input: Inquiry details, user contact information

[0149] Output: Query data received by the server

[0150] Step 2:

[0151] The server receives the query data. The server receives the query data sent by the user and saves it as a file.

[0152] Input: Inquiry data submitted by the user

[0153] Output: Received query data

[0154] Step 3:

[0155] The server preprocesses the query data. To convert the received query data into a parseable format, it removes unnecessary information (HTML tags, special characters, etc.) and converts it into text format.

[0156] Input: Received query data

[0157] Output: Preprocessed query data

[0158] Specific operation: Uses a regular expression library to remove HTML tags and special characters.

[0159] Step 4:

[0160] The server passes the pre-processed query data to a natural language processing (NLP) model for analysis. The NLP model (e.g., spaCy, transformers) is used to extract the intent and important keywords of the query.

[0161] Input: Preprocessed query data

[0162] Output: Analysis results (intent, important keywords)

[0163] Specific operation: Input data into the NLP model and obtain the analysis results.

[0164] Step 5:

[0165] The server classifies the inquiry based on the analysis results. The analyzed inquiry data is categorized into specific categories (e.g., general questions, technical support, complaints, etc.).

[0166] Input: Analysis results

[0167] Output: Classified query data

[0168] Specific operation: Based on the analysis results, conditional branching is performed, and data is distributed to each category.

[0169] Step 6:

[0170] The server inputs a prompt message to a generative artificial intelligence (AI) and generates an automated response. If applicable, the generative AI (e.g., GPT-4) creates a prompt message based on the analyzed query content and generates a response.

[0171] Input: Classified query data, prompt text

[0172] Output: Auto-generated answer

[0173] Specific operation: Input a prompt sentence to a generative AI model and generate an appropriate response.

[0174] Step 7:

[0175] The server sends the generated response to the user. The generated response is also sent to the user's contacts (e.g., email, chat).

[0176] Input: Automated response, user contact information

[0177] Output: Response sent to the user

[0178] Specific actions: Send responses using an email system or chat system.

[0179] Step 8:

[0180] The server saves the query content and the generated response to the database. All query content and its response are recorded in the database and can be referenced later.

[0181] Input: Inquiry details, analysis results, generated response

[0182] Output: Data stored in the database

[0183] Specific operation: Use a database management system to save data.

[0184] Step 9:

[0185] The server transfers inquiries it cannot handle to human operators. If the generative AI cannot process the inquiry, it sends a notification to the internal operations system.

[0186] Input: Classified query data

[0187] Output: Inquiry details handed over to the operator, notification sent.

[0188] Specific action: Send a notification to the internal operating system and display it on the operator's terminal.

[0189] (Application Example 1)

[0190] 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."

[0191] Current customer support systems fail to adequately optimize resources and improve customer satisfaction. In particular, there are problems with delayed inquiry processing and inability to provide appropriate responses. Furthermore, the limited availability of automated responses to inquiries places a heavy burden on human operators. This often leads to customer dissatisfaction and a potential decline in customer satisfaction.

[0192] 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.

[0193] In this invention, the server includes means for receiving inquiry data, means for natural language processing to analyze the content of the inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for transferring the inquiry to a human operator based on the analysis results, means for storing the inquiry content and its response in a database, means including a smartphone application for receiving inquiries, means for performing automated responses using generative artificial intelligence within the application, and means for notifying the operator if the inquiry is complex. This enables improved resource efficiency and faster response times, and is expected to improve customer satisfaction.

[0194] "Inquiry data" refers to questions, requests, and related information sent by users to the support center.

[0195] "Natural language processing" is the technology that enables computers to understand and analyze human language.

[0196] "Generative artificial intelligence" refers to a system that uses artificial intelligence technology to automatically generate answers to user inquiries.

[0197] A "smartphone application" is software specifically designed to run on a smartphone.

[0198] "Automated response" refers to a function that automatically provides pre-programmed answers using generative artificial intelligence.

[0199] "Preprocessing" refers to the process of removing unnecessary information from data and converting it into an analyzable format.

[0200] A "database" is a system for organizing and efficiently managing data.

[0201] "Notify operator" refers to the system sending a notification to a human operator to request assistance when the inquiry is complex.

[0202] This invention aims to streamline customer support systems and improve customer satisfaction. The details of a system that receives inquiry data, generates automated responses, and, when necessary, transfers the inquiry to a human operator are described below.

[0203] Hardware and software to use

[0204] This system consists of a server, a smartphone, and the necessary software.

[0205] Hardware: Servers, cloud infrastructure (AWS®, GCP, Azure®), smartphones

[0206] Software: Flask, spaCy, transformers, SQLite

[0207] Receiving and preprocessing query data

[0208] 1. Received:

[0209] Users submit questions via chat, email, or contact forms through a smartphone application. The server then receives the inquiry data.

[0210] 2. Pre-processing:

[0211] The server preprocesses the query data it receives. SpaCy is used to remove unnecessary information (e.g., HTML tags and special characters) and convert it into a format that can be parsed as text.

[0212] Analysis and classification

[0213] 3. Natural Language Processing:

[0214] The server passes pre-processed data to a natural language processing (NLP) model to analyze the query. Libraries such as spaCy and transformers are used for this process.

[0215] 4. Classification:

[0216] The server classifies the inquiry based on the analysis results. For example, it might classify it as a general question, a request for technical support, or an emotional complaint.

[0217] Automatic response generation

[0218] 5. Generative Artificial Intelligence:

[0219] The server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-3(registered trademark).5-turbo) and generates a response. If an automated response is applicable, this response is generated.

[0220] 6. Submit your response:

[0221] The server sends the generated response to the user. For example, in response to the inquiry, "I want to know about the new mobile phone plans," the server sends the response, "The following types of mobile phone plans are currently available..."

[0222] Handover to operator

[0223] 7. Handover:

[0224] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends notifications to transfer these inquiries to human operators. The notifications include the inquiry history and related information.

[0225] Data storage

[0226] 8. Save:

[0227] The server saves all inquiry details and generated responses to a database (SQLite). This allows for future inquiry handling and analysis.

[0228] As a concrete example of a prompt message, if a user asks, "I want to know about the new mobile phone plan,"

[0229] I want to know about the new mobile phone plans.

[0230] By inputting this sentence into a generative artificial intelligence system, an appropriate answer will be automatically generated.

[0231] As described above, a system is realized that enables improved resource efficiency and faster response times, leading to increased customer satisfaction.

[0232] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0233] Step 1:

[0234] Users submit inquiries through a smartphone application. In this process, users enter their inquiry details via various means, such as chat, email, or a contact form, and then press the submit button. The input data includes the user's question or request, as well as contact information. The server receives this inquiry data.

[0235] Step 2:

[0236] The server preprocesses the query data it receives. This preprocessing uses spaCy to remove unnecessary information (such as HTML tags and special characters) from the data and convert it into a parseable format. The output of this preprocessing is cleaned text data.

[0237] Step 3:

[0238] The server passes the pre-processed text data to a natural language processing (NLP) model. The NLP model (e.g., spaCy or transformers) analyzes the text and extracts important keywords and intent. The analysis results are then output.

[0239] Step 4:

[0240] The server classifies the inquiry based on the analysis results from the NLP model. For example, it might be classified as a general question, technical support, or an emotional complaint. Based on this classification, it determines whether an automated response is possible.

[0241] Step 5:

[0242] If automated response is applicable, the server passes the analysis results as an input prompt to a generative artificial intelligence (e.g., GPT-3.5-turbo) to generate a response. The output of this generative AI is the text response to the user's inquiry.

[0243] Step 6:

[0244] The server sends the generated response to the user. The response is sent via chat, email, or notification functions through a smartphone application. At this time, an automatically generated response to the user's inquiry is output.

[0245] Step 7:

[0246] For inquiries that are difficult for generative artificial intelligence to handle, the server will transfer the inquiry to a human operator. The server will notify the operator of the inquiry details and send the history and related information along with the inquiry. This transfer to an operator makes it possible to handle even complex inquiries.

[0247] Step 8:

[0248] The server saves all inquiry details, generated responses, and response history to a database (SQLite). This data is used for future inquiry handling and analysis. The saved data includes inquiry text, response text, user information, and related metadata.

[0249] 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.

[0250] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence and an emotion engine. This system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The embodiments for carrying out this invention are described in detail below.

[0251] 1. Receiving inquiry data

[0252] When a user submits an inquiry via a web form, email, or chat system,

[0253] The server receives this.

[0254] The data received includes inquiry details and user contact information.

[0255] 2. Analysis and classification of inquiry content

[0256] The server preprocesses the query data it receives.

[0257] During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0258] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[0259] NLP models are implemented using libraries such as spaCy or transformers.

[0260] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[0261] 3. Emotion recognition by an emotion engine

[0262] The server passes the analysis results and pre-processed data to the emotion engine.

[0263] The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[0264] The output of the emotion engine includes data that quantifies the user's emotions (e.g., anger level 70%).

[0265] The server uses the emotions recognized by the emotion engine to determine the urgency of the inquiry and the appropriate course of action.

[0266] 4. Automated responses using generative AI

[0267] Extract queries that the server can automatically respond to.

[0268] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[0269] The generative artificial intelligence generates the answer and returns it to the server.

[0270] The server sends the generated response to the user.

[0271] For example, in response to an inquiry such as "Please tell me the details of the data plan," the response would be "The data plans currently available to you are as follows..."

[0272] The server saves the query content and the generated response to the database.

[0273] 5. Handover to a human operator

[0274] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[0275] In particular, if the emotion engine identifies the user's emotion as "negative," the server automatically initiates a process to hand over the user to a human operator.

[0276] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0277] In this process, the server provides the operator with the necessary inquiry history and related information.

[0278] Perform specific responses to inquiries taken over by a human operator.

[0279] As an example, consider the processing flow when a user inquires "I am really disappointed with this service. Please respond."

[0280] 1. The user sends an inquiry from the chat system.

[0281] 2. The server receives this and performs preprocessing.

[0282] 3. The server analyzes it with an NLP model and classifies the inquiry content as an "emotional claim".

[0283] 4. The server passes the preprocessed data to the emotion engine to identify the user's emotion.

[0284] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[0285] 6. The server receives this result and passes the inquiry to the appropriate claim operator.

[0286] 7. The human operator directly interacts with the user for the inquiry taken over and performs appropriate responses.

[0287] In this way, by combining the emotion engine, it becomes possible to make appropriate and prompt responses according to the user's emotion, and especially to make appropriate responses even to users with negative emotions. With this system, the efficiency and quality of customer support can be greatly improved, and an improvement in customer satisfaction can be expected.

[0288] The following explains the processing flow.

[0289] Step 1:

[0290] A user submits an inquiry. Inquiries can be submitted via web form, email, or chat system.

[0291] Step 2:

[0292] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information.

[0293] Step 3:

[0294] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0295] Step 4:

[0296] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[0297] Step 5:

[0298] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[0299] Step 6:

[0300] The server passes the analysis results to the emotion engine. The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[0301] Step 7:

[0302] The emotion engine quantifies the user's emotions and returns the result (e.g., anger level 70%) to the server.

[0303] Step 8:

[0304] The server receives the output result of the emotion engine and determines the urgency of the inquiry and specific response methods.

[0305] Step 9:

[0306] Based on the analysis result and the result of the emotion engine, the server classifies the inquiry content. Classification categories include general questions, requests for technical support, emotional claims, etc.

[0307] Step 10:

[0308] The server extracts the inquiries classified as "automatically answerable".

[0309] Step 11:

[0310] The server inputs the inquiry content into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[0311] Step 12:

[0312] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[0313] Step 13:

[0314] The server sends the generated answer to the user. For example, in response to an inquiry such as "Please tell me the details of the data plan", it responds with "The data plans currently available to you are as follows...".

[0315] Step 14:

[0316] The server saves the inquiry content and the generated answer in the database. The saved content includes the inquiry content, the generated answer, user information, timestamp, etc.

[0317] Step 15:

[0318] The server extracts inquiries that it has classified as requiring human intervention. In particular, if the sentiment engine identifies the user's emotion as "negative," this is automatically applied.

[0319] Step 16:

[0320] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0321] Step 17:

[0322] The server provides the operator with the inquiry history and related information.

[0323] Step 18:

[0324] For inquiries taken over by human operators, specific actions are taken. For example, users with negative feelings may receive follow-up emails or phone calls.

[0325] As a concrete example, let's consider the flow of a case where a user contacts us saying, "I am truly disappointed with this service. Please take action."

[0326] 1. The user submits an inquiry through the chat system.

[0327] 2. The server receives this and performs preprocessing.

[0328] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[0329] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[0330] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[0331] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[0332] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[0333] In this way, this system improves the efficiency and quality of customer support by combining automated responses with appropriate handover based on emotion recognition.

[0334] (Example 2)

[0335] 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".

[0336] To improve efficiency and customer satisfaction in customer support, a system that processes inquiry data quickly and accurately is necessary. However, conventional systems have the problem of being unable to keep up with processing a large volume of inquiries, leading to increased customer dissatisfaction. Furthermore, they do not adequately recognize emotions, making it difficult to provide appropriate responses that are in line with the customer's feelings. This invention aims to solve these problems.

[0337] 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.

[0338] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for performing natural language processing using the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for performing emotion recognition, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for handing over the inquiry to a human operator based on the analysis results and emotion recognition results, and means for storing the inquiry data and the generated response in a database. This enables rapid and accurate processing of inquiry data and further realizes appropriate responses that are in line with the customer's emotions.

[0339] "Inquiry data" refers to data containing information such as questions, requests, and complaints sent by users to customer support.

[0340] "Preprocessing" refers to the process of removing unnecessary information from received query data and converting it into a format that can be processed by natural language processing.

[0341] "Natural language processing" is a technology that uses computers to analyze human language and understand its meaning.

[0342] "Emotion recognition" is a technology that identifies and quantifies emotions such as joy, anger, and sadness from the content of user inquiries.

[0343] "Generative artificial intelligence" refers to machine learning models used to generate appropriate answers based on the content of inquiries.

[0344] "Automated response" refers to the process in which a server sends a response generated by a generative artificial intelligence system to the user.

[0345] A "human operator" is a human worker who directly handles inquiries that the server cannot process.

[0346] A "database" is an information management system used to store query data and generated responses.

[0347] "Analysis results" refers to the analysis of inquiry data obtained through natural language processing and sentiment recognition.

[0348] "Handing over" refers to the process by which a server notifies a human operator of an inquiry requiring attention and entrusts them with handling it.

[0349] This invention is a system aimed at improving the efficiency of resources in customer support and enhancing customer satisfaction, and it automatically processes inquiry data using generative artificial intelligence and an emotion recognition engine. The embodiments of this system are described in detail below.

[0350] First, the user submits an inquiry via a web form, email, or chat system. For example, consider a case where a user submits a request via a web form asking, "How do I use this product?" This inquiry data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address.

[0351] Next, the server receives the transmitted data at the specified endpoint. This received data is preprocessed using regular expressions and text cleaning libraries (e.g., BeautifulSoup4 in Python) to remove HTML tags, special characters, and unnecessary whitespace. The preprocessed text is then converted into a parseable format.

[0352] The server passes the pre-processed query data to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers, and it analyzes the data to extract the query's intent and important keywords. For example, a query like "Please tell me how to use the product" would be classified as "Information Provision."

[0353] Furthermore, the server passes the analysis results and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob, and identifies and quantifies emotions such as joy, anger, and sadness from the user's text. For example, it outputs quantified data such as "Anger level 70%". The server receives the output from the emotion engine and determines the urgency of the inquiry and how to respond.

[0354] The server passes an automated response-capable inquiry to a generative artificial intelligence (e.g., GPT-4) to generate a response. An example prompt might be: "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response." The generative AI analyzes the text and generates an appropriate response. The generated response is returned to the server, which then sends it to the user. For example, a specific response such as "To use this product, please follow these steps..." might be generated.

[0355] Furthermore, the server stores the query content and the generated response in a database. This allows for future reference and analysis.

[0356] On the other hand, the server also extracts inquiries that are difficult for generative artificial intelligence to handle and hands them over to human operators. In particular, if the emotion recognition engine highly values ​​"negative emotions," the server automatically hands over the inquiry to a human operator. In this process, notifications are sent via the internal operating system, providing the necessary inquiry history and related information.

[0357] This enables the rapid and accurate processing of inquiry data, as well as appropriate responses that are tailored to the customer's emotions. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[0358] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0359] Step 1: Receiving inquiry data

[0360] A user submits an inquiry via a web form, email, or chat system. For example, a user might submit "How do I use this product?" through a web form. The server receives this at a specified endpoint. The input data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address. The server performs initial database processing to store this data.

[0361] Step 2: Pre-processing of inquiry content

[0362] The server preprocesses the query data it receives. The input is the query data received in step 1. Specific preprocessing actions include removing HTML tags, special characters, and unnecessary whitespace. This is done using regular expressions and the Python BeautifulSoup4 library, among others. The output is data with the text and metadata cleaned and converted into a parseable format.

[0363] Step 3: Analysis and classification of inquiries

[0364] The server passes pre-processed data as input to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers. The server runs the NLP model to extract the intent of the query and important keywords. For example, the query "Please tell me how to use the product" would be classified as "Information Provision." The output is data containing the analysis results and classification results.

[0365] Step 4: Emotion recognition by the emotion engine

[0366] The server passes the analysis results of the NLP model and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob. Here, the input is the analysis results and classification results obtained in step 3. The emotion engine identifies emotions such as joy, anger, and sadness from the user's text and outputs quantified data (e.g., "Anger level 70%)". The server receives this output and makes decisions regarding the urgency of the response and how to respond.

[0367] Step 5: Automated response by generative AI

[0368] The server passes a query that can be automatically answered to a generative artificial intelligence (e.g., GPT-4). The input is pre-processed query data and analysis results. The generative AI is given prompts to generate a response. For example, a prompt such as "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response" can be used. The generative AI generates a response (e.g., "For instructions on how to use this product, please follow these steps...") and returns it to the server. The server receives the generated response and sends it to the user. The output is the response text sent to the user.

[0369] Step 6: Handover to human operator

[0370] The server extracts inquiries that are difficult for generative artificial intelligence to handle. In particular, if the emotion engine identifies the user's emotion as "negative," the server initiates a process to hand over the inquiry to a human operator. The input here is the output from the emotion recognition engine and the content of the inquiry that the generative artificial intelligence could not handle. The server sends a notification to the internal operating system to hand over these inquiries to human operators. The output is the notification to the human operator and related inquiry data.

[0371] As described above, this system processes inquiry data step by step, efficiently managing and processing it. This enables improved customer satisfaction and more efficient customer support.

[0372] (Application Example 2)

[0373] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0374] Current customer support systems consume a significant amount of resources in handling inquiries, necessitating greater resource efficiency. Furthermore, accurately understanding customer emotions and providing appropriate responses is challenging, hindering customer satisfaction. Additionally, in real-time customer service at physical stores, staff may not always be able to provide the most suitable answer immediately, potentially detracting from the customer experience.

[0375] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for receiving inquiry data, means for natural language processing for analyzing the content of the inquiry data, means for classifying the content of the inquiry based on the analysis results, means for generating an answer to the inquiry using generative artificial intelligence, means for transmitting the generated answer, means for transferring the inquiry to a human operator based on the analysis results, means for emotion recognition, means for determining the urgency of the inquiry based on the output result of the emotion recognition means, and means for displaying the generated answer to customer service in the store. This enables efficient use of resources in customer support and improved customer satisfaction. Furthermore, it enables smooth real-time customer service in physical stores and improves the customer experience.

[0376] "Inquiry data" refers to data that includes the content of questions and requests submitted by customers.

[0377] "Means of receiving" refers to the means by which the system takes in query data.

[0378] "Natural language processing means" refers to technologies used to analyze query data and understand its content.

[0379] "Generative artificial intelligence" refers to AI technology that automatically generates answers to inquiries.

[0380] "Emotion recognition means" refers to technology used to identify customer emotions from inquiry data.

[0381] "Means of classification" refers to methods for categorizing inquiry content based on analysis results.

[0382] "Means for determining urgency" refers to means for determining the urgency of responding to an inquiry based on the output results of the emotion recognition means.

[0383] "Means of transmission" refers to the means of communicating the generated response to the customer.

[0384] "Means of transferring inquiries to human operators" refers to a method of automatically transferring inquiries to human operators when action is required.

[0385] "Means for displaying responses generated in response to customer inquiries within a store" refers to means for displaying responses generated in response to customer inquiries within a physical store.

[0386] This invention is a real-time customer support system using generative artificial intelligence and emotion recognition technology. The specific implementation of this system is described below.

[0387] System Configuration

[0388] hardware

[0389] Servers: Multiple servers are used to receive, analyze, classify, and generate responses to query data.

[0390] Terminal: A device used to access the system, including smartphones and smart glasses.

[0391] Users include customers of physical stores and end users who utilize customer support.

[0392] software

[0393] Natural language processing (NLP) libraries such as spaCy and transformers will be used.

[0394] Generative artificial intelligence: Uses generative AI models such as GPT-4.

[0395] Sentiment recognition engine: Uses sentiment analysis tools that utilize dictionary-based or machine learning-based algorithms.

[0396] Database: Use an RDBMS or NoSQL database to store query data and generated responses.

[0397] System operation

[0398] 1. Receiving inquiry data:

[0399] Users can make inquiries in real time at physical stores via their smartphones or smart glasses.

[0400] The terminal sends this to the server, and the query data is received.

[0401] 2. Analysis and classification of inquiries:

[0402] The server preprocesses the received query data and removes unnecessary information.

[0403] We use NLP models to analyze the content of inquiries and extract their intent and keywords.

[0404] 3. Emotion recognition:

[0405] The server uses an emotion recognition engine to identify the user's emotions from the inquiry data.

[0406] For example, a user's emotions might be quantified as "70% anger."

[0407] 4. Automated responses using generative artificial intelligence:

[0408] The server passes the inquiry details to a generative artificial intelligence system based on the analysis results and sentiment data, and generates an automated response.

[0409] 5. Sending and displaying responses:

[0410] The server sends the generated response to the terminal, which then displays it to the user.

[0411] 6. Handover to a human operator:

[0412] In particular, if the emotion recognition engine determines that the user's emotions are negative, the server automatically hands over the inquiry to a human operator.

[0413] Specific example

[0414] Example 1:

[0415] Inquiry: "Where can I use Wi-Fi in the store?"

[0416] Generated response: "Wi-Fi is available in all areas of the store."

[0417] Example 2:

[0418] Inquiry: "I am truly dissatisfied with the service at your store."

[0419] Emotional analysis result: Anger level 80%

[0420] Countermeasure: The case will be immediately handed over to a human operator.

[0421] Example of a prompt:

[0422] Customer inquiry: "Where can I use Wi-Fi in the store?"

[0423] AI response: "Wi-Fi is available in all areas of the store."

[0424] or

[0425] Customer inquiry: "I am truly dissatisfied with the service at your store."

[0426] AI response: "First, to resolve your concerns, our store manager will speak with you."

[0427] This system will significantly improve the efficiency of customer support and increase customer satisfaction. It will also enable smoother, real-time customer service in physical stores.

[0428] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0429] Step 1:

[0430] A user uses a smartphone or smart glasses to make an inquiry within a physical store. For example, a customer might ask, "Where can I get Wi-Fi?" The device receives this inquiry data and sends it to a server. The input includes the text data of the inquiry and the user's basic information. The output is the inquiry data sent to the server.

[0431] Step 2:

[0432] The server preprocesses the received query data. During this process, unnecessary information (e.g., HTML tags, special characters) is removed, and the data is converted into a format that can be parsed as text. The input includes the original query text. The output is clean text data.

[0433] Step 3:

[0434] The server inputs pre-processed data into a natural language processing (NLP) model. The server uses this model to analyze the query content and extract its intent and keywords. For example, libraries such as spaCy or transformers are used. Clean text data is passed to the NLP model as input. The output is the analyzed intent and keywords.

[0435] Step 4:

[0436] The server inputs the analysis results and pre-processed data into the emotion recognition engine. This engine identifies the user's emotion (e.g., joy, anger, sadness) from the query text. The input includes the analysis results and clean text data. The output is numerical data representing the user's emotion (e.g., anger level 70%).

[0437] Step 5:

[0438] The server determines the urgency of an inquiry based on emotion data obtained from an emotion recognition engine. For example, if the "anger level" exceeds a certain threshold, the inquiry is classified as highly urgent. The input includes emotion data. The output is a classification of the inquiry's urgency.

[0439] Step 6:

[0440] The server determines whether it is possible to generate an automated response to the inquiry. If an automated response is possible, the server passes the inquiry details to a generative artificial intelligence (e.g., GPT-4) to generate a response. The input includes the NLP analysis results and the urgency assessment results. The output is the generated response.

[0441] Step 7:

[0442] The server sends the generated response to the terminal, which then displays it to the user. The input includes the generated response and the user's query data. The output provides the response that is displayed to the user.

[0443] Step 8:

[0444] The server handles the transfer of high-priority or emotionally charged inquiries to human operators. The server provides the operator with necessary inquiry history and relevant information. Inputs include the urgency assessment result and the user's inquiry data. Outputs provide the inquiry information to be transferred to the human operator.

[0445] 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.

[0446] 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.

[0447] 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.

[0448] [Second Embodiment]

[0449] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0450] 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.

[0451] 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).

[0452] 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.

[0453] 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.

[0454] 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).

[0455] 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.

[0456] 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.

[0457] 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.

[0458] 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.

[0459] 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.

[0460] 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".

[0461] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence. The system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The necessary embodiments for carrying out this invention are described in detail below.

[0462] 1. Receiving inquiry data

[0463] When a user submits an inquiry via a web form, email, or chat system,

[0464] The server receives this.

[0465] The data received includes inquiry details and user contact information.

[0466] 2. Analysis and classification of inquiry content

[0467] The server preprocesses the query data it receives.

[0468] During preprocessing, unnecessary information (such as HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0469] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[0470] NLP models are implemented using libraries such as spaCy or transformers.

[0471] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[0472] The server classifies the query content based on the analysis results.

[0473] For example, these can be categorized into general questions, requests for technical support, and emotional complaints.

[0474] 3. Automated responses using generative AI

[0475] Extract queries that the server can automatically respond to.

[0476] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[0477] The generative artificial intelligence generates the answer and returns it to the server.

[0478] The server sends the generated response to the user.

[0479] For example, in response to an inquiry such as "Please tell me more about the data plans," the response would be in the format of "The following types of data plans are currently available..."

[0480] The server saves the query content and the generated response to the database.

[0481] 4. Handover to a human operator

[0482] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[0483] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0484] In this process, the server provides the operator with the necessary inquiry history and related information.

[0485] We will provide appropriate responses to inquiries that have been handed over to human operators.

[0486] As an example, let's consider the processing flow when a user inquires, "Please tell me about the latest mobile phone plans."

[0487] 1. The user submits an inquiry via a web form.

[0488] 2. The server receives this and performs preprocessing.

[0489] 3. The server analyzes the inquiry using an NLP model and classifies it as "capable of automated response."

[0490] 4. The server sends the question to the generative AI and instructs it to generate an answer.

[0491] 5. The generative artificial intelligence generates the answer and returns it to the server.

[0492] 6. The server sends the response to the user.

[0493] 7. The server saves all interactions to the database.

[0494] This allows for quick and accurate responses to user inquiries, as well as the ability to flexibly transfer them to human operators when necessary. This system is expected to improve the efficiency of customer support operations and enhance customer satisfaction.

[0495] The following describes the processing flow.

[0496] Step 1:

[0497] A user submits an inquiry. Inquiries may be submitted via a web form, email, or chat system.

[0498] Step 2:

[0499] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information (email address and phone number).

[0500] Step 3:

[0501] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0502] Step 4:

[0503] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[0504] Step 5:

[0505] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[0506] Step 6:

[0507] The server classifies the inquiry based on the analysis results. Classification categories include general questions, requests for technical support, and emotional complaints.

[0508] Step 7:

[0509] The server extracts queries that it has classified as "capable of automatic response."

[0510] Step 8:

[0511] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[0512] Step 9:

[0513] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[0514] Step 10:

[0515] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[0516] Step 11:

[0517] The server saves all interactions to a database. This database includes query details, generated responses, user information, and timestamps.

[0518] Step 12:

[0519] The server extracts inquiries that it has classified as requiring human intervention.

[0520] Step 13:

[0521] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0522] Step 14:

[0523] The server provides the operator with the inquiry history and related information.

[0524] Step 15:

[0525] For inquiries taken over by human operators, specific actions will be taken.

[0526] These steps streamline customer support operations by automatically classifying and processing inquiries, and only handing them over to human operators when necessary.

[0527] (Example 1)

[0528] 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".

[0529] Conventional customer support systems have often failed to adequately streamline inquiry processing, leading to decreased customer satisfaction. Furthermore, automated response systems struggled to handle complex inquiries, placing a burden on human operators. This invention aims to solve these problems by providing an efficient inquiry processing system utilizing generative artificial intelligence.

[0530] 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.

[0531] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for natural language processing to analyze the content of the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for storing the inquiry data and the generated response in a database, means for transferring the inquiry to a human operator based on the analysis results, means for sending a notification to an internal operation system, and means for removing unnecessary information in order to convert the inquiry content into an analyzable format. This enables increased efficiency in customer support operations and improved customer satisfaction.

[0532] "Inquiry data" refers to information that users send to the customer support system, including questions, requests, and complaints.

[0533] "Preprocessing" is the process of removing unnecessary information in order to convert query data into a parseable format.

[0534] "Natural language processing (NLP) techniques" refer to technologies for analyzing received query data and understanding its content, and include text analysis and keyword extraction.

[0535] A "classification method" is a process that classifies query content into a specific category based on the results of analysis using natural language processing methods.

[0536] "Generative artificial intelligence (AI)" is an artificial intelligence technology that automatically generates responses based on the content of inquiries.

[0537] "Methods for generating answers" refers to the process of generating appropriate answers to inquiries using generative artificial intelligence.

[0538] "Means of submitting responses" refers to the process of sending the generated responses to the user.

[0539] A "database" is a digital storage system for storing query data and generated responses.

[0540] "Methods for transferring inquiries" refers to the process of transferring inquiries that are difficult for generative artificial intelligence to handle to human operators.

[0541] "Methods for sending notifications" refers to the process of notifying human operators of important inquiries or inquiries that are difficult for generative artificial intelligence to handle.

[0542] "Converting to a parsable format" means transforming query data into a form that is easy to process, and this includes removing unnecessary information.

[0543] This invention is a system aimed at improving the efficiency of inquiry processing and customer satisfaction in customer support systems. Specific embodiments are described in detail below.

[0544] System Configuration

[0545] This system is primarily composed of the following hardware and software.

[0546] Server: Receives, processes, analyzes, generates responses, stores data, and sends notifications.

[0547] Device: The device that the user uses to submit an inquiry (e.g., a personal computer or smartphone).

[0548] software:

[0549] Natural Language Processing (NLP) libraries (e.g., spaCy, transformers)

[0550] Generative artificial intelligence (AI) models (e.g., GPT-4)

[0551] Database management systems (e.g., MySQL, PostgreSQL)

[0552] Description of specific embodiments

[0553] Receiving inquiry data

[0554] The user submits an inquiry using their device via a web form, email, or chat system. During this process, data containing the inquiry details and the user's contact information is generated. The server receives this data.

[0555] Pre-processing of inquiry content

[0556] The query data received by the server is preprocessed. Specifically, the software's regular expression library and string manipulation functions are used to remove unnecessary information (e.g., HTML tags, special characters, unnecessary line breaks and spaces) and convert it into a parseable format.

[0557] Analysis and classification of inquiry content

[0558] The server passes pre-processed query data to a natural language processing (NLP) model (e.g., spaCy, transformers) for analysis. The NLP model extracts the intent and key keywords of the query. Based on this analysis, the server classifies the query into a specific category (e.g., general question, request for technical support, emotional complaint, etc.).

[0559] Generation of automated responses using generative AI

[0560] The server extracts queries that can be answered automatically. If an automatic response is applicable, the server inputs the query details into a generative artificial intelligence (e.g., GPT-4) to create an appropriate prompt. For example, it might use a prompt like this:

[0561] "Please tell me about the latest mobile phone plans."

[0562] The server inputs this prompt into a generative AI model, which then generates a response. The generative artificial intelligence generates an appropriate response and returns it to the server.

[0563] Submitting responses and saving data

[0564] The server sends the generated response to the user. An example of the response might be, "The following data plans are currently available to you..." The server also saves the inquiry and the generated response to a database. The information saved includes the inquiry, analysis results, generated response, user contact information, date and time, etc.

[0565] Handover to human operator

[0566] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends a notification to the internal operation system to transfer these inquiries to human operators. The server provides the operators with the necessary inquiry history and related information. Human operators then take appropriate action on the inquiries they receive.

[0567] The above describes a specific embodiment of the present invention. This system enables increased efficiency in customer support operations and improved customer satisfaction.

[0568] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0569] Step 1:

[0570] The user submits an inquiry. The user enters and submits their inquiry and contact information using a web form, email, or chat system.

[0571] Input: Inquiry details, user contact information

[0572] Output: Query data received by the server

[0573] Step 2:

[0574] The server receives the query data. The server receives the query data sent by the user and saves it as a file.

[0575] Input: Inquiry data submitted by the user

[0576] Output: Received query data

[0577] Step 3:

[0578] The server preprocesses the query data. To convert the received query data into a parseable format, it removes unnecessary information (HTML tags, special characters, etc.) and converts it into text format.

[0579] Input: Received query data

[0580] Output: Preprocessed query data

[0581] Specific operation: Uses a regular expression library to remove HTML tags and special characters.

[0582] Step 4:

[0583] The server passes the pre-processed query data to a natural language processing (NLP) model for analysis. The NLP model (e.g., spaCy, transformers) is used to extract the intent and key keywords of the query.

[0584] Input: Preprocessed query data

[0585] Output: Analysis results (intent, important keywords)

[0586] Specific operation: Input data into the NLP model and obtain the analysis results.

[0587] Step 5:

[0588] The server classifies the inquiry based on the analysis results. The analyzed inquiry data is categorized into specific categories (e.g., general questions, technical support, complaints, etc.).

[0589] Input: Analysis results

[0590] Output: Classified query data

[0591] Specific operation: Based on the analysis results, conditional branching is performed, and data is distributed to each category.

[0592] Step 6:

[0593] The server inputs a prompt message to a generative artificial intelligence (AI) and generates an automated response. If applicable, the generative AI (e.g., GPT-4) creates a prompt message based on the analyzed query content and generates a response.

[0594] Input: Classified query data, prompt text

[0595] Output: Auto-generated answer

[0596] Specific operation: Input a prompt sentence to a generative AI model and generate an appropriate response.

[0597] Step 7:

[0598] The server sends the generated response to the user. The generated response is also sent to the user's contacts (e.g., email, chat).

[0599] Input: Automated response, user contact information

[0600] Output: Response sent to the user

[0601] Specific actions: Send responses using an email system or chat system.

[0602] Step 8:

[0603] The server saves the query content and the generated response to the database. All query content and its response are recorded in the database and can be referenced later.

[0604] Input: Inquiry details, analysis results, generated response

[0605] Output: Data stored in the database

[0606] Specific operation: Use a database management system to save data.

[0607] Step 9:

[0608] The server transfers inquiries it cannot handle to human operators. If the generative AI cannot process the inquiry, it sends a notification to the internal operations system.

[0609] Input: Classified query data

[0610] Output: Inquiry details handed over to the operator, notification sent.

[0611] Specific action: Send a notification to the internal operating system and display it on the operator's terminal.

[0612] (Application Example 1)

[0613] 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."

[0614] Current customer support systems fail to adequately optimize resources and improve customer satisfaction. In particular, there are problems with delayed inquiry processing and inability to provide appropriate responses. Furthermore, the limited availability of automated responses to inquiries places a heavy burden on human operators. This often leads to customer dissatisfaction and a potential decline in customer satisfaction.

[0615] 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.

[0616] In this invention, the server includes means for receiving inquiry data, means for natural language processing to analyze the content of the inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for transferring the inquiry to a human operator based on the analysis results, means for storing the inquiry content and its response in a database, means including a smartphone application for receiving inquiries, means for performing automated responses using generative artificial intelligence within the application, and means for notifying the operator if the inquiry is complex. This enables improved resource efficiency and faster response times, and is expected to improve customer satisfaction.

[0617] "Inquiry data" refers to questions, requests, and related information sent by users to the support center.

[0618] "Natural language processing" is the technology that enables computers to understand and analyze human language.

[0619] "Generative artificial intelligence" refers to a system that uses artificial intelligence technology to automatically generate answers to user inquiries.

[0620] A "smartphone application" is software specifically designed to run on a smartphone.

[0621] "Automated response" refers to a function that automatically provides pre-programmed answers using generative artificial intelligence.

[0622] "Preprocessing" refers to the process of removing unnecessary information from data and converting it into an analyzable format.

[0623] A "database" is a system for organizing and efficiently managing data.

[0624] "Notify operator" refers to the system sending a notification to a human operator to request assistance when the inquiry is complex.

[0625] This invention aims to streamline customer support systems and improve customer satisfaction. The details of a system that receives inquiry data, generates automated responses, and, when necessary, transfers the inquiry to a human operator are described below.

[0626] Hardware and software to use

[0627] This system consists of a server, a smartphone, and the necessary software.

[0628] Hardware: Servers, cloud infrastructure (AWS, GCP, Azure), smartphones

[0629] Software: Flask, spaCy, transformers, SQLite

[0630] Receiving and preprocessing query data

[0631] 1. Received:

[0632] Users submit questions via chat, email, or contact forms through a smartphone application. The server then receives the inquiry data.

[0633] 2. Pre-processing:

[0634] The server preprocesses the query data it receives. SpaCy is used to remove unnecessary information (e.g., HTML tags and special characters) and convert it into a format that can be parsed as text.

[0635] Analysis and classification

[0636] 3. Natural Language Processing:

[0637] The server passes pre-processed data to a natural language processing (NLP) model to analyze the query. Libraries such as spaCy and transformers are used for this process.

[0638] 4. Classification:

[0639] The server classifies the inquiry based on the analysis results. For example, it might classify it as a general question, a request for technical support, or an emotional complaint.

[0640] Automatic response generation

[0641] 5. Generative Artificial Intelligence:

[0642] The server inputs the query details into a generative artificial intelligence (e.g., GPT-3.5-turbo) and generates a response. If an automated response is applicable, this response is generated.

[0643] 6. Submit your response:

[0644] The server sends the generated response to the user. For example, in response to the inquiry, "I want to know about the new mobile phone plans," the server sends the response, "The following types of mobile phone plans are currently available..."

[0645] Handover to operator

[0646] 7. Handover:

[0647] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends notifications to transfer these inquiries to human operators. The notifications include the inquiry history and related information.

[0648] Data storage

[0649] 8. Save:

[0650] The server saves all inquiry details and generated responses to a database (SQLite). This allows for future inquiry handling and analysis.

[0651] As a concrete example of a prompt message, if a user asks, "I want to know about the new mobile phone plan,"

[0652] I want to know about the new mobile phone plans.

[0653] By inputting this sentence into a generative artificial intelligence system, an appropriate answer will be automatically generated.

[0654] As described above, a system is realized that enables improved resource efficiency and faster response times, leading to increased customer satisfaction.

[0655] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0656] Step 1:

[0657] Users submit inquiries through a smartphone application. In this process, users enter their inquiry details via various means, such as chat, email, or a contact form, and then press the submit button. The input data includes the user's question or request, as well as contact information. The server receives this inquiry data.

[0658] Step 2:

[0659] The server preprocesses the query data it receives. This preprocessing uses spaCy to remove unnecessary information (such as HTML tags and special characters) from the data and convert it into a parseable format. The output of this preprocessing is cleaned text data.

[0660] Step 3:

[0661] The server passes the pre-processed text data to a natural language processing (NLP) model. The NLP model (e.g., spaCy or transformers) analyzes the text and extracts important keywords and intent. The analysis results are then output.

[0662] Step 4:

[0663] The server classifies the inquiry based on the analysis results from the NLP model. For example, it might be classified as a general question, technical support, or an emotional complaint. Based on this classification, it determines whether an automated response is possible.

[0664] Step 5:

[0665] If automated response is applicable, the server passes the analysis results as an input prompt to a generative artificial intelligence (e.g., GPT-3.5-turbo) to generate a response. The output of this generative AI is the text response to the user's inquiry.

[0666] Step 6:

[0667] The server sends the generated response to the user. The response is sent via chat, email, or notification functions through a smartphone application. At this time, an automatically generated response to the user's inquiry is output.

[0668] Step 7:

[0669] For inquiries that are difficult for generative artificial intelligence to handle, the server will transfer the inquiry to a human operator. The server will notify the operator of the inquiry details and send the history and related information along with it. This transfer to an operator makes it possible to handle even complex inquiries.

[0670] Step 8:

[0671] The server saves all inquiry details, generated responses, and response history to a database (SQLite). This data is used for future inquiry handling and analysis. The saved data includes inquiry text, response text, user information, and related metadata.

[0672] 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.

[0673] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence and an emotion engine. This system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The embodiments for carrying out this invention are described in detail below.

[0674] 1. Receiving inquiry data

[0675] When a user submits an inquiry via a web form, email, or chat system,

[0676] The server receives this.

[0677] The data received includes inquiry details and user contact information.

[0678] 2. Analysis and classification of inquiry content

[0679] The server preprocesses the query data it receives.

[0680] During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0681] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[0682] NLP models are implemented using libraries such as spaCy or transformers.

[0683] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[0684] 3. Emotion recognition by an emotion engine

[0685] The server passes the analysis results and pre-processed data to the emotion engine.

[0686] The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[0687] The output of the emotion engine includes data that quantifies the user's emotions (e.g., anger level 70%).

[0688] The server uses the emotions recognized by the emotion engine to determine the urgency of the inquiry and the appropriate course of action.

[0689] 4. Automated responses using generative AI

[0690] Extract queries that the server can automatically respond to.

[0691] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[0692] The generative artificial intelligence generates the answer and returns it to the server.

[0693] The server sends the generated response to the user.

[0694] For example, in response to an inquiry such as "Please tell me the details of the data plan," the response would be "The data plans currently available to you are as follows..."

[0695] The server saves the query content and the generated response to the database.

[0696] 5. Handover to a human operator

[0697] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[0698] In particular, if the emotion engine identifies the user's emotion as "negative," the server automatically initiates a process to hand over the user to a human operator.

[0699] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0700] In this process, the server provides the operator with the necessary inquiry history and related information.

[0701] For inquiries taken over by human operators, specific actions will be taken.

[0702] As an example, let's consider the processing flow when a user contacts us saying, "I am truly disappointed with this service. Please take action."

[0703] 1. The user submits an inquiry through the chat system.

[0704] 2. The server receives this and performs preprocessing.

[0705] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[0706] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[0707] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[0708] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[0709] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[0710] By combining this emotion engine, it becomes possible to provide appropriate and rapid responses tailored to the user's emotions, especially to users experiencing negative emotions. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[0711] The following describes the processing flow.

[0712] Step 1:

[0713] A user submits an inquiry. Inquiries can be submitted via web form, email, or chat system.

[0714] Step 2:

[0715] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information.

[0716] Step 3:

[0717] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0718] Step 4:

[0719] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[0720] Step 5:

[0721] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[0722] Step 6:

[0723] The server passes the analysis results to the emotion engine. The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[0724] Step 7:

[0725] The emotion engine quantifies the user's emotions and returns the result (e.g., anger level 70%) to the server.

[0726] Step 8:

[0727] The server receives the output from the emotion engine and determines the urgency of the inquiry and the appropriate course of action.

[0728] Step 9:

[0729] The server classifies the inquiry based on the analysis results and the sentiment engine's findings. Classification categories include general questions, requests for technical support, and emotional complaints.

[0730] Step 10:

[0731] The server extracts queries that it has classified as "capable of automatic response."

[0732] Step 11:

[0733] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[0734] Step 12:

[0735] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[0736] Step 13:

[0737] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[0738] Step 14:

[0739] The server saves the query content and the generated response to a database. The saved content includes the query content, the generated response, user information, and a timestamp.

[0740] Step 15:

[0741] The server extracts inquiries that it has classified as requiring human intervention. In particular, if the sentiment engine identifies the user's sentiment as "negative," this is automatically applied.

[0742] Step 16:

[0743] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0744] Step 17:

[0745] The server provides the operator with the inquiry history and related information.

[0746] Step 18:

[0747] For inquiries taken over by human operators, specific actions are taken. For example, for users with negative feelings, follow-up emails or phone calls are provided.

[0748] As a concrete example, let's consider the flow of a case where a user contacts us saying, "I am truly disappointed with this service. Please take action."

[0749] 1. The user submits an inquiry through the chat system.

[0750] 2. The server receives this and performs preprocessing.

[0751] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[0752] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[0753] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[0754] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[0755] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[0756] In this way, this system improves the efficiency and quality of customer support by combining automated responses with appropriate handover based on emotion recognition.

[0757] (Example 2)

[0758] 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".

[0759] To improve efficiency and customer satisfaction in customer support, a system that processes inquiry data quickly and accurately is necessary. However, conventional systems have the problem of being unable to keep up with processing a large volume of inquiries, leading to increased customer dissatisfaction. Furthermore, they do not adequately recognize emotions, making it difficult to provide appropriate responses that are in line with the customer's feelings. This invention aims to solve these problems.

[0760] 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.

[0761] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for performing natural language processing using the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for performing emotion recognition, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for handing over the inquiry to a human operator based on the analysis results and emotion recognition results, and means for storing the inquiry data and the generated response in a database. This enables rapid and accurate processing of inquiry data and further realizes appropriate responses that are in line with the customer's emotions.

[0762] "Inquiry data" refers to data containing information such as questions, requests, and complaints sent by users to customer support.

[0763] "Preprocessing" refers to the process of removing unnecessary information from received query data and converting it into a format that can be processed by natural language processing.

[0764] "Natural language processing" is a technology that uses computers to analyze human language and understand its meaning.

[0765] "Emotion recognition" is a technology that identifies and quantifies emotions such as joy, anger, and sadness from the content of user inquiries.

[0766] "Generative artificial intelligence" refers to machine learning models used to generate appropriate answers based on the content of inquiries.

[0767] "Automated response" refers to the process in which a server sends a response generated by a generative artificial intelligence system to the user.

[0768] A "human operator" is a human worker who directly handles inquiries that the server cannot process.

[0769] A "database" is an information management system used to store query data and generated responses.

[0770] "Analysis results" refers to the analysis of inquiry data obtained through natural language processing and sentiment recognition.

[0771] "Handing over" refers to the process by which a server notifies a human operator of an inquiry requiring attention and entrusts them with handling it.

[0772] This invention is a system aimed at improving the efficiency of resources in customer support and enhancing customer satisfaction, and it automatically processes inquiry data using generative artificial intelligence and an emotion recognition engine. Embodiments of this system are described in detail below.

[0773] First, the user submits an inquiry via a web form, email, or chat system. For example, consider a case where a user submits a request via a web form asking, "How do I use this product?" This inquiry data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address.

[0774] Next, the server receives the transmitted data at the specified endpoint. This received data is preprocessed using regular expressions and text cleaning libraries (e.g., BeautifulSoup4 in Python) to remove HTML tags, special characters, and unnecessary whitespace. The preprocessed text is then converted into a parseable format.

[0775] The server passes the pre-processed query data to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers, and it analyzes the data to extract the query's intent and important keywords. For example, a query like "Please tell me how to use the product" would be classified as "Information Provision."

[0776] Furthermore, the server passes the analysis results and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob, and identifies and quantifies emotions such as joy, anger, and sadness from the user's text. For example, it outputs quantified data such as "Anger level 70%". The server receives the output from the emotion engine and determines the urgency of the inquiry and how to respond.

[0777] The server passes an automated response-capable inquiry to a generative artificial intelligence (e.g., GPT-4) to generate a response. An example prompt might be: "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response." The generative AI analyzes the text and generates an appropriate response. The generated response is returned to the server, which then sends it to the user. For example, a specific response such as "To use this product, please follow these steps..." might be generated.

[0778] Furthermore, the server stores the query content and the generated response in a database. This allows for future reference and analysis.

[0779] On the other hand, the server also extracts inquiries that are difficult for generative artificial intelligence to handle and hands them over to human operators. In particular, if the emotion recognition engine highly values ​​"negative emotions," the server automatically hands over the inquiry to a human operator. In this process, notifications are sent via the internal operating system, providing the necessary inquiry history and related information.

[0780] This enables the rapid and accurate processing of inquiry data, as well as appropriate responses that are tailored to the customer's feelings. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[0781] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0782] Step 1: Receiving inquiry data

[0783] A user submits an inquiry via a web form, email, or chat system. For example, a user might submit "How do I use this product?" through a web form. The server receives this at a specified endpoint. The input data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address. The server performs initial database processing to store this data.

[0784] Step 2: Pre-processing of inquiry content

[0785] The server preprocesses the query data it receives. The input is the query data received in step 1. Specific preprocessing actions include removing HTML tags, special characters, and unnecessary whitespace. This is done using regular expressions and the Python BeautifulSoup4 library, among others. The output is data with the text and metadata cleaned and converted into a parseable format.

[0786] Step 3: Analysis and classification of inquiries

[0787] The server passes pre-processed data as input to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers. The server runs the NLP model to extract the intent of the query and important keywords. For example, the query "Please tell me how to use the product" would be classified as "Information Provision." The output is data containing the analysis results and classification results.

[0788] Step 4: Emotion recognition by the emotion engine

[0789] The server passes the analysis results of the NLP model and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob. Here, the input is the analysis results and classification results obtained in step 3. The emotion engine identifies emotions such as joy, anger, and sadness from the user's text and outputs quantified data (e.g., "Anger level 70%)". The server receives this output and makes decisions regarding the urgency of the response and how to respond.

[0790] Step 5: Automated response by generative AI

[0791] The server passes a query that can be automatically answered to a generative artificial intelligence (e.g., GPT-4). The input is pre-processed query data and analysis results. The generative AI is given prompts to generate a response. For example, a prompt such as "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response" can be used. The generative AI generates a response (e.g., "For instructions on how to use this product, please follow these steps...") and returns it to the server. The server receives the generated response and sends it to the user. The output is the response text sent to the user.

[0792] Step 6: Handover to human operator

[0793] The server extracts inquiries that are difficult for generative artificial intelligence to handle. In particular, if the emotion engine identifies the user's emotion as "negative," the server initiates a process to hand over the inquiry to a human operator. The input here is the output from the emotion recognition engine and the content of the inquiry that the generative artificial intelligence could not handle. The server sends a notification to the internal operating system to hand over these inquiries to human operators. The output is the notification to the human operator and related inquiry data.

[0794] As described above, this system processes inquiry data step by step, efficiently managing and processing it. This enables improved customer satisfaction and more efficient customer support.

[0795] (Application Example 2)

[0796] 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."

[0797] Current customer support systems consume a significant amount of resources in handling inquiries, necessitating greater resource efficiency. Furthermore, accurately understanding customer emotions and providing appropriate responses is challenging, hindering customer satisfaction. Additionally, in real-time customer service at physical stores, staff may not always be able to provide the most suitable answer immediately, potentially detracting from the customer experience.

[0798] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for receiving inquiry data, means for natural language processing for analyzing the content of the inquiry data, means for classifying the content of the inquiry based on the analysis results, means for generating an answer to the inquiry using generative artificial intelligence, means for transmitting the generated answer, means for transferring the inquiry to a human operator based on the analysis results, means for emotion recognition, means for determining the urgency of the inquiry based on the output result of the emotion recognition means, and means for displaying the generated answer to customer service in the store. This enables efficient use of resources in customer support and improved customer satisfaction. Furthermore, it enables smooth real-time customer service in physical stores and improves the customer experience.

[0799] "Inquiry data" refers to data that includes the content of questions and requests submitted by customers.

[0800] "Means of receiving" refers to the means by which the system takes in query data.

[0801] "Natural language processing means" refers to technologies used to analyze query data and understand its content.

[0802] "Generative artificial intelligence" refers to AI technology that automatically generates answers to inquiries.

[0803] "Emotion recognition means" refers to technology used to identify customer emotions from inquiry data.

[0804] "Means of classification" refers to methods for categorizing inquiry content based on analysis results.

[0805] "Means for determining urgency" refers to means for determining the urgency of responding to an inquiry based on the output results of the emotion recognition means.

[0806] "Means of transmission" refers to the means of communicating the generated response to the customer.

[0807] "Means of transferring inquiries to human operators" refers to a method of automatically transferring inquiries to human operators when action is required.

[0808] "Means for displaying responses generated in response to customer inquiries within a store" refers to means for displaying responses generated in response to customer inquiries within a physical store.

[0809] This invention is a real-time customer support system using generative artificial intelligence and emotion recognition technology. The specific implementation of this system is described below.

[0810] System Configuration

[0811] hardware

[0812] Servers: Multiple servers are used to receive, analyze, classify, and generate responses to query data.

[0813] Terminal: A device used to access the system, including smartphones and smart glasses.

[0814] Users include customers of physical stores and end users who utilize customer support.

[0815] software

[0816] Natural language processing (NLP) libraries such as spaCy and transformers will be used.

[0817] Generative artificial intelligence: Uses generative AI models such as GPT-4.

[0818] Sentiment recognition engine: Uses sentiment analysis tools that utilize dictionary-based or machine learning-based algorithms.

[0819] Database: Use an RDBMS or NoSQL database to store query data and generated responses.

[0820] System operation

[0821] 1. Receiving inquiry data:

[0822] Users can make inquiries in real time at physical stores via their smartphones or smart glasses.

[0823] The terminal sends this to the server, and the query data is received.

[0824] 2. Analysis and classification of inquiries:

[0825] The server preprocesses the received query data and removes unnecessary information.

[0826] We use NLP models to analyze the content of inquiries and extract their intent and keywords.

[0827] 3. Emotion recognition:

[0828] The server uses an emotion recognition engine to identify the user's emotions from the inquiry data.

[0829] For example, a user's emotions might be quantified as "70% anger."

[0830] 4. Automated responses using generative artificial intelligence:

[0831] The server passes the query details to a generative artificial intelligence system based on the analysis results and sentiment data, and generates an automated response.

[0832] 5. Sending and displaying responses:

[0833] The server sends the generated response to the terminal, which then displays it to the user.

[0834] 6. Handover to a human operator:

[0835] In particular, if the emotion recognition engine determines that the user's emotions are negative, the server automatically hands over the inquiry to a human operator.

[0836] Specific example

[0837] Example 1:

[0838] Inquiry: "Where can I use Wi-Fi in the store?"

[0839] Generated response: "Wi-Fi is available in all areas of the store."

[0840] Example 2:

[0841] Inquiry: "I am truly dissatisfied with the service at your store."

[0842] Emotional analysis result: Anger level 80%

[0843] Countermeasure: The case will be immediately handed over to a human operator.

[0844] Example of a prompt:

[0845] Customer inquiry: "Where can I use Wi-Fi in the store?"

[0846] AI response: "Wi-Fi is available in all areas of the store."

[0847] or

[0848] Customer inquiry: "I am truly dissatisfied with the service at your store."

[0849] AI response: "First, to resolve your concerns, our store manager will speak with you."

[0850] This system will significantly improve the efficiency of customer support and increase customer satisfaction. It will also enable smoother, real-time customer service in physical stores.

[0851] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0852] Step 1:

[0853] A user uses a smartphone or smart glasses to make an inquiry within a physical store. For example, a customer might ask, "Where can I get Wi-Fi?" The device receives this inquiry data and sends it to a server. The input includes the text data of the inquiry and the user's basic information. The output is the inquiry data sent to the server.

[0854] Step 2:

[0855] The server preprocesses the received query data. During this process, unnecessary information (e.g., HTML tags, special characters) is removed, and the data is converted into a format that can be parsed as text. The input includes the original query text. The output is clean text data.

[0856] Step 3:

[0857] The server inputs pre-processed data into a natural language processing (NLP) model. The server uses this model to analyze the query content and extract its intent and keywords. For example, libraries such as spaCy or transformers are used. Clean text data is passed to the NLP model as input. The output is the analyzed intent and keywords.

[0858] Step 4:

[0859] The server inputs the analysis results and pre-processed data into the emotion recognition engine. This engine identifies the user's emotion (e.g., joy, anger, sadness) from the query text. The input includes the analysis results and clean text data. The output is numerical data representing the user's emotion (e.g., anger level 70%).

[0860] Step 5:

[0861] The server determines the urgency of an inquiry based on emotion data obtained from an emotion recognition engine. For example, if the "anger level" exceeds a certain threshold, the inquiry is classified as highly urgent. The input includes emotion data. The output is a classification of the inquiry's urgency.

[0862] Step 6:

[0863] The server determines whether it is possible to generate an automated response to the inquiry. If an automated response is possible, the server passes the inquiry details to a generative artificial intelligence (e.g., GPT-4) to generate a response. The input includes the NLP analysis results and the urgency assessment results. The output is the generated response.

[0864] Step 7:

[0865] The server sends the generated response to the terminal, which then displays it to the user. The input includes the generated response and the user's query data. The output provides the response that is displayed to the user.

[0866] Step 8:

[0867] The server handles the transfer of high-priority or emotionally charged inquiries to human operators. The server provides the operator with necessary inquiry history and relevant information. Inputs include the urgency assessment result and the user's inquiry data. Outputs provide the inquiry information to be transferred to the human operator.

[0868] 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.

[0869] 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.

[0870] 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.

[0871] [Third Embodiment]

[0872] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0873] 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.

[0874] 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).

[0875] 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.

[0876] 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.

[0877] 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).

[0878] 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.

[0879] 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.

[0880] 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.

[0881] 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.

[0882] 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.

[0883] 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".

[0884] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence. The system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The necessary embodiments for carrying out this invention are described in detail below.

[0885] 1. Receiving inquiry data

[0886] When a user submits an inquiry via a web form, email, or chat system,

[0887] The server receives this.

[0888] The data received includes inquiry details and user contact information.

[0889] 2. Analysis and classification of inquiry content

[0890] The server preprocesses the query data it receives.

[0891] During preprocessing, unnecessary information (such as HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0892] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[0893] NLP models are implemented using libraries such as spaCy or transformers.

[0894] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[0895] The server classifies the query content based on the analysis results.

[0896] For example, these can be categorized into general questions, requests for technical support, and emotional complaints.

[0897] 3. Automated responses using generative AI

[0898] Extract queries that the server can automatically respond to.

[0899] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[0900] The generative artificial intelligence generates the answer and returns it to the server.

[0901] The server sends the generated response to the user.

[0902] For example, in response to an inquiry such as "Please tell me more about the data plans," the response would be in the format of "The following types of data plans are currently available..."

[0903] The server saves the query content and the generated response to the database.

[0904] 4. Handover to a human operator

[0905] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[0906] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0907] In this process, the server provides the operator with the necessary inquiry history and related information.

[0908] We will provide appropriate responses to inquiries that have been handed over to human operators.

[0909] As an example, let's consider the processing flow when a user inquires, "Please tell me about the latest mobile phone plans."

[0910] 1. The user submits an inquiry via a web form.

[0911] 2. The server receives this and performs preprocessing.

[0912] 3. The server analyzes the inquiry using an NLP model and classifies it as "capable of automated response."

[0913] 4. The server sends the question to the generative AI and instructs it to generate an answer.

[0914] 5. The generative artificial intelligence generates the answer and returns it to the server.

[0915] 6. The server sends the response to the user.

[0916] 7. The server saves all interactions to the database.

[0917] This allows for quick and accurate responses to user inquiries, as well as the ability to flexibly transfer them to human operators when necessary. This system is expected to improve the efficiency of customer support operations and enhance customer satisfaction.

[0918] The following describes the processing flow.

[0919] Step 1:

[0920] A user submits an inquiry. Inquiries may be submitted via a web form, email, or chat system.

[0921] Step 2:

[0922] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information (email address and phone number).

[0923] Step 3:

[0924] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[0925] Step 4:

[0926] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[0927] Step 5:

[0928] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[0929] Step 6:

[0930] The server classifies the inquiry based on the analysis results. Classification categories include general questions, requests for technical support, and emotional complaints.

[0931] Step 7:

[0932] The server extracts queries that it has classified as "capable of automatic response."

[0933] Step 8:

[0934] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[0935] Step 9:

[0936] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[0937] Step 10:

[0938] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[0939] Step 11:

[0940] The server saves all interactions to a database. This database includes query details, generated responses, user information, and timestamps.

[0941] Step 12:

[0942] The server extracts inquiries that it has classified as requiring human intervention.

[0943] Step 13:

[0944] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[0945] Step 14:

[0946] The server provides the operator with the inquiry history and related information.

[0947] Step 15:

[0948] For inquiries taken over by human operators, specific actions will be taken.

[0949] These steps streamline customer support operations by automatically classifying and processing inquiries, and only handing them over to human operators when necessary.

[0950] (Example 1)

[0951] 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."

[0952] Conventional customer support systems have often failed to adequately streamline inquiry processing, leading to decreased customer satisfaction. Furthermore, automated response systems struggled to handle complex inquiries, placing a burden on human operators. This invention aims to solve these problems by providing an efficient inquiry processing system utilizing generative artificial intelligence.

[0953] 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.

[0954] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for natural language processing to analyze the content of the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for storing the inquiry data and the generated response in a database, means for transferring the inquiry to a human operator based on the analysis results, means for sending a notification to an internal operation system, and means for removing unnecessary information in order to convert the inquiry content into an analyzable format. This enables increased efficiency in customer support operations and improved customer satisfaction.

[0955] "Inquiry data" refers to information that users send to the customer support system, including questions, requests, and complaints.

[0956] "Preprocessing" is the process of removing unnecessary information in order to convert query data into a parseable format.

[0957] "Natural language processing (NLP) techniques" refer to technologies for analyzing received query data and understanding its content, and include text analysis and keyword extraction.

[0958] A "classification method" is a process that classifies query content into a specific category based on the results of analysis using natural language processing methods.

[0959] "Generative artificial intelligence (AI)" is an artificial intelligence technology that automatically generates responses based on the content of inquiries.

[0960] "Methods for generating answers" refers to the process of generating appropriate answers to inquiries using generative artificial intelligence.

[0961] "Means of submitting responses" refers to the process of sending the generated responses to the user.

[0962] A "database" is a digital storage system for storing query data and generated responses.

[0963] "Methods for transferring inquiries" refers to the process of transferring inquiries that are difficult for generative artificial intelligence to handle to human operators.

[0964] "Methods for sending notifications" refers to the process of notifying human operators of important inquiries or inquiries that are difficult for generative artificial intelligence to handle.

[0965] "Converting to a parsable format" means transforming query data into a form that is easy to process, and this includes removing unnecessary information.

[0966] This invention is a system aimed at improving the efficiency of inquiry processing and customer satisfaction in customer support systems. Specific embodiments are described in detail below.

[0967] System Configuration

[0968] This system is primarily composed of the following hardware and software.

[0969] Server: Receives, processes, analyzes, generates responses, stores data, and sends notifications.

[0970] Device: The device that the user uses to submit an inquiry (e.g., a personal computer or smartphone).

[0971] software:

[0972] Natural Language Processing (NLP) libraries (e.g., spaCy, transformers)

[0973] Generative artificial intelligence (AI) models (e.g., GPT-4)

[0974] Database management systems (e.g., MySQL, PostgreSQL)

[0975] Description of specific embodiments

[0976] Receiving inquiry data

[0977] The user submits an inquiry using their device via a web form, email, or chat system. During this process, data containing the inquiry details and the user's contact information is generated. The server receives this data.

[0978] Pre-processing of inquiry content

[0979] The query data received by the server is preprocessed. Specifically, the software's regular expression library and string manipulation functions are used to remove unnecessary information (e.g., HTML tags, special characters, unnecessary line breaks and spaces) and convert it into a parseable format.

[0980] Analysis and classification of inquiry content

[0981] The server passes pre-processed query data to a natural language processing (NLP) model (e.g., spaCy, transformers) for analysis. The NLP model extracts the intent and key keywords of the query. Based on this analysis, the server classifies the query into a specific category (e.g., general question, request for technical support, emotional complaint, etc.).

[0982] Generation of automated responses using generative AI

[0983] The server extracts queries that can be answered automatically. If an automatic response is applicable, the server inputs the query details into a generative artificial intelligence (e.g., GPT-4) to create an appropriate prompt. For example, it might use a prompt like this:

[0984] "Please tell me about the latest mobile phone plans."

[0985] The server inputs this prompt into a generative AI model, which then generates a response. The generative artificial intelligence generates an appropriate response and returns it to the server.

[0986] Submitting responses and saving data

[0987] The server sends the generated response to the user. An example of the response might be, "The following data plans are currently available to you..." The server also saves the inquiry and the generated response to a database. The information saved includes the inquiry, analysis results, generated response, user contact information, date and time, etc.

[0988] Handover to human operator

[0989] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends a notification to the internal operation system to transfer these inquiries to human operators. The server provides the operators with the necessary inquiry history and related information. Human operators then take appropriate action on the inquiries they receive.

[0990] The above describes a specific embodiment of the present invention. This system enables increased efficiency in customer support operations and improved customer satisfaction.

[0991] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0992] Step 1:

[0993] The user submits an inquiry. The user enters and submits their inquiry and contact information using a web form, email, or chat system.

[0994] Input: Inquiry details, user contact information

[0995] Output: Query data received by the server

[0996] Step 2:

[0997] The server receives the query data. The server receives the query data sent by the user and saves it as a file.

[0998] Input: Inquiry data submitted by the user

[0999] Output: Received query data

[1000] Step 3:

[1001] The server preprocesses the query data. To convert the received query data into a parseable format, it removes unnecessary information (HTML tags, special characters, etc.) and converts it into text format.

[1002] Input: Received query data

[1003] Output: Preprocessed query data

[1004] Specific operation: Uses a regular expression library to remove HTML tags and special characters.

[1005] Step 4:

[1006] The server passes the pre-processed query data to a natural language processing (NLP) model for analysis. The NLP model (e.g., spaCy, transformers) is used to extract the intent and key keywords of the query.

[1007] Input: Preprocessed query data

[1008] Output: Analysis results (intent, important keywords)

[1009] Specific operation: Input data into the NLP model and obtain the analysis results.

[1010] Step 5:

[1011] The server classifies the inquiry based on the analysis results. The analyzed inquiry data is categorized into specific categories (e.g., general questions, technical support, complaints, etc.).

[1012] Input: Analysis results

[1013] Output: Classified query data

[1014] Specific operation: Based on the analysis results, conditional branching is performed, and data is distributed to each category.

[1015] Step 6:

[1016] The server inputs a prompt message to a generative artificial intelligence (AI) and generates an automated response. If applicable, the generative AI (e.g., GPT-4) creates a prompt message based on the analyzed query content and generates a response.

[1017] Input: Classified query data, prompt text

[1018] Output: Auto-generated answer

[1019] Specific operation: Input a prompt sentence to a generative AI model and generate an appropriate response.

[1020] Step 7:

[1021] The server sends the generated response to the user. The generated response is also sent to the user's contacts (e.g., email, chat).

[1022] Input: Automated response, user contact information

[1023] Output: Response sent to the user

[1024] Specific actions: Send responses using an email system or chat system.

[1025] Step 8:

[1026] The server saves the query content and the generated response to the database. All query content and its response are recorded in the database and can be referenced later.

[1027] Input: Inquiry details, analysis results, generated response

[1028] Output: Data stored in the database

[1029] Specific operation: Use a database management system to save data.

[1030] Step 9:

[1031] The server transfers inquiries it cannot handle to human operators. If the generative AI cannot process the inquiry, it sends a notification to the internal operations system.

[1032] Input: Classified query data

[1033] Output: Inquiry details handed over to the operator, notification sent.

[1034] Specific action: Send a notification to the internal operating system and display it on the operator's terminal.

[1035] (Application Example 1)

[1036] 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."

[1037] Current customer support systems fail to adequately optimize resources and improve customer satisfaction. In particular, there are problems with delayed inquiry processing and inability to provide appropriate responses. Furthermore, the limited availability of automated responses to inquiries places a heavy burden on human operators. This often leads to customer dissatisfaction and a potential decline in customer satisfaction.

[1038] 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.

[1039] In this invention, the server includes means for receiving inquiry data, means for natural language processing to analyze the content of the inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for transferring the inquiry to a human operator based on the analysis results, means for storing the inquiry content and its response in a database, means including a smartphone application for receiving inquiries, means for performing automated responses using generative artificial intelligence within the application, and means for notifying the operator if the inquiry is complex. This enables improved resource efficiency and faster response times, and is expected to improve customer satisfaction.

[1040] "Inquiry data" refers to questions, requests, and related information sent by users to the support center.

[1041] "Natural language processing" is the technology that enables computers to understand and analyze human language.

[1042] "Generative artificial intelligence" refers to a system that uses artificial intelligence technology to automatically generate answers to user inquiries.

[1043] A "smartphone application" is software specifically designed to run on a smartphone.

[1044] "Automated response" refers to a function that automatically provides pre-programmed answers using generative artificial intelligence.

[1045] "Preprocessing" refers to the process of removing unnecessary information from data and converting it into an analyzable format.

[1046] A "database" is a system for organizing and efficiently managing data.

[1047] "Notify operator" refers to the system sending a notification to a human operator to request assistance when the inquiry is complex.

[1048] This invention aims to streamline customer support systems and improve customer satisfaction. The details of a system that receives inquiry data, generates automated responses, and, when necessary, transfers the inquiry to a human operator are described below.

[1049] Hardware and software to use

[1050] This system consists of a server, a smartphone, and the necessary software.

[1051] Hardware: Servers, cloud infrastructure (AWS, GCP, Azure), smartphones

[1052] Software: Flask, spaCy, transformers, SQLite

[1053] Receiving and preprocessing query data

[1054] 1. Received:

[1055] Users submit questions via chat, email, or contact forms through a smartphone application. The server then receives the inquiry data.

[1056] 2. Pre-processing:

[1057] The server preprocesses the query data it receives. SpaCy is used to remove unnecessary information (e.g., HTML tags and special characters) and convert it into a format that can be parsed as text.

[1058] Analysis and classification

[1059] 3. Natural Language Processing:

[1060] The server passes pre-processed data to a natural language processing (NLP) model to analyze the query. Libraries such as spaCy and transformers are used for this process.

[1061] 4. Classification:

[1062] The server classifies the inquiry based on the analysis results. For example, it might classify it as a general question, a request for technical support, or an emotional complaint.

[1063] Automatic response generation

[1064] 5. Generative Artificial Intelligence:

[1065] The server inputs the query details into a generative artificial intelligence (e.g., GPT-3.5-turbo) and generates a response. If an automated response is applicable, this response is generated.

[1066] 6. Submit your response:

[1067] The server sends the generated response to the user. For example, in response to the inquiry, "I want to know about the new mobile phone plans," the server sends the response, "The following types of mobile phone plans are currently available..."

[1068] Handover to operator

[1069] 7. Handover:

[1070] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends notifications to transfer these inquiries to human operators. The notifications include the inquiry history and related information.

[1071] Data storage

[1072] 8. Save:

[1073] The server saves all inquiry details and generated responses to a database (SQLite). This allows for future inquiry handling and analysis.

[1074] As a concrete example of a prompt message, if a user asks, "I want to know about the new mobile phone plan,"

[1075] I want to know about the new mobile phone plans.

[1076] By inputting this sentence into a generative artificial intelligence system, an appropriate answer will be automatically generated.

[1077] As described above, a system is realized that enables improved resource efficiency and faster response times, leading to increased customer satisfaction.

[1078] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1079] Step 1:

[1080] Users submit inquiries through a smartphone application. In this process, users enter their inquiry details via various means, such as chat, email, or a contact form, and then press the submit button. The input data includes the user's question or request, as well as contact information. The server receives this inquiry data.

[1081] Step 2:

[1082] The server preprocesses the query data it receives. This preprocessing uses spaCy to remove unnecessary information (such as HTML tags and special characters) from the data and convert it into a parseable format. The output of this preprocessing is cleaned text data.

[1083] Step 3:

[1084] The server passes the pre-processed text data to a natural language processing (NLP) model. The NLP model (e.g., spaCy or transformers) analyzes the text and extracts important keywords and intent. The analysis results are then output.

[1085] Step 4:

[1086] The server classifies the inquiry based on the analysis results from the NLP model. For example, it might be classified as a general question, technical support, or an emotional complaint. Based on this classification, it determines whether an automated response is possible.

[1087] Step 5:

[1088] If automated response is applicable, the server passes the analysis results as an input prompt to a generative artificial intelligence (e.g., GPT-3.5-turbo) to generate a response. The output of this generative AI is the text response to the user's inquiry.

[1089] Step 6:

[1090] The server sends the generated response to the user. The response is sent via chat, email, or notification functions through a smartphone application. At this time, an automatically generated response to the user's inquiry is output.

[1091] Step 7:

[1092] For inquiries that are difficult for generative artificial intelligence to handle, the server will transfer the inquiry to a human operator. The server will notify the operator of the inquiry details and send the history and related information along with it. This transfer to an operator makes it possible to handle even complex inquiries.

[1093] Step 8:

[1094] The server saves all inquiry details, generated responses, and response history to a database (SQLite). This data is used for future inquiry handling and analysis. The saved data includes inquiry text, response text, user information, and related metadata.

[1095] 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.

[1096] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence and an emotion engine. This system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The embodiments for carrying out this invention are described in detail below.

[1097] 1. Receiving inquiry data

[1098] When a user submits an inquiry via a web form, email, or chat system,

[1099] The server receives this.

[1100] The data received includes inquiry details and user contact information.

[1101] 2. Analysis and classification of inquiry content

[1102] The server preprocesses the query data it receives.

[1103] During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[1104] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[1105] NLP models are implemented using libraries such as spaCy or transformers.

[1106] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[1107] 3. Emotion recognition by an emotion engine

[1108] The server passes the analysis results and pre-processed data to the emotion engine.

[1109] The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[1110] The output of the emotion engine includes data that quantifies the user's emotions (e.g., anger level 70%).

[1111] The server uses the emotions recognized by the emotion engine to determine the urgency of the inquiry and the appropriate course of action.

[1112] 4. Automated responses using generative AI

[1113] Extract queries that the server can automatically respond to.

[1114] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[1115] The generative artificial intelligence generates the answer and returns it to the server.

[1116] The server sends the generated response to the user.

[1117] For example, in response to an inquiry such as "Please tell me the details of the data plan," the response would be "The data plans currently available to you are as follows..."

[1118] The server saves the query content and the generated response to the database.

[1119] 5. Handover to a human operator

[1120] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[1121] In particular, if the emotion engine identifies the user's emotion as "negative," the server automatically initiates a process to hand over the user to a human operator.

[1122] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[1123] In this process, the server provides the operator with the necessary inquiry history and related information.

[1124] For inquiries taken over by human operators, specific actions will be taken.

[1125] As an example, let's consider the processing flow when a user contacts us saying, "I am truly disappointed with this service. Please take action."

[1126] 1. The user submits an inquiry through the chat system.

[1127] 2. The server receives this and performs preprocessing.

[1128] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[1129] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[1130] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[1131] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[1132] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[1133] By combining this emotion engine, it becomes possible to provide appropriate and rapid responses tailored to the user's emotions, especially to users experiencing negative emotions. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[1134] The following describes the processing flow.

[1135] Step 1:

[1136] A user submits an inquiry. Inquiries can be submitted via web form, email, or chat system.

[1137] Step 2:

[1138] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information.

[1139] Step 3:

[1140] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[1141] Step 4:

[1142] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[1143] Step 5:

[1144] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[1145] Step 6:

[1146] The server passes the analysis results to the emotion engine. The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[1147] Step 7:

[1148] The emotion engine quantifies the user's emotions and returns the result (e.g., anger level 70%) to the server.

[1149] Step 8:

[1150] The server receives the output from the emotion engine and determines the urgency of the inquiry and the appropriate course of action.

[1151] Step 9:

[1152] The server classifies the inquiry based on the analysis results and the sentiment engine's findings. Classification categories include general questions, requests for technical support, and emotional complaints.

[1153] Step 10:

[1154] The server extracts queries that it has classified as "capable of automatic response."

[1155] Step 11:

[1156] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[1157] Step 12:

[1158] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[1159] Step 13:

[1160] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[1161] Step 14:

[1162] The server saves the query content and the generated response to a database. The saved content includes the query content, the generated response, user information, and a timestamp.

[1163] Step 15:

[1164] The server extracts inquiries that it has classified as requiring human intervention. In particular, if the sentiment engine identifies the user's sentiment as "negative," this is automatically applied.

[1165] Step 16:

[1166] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[1167] Step 17:

[1168] The server provides the operator with the inquiry history and related information.

[1169] Step 18:

[1170] For inquiries taken over by human operators, specific actions are taken. For example, for users with negative feelings, follow-up emails or phone calls are provided.

[1171] As a concrete example, let's consider the flow of a case where a user contacts us saying, "I am truly disappointed with this service. Please take action."

[1172] 1. The user submits an inquiry through the chat system.

[1173] 2. The server receives this and performs preprocessing.

[1174] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[1175] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[1176] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[1177] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[1178] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[1179] In this way, this system improves the efficiency and quality of customer support by combining automated responses with appropriate handover based on emotion recognition.

[1180] (Example 2)

[1181] 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."

[1182] To improve efficiency and customer satisfaction in customer support, a system that processes inquiry data quickly and accurately is necessary. However, conventional systems have the problem of being unable to keep up with processing a large volume of inquiries, leading to increased customer dissatisfaction. Furthermore, they do not adequately recognize emotions, making it difficult to provide appropriate responses that are in line with the customer's feelings. This invention aims to solve these problems.

[1183] 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.

[1184] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for performing natural language processing using the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for performing emotion recognition, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for handing over the inquiry to a human operator based on the analysis results and emotion recognition results, and means for storing the inquiry data and the generated response in a database. This enables rapid and accurate processing of inquiry data and further realizes appropriate responses that are in line with the customer's emotions.

[1185] "Inquiry data" refers to data containing information such as questions, requests, and complaints sent by users to customer support.

[1186] "Preprocessing" refers to the process of removing unnecessary information from received query data and converting it into a format that can be processed by natural language processing.

[1187] "Natural language processing" is a technology that uses computers to analyze human language and understand its meaning.

[1188] "Emotion recognition" is a technology that identifies and quantifies emotions such as joy, anger, and sadness from the content of user inquiries.

[1189] "Generative artificial intelligence" refers to machine learning models used to generate appropriate answers based on the content of inquiries.

[1190] "Automated response" refers to the process in which a server sends a response generated by a generative artificial intelligence system to the user.

[1191] A "human operator" is a human worker who directly handles inquiries that the server cannot process.

[1192] A "database" is an information management system used to store query data and generated responses.

[1193] "Analysis results" refers to the analysis of inquiry data obtained through natural language processing and sentiment recognition.

[1194] "Handing over" refers to the process by which a server notifies a human operator of an inquiry requiring attention and entrusts them with handling it.

[1195] This invention is a system aimed at improving the efficiency of resources in customer support and enhancing customer satisfaction, and it automatically processes inquiry data using generative artificial intelligence and an emotion recognition engine. Embodiments of this system are described in detail below.

[1196] First, the user submits an inquiry via a web form, email, or chat system. For example, consider a case where a user submits a request via a web form asking, "How do I use this product?" This inquiry data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address.

[1197] Next, the server receives the transmitted data at the specified endpoint. This received data is preprocessed using regular expressions and text cleaning libraries (e.g., BeautifulSoup4 in Python) to remove HTML tags, special characters, and unnecessary whitespace. The preprocessed text is then converted into a parseable format.

[1198] The server passes the pre-processed query data to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers, and it analyzes the data to extract the query's intent and important keywords. For example, a query like "Please tell me how to use the product" would be classified as "Information Provision."

[1199] Furthermore, the server passes the analysis results and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob, and identifies and quantifies emotions such as joy, anger, and sadness from the user's text. For example, it outputs quantified data such as "Anger level 70%". The server receives the output from the emotion engine and determines the urgency of the inquiry and how to respond.

[1200] The server passes an automated response-capable inquiry to a generative artificial intelligence (e.g., GPT-4) to generate a response. An example prompt might be: "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response." The generative AI analyzes the text and generates an appropriate response. The generated response is returned to the server, which then sends it to the user. For example, a specific response such as "To use this product, please follow these steps..." might be generated.

[1201] Furthermore, the server stores the query content and the generated response in a database. This allows for future reference and analysis.

[1202] On the other hand, the server also extracts inquiries that are difficult for generative artificial intelligence to handle and hands them over to human operators. In particular, if the emotion recognition engine highly values ​​"negative emotions," the server automatically hands over the inquiry to a human operator. In this process, notifications are sent via the internal operating system, providing the necessary inquiry history and related information.

[1203] This enables the rapid and accurate processing of inquiry data, as well as appropriate responses that are tailored to the customer's feelings. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[1204] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1205] Step 1: Receiving inquiry data

[1206] A user submits an inquiry via a web form, email, or chat system. For example, a user might submit "How do I use this product?" through a web form. The server receives this at a specified endpoint. The input data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address. The server performs initial database processing to store this data.

[1207] Step 2: Pre-processing of inquiry content

[1208] The server preprocesses the query data it receives. The input is the query data received in step 1. Specific preprocessing actions include removing HTML tags, special characters, and unnecessary whitespace. This is done using regular expressions and the Python BeautifulSoup4 library, among others. The output is data with the text and metadata cleaned and converted into a parseable format.

[1209] Step 3: Analysis and classification of inquiries

[1210] The server passes pre-processed data as input to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers. The server runs the NLP model to extract the intent of the query and important keywords. For example, the query "Please tell me how to use the product" would be classified as "Information Provision." The output is data containing the analysis results and classification results.

[1211] Step 4: Emotion recognition by the emotion engine

[1212] The server passes the analysis results of the NLP model and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob. Here, the input is the analysis results and classification results obtained in step 3. The emotion engine identifies emotions such as joy, anger, and sadness from the user's text and outputs quantified data (e.g., "Anger level 70%)". The server receives this output and makes decisions regarding the urgency of the response and how to respond.

[1213] Step 5: Automated response by generative AI

[1214] The server passes a query that can be automatically answered to a generative artificial intelligence (e.g., GPT-4). The input is pre-processed query data and analysis results. The generative AI is given prompts to generate a response. For example, a prompt such as "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response" can be used. The generative AI generates a response (e.g., "For instructions on how to use this product, please follow these steps...") and returns it to the server. The server receives the generated response and sends it to the user. The output is the response text sent to the user.

[1215] Step 6: Handover to human operator

[1216] The server extracts inquiries that are difficult for generative artificial intelligence to handle. In particular, if the emotion engine identifies the user's emotion as "negative," the server initiates a process to hand over the inquiry to a human operator. The input here is the output from the emotion recognition engine and the content of the inquiry that the generative artificial intelligence could not handle. The server sends a notification to the internal operating system to hand over these inquiries to human operators. The output is the notification to the human operator and related inquiry data.

[1217] As described above, this system processes inquiry data step by step, efficiently managing and processing it. This enables improved customer satisfaction and more efficient customer support.

[1218] (Application Example 2)

[1219] 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."

[1220] Current customer support systems consume a significant amount of resources in handling inquiries, necessitating greater resource efficiency. Furthermore, accurately understanding customer emotions and providing appropriate responses is challenging, hindering customer satisfaction. Additionally, in real-time customer service at physical stores, staff may not always be able to provide the most suitable answer immediately, potentially detracting from the customer experience.

[1221] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for receiving inquiry data, means for natural language processing for analyzing the content of the inquiry data, means for classifying the content of the inquiry based on the analysis results, means for generating an answer to the inquiry using generative artificial intelligence, means for transmitting the generated answer, means for transferring the inquiry to a human operator based on the analysis results, means for emotion recognition, means for determining the urgency of the inquiry based on the output result of the emotion recognition means, and means for displaying the generated answer to customer service in the store. This enables efficient use of resources in customer support and improved customer satisfaction. Furthermore, it enables smooth real-time customer service in physical stores and improves the customer experience.

[1222] "Inquiry data" refers to data that includes the content of questions and requests submitted by customers.

[1223] "Means of receiving" refers to the means by which the system takes in query data.

[1224] "Natural language processing means" refers to technologies used to analyze query data and understand its content.

[1225] "Generative artificial intelligence" refers to AI technology that automatically generates answers to inquiries.

[1226] "Emotion recognition means" refers to technology used to identify customer emotions from inquiry data.

[1227] "Means of classification" refers to methods for categorizing inquiry content based on analysis results.

[1228] "Means for determining urgency" refers to means for determining the urgency of responding to an inquiry based on the output results of the emotion recognition means.

[1229] "Means of transmission" refers to the means of communicating the generated response to the customer.

[1230] "Means of transferring inquiries to human operators" refers to a method of automatically transferring inquiries to human operators when action is required.

[1231] "Means for displaying responses generated in response to customer inquiries within a store" refers to means for displaying responses generated in response to customer inquiries within a physical store.

[1232] This invention is a real-time customer support system using generative artificial intelligence and emotion recognition technology. The specific implementation of this system is described below.

[1233] System Configuration

[1234] hardware

[1235] Servers: Multiple servers are used to receive, analyze, classify, and generate responses to query data.

[1236] Terminal: A device used to access the system, including smartphones and smart glasses.

[1237] Users include customers of physical stores and end users who utilize customer support.

[1238] software

[1239] Natural language processing (NLP) libraries such as spaCy and transformers will be used.

[1240] Generative artificial intelligence: Uses generative AI models such as GPT-4.

[1241] Sentiment recognition engine: Uses sentiment analysis tools that utilize dictionary-based or machine learning-based algorithms.

[1242] Database: Use an RDBMS or NoSQL database to store query data and generated responses.

[1243] System operation

[1244] 1. Receiving inquiry data:

[1245] Users can make inquiries in real time at physical stores via their smartphones or smart glasses.

[1246] The terminal sends this to the server, and the query data is received.

[1247] 2. Analysis and classification of inquiries:

[1248] The server preprocesses the received query data and removes unnecessary information.

[1249] We use NLP models to analyze the content of inquiries and extract their intent and keywords.

[1250] 3. Emotion recognition:

[1251] The server uses an emotion recognition engine to identify the user's emotions from the inquiry data.

[1252] For example, a user's emotions might be quantified as "70% anger."

[1253] 4. Automated responses using generative artificial intelligence:

[1254] The server passes the query details to a generative artificial intelligence system based on the analysis results and sentiment data, and generates an automated response.

[1255] 5. Sending and displaying responses:

[1256] The server sends the generated response to the terminal, which then displays it to the user.

[1257] 6. Handover to a human operator:

[1258] In particular, if the emotion recognition engine determines that the user's emotions are negative, the server automatically hands over the inquiry to a human operator.

[1259] Specific example

[1260] Example 1:

[1261] Inquiry: "Where can I use Wi-Fi in the store?"

[1262] Generated response: "Wi-Fi is available in all areas of the store."

[1263] Example 2:

[1264] Inquiry: "I am truly dissatisfied with the service at your store."

[1265] Emotional analysis result: Anger level 80%

[1266] Countermeasure: The case will be immediately handed over to a human operator.

[1267] Example of a prompt:

[1268] Customer inquiry: "Where can I use Wi-Fi in the store?"

[1269] AI response: "Wi-Fi is available in all areas of the store."

[1270] or

[1271] Customer inquiry: "I am truly dissatisfied with the service at your store."

[1272] AI response: "First, to resolve your concerns, our store manager will speak with you."

[1273] This system will significantly improve the efficiency of customer support and increase customer satisfaction. It will also enable smoother, real-time customer service in physical stores.

[1274] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1275] Step 1:

[1276] A user uses a smartphone or smart glasses to make an inquiry within a physical store. For example, a customer might ask, "Where can I get Wi-Fi?" The device receives this inquiry data and sends it to a server. The input includes the text data of the inquiry and the user's basic information. The output is the inquiry data sent to the server.

[1277] Step 2:

[1278] The server preprocesses the received query data. During this process, unnecessary information (e.g., HTML tags, special characters) is removed, and the data is converted into a format that can be parsed as text. The input includes the original query text. The output is clean text data.

[1279] Step 3:

[1280] The server inputs pre-processed data into a natural language processing (NLP) model. The server uses this model to analyze the query content and extract its intent and keywords. For example, libraries such as spaCy or transformers are used. Clean text data is passed to the NLP model as input. The output is the analyzed intent and keywords.

[1281] Step 4:

[1282] The server inputs the analysis results and pre-processed data into the emotion recognition engine. This engine identifies the user's emotion (e.g., joy, anger, sadness) from the query text. The input includes the analysis results and clean text data. The output is numerical data representing the user's emotion (e.g., anger level 70%).

[1283] Step 5:

[1284] The server determines the urgency of an inquiry based on emotion data obtained from an emotion recognition engine. For example, if the "anger level" exceeds a certain threshold, the inquiry is classified as highly urgent. The input includes emotion data. The output is a classification of the inquiry's urgency.

[1285] Step 6:

[1286] The server determines whether it is possible to generate an automated response to the inquiry. If an automated response is possible, the server passes the inquiry details to a generative artificial intelligence (e.g., GPT-4) to generate a response. The input includes the NLP analysis results and the urgency assessment results. The output is the generated response.

[1287] Step 7:

[1288] The server sends the generated response to the terminal, which then displays it to the user. The input includes the generated response and the user's query data. The output provides the response that is displayed to the user.

[1289] Step 8:

[1290] The server handles the transfer of high-priority or emotionally charged inquiries to human operators. The server provides the operator with necessary inquiry history and relevant information. Inputs include the urgency assessment result and the user's inquiry data. Outputs provide the inquiry information to be transferred to the human operator.

[1291] 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.

[1292] 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.

[1293] 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.

[1294] [Fourth Embodiment]

[1295] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1296] 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.

[1297] 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).

[1298] 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.

[1299] 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.

[1300] 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).

[1301] 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.

[1302] 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.

[1303] 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.

[1304] 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.

[1305] 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.

[1306] 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.

[1307] 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".

[1308] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence. The system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The necessary embodiments for carrying out this invention are described in detail below.

[1309] 1. Receiving inquiry data

[1310] When a user submits an inquiry via a web form, email, or chat system,

[1311] The server receives this.

[1312] The data received includes inquiry details and user contact information.

[1313] 2. Analysis and classification of inquiry content

[1314] The server preprocesses the query data it receives.

[1315] During preprocessing, unnecessary information (such as HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[1316] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[1317] NLP models are implemented using libraries such as spaCy or transformers.

[1318] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[1319] The server classifies the query content based on the analysis results.

[1320] For example, these can be categorized into general questions, requests for technical support, and emotional complaints.

[1321] 3. Automated responses using generative AI

[1322] Extract queries that the server can automatically respond to.

[1323] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[1324] The generative artificial intelligence generates the answer and returns it to the server.

[1325] The server sends the generated response to the user.

[1326] For example, in response to an inquiry such as "Please tell me more about the data plans," the response would be in the format of "The following types of data plans are currently available..."

[1327] The server saves the query content and the generated response to the database.

[1328] 4. Handover to a human operator

[1329] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[1330] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[1331] In this process, the server provides the operator with the necessary inquiry history and related information.

[1332] We will provide appropriate responses to inquiries that have been handed over to human operators.

[1333] As an example, let's consider the processing flow when a user inquires, "Please tell me about the latest mobile phone plans."

[1334] 1. The user submits an inquiry via a web form.

[1335] 2. The server receives this and performs preprocessing.

[1336] 3. The server analyzes the inquiry using an NLP model and classifies it as "capable of automated response."

[1337] 4. The server sends the question to the generative AI and instructs it to generate an answer.

[1338] 5. The generative artificial intelligence generates the answer and returns it to the server.

[1339] 6. The server sends the response to the user.

[1340] 7. The server saves all interactions to the database.

[1341] This allows for quick and accurate responses to user inquiries, as well as the ability to flexibly transfer them to human operators when necessary. This system is expected to improve the efficiency of customer support operations and enhance customer satisfaction.

[1342] The following describes the processing flow.

[1343] Step 1:

[1344] A user submits an inquiry. Inquiries may be submitted via a web form, email, or chat system.

[1345] Step 2:

[1346] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information (email address and phone number).

[1347] Step 3:

[1348] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[1349] Step 4:

[1350] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[1351] Step 5:

[1352] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[1353] Step 6:

[1354] The server classifies the inquiry based on the analysis results. Classification categories include general questions, requests for technical support, and emotional complaints.

[1355] Step 7:

[1356] The server extracts queries that it has classified as "capable of automatic response."

[1357] Step 8:

[1358] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[1359] Step 9:

[1360] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[1361] Step 10:

[1362] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[1363] Step 11:

[1364] The server saves all interactions to a database. This database includes query details, generated responses, user information, and timestamps.

[1365] Step 12:

[1366] The server extracts inquiries that it has classified as requiring human intervention.

[1367] Step 13:

[1368] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[1369] Step 14:

[1370] The server provides the operator with the inquiry history and related information.

[1371] Step 15:

[1372] For inquiries taken over by human operators, specific actions will be taken.

[1373] These steps streamline customer support operations by automatically classifying and processing inquiries, and only handing them over to human operators when necessary.

[1374] (Example 1)

[1375] 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".

[1376] Conventional customer support systems have often failed to adequately streamline inquiry processing, leading to decreased customer satisfaction. Furthermore, automated response systems struggled to handle complex inquiries, placing a burden on human operators. This invention aims to solve these problems by providing an efficient inquiry processing system utilizing generative artificial intelligence.

[1377] 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.

[1378] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for natural language processing to analyze the content of the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for storing the inquiry data and the generated response in a database, means for transferring the inquiry to a human operator based on the analysis results, means for sending a notification to an internal operation system, and means for removing unnecessary information in order to convert the inquiry content into an analyzable format. This enables increased efficiency in customer support operations and improved customer satisfaction.

[1379] "Inquiry data" refers to information that users send to the customer support system, including questions, requests, and complaints.

[1380] "Preprocessing" is the process of removing unnecessary information in order to convert query data into a parseable format.

[1381] "Natural language processing (NLP) techniques" refer to technologies for analyzing received query data and understanding its content, and include text analysis and keyword extraction.

[1382] A "classification method" is a process that classifies query content into a specific category based on the results of analysis using natural language processing methods.

[1383] "Generative artificial intelligence (AI)" is an artificial intelligence technology that automatically generates responses based on the content of inquiries.

[1384] "Methods for generating answers" refers to the process of generating appropriate answers to inquiries using generative artificial intelligence.

[1385] "Means of submitting responses" refers to the process of sending the generated responses to the user.

[1386] A "database" is a digital storage system for storing query data and generated responses.

[1387] "Methods for transferring inquiries" refers to the process of transferring inquiries that are difficult for generative artificial intelligence to handle to human operators.

[1388] "Methods for sending notifications" refers to the process of notifying human operators of important inquiries or inquiries that are difficult for generative artificial intelligence to handle.

[1389] "Converting to a parsable format" means transforming query data into a form that is easy to process, and this includes removing unnecessary information.

[1390] This invention is a system aimed at improving the efficiency of inquiry processing and customer satisfaction in customer support systems. Specific embodiments are described in detail below.

[1391] System Configuration

[1392] This system is primarily composed of the following hardware and software.

[1393] Server: Receives, processes, analyzes, generates responses, stores data, and sends notifications.

[1394] Device: The device that the user uses to submit an inquiry (e.g., a personal computer or smartphone).

[1395] software:

[1396] Natural Language Processing (NLP) libraries (e.g., spaCy, transformers)

[1397] Generative artificial intelligence (AI) models (e.g., GPT-4)

[1398] Database management systems (e.g., MySQL, PostgreSQL)

[1399] Description of specific embodiments

[1400] Receiving inquiry data

[1401] The user submits an inquiry using their device via a web form, email, or chat system. During this process, data containing the inquiry details and the user's contact information is generated. The server receives this data.

[1402] Pre-processing of inquiry content

[1403] The query data received by the server is preprocessed. Specifically, the software's regular expression library and string manipulation functions are used to remove unnecessary information (e.g., HTML tags, special characters, unnecessary line breaks and spaces) and convert it into a parseable format.

[1404] Analysis and classification of inquiry content

[1405] The server passes pre-processed query data to a natural language processing (NLP) model (e.g., spaCy, transformers) for analysis. The NLP model extracts the intent and key keywords of the query. Based on this analysis, the server classifies the query into a specific category (e.g., general question, request for technical support, emotional complaint, etc.).

[1406] Generation of automated responses using generative AI

[1407] The server extracts queries that can be answered automatically. If an automatic response is applicable, the server inputs the query details into a generative artificial intelligence (e.g., GPT-4) to create an appropriate prompt. For example, it might use a prompt like this:

[1408] "Please tell me about the latest mobile phone plans."

[1409] The server inputs this prompt into a generative AI model, which then generates a response. The generative artificial intelligence generates an appropriate response and returns it to the server.

[1410] Submitting responses and saving data

[1411] The server sends the generated response to the user. An example of the response might be, "The following data plans are currently available to you..." The server also saves the inquiry and the generated response to a database. The information saved includes the inquiry, analysis results, generated response, user contact information, date and time, etc.

[1412] Handover to human operator

[1413] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends a notification to the internal operation system to transfer these inquiries to human operators. The server provides the operators with the necessary inquiry history and related information. Human operators then take appropriate action on the inquiries they receive.

[1414] The above describes a specific embodiment of the present invention. This system enables increased efficiency in customer support operations and improved customer satisfaction.

[1415] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1416] Step 1:

[1417] The user submits an inquiry. The user enters and submits their inquiry and contact information using a web form, email, or chat system.

[1418] Input: Inquiry details, user contact information

[1419] Output: Query data received by the server

[1420] Step 2:

[1421] The server receives the query data. The server receives the query data sent by the user and saves it as a file.

[1422] Input: Inquiry data submitted by the user

[1423] Output: Received query data

[1424] Step 3:

[1425] The server preprocesses the query data. To convert the received query data into a parseable format, it removes unnecessary information (HTML tags, special characters, etc.) and converts it into text format.

[1426] Input: Received query data

[1427] Output: Preprocessed query data

[1428] Specific operation: Uses a regular expression library to remove HTML tags and special characters.

[1429] Step 4:

[1430] The server passes the pre-processed query data to a natural language processing (NLP) model for analysis. The NLP model (e.g., spaCy, transformers) is used to extract the intent and important keywords of the query.

[1431] Input: Preprocessed query data

[1432] Output: Analysis results (intent, important keywords)

[1433] Specific operation: Input data into the NLP model and obtain the analysis results.

[1434] Step 5:

[1435] The server classifies the inquiry based on the analysis results. The analyzed inquiry data is categorized into specific categories (e.g., general questions, technical support, complaints, etc.).

[1436] Input: Analysis results

[1437] Output: Classified query data

[1438] Specific operation: Based on the analysis results, conditional branching is performed, and data is distributed to each category.

[1439] Step 6:

[1440] The server inputs a prompt message to a generative artificial intelligence (AI) and generates an automated response. If applicable, the generative AI (e.g., GPT-4) creates a prompt message based on the analyzed query content and generates a response.

[1441] Input: Classified query data, prompt text

[1442] Output: Auto-generated answer

[1443] Specific operation: Input a prompt sentence to a generative AI model and generate an appropriate response.

[1444] Step 7:

[1445] The server sends the generated response to the user. The generated response is also sent to the user's contacts (e.g., email, chat).

[1446] Input: Automated response, user contact information

[1447] Output: Response sent to the user

[1448] Specific actions: Send responses using an email system or chat system.

[1449] Step 8:

[1450] The server saves the query content and the generated response to the database. All query content and its response are recorded in the database and can be referenced later.

[1451] Input: Inquiry details, analysis results, generated response

[1452] Output: Data stored in the database

[1453] Specific operation: Use a database management system to save data.

[1454] Step 9:

[1455] The server transfers inquiries it cannot handle to human operators. If the generative AI cannot process the inquiry, it sends a notification to the internal operations system.

[1456] Input: Classified query data

[1457] Output: Inquiry details handed over to the operator, notification sent.

[1458] Specific action: Send a notification to the internal operating system and display it on the operator's terminal.

[1459] (Application Example 1)

[1460] 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".

[1461] Current customer support systems fail to adequately optimize resources and improve customer satisfaction. In particular, there are problems with delayed inquiry processing and inability to provide appropriate responses. Furthermore, the limited availability of automated responses to inquiries places a heavy burden on human operators. This often leads to customer dissatisfaction and a potential decline in customer satisfaction.

[1462] 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.

[1463] In this invention, the server includes means for receiving inquiry data, means for natural language processing to analyze the content of the inquiry data, means for classifying the inquiry content based on the analysis results, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for transferring the inquiry to a human operator based on the analysis results, means for storing the inquiry content and its response in a database, means including a smartphone application for receiving inquiries, means for performing automated responses using generative artificial intelligence within the application, and means for notifying the operator if the inquiry is complex. This enables improved resource efficiency and faster response times, and is expected to improve customer satisfaction.

[1464] "Inquiry data" refers to questions, requests, and related information sent by users to the support center.

[1465] "Natural language processing" is the technology that enables computers to understand and analyze human language.

[1466] "Generative artificial intelligence" refers to a system that uses artificial intelligence technology to automatically generate answers to user inquiries.

[1467] A "smartphone application" is software specifically designed to run on a smartphone.

[1468] "Automated response" refers to a function that automatically provides pre-programmed answers using generative artificial intelligence.

[1469] "Preprocessing" refers to the process of removing unnecessary information from data and converting it into an analyzable format.

[1470] A "database" is a system for organizing and efficiently managing data.

[1471] "Notify operator" refers to the system sending a notification to a human operator to request assistance when the inquiry is complex.

[1472] This invention aims to streamline customer support systems and improve customer satisfaction. The details of a system that receives inquiry data, generates automated responses, and, when necessary, transfers the inquiry to a human operator are described below.

[1473] Hardware and software to use

[1474] This system consists of a server, a smartphone, and the necessary software.

[1475] Hardware: Servers, cloud infrastructure (AWS, GCP, Azure), smartphones

[1476] Software: Flask, spaCy, transformers, SQLite

[1477] Receiving and preprocessing query data

[1478] 1. Received:

[1479] Users submit questions via chat, email, or contact forms through a smartphone application. The server then receives the inquiry data.

[1480] 2. Pre-processing:

[1481] The server preprocesses the query data it receives. SpaCy is used to remove unnecessary information (e.g., HTML tags and special characters) and convert it into a format that can be parsed as text.

[1482] Analysis and classification

[1483] 3. Natural Language Processing:

[1484] The server passes pre-processed data to a natural language processing (NLP) model to analyze the query. Libraries such as spaCy and transformers are used for this process.

[1485] 4. Classification:

[1486] The server classifies the inquiry based on the analysis results. For example, it might classify it as a general question, a request for technical support, or an emotional complaint.

[1487] Automatic response generation

[1488] 5. Generative Artificial Intelligence:

[1489] The server inputs the query into a generative artificial intelligence (e.g., GPT-3.5-turbo) and generates a response. If an automated response is applicable, this response is generated.

[1490] 6. Submit your response:

[1491] The server sends the generated response to the user. For example, in response to the inquiry, "I want to know about the new mobile phone plans," the server sends the response, "The following types of mobile phone plans are currently available..."

[1492] Handover to operator

[1493] 7. Handover:

[1494] The server extracts inquiries that are difficult for generative artificial intelligence to handle and sends notifications to transfer these inquiries to human operators. The notifications include the inquiry history and related information.

[1495] Data storage

[1496] 8. Save:

[1497] The server saves all inquiry details and generated responses to a database (SQLite). This allows for future inquiry handling and analysis.

[1498] As a concrete example of a prompt message, if a user asks, "I want to know about the new mobile phone plan,"

[1499] I want to know about the new mobile phone plans.

[1500] By inputting this sentence into a generative artificial intelligence system, an appropriate answer will be automatically generated.

[1501] As described above, a system is realized that enables improved resource efficiency and faster response times, leading to increased customer satisfaction.

[1502] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1503] Step 1:

[1504] Users submit inquiries through a smartphone application. In this process, users enter their inquiry details via various means, such as chat, email, or a contact form, and then press the submit button. The input data includes the user's question or request, as well as contact information. The server receives this inquiry data.

[1505] Step 2:

[1506] The server preprocesses the query data it receives. This preprocessing uses spaCy to remove unnecessary information (such as HTML tags and special characters) from the data and convert it into a parseable format. The output of this preprocessing is cleaned text data.

[1507] Step 3:

[1508] The server passes the pre-processed text data to a natural language processing (NLP) model. The NLP model (e.g., spaCy or transformers) analyzes the text and extracts important keywords and intent. The analysis results are then output.

[1509] Step 4:

[1510] The server classifies the inquiry based on the analysis results from the NLP model. For example, it might be classified as a general question, technical support, or an emotional complaint. Based on this classification, it determines whether an automated response is possible.

[1511] Step 5:

[1512] If automated response is applicable, the server passes the analysis results as an input prompt to a generative artificial intelligence (e.g., GPT-3.5-turbo) to generate a response. The output of this generative AI is the text response to the user's inquiry.

[1513] Step 6:

[1514] The server sends the generated response to the user. The response is sent via chat, email, or notification functions through a smartphone application. At this time, an automatically generated response to the user's inquiry is output.

[1515] Step 7:

[1516] For inquiries that are difficult for generative artificial intelligence to handle, the server will transfer the inquiry to a human operator. The server will notify the operator of the inquiry details and send the history and related information along with the inquiry. This transfer to an operator makes it possible to handle even complex inquiries.

[1517] Step 8:

[1518] The server saves all inquiry details, generated responses, and response history to a database (SQLite). This data is used for future inquiry handling and analysis. The saved data includes inquiry text, response text, user information, and related metadata.

[1519] 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.

[1520] This invention aims to improve the efficiency of customer support resources and enhance customer satisfaction by providing a system that automatically processes inquiry data using generative artificial intelligence and an emotion engine. This system receives inquiry data, analyzes its content, generates an automated response, and, if necessary, hands it over to a human operator. The embodiments for carrying out this invention are described in detail below.

[1521] 1. Receiving inquiry data

[1522] When a user submits an inquiry via a web form, email, or chat system,

[1523] The server receives this.

[1524] The data received includes inquiry details and user contact information.

[1525] 2. Analysis and classification of inquiry content

[1526] The server preprocesses the query data it receives.

[1527] During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[1528] Next, the server passes the pre-processed data to a natural language processing (NLP) model.

[1529] NLP models are implemented using libraries such as spaCy or transformers.

[1530] The server receives the output from the NLP model and analyzes the query intent and important keywords based on those results.

[1531] 3. Emotion recognition by an emotion engine

[1532] The server passes the analysis results and pre-processed data to the emotion engine.

[1533] The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[1534] The output of the emotion engine includes data that quantifies the user's emotions (e.g., anger level 70%).

[1535] The server determines the urgency of an inquiry and the appropriate course of action based on the emotions recognized by the emotion engine.

[1536] 4. Automated responses using generative AI

[1537] Extract queries that the server can automatically respond to.

[1538] If automated responses are applicable, the server inputs the inquiry details into a generative artificial intelligence (e.g., GPT-4) and has it generate a response.

[1539] The generative artificial intelligence generates the answer and returns it to the server.

[1540] The server sends the generated response to the user.

[1541] For example, in response to an inquiry such as "Please tell me the details of the data plan," the response would be "The data plans currently available to you are as follows..."

[1542] The server saves the query content and the generated response to the database.

[1543] 5. Handover to a human operator

[1544] The server extracts inquiries that are difficult for generative artificial intelligence to handle.

[1545] In particular, if the emotion engine identifies the user's emotion as "negative," the server automatically initiates a process to hand over the user to a human operator.

[1546] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[1547] In this process, the server provides the operator with the necessary inquiry history and related information.

[1548] For inquiries taken over by human operators, specific actions will be taken.

[1549] As an example, let's consider the processing flow when a user contacts us saying, "I am truly disappointed with this service. Please take action."

[1550] 1. The user submits an inquiry through the chat system.

[1551] 2. The server receives this and performs preprocessing.

[1552] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[1553] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[1554] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[1555] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[1556] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[1557] By combining this emotion engine, it becomes possible to provide appropriate and rapid responses tailored to the user's emotions, especially to users experiencing negative emotions. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[1558] The following describes the processing flow.

[1559] Step 1:

[1560] A user submits an inquiry. Inquiries can be submitted via web form, email, or chat system.

[1561] Step 2:

[1562] The server receives the inquiry data. The received data includes the inquiry details and the user's contact information.

[1563] Step 3:

[1564] The server preprocesses the query data. During preprocessing, unnecessary information (e.g., HTML tags and special characters) is removed, and the data is converted into a format that can be parsed as text.

[1565] Step 4:

[1566] The server passes the pre-processed text to a natural language processing (NLP) model. The NLP model is implemented using libraries such as spaCy or transformers.

[1567] Step 5:

[1568] The server receives the analysis results output from the NLP model. The analysis results include the query intent and important keywords.

[1569] Step 6:

[1570] The server passes the analysis results to the emotion engine. The emotion engine identifies emotions (e.g., joy, anger, sadness, etc.) from the user's text.

[1571] Step 7:

[1572] The emotion engine quantifies the user's emotions and returns the result (e.g., anger level 70%) to the server.

[1573] Step 8:

[1574] The server receives the output from the emotion engine and determines the urgency of the inquiry and the appropriate course of action.

[1575] Step 9:

[1576] The server classifies the inquiry based on the analysis results and the sentiment engine's findings. Classification categories include general questions, requests for technical support, and emotional complaints.

[1577] Step 10:

[1578] The server extracts queries that it has classified as "capable of automatic response."

[1579] Step 11:

[1580] The server inputs the query details into a generative artificial intelligence (e.g., GPT-4) and instructs it to generate an answer.

[1581] Step 12:

[1582] The generative artificial intelligence generates an answer to the inquiry and returns it to the server.

[1583] Step 13:

[1584] The server sends the generated response to the user. For example, in response to the inquiry, "Please tell me the details of the data plan," it will respond, "The data plans currently available to you are as follows..."

[1585] Step 14:

[1586] The server saves the query content and the generated response to a database. The saved content includes the query content, the generated response, user information, and a timestamp.

[1587] Step 15:

[1588] The server extracts inquiries that it has classified as requiring human intervention. In particular, if the sentiment engine identifies the user's emotion as "negative," this is automatically applied.

[1589] Step 16:

[1590] The server sends a notification to its internal operations system to forward these inquiries to a human operator.

[1591] Step 17:

[1592] The server provides the operator with the inquiry history and related information.

[1593] Step 18:

[1594] For inquiries taken over by human operators, specific actions are taken. For example, for users with negative feelings, follow-up emails or phone calls are provided.

[1595] As a concrete example, let's consider the flow of a case where a user contacts us saying, "I am truly disappointed with this service. Please take action."

[1596] 1. The user submits an inquiry through the chat system.

[1597] 2. The server receives this and performs preprocessing.

[1598] 3. The server analyzes the inquiry using an NLP model and classifies it as an "emotional complaint."

[1599] 4. The server passes pre-processed data to the emotion engine to identify the user's emotions.

[1600] 5. The emotion engine recognizes a "negative emotion" (e.g., anger level 80%) and notifies the server.

[1601] 6. The server receives this result and transfers the inquiry to the appropriate claims operator.

[1602] 7. For inquiries taken over by human operators, the customer will directly interact with the user and provide appropriate support.

[1603] In this way, this system improves the efficiency and quality of customer support by combining automated responses with appropriate handover based on emotion recognition.

[1604] (Example 2)

[1605] 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".

[1606] To improve efficiency and customer satisfaction in customer support, a system that processes inquiry data quickly and accurately is necessary. However, conventional systems have the problem of being unable to keep up with processing a large volume of inquiries, leading to increased customer dissatisfaction. Furthermore, they do not adequately recognize emotions, making it difficult to provide appropriate responses that are in line with the customer's feelings. This invention aims to solve these problems.

[1607] 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.

[1608] In this invention, the server includes means for receiving inquiry data, means for preprocessing the inquiry data, means for performing natural language processing using the preprocessed inquiry data, means for classifying the inquiry content based on the analysis results, means for performing emotion recognition, means for generating a response to the inquiry using generative artificial intelligence, means for transmitting the generated response, means for handing over the inquiry to a human operator based on the analysis results and emotion recognition results, and means for storing the inquiry data and the generated response in a database. This enables rapid and accurate processing of inquiry data and further realizes appropriate responses that are in line with the customer's emotions.

[1609] "Inquiry data" refers to data containing information such as questions, requests, and complaints sent by users to customer support.

[1610] "Preprocessing" refers to the process of removing unnecessary information from received query data and converting it into a format that can be processed by natural language processing.

[1611] "Natural language processing" is a technology that uses computers to analyze human language and understand its meaning.

[1612] "Emotion recognition" is a technology that identifies and quantifies emotions such as joy, anger, and sadness from the content of user inquiries.

[1613] "Generative artificial intelligence" refers to machine learning models used to generate appropriate answers based on the content of inquiries.

[1614] "Automated response" refers to the process in which a server sends a response generated by a generative artificial intelligence system to the user.

[1615] A "human operator" is a human worker who directly handles inquiries that the server cannot process.

[1616] A "database" is an information management system used to store query data and generated responses.

[1617] "Analysis results" refers to the analysis of inquiry data obtained through natural language processing and sentiment recognition.

[1618] "Handing over" refers to the process by which a server notifies a human operator of an inquiry requiring attention and entrusts them with handling it.

[1619] This invention is a system aimed at improving the efficiency of resources in customer support and enhancing customer satisfaction, and it automatically processes inquiry data using generative artificial intelligence and an emotion recognition engine. Embodiments of this system are described in detail below.

[1620] First, the user submits an inquiry via a web form, email, or chat system. For example, consider a case where a user submits a request via a web form asking, "How do I use this product?" This inquiry data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address.

[1621] Next, the server receives the transmitted data at the specified endpoint. This received data is preprocessed using regular expressions and text cleaning libraries (e.g., BeautifulSoup4 in Python) to remove HTML tags, special characters, and unnecessary whitespace. The preprocessed text is then converted into a parseable format.

[1622] The server passes the pre-processed query data to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers, and it analyzes the data to extract the query's intent and important keywords. For example, a query like "Please tell me how to use the product" would be classified as "Information Provision."

[1623] Furthermore, the server passes the analysis results and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob, and identifies and quantifies emotions such as joy, anger, and sadness from the user's text. For example, it outputs quantified data such as "Anger level 70%". The server receives the output from the emotion engine and determines the urgency of the inquiry and how to respond.

[1624] The server passes an automated response-capable inquiry to a generative artificial intelligence (e.g., GPT-4) to generate a response. An example prompt might be: "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response." The generative AI analyzes the text and generates an appropriate response. The generated response is returned to the server, which then sends it to the user. For example, a specific response such as "To use this product, please follow these steps..." might be generated.

[1625] Furthermore, the server stores the query content and the generated response in a database. This allows for future reference and analysis.

[1626] On the other hand, the server also extracts inquiries that are difficult for generative artificial intelligence to handle and hands them over to human operators. In particular, if the emotion recognition engine highly values ​​"negative emotions," the server automatically hands over the inquiry to a human operator. In this process, notifications are sent via the internal operating system, providing the necessary inquiry history and related information.

[1627] This enables the rapid and accurate processing of inquiry data, as well as appropriate responses that are tailored to the customer's emotions. This system is expected to significantly improve the efficiency and quality of customer support, leading to increased customer satisfaction.

[1628] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1629] Step 1: Receiving inquiry data

[1630] A user submits an inquiry via a web form, email, or chat system. For example, a user might submit "How do I use this product?" through a web form. The server receives this at a specified endpoint. The input data includes the inquiry content, the user's contact information, and metadata such as a timestamp and IP address. The server performs initial database processing to store this data.

[1631] Step 2: Pre-processing of inquiry content

[1632] The server preprocesses the query data it receives. The input is the query data received in step 1. Specific preprocessing actions include removing HTML tags, special characters, and unnecessary whitespace. This is done using regular expressions and the Python BeautifulSoup4 library, among others. The output is data with the text and metadata cleaned and converted into a parseable format.

[1633] Step 3: Analysis and classification of inquiries

[1634] The server passes pre-processed data as input to a natural language processing (NLP) model. This NLP model is implemented using libraries such as spaCy and Transformers. The server runs the NLP model to extract the intent of the query and important keywords. For example, the query "Please tell me how to use the product" would be classified as "Information Provision." The output is data containing the analysis results and classification results.

[1635] Step 4: Emotion recognition by the emotion engine

[1636] The server passes the analysis results of the NLP model and pre-processed data to the emotion recognition engine. The emotion recognition engine is implemented using libraries such as NLTK and TextBlob. Here, the input is the analysis results and classification results obtained in step 3. The emotion engine identifies emotions such as joy, anger, and sadness from the user's text and outputs quantified data (e.g., "Anger level 70%)". The server receives this output and makes decisions regarding the urgency of the response and how to respond.

[1637] Step 5: Automated response by generative AI

[1638] The server passes a query that can be automatically answered to a generative artificial intelligence (e.g., GPT-4). The input is pre-processed query data and analysis results. The generative AI is given prompts to generate a response. For example, a prompt such as "A customer has submitted the following inquiry: 'How do I use this product?' Please generate an appropriate response" can be used. The generative AI generates a response (e.g., "For instructions on how to use this product, please follow these steps...") and returns it to the server. The server receives the generated response and sends it to the user. The output is the response text sent to the user.

[1639] Step 6: Handover to human operator

[1640] The server extracts inquiries that are difficult for generative artificial intelligence to handle. In particular, if the emotion engine identifies the user's emotion as "negative," the server initiates a process to hand over the inquiry to a human operator. The input here is the output from the emotion recognition engine and the content of the inquiry that the generative artificial intelligence could not handle. The server sends a notification to the internal operating system to hand over these inquiries to human operators. The output is the notification to the human operator and related inquiry data.

[1641] As described above, this system processes inquiry data step by step, efficiently managing and processing it. This enables improved customer satisfaction and more efficient customer support.

[1642] (Application Example 2)

[1643] 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".

[1644] Current customer support systems consume a significant amount of resources in handling inquiries, necessitating greater resource efficiency. Furthermore, accurately understanding customer emotions and providing appropriate responses is challenging, hindering customer satisfaction. Additionally, in real-time customer service at physical stores, staff may not always be able to provide the most suitable answer immediately, potentially detracting from the customer experience.

[1645] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for receiving inquiry data, means for natural language processing for analyzing the content of the inquiry data, means for classifying the content of the inquiry based on the analysis results, means for generating an answer to the inquiry using generative artificial intelligence, means for transmitting the generated answer, means for transferring the inquiry to a human operator based on the analysis results, means for emotion recognition, means for determining the urgency of the inquiry based on the output result of the emotion recognition means, and means for displaying the generated answer to customer service in the store. This enables efficient use of resources in customer support and improved customer satisfaction. Furthermore, it enables smooth real-time customer service in physical stores and improves the customer experience.

[1646] "Inquiry data" refers to data that includes the content of questions and requests submitted by customers.

[1647] "Means of receiving" refers to the means by which the system takes in query data.

[1648] "Natural language processing means" refers to technologies used to analyze query data and understand its content.

[1649] "Generative artificial intelligence" refers to AI technology that automatically generates answers to inquiries.

[1650] "Emotion recognition means" refers to technology used to identify customer emotions from inquiry data.

[1651] "Means of classification" refers to methods for categorizing inquiry content based on analysis results.

[1652] "Means for determining urgency" refers to means for determining the urgency of responding to an inquiry based on the output results of the emotion recognition means.

[1653] "Means of transmission" refers to the means of communicating the generated response to the customer.

[1654] "Means of transferring inquiries to human operators" refers to a method of automatically transferring inquiries to human operators when action is required.

[1655] "A means of displaying responses generated for customer service within a store" refers to a means of displaying responses generated in response to customer inquiries within a physical store.

[1656] This invention is a real-time customer support system using generative artificial intelligence and emotion recognition technology. The specific implementation of this system is described below.

[1657] System Configuration

[1658] hardware

[1659] Servers: Multiple servers are used to receive, analyze, classify, and generate responses to query data.

[1660] Terminal: A device used to access the system, including smartphones and smart glasses.

[1661] Users include customers of physical stores and end users who utilize customer support.

[1662] software

[1663] Natural language processing (NLP) libraries such as spaCy and transformers will be used.

[1664] Generative artificial intelligence: Uses generative AI models such as GPT-4.

[1665] Sentiment recognition engine: Uses sentiment analysis tools that utilize dictionary-based or machine learning-based algorithms.

[1666] Database: Use an RDBMS or NoSQL database to store query data and generated responses.

[1667] System operation

[1668] 1. Receiving inquiry data:

[1669] Users can make inquiries in real time at physical stores via their smartphones or smart glasses.

[1670] The terminal sends this to the server, and the query data is received.

[1671] 2. Analysis and classification of inquiries:

[1672] The server preprocesses the received query data and removes unnecessary information.

[1673] We use NLP models to analyze the content of inquiries and extract their intent and keywords.

[1674] 3. Emotion recognition:

[1675] The server uses an emotion recognition engine to identify the user's emotions from the inquiry data.

[1676] For example, a user's emotions might be quantified as "70% anger."

[1677] 4. Automated responses using generative artificial intelligence:

[1678] The server passes the inquiry details to a generative artificial intelligence system based on the analysis results and sentiment data, and generates an automated response.

[1679] 5. Sending and displaying responses:

[1680] The server sends the generated response to the terminal, which then displays it to the user.

[1681] 6. Handover to a human operator:

[1682] In particular, if the emotion recognition engine determines that the user's emotions are negative, the server automatically hands over the inquiry to a human operator.

[1683] Specific example

[1684] Example 1:

[1685] Inquiry: "Where can I use Wi-Fi in the store?"

[1686] Generated response: "Wi-Fi is available in all areas of the store."

[1687] Example 2:

[1688] Inquiry: "I am truly dissatisfied with the service at your store."

[1689] Emotional analysis result: Anger level 80%

[1690] Countermeasure: The case will be immediately handed over to a human operator.

[1691] Example of a prompt:

[1692] Customer inquiry: "Where can I use Wi-Fi in the store?"

[1693] AI response: "Wi-Fi is available in all areas of the store."

[1694] or

[1695] Customer inquiry: "I am truly dissatisfied with the service at your store."

[1696] AI response: "First, to resolve your concerns, our store manager will speak with you."

[1697] This system will significantly improve the efficiency of customer support and increase customer satisfaction. It will also enable smoother, real-time customer service in physical stores.

[1698] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1699] Step 1:

[1700] A user uses a smartphone or smart glasses to make an inquiry within a physical store. For example, a customer might ask, "Where can I get Wi-Fi?" The device receives this inquiry data and sends it to a server. The input includes the text data of the inquiry and the user's basic information. The output is the inquiry data sent to the server.

[1701] Step 2:

[1702] The server preprocesses the received query data. During this process, unnecessary information (e.g., HTML tags, special characters) is removed, and the data is converted into a format that can be parsed as text. The input includes the original query text. The output is clean text data.

[1703] Step 3:

[1704] The server inputs pre-processed data into a natural language processing (NLP) model. The server uses this model to analyze the query content and extract its intent and keywords. For example, libraries such as spaCy or transformers are used. Clean text data is passed to the NLP model as input. The output is the analyzed intent and keywords.

[1705] Step 4:

[1706] The server inputs the analysis results and pre-processed data into the emotion recognition engine. This engine identifies the user's emotion (e.g., joy, anger, sadness) from the query text. The input includes the analysis results and clean text data. The output is numerical data representing the user's emotion (e.g., anger level 70%).

[1707] Step 5:

[1708] The server determines the urgency of an inquiry based on emotion data obtained from an emotion recognition engine. For example, if the "anger level" exceeds a certain threshold, the inquiry is classified as highly urgent. The input includes emotion data. The output is a classification of the inquiry's urgency.

[1709] Step 6:

[1710] The server determines whether it is possible to generate an automated response to the inquiry. If an automated response is possible, the server passes the inquiry details to a generative artificial intelligence (e.g., GPT-4) to generate a response. The input includes the NLP analysis results and the urgency assessment results. The output is the generated response.

[1711] Step 7:

[1712] The server sends the generated response to the terminal, which then displays it to the user. The input includes the generated response and the user's query data. The output provides the response that is displayed to the user.

[1713] Step 8:

[1714] The server handles the transfer of high-priority or emotionally charged inquiries to human operators. The server provides the operator with necessary inquiry history and relevant information. Inputs include the urgency assessment result and the user's inquiry data. Outputs provide the inquiry information to be transferred to the human operator.

[1715] 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.

[1716] 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.

[1717] 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.

[1718] 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.

[1719] 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.

[1720] 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.

[1721] 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.

[1722] 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.

[1723] 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."

[1724] 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.

[1725] 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.

[1726] 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.

[1727] 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.

[1728] 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.

[1729] 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.

[1730] 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.

[1731] 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.

[1732] 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.

[1733] 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.

[1734] 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.

[1735] 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.

[1736] The following is further disclosed regarding the embodiments described above.

[1737] (Claim 1)

[1738] Means for receiving inquiry data,

[1739] A natural language processing means for analyzing the content of the aforementioned query data,

[1740] A means for classifying the content of inquiries based on the aforementioned analysis results,

[1741] A means of generating responses to inquiries using generative artificial intelligence,

[1742] Means for transmitting the generated response,

[1743] A means of transferring the inquiry to a human operator based on the aforementioned analysis results,

[1744] A system that includes this.

[1745] (Claim 2)

[1746] Means for preprocessing the aforementioned query data,

[1747] The system according to claim 1, further comprising means for performing natural language processing using the pre-processed query data.

[1748] (Claim 3)

[1749] The system according to claim 1, further comprising means for storing the aforementioned inquiry data and the generated response in a database.

[1750] "Example 1"

[1751] (Claim 1)

[1752] Means for receiving inquiry data,

[1753] Means for preprocessing the aforementioned query data,

[1754] A natural language processing means for analyzing the content of the pre-processed query data,

[1755] A means for classifying the content of inquiries based on the aforementioned analysis results,

[1756] A means of generating responses to inquiries using generative artificial intelligence,

[1757] Means for transmitting the generated response,

[1758] Means for storing the aforementioned inquiry data and generated responses in a database,

[1759] A means of transferring the inquiry to a human operator based on the aforementioned analysis results,

[1760] A system that includes this.

[1761] (Claim 2)

[1762] The system according to claim 1, further comprising means for sending a notification to an internal operating system.

[1763] (Claim 3)

[1764] The system according to claim 1, further comprising means for removing unnecessary information in order to convert the content of the inquiry into a parseable format.

[1765] "Application Example 1"

[1766] (Claim 1)

[1767] Means for receiving inquiry data,

[1768] A natural language processing means for analyzing the content of the aforementioned query data,

[1769] A means for classifying the content of inquiries based on the aforementioned analysis results,

[1770] A means of generating responses to inquiries using generative artificial intelligence,

[1771] Means for transmitting the generated response,

[1772] A means of transferring the inquiry to a human operator based on the aforementioned analysis results,

[1773] A means of storing the content of the inquiry and its response in a database,

[1774] Means including a smartphone application for receiving inquiries,

[1775] The aforementioned application includes means for performing automated responses using generative artificial intelligence,

[1776] A means of notifying the operator when the aforementioned inquiry is complex,

[1777] A system that includes this.

[1778] (Claim 2)

[1779] Means for preprocessing the aforementioned query data,

[1780] The system according to claim 1, further comprising means for performing natural language processing using the pre-processed query data.

[1781] (Claim 3)

[1782] The system according to claim 1, further comprising means for storing the query data and the generated response in a database on the device.

[1783] "Example 2 of combining an emotion engine"

[1784] (Claim 1)

[1785] Means for receiving inquiry data,

[1786] Means for preprocessing the aforementioned query data,

[1787] A means for performing natural language processing using the aforementioned preprocessed query data,

[1788] A means for classifying the content of inquiries based on the aforementioned analysis results,

[1789] Means of recognizing emotions,

[1790] A means of generating responses to inquiries using generative artificial intelligence,

[1791] Means for transmitting the generated response,

[1792] A means for handing over the inquiry to a human operator based on the aforementioned analysis results and emotion recognition results,

[1793] Means for storing the aforementioned inquiry data and the generated response in a database,

[1794] A system that includes this.

[1795] (Claim 2)

[1796] The system according to claim 1, further comprising means for removing unnecessary information contained in the query data.

[1797] (Claim 3)

[1798] The system according to claim 1, further comprising means for performing the automated response if the generated response is capable of automated response.

[1799] (Claim 4)

[1800] The system according to claim 1, further comprising means for determining the priority of a response based on the urgency of the emotion recognized by the emotion recognition means.

[1801] "Application example 2 when combining with an emotional engine"

[1802] (Claim 1)

[1803] Means for receiving inquiry data,

[1804] A natural language processing means for analyzing the content of the aforementioned query data,

[1805] A means for classifying the content of inquiries based on the aforementioned analysis results,

[1806] A means of generating responses to inquiries using generative artificial intelligence,

[1807] Means for transmitting the generated response,

[1808] A means of transferring the inquiry to a human operator based on the aforementioned analysis results,

[1809] Means of recognizing emotions,

[1810] A means for determining the urgency of an inquiry based on the output result of the emotion recognition means,

[1811] A means of displaying the generated response for customer service within the store,

[1812] A system that includes this.

[1813] (Claim 2)

[1814] Means for preprocessing the aforementioned query data,

[1815] The system according to claim 1, further comprising means for performing natural language processing using the pre-processed query data.

[1816] (Claim 3)

[1817] The system according to claim 1, further comprising means for storing the aforementioned inquiry data and the generated response in a database. [Explanation of Symbols]

[1818] 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. Means for receiving inquiry data, A natural language processing means for analyzing the content of the aforementioned query data, A means for classifying the content of inquiries based on the aforementioned analysis results, A means of generating responses to inquiries using generative artificial intelligence, Means for transmitting the generated response, A means of transferring the inquiry to a human operator based on the aforementioned analysis results, A system that includes this.

2. Means for preprocessing the aforementioned query data, The system according to claim 1, further comprising means for performing natural language processing using the pre-processed query data.

3. The system according to claim 1, further comprising means for storing the inquiry data and the generated response in a database.

Images