system

Abstract

We provide the system. [Solution] A means for receiving inquiries via multiple communication methods and converting them into a standardized format, A means of analyzing the content of a query and identifying its intent using a generated natural language processing model, Means for collecting necessary information from relevant sources based on identified intentions, A means of generating a response in natural language based on the collected information, A means for transmitting the generated response via the original communication means, A means of analyzing user feedback and evaluating the effectiveness of responses, A system that includes means to improve the quality of responses by learning a generative model based on evaluation results.

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, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of the chatbot's character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance as a response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the sales profession, the personal nature of work and the difficulty of obtaining paid leave are problems, and these often lead to situations where salespersons feel stressed even during holidays. In such situations, there is a risk that customer satisfaction will decrease due to delays in customer response. Therefore, there is a need for a system that automates customer response and allows salespersons to obtain paid leave with peace of mind.

Means for Solving the Problems

[0005] This invention provides means for receiving inquiries via multiple communication methods and converting them into a standardized format. Subsequently, it includes means for analyzing the inquiry content using a generated natural language processing model and identifying the intent. Furthermore, it includes means for collecting necessary information from relevant information sources based on the identified intent. This enables customer service by providing means for generating responses in natural language based on the collected information. The entirety of these means maintains customer service by sending the generated response to the customer via the original communication method, analyzes user feedback, and evaluates the effectiveness of the response. It also includes a mechanism for learning the generation model based on the evaluation results and improving the quality of the response, thereby improving the work efficiency of sales staff while making it easier for them to take paid leave.

[0006] An "inquiry" refers to a question or request that a customer or user sends to a company to seek information or support about a product or service.

[0007] "Communication methods" refer to the media and protocols used to exchange information through voice, text, digital data, etc., and include chat, email, and social networking services (SNS).

[0008] A "standardized format" refers to a data format that converts data from different formats into a unified format based on certain rules, thereby facilitating processing.

[0009] A "generated natural language processing model" refers to an algorithm or program that has the ability to understand and analyze human language based on large amounts of data.

[0010] "Intention" refers to the main purpose or motivation a user has when submitting an inquiry, and is the key information that the system should understand.

[0011] "Information sources" refer to databases and external systems that provide the necessary data and knowledge, through which answers to inquiries can be obtained.

[0012] "Generating responses in natural language" refers to the process by which AI or programs generate answers using language in a format that is easily understood by humans.

[0013] "Feedback" refers to the evaluations and comments that users provide regarding the responses they receive, and is information used to improve the system.

[0014] "Learning" refers to the continuous process of improving the accuracy of algorithms and responses based on data and feedback collected by the system. [Brief explanation of the drawing]

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

Mode for Carrying Out the Invention

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

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

[0018] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include 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.

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

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

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

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

[0023] [First Embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0036] This invention utilizes a multi-functional customer service system to improve the work efficiency of sales staff and promote the use of paid leave. This system can receive user inquiries through multiple communication channels, process them effectively, and return responses.

[0037] First, the user submits an inquiry via chat, voice assistant, email, or social media. The server receives these inquiries and converts them into a standardized format. This conversion makes it easy for the server to parse data from different channels.

[0038] Next, the server uses the generated natural language processing model to analyze the query and identify the user's intent. Based on this analysis, the server categorizes the query into relevant categories and determines what information needs to be retrieved.

[0039] The server queries FAQ databases, customer information systems, and external APIs to gather information related to the identified intent. Based on this information, the server generates a response in natural language and ensures that its content is clear to the user.

[0040] The generated response is sent from the server to the user using the original communication method. The user receiving and reviewing this response facilitates problem resolution and further support.

[0041] Furthermore, the server evaluates user feedback and verifies the effectiveness of its responses. This improves the quality of responses to future inquiries. The collected feedback is used to improve the generative AI model, thus increasing the overall response accuracy of the system.

[0042] As a concrete example, consider a case where a user asks "What is the latest product price?" via chat. In this case, the server analyzes the inquiry and determines it is a "request for product information." It retrieves information from the FAQ database, which includes pricing information, and generates a response in natural language such as "The latest product price is XX yen," which it then sends to the user. The appropriateness of this response is then further improved based on the user's reaction and feedback.

[0043] This invention aims to improve the quality of customer service by providing efficient support for sales personnel through these processes.

[0044] The following describes the processing flow.

[0045] Step 1:

[0046] A user submits an inquiry via chat, voice assistant, email, or social media. This initiates an information request to the customer support system.

[0047] Step 2:

[0048] The server receives the query. Since the received information comes from multiple communication methods, the server converts each piece of data into a standardized format and prepares it for analysis.

[0049] Step 3:

[0050] The server uses a generated natural language processing model to analyze the query. Through this analysis, the server identifies the user's intent and determines the appropriate response category.

[0051] Step 4:

[0052] Based on the identified intent, the server accesses FAQ databases, customer information systems, and external APIs to collect necessary information in order to obtain relevant data.

[0053] Step 5:

[0054] Based on the information collected by the server, a response is generated in natural language. In this process, the generative AI summarizes the key points of the information and creates a clear answer.

[0055] Step 6:

[0056] The server sends the generated response to the user via the communication method used for the initial inquiry. This is done according to the method the user used, such as chat or email.

[0057] Step 7:

[0058] The user receives and reviews the response. The user can ask additional questions or provide feedback on the response as needed.

[0059] Step 8:

[0060] The server receives and analyzes user feedback. Based on this analysis, the effectiveness of the response is evaluated, and the training database of the generative AI model is updated, contributing to future improvements in responses.

[0061] (Example 1)

[0062] 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."

[0063] There is a need to improve the work efficiency of sales staff, ensure uninterrupted customer support even when the assigned representative is absent, and maintain customer satisfaction. Furthermore, it is necessary to continuously improve the quality of responses to user inquiries and enhance the accuracy of those inquiries. To address these challenges, there is a lack of efficient and flexible customer support systems.

[0064] 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.

[0065] In this invention, the server includes means for acquiring queries via multiple communication technologies and converting them into a standardized data format, means for analyzing query information and identifying intent using a natural language processing device, and means for collecting necessary information from relevant information sources based on the identified intent. This makes it possible to provide efficient and flexible customer service and continuously improve the quality of responses.

[0066] "Inquiry" refers to requests for information or support submitted by users.

[0067] "Communication technology" refers to the means and methods for sending and receiving data, and includes chat, voice, email, and social networking services (SNS).

[0068] A "standardized data format" refers to a method of representing data based on conventions used to unify data in different formats.

[0069] A "natural language processing system" refers to the technology and devices that enable computers to understand and analyze human language.

[0070] "Identifying intent" refers to the act of analyzing a user's inquiry to determine its main point and purpose.

[0071] "Information source" refers to a database, system, or external service accessed to retrieve the necessary information.

[0072] A "generation device" refers to a technology or system used to create a response in natural language based on collected information.

[0073] "Response" refers to the feedback or reply that a user gives to the information or service they receive.

[0074] This invention provides a system in which a server plays a central role in receiving user inquiries through multiple communication technologies. This system utilizes communication methods such as chat, voice assistants, email, and social networking services (SNS), and the server converts the inquiries received through these means into a standardized data format. For example, formats such as XML or JSON are used for conversion, and appropriate libraries are employed for data formatting.

[0075] The server, acting as a generative AI model, uses a natural language processing unit (NLP) to analyze and identify the user's intent. Specifically, existing technologies such as "GPT" and "BERT" are applied. Based on the analysis results, the server collects necessary information from sources related to the identified intent. This information collection process includes access to database management systems (DBMS) and external APIs.

[0076] Based on the information collected, the server uses a generator to produce a response in natural language. This generation process uses prompts such as, "Create an appropriate response to the user's inquiry." The generated response is then sent to the user via the original communication technology.

[0077] For example, if a user asks in a chat, "Please tell me the latest product price," the server analyzes the message and identifies the intent as a "request for product information." The server then queries sources containing price information, generates a response such as "The latest product price is XX yen," and sends it back to the user's chat.

[0078] Through the process described above, this invention aims to efficiently support sales personnel and improve the quality of customer service.

[0079] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0080] Step 1:

[0081] Users submit inquiries via chat, voice assistant, email, or social media. The input is the user's inquiry message, which is then captured by the system. The server receives messages from these different communication technologies and converts the inquiries into a unified data format. For example, upon receiving a text message, the server converts it to JSON format for standardization. The output is formatted and parseable data.

[0082] Step 2:

[0083] The server analyzes standardized data using a generative AI model to identify the user's intent. The input is standardized query data. The server uses a natural language processing unit to analyze the query content and apply a generative AI model to identify the intent. In this process, tokenization and grammatical analysis are performed, and the output is the result with the intent identified.

[0084] Step 3:

[0085] Based on the identified intent, the server queries relevant information sources to collect the necessary information. The input is the result of the identified intent. The server accesses database management systems and external APIs to retrieve the relevant information. For example, if the request is for price information, it sends an SQL query to the product database to retrieve the corresponding price information. The output of this process is the collected relevant data.

[0086] Step 4:

[0087] The server uses a generator to produce a natural language response based on the collected information. The input is the collected information data. The generating AI model is given a prompt sentence, instructing it to "generate an appropriate response based on this information." The output is a response message in the form of text, which is then formatted in language that is easy for the user to understand.

[0088] Step 5:

[0089] The server sends the generated response to the user via the original communication method. The input is the generated response message. The server sends the response back to the user using a chat API, mail server, or SNS platform. As output, the user receives the corresponding message, reviews it, and takes appropriate action.

[0090] Step 6:

[0091] The server collects user feedback, analyzes it, and evaluates the effectiveness of its responses. The input is user feedback. The server uses feedback analysis tools to analyze the feedback and identify areas for improvement. This process yields an evaluation as output, which can be used to improve the generated AI model.

[0092] (Application Example 1)

[0093] 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."

[0094] In today's business environment, there is a demand for reduced workload for sales staff and more efficient customer service. However, conventional systems often experience delays in service when sales staff are absent, leading to decreased customer satisfaction and reduced operational efficiency. Furthermore, it is difficult to improve the system using feedback, making it challenging to adequately improve the quality of service. This invention aims to solve these problems and achieve more effective customer service and system improvement.

[0095] 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.

[0096] In this invention, the server includes means for receiving inquiries via multiple communication technologies and converting them into a unified format, means for analyzing the content of the inquiries and identifying their intent using the generated natural language processing technology, and means for providing a function to immediately respond in natural language to inquiries made by users through voice and visual devices. This enables stable customer support even when sales staff are absent and allows for continuous improvement through the use of feedback.

[0097] An "inquiry" is a request from a user for information or support to resolve a problem.

[0098] "Communication technology" refers to the technical means of transmitting information through different channels.

[0099] A "unified format" is a transformation of information from different formats into a consistent state, intended to facilitate analysis.

[0100] "Natural language processing technology" refers to technologies for analyzing, understanding, and generating human language.

[0101] "Intention" refers to the purpose or request that the user is trying to obtain through their inquiry.

[0102] "Information resources" is a general term for databases and external information sources that are referenced in order to provide answers.

[0103] "Responding in natural language" refers to a response generated in a format that is intuitively understandable to the user.

[0104] "Audio and visual devices" are hardware devices that allow users to exchange information through sound and vision.

[0105] "Feedback" refers to the responses, evaluations, and opinions that users provide regarding services.

[0106] "Training" is the process of repeatedly learning by utilizing evaluation results to improve the performance of a system or model.

[0107] This invention realizes a system that receives inquiries using various communication technologies and generates responses using natural language processing technology. The following describes specific embodiments of the invention.

[0108] The server receives inquiries from users via devices such as smartphones and smart glasses. These inquiries are converted into a standardized format and analyzed using natural language processing techniques. Based on the analysis, the server identifies the user's intent and collects necessary data from relevant information resources. This system utilizes generative AI models, particularly OpenAI, to generate natural language responses that are intuitively understandable to the user.

[0109] The generated response is sent to the user via the original communication technology. For example, if a user makes a voice inquiry saying, "I want to check my usage details for this month," the server understands the inquiry, immediately collects the information to display the latest usage details, and presents that information on the smart device.

[0110] User feedback verifies the quality of the system's responses and is used as training data to improve the model. This mechanism allows the accuracy and reliability of responses to improve over time.

[0111] As a concrete example, we will use the following example prompt:

[0112] "Generate an appropriate response when a user requests their monthly billing statement."

[0113] This system makes it possible to maintain consistent customer service and improve the user experience even when sales staff are absent.

[0114] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0115] Step 1:

[0116] The user sends an inquiry via their device. The inquiry can be sent as voice, text, or visual input. The device forwards this to the server, providing initial data for processing.

[0117] Step 2:

[0118] The server prepares to parse the received query. First, it converts the query into a unified format, even if it contains data in different formats. This conversion ensures data consistency and facilitates subsequent processing. The input is the raw query data, and the output is data in a unified format.

[0119] Step 3:

[0120] The server analyzes the query data transformed using a generative AI model. During the analysis, the generative AI model utilizes natural language processing to understand the subject and purpose of the query. The input is data in a unified format, and the output is information indicating the intent of the query. For example, if a user makes a query about pricing, the related intent is identified.

[0121] Step 4:

[0122] Based on the identified intent, the server queries relevant information resources. In this step, the server retrieves data related to the query from databases, external APIs, etc. The input is the query intent, and the output is data containing the necessary information. For example, the most recent billing information is retrieved for a billing inquiry intent.

[0123] Step 5:

[0124] Based on the collected information, the server generates a response in natural language. It utilizes a generative AI model to construct answers that are intuitively understandable to the user. The input is the collected information, and the output is a natural language response. Specifically, it is provided in the form of "Your latest usage details are XX."

[0125] Step 6:

[0126] The server sends the generated response to the user via the original communication method. The terminal displays or speaks this response to the user, conveying the information. The input is the generated response statement, and the output is the information provided to the user.

[0127] Step 7:

[0128] Users provide feedback on the response. This feedback is collected and analyzed by the server. This information is used to improve the model, aiming to increase the accuracy of future responses. The input is user feedback, and the output is data intended for improvement.

[0129] 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.

[0130] This invention aims to improve communication with customers by combining emotion recognition with an automated response system for inquiries. This system centrally processes inquiries received through multiple communication channels, identifies the user's emotions using an emotion engine, and generates a response based on those emotions.

[0131] First, users submit inquiries through any communication method they choose, such as chat, voice assistants, email, or social media. The server receives these inquiries and standardizes the data format. This integrates data from different channels, making it easier to analyze.

[0132] Next, the server uses a natural language processing model to analyze the query and identify the user's intent. Furthermore, it uses an emotion engine to analyze the emotions contained in the text and voice. Emotions can, for example, identify whether the user is angry, confused, or satisfied.

[0133] Based on identified intentions and emotions, the server accesses FAQ databases, customer information systems, and external services to collect relevant information. This information is then comprehensively evaluated in light of newly acquired emotional information.

[0134] The server then generates an optimized response in natural language based on the collected information and sentiment analysis. The response is tailored to the user's emotions and adjusted as needed. For example, an angry user will be provided with a formal response emphasizing quick problem resolution.

[0135] The generated response is sent back to the original communication method via the server and then sent to the user. This allows the user to receive the response quickly.

[0136] Furthermore, user responses and feedback are analyzed by the server, and the effectiveness of the responses is evaluated. This feedback and sentiment analysis results are added to the training data of the generative model and used to improve the sentiment relevance of future responses.

[0137] As a concrete example, consider a scenario where a user sends a message via chat saying, "My ordered item hasn't arrived yet, and I'm very unhappy." In response to this inquiry, the server generates a response that demonstrates understanding and prompt action, such as, "We understand you're distressed that your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." This process aims to improve customer satisfaction by providing a response that emotionally satisfies the user.

[0138] The following describes the processing flow.

[0139] Step 1:

[0140] Users send inquiries using various communication methods (chat, voice assistant, email, social media, etc.). These messages may contain emotional content.

[0141] Step 2:

[0142] The server receives the query. It converts the received data into a standardized format and prepares it for data analysis. This ensures consistent processing of data from different input channels.

[0143] Step 3:

[0144] The server uses a generated natural language processing (NLP) model to analyze the query and identify the user's intent. This helps it understand what kind of information needs to be returned.

[0145] Step 4:

[0146] The server activates the emotion engine and identifies the user's emotions from the analyzed text data. In this step, the emotion engine determines whether the user is angry, excited, calm, etc.

[0147] Step 5:

[0148] The server collects relevant information based on the identified intent and sentiment. This information is obtained by the server using FAQ databases, customer information systems, and external APIs.

[0149] Step 6:

[0150] The server generates responses in natural language based on collected information and sentiment data. These responses are sensitive to the user's emotions and are adjusted as needed. For example, if the user is angry, the responses may emphasize solutions.

[0151] Step 7:

[0152] The server sends the generated response to the user via the original communication method. This allows the user to receive service and support quickly.

[0153] Step 8:

[0154] The system sends feedback to the server regarding the response received by the user. This feedback includes the effectiveness of the response and the user's emotional reaction.

[0155] Step 9:

[0156] The server receives feedback, analyzes the user's response, and evaluates its effectiveness. Based on this evaluation, the server continuously learns and improves its emotion engine and generative AI models to improve the emotional relevance of future responses.

[0157] (Example 2)

[0158] 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".

[0159] The diversification of information and communication technologies has made it difficult to maintain effective communication with users. Accurately understanding user intent and providing responses that consider appropriate emotions is essential for improving customer satisfaction. However, current technology has a problem in that it is difficult to effectively utilize sentiment analysis in generating such responses.

[0160] 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.

[0161] In this invention, the server includes means for receiving queries via multiple information transmission means and converting them into a unified format; means for analyzing the content of the queries and identifying their purpose using generated natural language processing techniques; and means for generating responses in emotionally sensitive natural language based on the collected information and user sentiment analysis. This makes it possible to provide timely and appropriate responses that take user sentiment into account, even for queries transmitted through diverse information channels.

[0162] An "inquiry" is a message or request that a user sends seeking some kind of information or assistance.

[0163] "Information transmission means" refers to a variety of methods and platforms for users to communicate with a server, such as chat, voice assistants, email, and social networking services (SNS).

[0164] A "standardized format" is a data format that has been standardized to make it easier to analyze data received from different means of information transmission.

[0165] "Natural language processing technology" is a technology that allows computers to understand and analyze human language, and is used to grasp the user's intent.

[0166] "Purpose" refers to the objective or intention that a user expects when making an inquiry.

[0167] "Sentiment analysis" is a technology that identifies a user's emotional state from the words and expressions contained in their inquiries.

[0168] An "information source" is a database or external system that is referenced to obtain an appropriate answer in response to a user's inquiry.

[0169] "Generating responses in natural language" refers to the process by which a computer creates an answer to a user's inquiry using language that is understandable to humans.

[0170] "Response" refers to feedback from users, including their impressions and suggestions for improvement, regarding the response provided.

[0171] This invention relates to a system in which a server centrally receives and processes inquiries made by a user through multiple information transmission means, and generates responses using sentiment analysis. This system includes a server and a user terminal as hardware, and incorporates natural language processing technology and a sentiment analysis engine as software.

[0172] Users submit inquiries using various means of information transmission, such as chat, voice assistants, email, and social media. User devices have access to these information transmission platforms and the ability to send inquiries to the server. Upon receiving an inquiry, the server first converts the data into a standardized format. This standardizes data from different communication methods, ensuring consistency in analysis.

[0173] Next, the server analyzes the query using natural language processing techniques to identify the user's purpose. Various natural language processing platforms, such as GPT and BERT, which are generative AI models, are used here. In parallel, an emotion analysis engine (e.g., an emotion recognition API) analyzes the user's emotional state from the text information and identifies emotions such as anger, anxiety, and joy.

[0174] Based on this information, the server accesses relevant information sources (FAQ databases and external services) to collect the necessary information. Considering the collected information and the results of sentiment analysis, the server generates a natural language response that takes emotions into account. In this process, a generative AI model is used, and input in the form of text, such as prompts, is used. For example, a prompt might read: "Generate an appropriate response based on the user's emotions. Inquiry: 'I am dissatisfied because the product I ordered has not yet arrived.'"

[0175] The generated response is returned to the user's original information transmission method via the server and then sent back to the user. The user is expected to receive this response and experience greater satisfaction due to the prompt and emotionally appropriate response. This allows users to use the service with confidence and improves the quality of customer service provided by the company.

[0176] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0177] Step 1:

[0178] Users use their devices to submit inquiries through various means of communication, such as chat, voice assistants, email, and social media. Input is raw text or voice data. These inquiries are then forwarded to the server as output.

[0179] Step 2:

[0180] The server converts the query data received from the server from a format specific to each information transmission method to a standardized format. The input is query data in a heterogeneous format, and the output is a standardized data format. This process allows subsequent analysis to proceed linearly.

[0181] Step 3:

[0182] The server analyzes data in a standardized format using natural language processing techniques. The input is standardized text data, and the output consists of keywords and intents identified as the user's purpose. This analysis utilizes AI models (e.g., natural language processing platforms) to process the data and identify the user's intent.

[0183] Step 4:

[0184] Simultaneously, the server uses an emotion analysis engine to identify the user's emotions from the text. The input is standardized text, and the output is a set of emotional state labels (e.g., anger, anxiety, joy, etc.). This allows the user's psychological state to be recognized.

[0185] Step 5:

[0186] Based on identified intent and sentiment information, the server accesses relevant sources to collect the necessary information. The input is the user's intent and the type of information associated with it, while the output is a specific set of information to respond to the user's query. Databases and external APIs are utilized here.

[0187] Step 6:

[0188] Using the collected information and the results of sentiment analysis, the server generates a response in natural language through a generative AI model. The input is an information set and sentiment labels, and the output is a sentiment-sensitive natural language response. A prompt is used, and the AI ​​model generates the text.

[0189] Step 7:

[0190] The generated response is sent via the server to the user's original information transmission means. The input is the generated response, and the output is the response displayed or played back as audio on the user's terminal. Specifically, the response is provided to the user visually or audibly.

[0191] Step 8:

[0192] The server evaluates the effectiveness of its responses by receiving user responses or feedback. The input is user feedback data, and the output is evaluation information regarding the effectiveness of the responses. This data is used as feedback in the learning process of the generative AI model to improve the quality of the responses.

[0193] (Application Example 2)

[0194] 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".

[0195] Traditional automated response systems provide a uniform response to user inquiries, making it difficult to offer appropriate responses that reflect the user's emotions. Furthermore, the lack of emotionally responsive responses hindered user satisfaction. As a result, there was a need to improve the quality of communication and customer satisfaction.

[0196] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving queries via multiple communication means and converting them into a standardized format, means for analyzing the content of the query and identifying the intent using a generated natural language processing model, and means for identifying the user's emotions using an emotion analysis engine. This makes it possible to take the user's emotions into consideration and generate and transmit an emotion-appropriate response.

[0197] Point down

[0198] An "inquiry" refers to information such as questions or requests that users submit through communication methods such as chat, voice assistants, email, and social media.

[0199] "Communication methods" refer to the channels and technologies used to send or receive information, and specifically include chat, voice assistants, email, and social networking services (SNS).

[0200] A "standardized format" is the process of converting data received from different channels into a unified format that facilitates analysis.

[0201] A "natural language processing model" is a collection of algorithms and technologies for understanding and appropriately analyzing human language.

[0202] "Intent" refers to the purpose or request that is understood by identifying what the user is seeking in their inquiry.

[0203] A "sentiment analysis engine" is a collection of algorithms and technologies that identify a user's emotions from text or audio.

[0204] "Information source" refers to a database or system that is queried to collect relevant information based on the user's intent.

[0205] "Feedback" refers to evaluations and opinions that users provide regarding responses, and this information is used to improve the system's performance.

[0206] A "generative model" is a set of algorithms and techniques trained to create the best possible response to a user's inquiry.

[0207] The system implementing this invention consists of a server and multiple terminals working together. Users send inquiries via terminals using communication methods such as chat, voice assistants, email, and social networking services. The server receives these inquiries and first converts them into a standardized format. This unifies data from different channels and makes analysis easier.

[0208] Next, the server uses a natural language processing model to analyze the content of the query and identify the user's intent. Simultaneously, it uses an emotion analysis engine to identify the user's emotions from the text and voice. This emotion analysis determines the user's psychological state.

[0209] Subsequently, based on the identified intent and sentiment, the server collects relevant information from various sources. These sources include FAQ databases, customer information systems, and external services. This makes it possible to obtain accurate and relevant information in response to user inquiries.

[0210] Based on the collected information and emotional data, the server uses a generative AI model to generate a natural language-optimized response. This response takes the user's emotions into consideration and is adjusted as needed. For example, if the user is emotionally dissatisfied, the response may be designed to emphasize rapid problem resolution.

[0211] Finally, the server sends the generated response to the user via the original communication method. This allows the user to receive a quick and appropriate response. The user's feedback is also analyzed by the server to evaluate the effectiveness of the response. This feedback and sentiment analysis results are used as training data for the generative model to improve the sentiment relevance of future responses.

[0212] For example, if a user enters via their device, "My ordered item hasn't arrived yet, and I'm very unhappy," the server will generate a response such as, "It seems you're having trouble because your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." In this way, it becomes possible to provide appropriate responses tailored to the user's requests.

[0213] A concrete example of a prompt message for a generative AI model would be: "The user seems confused. They want to know more details about product delivery, so please generate a reassuring explanation."

[0214] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0215] Point down

[0216] Step 1:

[0217] The terminal receives inquiries from users. Users send inquiries using chat, voice assistants, email, social media, etc. The input at this time is the user's inquiry information, and the output is the data sent directly to the server. Specifically, the terminal sends data to the server when it receives an inquiry.

[0218] Step 2:

[0219] The server standardizes the received queries. The input is query data sent from the terminal, and the output is in a standardized data format. As part of the data processing, it converts queries in different formats into a unified format. Specifically, this involves unifying character encodings and removing unnecessary tags.

[0220] Step 3:

[0221] The server analyzes queries using a natural language processing model. The input is standardized query data, and the output is the result of identifying the user's intent. The data calculation involves analyzing the query text to identify what information and responses are needed. Specific operations include keyword extraction and intent classification.

[0222] Step 4:

[0223] The server identifies the user's emotions using an emotion analysis engine. The input is the analyzed query data, and the output is the user's emotional state (e.g., anger, confusion, satisfaction). As a data calculation, it analyzes emotions from text or audio to determine the priority and style of response. Specifically, it evaluates emotion-related words and expressions within the text.

[0224] Step 5:

[0225] The server collects relevant information from sources based on identified intentions and sentiments. The input is intention and sentiment information, and the output is a collection of relevant information. As a data operation, it accesses sources and picks out the information necessary to resolve the user's inquiry. Specific actions include accessing FAQ databases and calling external APIs.

[0226] Step 6:

[0227] The server generates responses in natural language based on collected information and emotional data. The input is relevant information and user emotions, and the output is the generated response message. A generative AI model is used to select expressions appropriate to the user's psychological state. Specifically, it synthesizes response sentences with appropriate wording and tone.

[0228] Step 7:

[0229] The server sends the generated response to the terminal via the original communication method. The input is the generated response message, and the output is the response that reaches the user. Specifically, the information is encoded and returned to the original channel.

[0230] Step 8:

[0231] The user can receive a response and send feedback back to the server. The server receives this feedback and evaluates the effectiveness of the response. The input is user feedback information, and the output is the feedback analysis result. As a data calculation, problems and areas for improvement are analyzed from the feedback and recorded as training data. Specifically, positive and negative analysis of the feedback is performed.

[0232] 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.

[0233] 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.

[0234] 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.

[0235] [Second Embodiment]

[0236] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0237] 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.

[0238] 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).

[0239] 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.

[0240] 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.

[0241] 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).

[0242] 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.

[0243] 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.

[0244] 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.

[0245] 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.

[0246] 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.

[0247] 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".

[0248] This invention utilizes a multi-functional customer service system to improve the work efficiency of sales staff and promote the use of paid leave. This system can receive user inquiries through multiple communication channels, process them effectively, and return responses.

[0249] First, the user submits an inquiry via chat, voice assistant, email, or social media. The server receives these inquiries and converts them into a standardized format. This conversion makes it easy for the server to parse data from different channels.

[0250] Next, the server uses the generated natural language processing model to analyze the query and identify the user's intent. Based on this analysis, the server categorizes the query into relevant categories and determines what information needs to be retrieved.

[0251] The server queries FAQ databases, customer information systems, and external APIs to gather information related to the identified intent. Based on this information, the server generates a response in natural language and ensures that its content is clear to the user.

[0252] The generated response is sent from the server to the user using the original communication method. The user receiving and reviewing this response facilitates problem resolution and further support.

[0253] Furthermore, the server evaluates user feedback and verifies the effectiveness of its responses. This improves the quality of responses to future inquiries. The collected feedback is used to improve the generative AI model, thus increasing the overall response accuracy of the system.

[0254] As a concrete example, consider a case where a user asks "What is the latest product price?" via chat. In this case, the server analyzes the inquiry and determines it is a "request for product information." It retrieves information from the FAQ database, which includes pricing information, and generates a response in natural language such as "The latest product price is XX yen," which it then sends to the user. The appropriateness of this response is then further improved based on the user's reaction and feedback.

[0255] This invention aims to improve the quality of customer service by providing efficient support for sales personnel through these processes.

[0256] The following describes the processing flow.

[0257] Step 1:

[0258] A user submits an inquiry via chat, voice assistant, email, or social media. This initiates an information request to the customer support system.

[0259] Step 2:

[0260] The server receives the query. Since the received information comes from multiple communication methods, the server converts each piece of data into a standardized format and prepares it for analysis.

[0261] Step 3:

[0262] The server uses a generated natural language processing model to analyze the query. Through this analysis, the server identifies the user's intent and determines the appropriate response category.

[0263] Step 4:

[0264] Based on the identified intent, the server accesses FAQ databases, customer information systems, and external APIs to collect necessary information in order to obtain relevant data.

[0265] Step 5:

[0266] Based on the information collected by the server, a response is generated in natural language. In this process, the generative AI summarizes the key points of the information and creates a clear answer.

[0267] Step 6:

[0268] The server sends the generated response to the user via the communication method used for the initial inquiry. This is done according to the method the user used, such as chat or email.

[0269] Step 7:

[0270] The user receives and reviews the response. The user can ask additional questions or provide feedback on the response as needed.

[0271] Step 8:

[0272] The server receives and analyzes user feedback. Based on this analysis, the effectiveness of the response is evaluated, and the training database of the generative AI model is updated, contributing to future improvements in responses.

[0273] (Example 1)

[0274] 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."

[0275] There is a need to improve the work efficiency of sales staff, ensure uninterrupted customer support even when the assigned representative is absent, and maintain customer satisfaction. Furthermore, it is necessary to continuously improve the quality of responses to user inquiries and enhance the accuracy of those inquiries. To address these challenges, there is a lack of efficient and flexible customer support systems.

[0276] 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.

[0277] In this invention, the server includes means for acquiring queries via multiple communication technologies and converting them into a standardized data format, means for analyzing query information and identifying intent using a natural language processing device, and means for collecting necessary information from relevant information sources based on the identified intent. This makes it possible to provide efficient and flexible customer service and continuously improve the quality of responses.

[0278] "Inquiry" refers to requests for information or support submitted by users.

[0279] "Communication technology" refers to the means and methods for sending and receiving data, including chat, voice, email, SNS, etc.

[0280] "Standardized data format" refers to the method of representing data based on the conventions used to unify different forms of data.

[0281] "Natural language processing device" refers to the technology and device for a computer to understand and analyze human language.

[0282] "Identifying the intention" refers to the act of analyzing the content of a user's inquiry to identify its main idea and purpose.

[0283] "Information source" refers to the database, system, or external service accessed to extract the necessary information.

[0284] "Generation device" refers to the technology and system used to create responses in natural language based on the collected information.

[0285] "Reaction" refers to the feedback or response shown to the information and services received by the user.

[0286] In this invention, a system is provided in which the server plays a central role and receives inquiries from users through multiple communication technologies. In this system, communication means such as chat, voice assistant, email, and SNS are utilized, and the server converts the inquiries received through these means into a standardized data format. For the conversion, formats such as XML or JSON are used, and appropriate libraries are used for data formatting.

[0287] The server, acting as a generative AI model, uses a natural language processing unit (NLP) to analyze and identify the user's intent. Specifically, existing technologies such as "GPT" and "BERT" are applied. Based on the analysis results, the server collects necessary information from sources related to the identified intent. This information collection process includes access to database management systems (DBMS) and external APIs.

[0288] Based on the information collected, the server uses a generator to produce a response in natural language. This generation process uses prompts such as, "Create an appropriate response to the user's inquiry." The generated response is then sent to the user via the original communication technology.

[0289] For example, if a user asks in a chat, "Please tell me the latest product price," the server analyzes the message and identifies the intent as a "request for product information." The server then queries sources containing price information, generates a response such as "The latest product price is XX yen," and sends it back to the user's chat.

[0290] Through the process described above, this invention aims to efficiently support sales personnel and improve the quality of customer service.

[0291] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0292] Step 1:

[0293] Users submit inquiries via chat, voice assistant, email, or social media. The input is the user's inquiry message, which is then captured by the system. The server receives messages from these different communication technologies and converts the inquiries into a unified data format. For example, upon receiving a text message, the server converts it to JSON format for standardization. The output is formatted and parseable data.

[0294] Step 2:

[0295] The server analyzes standardized data using a generative AI model to identify the user's intent. The input is standardized query data. The server uses a natural language processing unit to analyze the query content and apply a generative AI model to identify the intent. In this process, tokenization and grammatical analysis are performed, and the output is the result with the intent identified.

[0296] Step 3:

[0297] Based on the identified intent, the server queries relevant information sources to collect the necessary information. The input is the result of the identified intent. The server accesses database management systems and external APIs to retrieve the relevant information. For example, if the request is for price information, it sends an SQL query to the product database to retrieve the corresponding price information. The output of this process is the collected relevant data.

[0298] Step 4:

[0299] The server uses a generator to produce a natural language response based on the collected information. The input is the collected information data. The generating AI model is given a prompt sentence, instructing it to "generate an appropriate response based on this information." The output is a response message in the form of text, which is then formatted in language that is easy for the user to understand.

[0300] Step 5:

[0301] The server sends the generated response to the user via the original communication method. The input is the generated response message. The server sends the response back to the user using a chat API, mail server, or SNS platform. As output, the user receives the corresponding message, reviews it, and takes appropriate action.

[0302] Step 6:

[0303] The server collects feedback from users, analyzes it, and evaluates the effectiveness of responses. The input is the feedback from users. The server uses a feedback analysis tool to analyze and identify areas for improvement. Through this process, an evaluation is obtained as output, which can be used to improve the generated AI model.

[0304] (Application Example 1)

[0305] Next, Application Example 1 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart glasses 214 are referred to as the "terminal".

[0306] In the modern business environment, there is a demand to reduce the workload of sales staff and provide efficient customer service. However, in conventional systems, responses are often delayed when sales staff are absent, leading to a decrease in customer satisfaction and deterioration of business efficiency. Also, it is difficult to improve the system by utilizing feedback, and the quality of responses cannot be sufficiently improved. The object of this invention is to solve these problems and achieve more effective customer service and system improvement.

[0307] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0308] In this invention, the server includes means for receiving inquiries via a number of communication technologies and converting them into a unified format, means for analyzing the content of the inquiries using the generated natural language processing technology to identify the intent, and means for providing a function to immediately respond in natural language to inquiries through the user's voice or visual device. Thereby, stable customer service is possible even when sales staff are absent, and continuous improvement using feedback is possible.

[0309] An "inquiry" is a request for information or support that a user seeks to resolve.

[0310] "Communication technology" refers to the technical means of transmitting information through different channels.

[0311] A "unified format" is a transformation of information from different formats into a consistent state, intended to facilitate analysis.

[0312] "Natural language processing technology" refers to technologies for analyzing, understanding, and generating human language.

[0313] "Intention" refers to the purpose or request that the user is trying to obtain through their inquiry.

[0314] "Information resources" is a general term for databases and external information sources that are referenced in order to provide answers.

[0315] "Responding in natural language" refers to a response generated in a format that is intuitively understandable to the user.

[0316] "Audio and visual devices" are hardware devices that allow users to exchange information through sound and vision.

[0317] "Feedback" refers to the responses, evaluations, and opinions that users provide regarding services.

[0318] "Training" is the process of repeatedly learning by utilizing evaluation results to improve the performance of a system or model.

[0319] This invention realizes a system that receives inquiries using various communication technologies and generates responses using natural language processing technology. The following describes specific embodiments of the invention.

[0320] The server receives inquiries from users via devices such as smartphones and smart glasses. These inquiries are converted into a standardized format and analyzed using natural language processing techniques. Based on the analysis, the server identifies the user's intent and collects necessary data from relevant information resources. This system utilizes generative AI models, particularly OpenAI, to generate natural language responses that are intuitively understandable to the user.

[0321] The generated response is sent to the user via the original communication technology. For example, if a user makes a voice inquiry saying, "I want to check my usage details for this month," the server understands the inquiry, immediately collects the information to display the latest usage details, and presents that information on the smart device.

[0322] User feedback verifies the quality of the system's responses and is used as training data to improve the model. This mechanism allows the accuracy and reliability of responses to improve over time.

[0323] As a concrete example, we will use the following example prompt:

[0324] "Generate an appropriate response when a user requests their monthly billing statement."

[0325] This system makes it possible to maintain consistent customer service and improve the user experience even when sales staff are absent.

[0326] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0327] Step 1:

[0328] The user sends an inquiry via their device. The inquiry can be sent as voice, text, or visual input. The device forwards this to the server, providing initial data for processing.

[0329] Step 2:

[0330] The server prepares to parse the received query. First, it converts the query into a unified format, even if it contains data in different formats. This conversion ensures data consistency and facilitates subsequent processing. The input is the raw query data, and the output is data in a unified format.

[0331] Step 3:

[0332] The server analyzes the query data transformed using a generative AI model. During the analysis, the generative AI model utilizes natural language processing to understand the subject and purpose of the query. The input is data in a unified format, and the output is information indicating the intent of the query. For example, if a user makes a query about pricing, the related intent is identified.

[0333] Step 4:

[0334] Based on the identified intent, the server queries relevant information resources. In this step, the server retrieves data related to the query from databases, external APIs, etc. The input is the query intent, and the output is data containing the necessary information. For example, the most recent billing information is retrieved for a billing inquiry intent.

[0335] Step 5:

[0336] Based on the collected information, the server generates a response in natural language. It utilizes a generative AI model to construct answers that are intuitively understandable to the user. The input is the collected information, and the output is a natural language response. Specifically, it is provided in the form of "Your latest usage details are XX."

[0337] Step 6:

[0338] The server sends the generated response to the user via the original communication method. The terminal displays or speaks this response to the user, conveying the information. The input is the generated response statement, and the output is the information provided to the user.

[0339] Step 7:

[0340] Users provide feedback on the response. This feedback is collected and analyzed by the server. This information is used to improve the model, aiming to increase the accuracy of future responses. The input is user feedback, and the output is data intended for improvement.

[0341] 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.

[0342] This invention aims to improve communication with customers by combining emotion recognition with an automated response system for inquiries. This system centrally processes inquiries received through multiple communication channels, identifies the user's emotions using an emotion engine, and generates a response based on those emotions.

[0343] First, users submit inquiries through any communication method they choose, such as chat, voice assistants, email, or social media. The server receives these inquiries and standardizes the data format. This integrates data from different channels, making it easier to analyze.

[0344] Next, the server uses a natural language processing model to analyze the query and identify the user's intent. Furthermore, it uses an emotion engine to analyze the emotions contained in the text and voice. Emotions can, for example, identify whether the user is angry, confused, or satisfied.

[0345] Based on identified intentions and emotions, the server accesses FAQ databases, customer information systems, and external services to collect relevant information. This information is then comprehensively evaluated in light of newly acquired emotional information.

[0346] The server then generates an optimized response in natural language based on the collected information and sentiment analysis. The response is tailored to the user's emotions and adjusted as needed. For example, an angry user will be provided with a formal response emphasizing quick problem resolution.

[0347] The generated response is sent back to the original communication method via the server and then sent to the user. This allows the user to receive the response quickly.

[0348] Furthermore, user responses and feedback are analyzed by the server, and the effectiveness of the responses is evaluated. This feedback and sentiment analysis results are added to the training data of the generative model and used to improve the sentiment relevance of future responses.

[0349] As a concrete example, consider a scenario where a user sends a message via chat saying, "My ordered item hasn't arrived yet, and I'm very unhappy." In response to this inquiry, the server generates a response that demonstrates understanding and prompt action, such as, "We understand you're distressed that your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." This process aims to improve customer satisfaction by providing a response that emotionally satisfies the user.

[0350] The following describes the processing flow.

[0351] Step 1:

[0352] Users send inquiries using various communication methods (chat, voice assistant, email, social media, etc.). These messages may contain emotional content.

[0353] Step 2:

[0354] The server receives the query. It converts the received data into a standardized format and prepares it for data analysis. This ensures consistent processing of data from different input channels.

[0355] Step 3:

[0356] The server uses a generated natural language processing (NLP) model to analyze the query and identify the user's intent. This helps it understand what kind of information needs to be returned.

[0357] Step 4:

[0358] The server activates the emotion engine and identifies the user's emotions from the analyzed text data. In this step, the emotion engine determines whether the user is angry, excited, calm, etc.

[0359] Step 5:

[0360] The server collects relevant information based on the identified intent and sentiment. This information is obtained by the server using FAQ databases, customer information systems, and external APIs.

[0361] Step 6:

[0362] The server generates responses in natural language based on collected information and sentiment data. These responses are sensitive to the user's emotions and are adjusted as needed. For example, if the user is angry, the responses may emphasize solutions.

[0363] Step 7:

[0364] The server sends the generated response to the user via the original communication method. This allows the user to receive service and support quickly.

[0365] Step 8:

[0366] The system sends feedback to the server regarding the response received by the user. This feedback includes the effectiveness of the response and the user's emotional reaction.

[0367] Step 9:

[0368] The server receives feedback, analyzes the user's response, and evaluates its effectiveness. Based on this evaluation, the server continuously learns and improves its emotion engine and generative AI models to improve the emotional relevance of future responses.

[0369] (Example 2)

[0370] 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".

[0371] The diversification of information and communication technologies has made it difficult to maintain effective communication with users. Accurately understanding user intent and providing responses that consider appropriate emotions is essential for improving customer satisfaction. However, current technology has a problem in that it is difficult to effectively utilize sentiment analysis in generating such responses.

[0372] 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.

[0373] In this invention, the server includes means for receiving queries via multiple information transmission means and converting them into a unified format; means for analyzing the content of the queries and identifying their purpose using generated natural language processing techniques; and means for generating responses in emotionally sensitive natural language based on the collected information and user sentiment analysis. This makes it possible to provide timely and appropriate responses that take user sentiment into account, even for queries transmitted through diverse information channels.

[0374] An "inquiry" is a message or request that a user sends seeking some kind of information or assistance.

[0375] "Information transmission means" refers to a variety of methods and platforms for users to communicate with a server, such as chat, voice assistants, email, and social networking services (SNS).

[0376] A "standardized format" is a data format that has been standardized to make it easier to analyze data received from different means of information transmission.

[0377] "Natural language processing technology" is a technology that allows computers to understand and analyze human language, and is used to grasp the user's intent.

[0378] "Purpose" refers to the objective or intention that a user expects when making an inquiry.

[0379] "Sentiment analysis" is a technology that identifies a user's emotional state from the words and expressions contained in their inquiries.

[0380] An "information source" is a database or external system that is referenced to obtain an appropriate answer in response to a user's inquiry.

[0381] "Generating responses in natural language" refers to the process by which a computer creates an answer to a user's inquiry using language that is understandable to humans.

[0382] "Response" refers to feedback from users, including their impressions and suggestions for improvement, regarding the response provided.

[0383] This invention relates to a system in which a server centrally receives and processes inquiries made by a user through multiple information transmission means, and generates responses using sentiment analysis. This system includes a server and a user terminal as hardware, and incorporates natural language processing technology and a sentiment analysis engine as software.

[0384] Users submit inquiries using various means of information transmission, such as chat, voice assistants, email, and social media. User devices have access to these information transmission platforms and the ability to send inquiries to the server. Upon receiving an inquiry, the server first converts the data into a standardized format. This standardizes data from different communication methods, ensuring consistency in analysis.

[0385] Next, the server analyzes the query using natural language processing techniques to identify the user's purpose. Various natural language processing platforms, such as GPT and BERT, which are generative AI models, are used here. In parallel, an emotion analysis engine (e.g., an emotion recognition API) analyzes the user's emotional state from the text information and identifies emotions such as anger, anxiety, and joy.

[0386] Based on this information, the server accesses relevant information sources (FAQ databases and external services) to collect the necessary information. Considering the collected information and the results of sentiment analysis, the server generates a natural language response that takes emotions into account. In this process, a generative AI model is used, and input in the form of text, such as prompts, is used. For example, a prompt might read: "Generate an appropriate response based on the user's emotions. Inquiry: 'I am dissatisfied because the product I ordered has not yet arrived.'"

[0387] The generated response is returned to the user's original information transmission method via the server and then sent back to the user. The user is expected to receive this response and experience greater satisfaction due to the prompt and emotionally appropriate response. This allows users to use the service with confidence and improves the quality of customer service provided by the company.

[0388] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0389] Step 1:

[0390] Users use their devices to submit inquiries through various means of communication, such as chat, voice assistants, email, and social media. Input is raw text or voice data. These inquiries are then forwarded to the server as output.

[0391] Step 2:

[0392] The server converts the query data received from the server from a format specific to each information transmission method to a standardized format. The input is query data in a heterogeneous format, and the output is a standardized data format. This process allows subsequent analysis to proceed linearly.

[0393] Step 3:

[0394] The server analyzes data in a standardized format using natural language processing techniques. The input is standardized text data, and the output consists of keywords and intents identified as the user's purpose. This analysis utilizes AI models (e.g., natural language processing platforms) to process the data and identify the user's intent.

[0395] Step 4:

[0396] Simultaneously, the server uses an emotion analysis engine to identify the user's emotions from the text. The input is standardized text, and the output is a set of emotional state labels (e.g., anger, anxiety, joy, etc.). This allows the user's psychological state to be recognized.

[0397] Step 5:

[0398] Based on identified intent and sentiment information, the server accesses relevant sources to collect the necessary information. The input is the user's intent and the type of information associated with it, while the output is a specific set of information to respond to the user's query. Databases and external APIs are utilized here.

[0399] Step 6:

[0400] Using the collected information and the results of sentiment analysis, the server generates a response in natural language through a generative AI model. The input is an information set and sentiment labels, and the output is a sentiment-sensitive natural language response. A prompt is used, and the AI ​​model generates the text.

[0401] Step 7:

[0402] The generated response is sent via the server to the user's original information transmission means. The input is the generated response, and the output is the response displayed or played back as audio on the user's terminal. Specifically, the response is provided to the user visually or audibly.

[0403] Step 8:

[0404] The server evaluates the effectiveness of its responses by receiving user responses or feedback. The input is user feedback data, and the output is evaluation information regarding the effectiveness of the responses. This data is used as feedback in the learning process of the generative AI model to improve the quality of the responses.

[0405] (Application Example 2)

[0406] 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."

[0407] Traditional automated response systems provide a uniform response to user inquiries, making it difficult to offer appropriate responses that reflect the user's emotions. Furthermore, the lack of emotionally responsive responses hindered user satisfaction. As a result, there was a need to improve the quality of communication and customer satisfaction.

[0408] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving queries via multiple communication means and converting them into a standardized format, means for analyzing the content of the query and identifying the intent using a generated natural language processing model, and means for identifying the user's emotions using an emotion analysis engine. This makes it possible to take the user's emotions into consideration and generate and transmit an emotion-appropriate response.

[0409] Point down

[0410] An "inquiry" refers to information such as questions or requests that users submit through communication methods such as chat, voice assistants, email, and social media.

[0411] "Communication methods" refer to the channels and technologies used to send or receive information, and specifically include chat, voice assistants, email, and social networking services (SNS).

[0412] A "standardized format" is the process of converting data received from different channels into a unified format that facilitates analysis.

[0413] A "natural language processing model" is a collection of algorithms and technologies for understanding and appropriately analyzing human language.

[0414] "Intent" refers to the purpose or request that is understood by identifying what the user is seeking in their inquiry.

[0415] A "sentiment analysis engine" is a collection of algorithms and technologies that identify a user's emotions from text or audio.

[0416] "Information source" refers to a database or system that is queried to collect relevant information based on the user's intent.

[0417] "Feedback" refers to evaluations and opinions that users provide regarding responses, and this information is used to improve the system's performance.

[0418] A "generative model" is a set of algorithms and techniques trained to create the best possible response to a user's inquiry.

[0419] The system implementing this invention consists of a server and multiple terminals working together. Users send inquiries via terminals using communication methods such as chat, voice assistants, email, and social networking services. The server receives these inquiries and first converts them into a standardized format. This unifies data from different channels and makes analysis easier.

[0420] Next, the server uses a natural language processing model to analyze the content of the query and identify the user's intent. Simultaneously, it uses an emotion analysis engine to identify the user's emotions from the text and voice. This emotion analysis determines the user's psychological state.

[0421] Subsequently, based on the identified intent and sentiment, the server collects relevant information from various sources. These sources include FAQ databases, customer information systems, and external services. This makes it possible to obtain accurate and relevant information in response to user inquiries.

[0422] Based on the collected information and emotional data, the server uses a generative AI model to generate a natural language-optimized response. This response takes the user's emotions into consideration and is adjusted as needed. For example, if the user is emotionally dissatisfied, the response may be designed to emphasize rapid problem resolution.

[0423] Finally, the server sends the generated response to the user via the original communication method. This allows the user to receive a quick and appropriate response. The user's feedback is also analyzed by the server to evaluate the effectiveness of the response. This feedback and sentiment analysis results are used as training data for the generative model to improve the sentiment relevance of future responses.

[0424] For example, if a user enters via their device, "My ordered item hasn't arrived yet, and I'm very unhappy," the server will generate a response such as, "It seems you're having trouble because your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." In this way, it becomes possible to provide appropriate responses tailored to the user's requests.

[0425] A concrete example of a prompt message for a generative AI model would be: "The user seems confused. They want to know more details about product delivery, so please generate a reassuring explanation."

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

[0427] Point down

[0428] Step 1:

[0429] The terminal receives inquiries from users. Users send inquiries using chat, voice assistants, email, social media, etc. The input at this time is the user's inquiry information, and the output is the data sent directly to the server. Specifically, the terminal sends data to the server when it receives an inquiry.

[0430] Step 2:

[0431] The server standardizes the received queries. The input is query data sent from the terminal, and the output is in a standardized data format. As part of the data processing, it converts queries in different formats into a unified format. Specifically, this involves unifying character encodings and removing unnecessary tags.

[0432] Step 3:

[0433] The server analyzes queries using a natural language processing model. The input is standardized query data, and the output is the result of identifying the user's intent. The data calculation involves analyzing the query text to identify what information and responses are needed. Specific operations include keyword extraction and intent classification.

[0434] Step 4:

[0435] The server identifies the user's emotions using an emotion analysis engine. The input is the analyzed query data, and the output is the user's emotional state (e.g., anger, confusion, satisfaction). As a data calculation, it analyzes emotions from text or audio to determine the priority and style of response. Specifically, it evaluates emotion-related words and expressions within the text.

[0436] Step 5:

[0437] The server collects relevant information from sources based on identified intentions and sentiments. The input is intention and sentiment information, and the output is a collection of relevant information. As a data operation, it accesses sources and picks out the information necessary to resolve the user's inquiry. Specific actions include accessing FAQ databases and calling external APIs.

[0438] Step 6:

[0439] The server generates responses in natural language based on collected information and emotional data. The input is relevant information and user emotions, and the output is the generated response message. A generative AI model is used to select expressions appropriate to the user's psychological state. Specifically, it synthesizes response sentences with appropriate wording and tone.

[0440] Step 7:

[0441] The server sends the generated response to the terminal via the original communication method. The input is the generated response message, and the output is the response that reaches the user. Specifically, the information is encoded and returned to the original channel.

[0442] Step 8:

[0443] The user can receive a response and send feedback back to the server. The server receives this feedback and evaluates the effectiveness of the response. The input is user feedback information, and the output is the feedback analysis result. As a data calculation, problems and areas for improvement are analyzed from the feedback and recorded as training data. Specifically, positive and negative analysis of the feedback is performed.

[0444] 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.

[0445] 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.

[0446] 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.

[0447] [Third Embodiment]

[0448] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0449] 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.

[0450] 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).

[0451] 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.

[0452] 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.

[0453] 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).

[0454] 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.

[0455] 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.

[0456] 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.

[0457] 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.

[0458] 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.

[0459] 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".

[0460] This invention utilizes a multi-functional customer service system to improve the work efficiency of sales staff and promote the use of paid leave. This system can receive user inquiries through multiple communication channels, process them effectively, and return responses.

[0461] First, the user submits an inquiry via chat, voice assistant, email, or social media. The server receives these inquiries and converts them into a standardized format. This conversion makes it easy for the server to parse data from different channels.

[0462] Next, the server uses the generated natural language processing model to analyze the query and identify the user's intent. Based on this analysis, the server categorizes the query into relevant categories and determines what information needs to be retrieved.

[0463] The server queries FAQ databases, customer information systems, and external APIs to gather information related to the identified intent. Based on this information, the server generates a response in natural language and ensures that its content is clear to the user.

[0464] The generated response is sent from the server to the user using the original communication method. The user receiving and reviewing this response facilitates problem resolution and further support.

[0465] Furthermore, the server evaluates user feedback and verifies the effectiveness of its responses. This improves the quality of responses to future inquiries. The collected feedback is used to improve the generative AI model, thus increasing the overall response accuracy of the system.

[0466] As a concrete example, consider a case where a user asks "What is the latest product price?" via chat. In this case, the server analyzes the inquiry and determines it is a "request for product information." It retrieves information from the FAQ database, which includes pricing information, and generates a response in natural language such as "The latest product price is XX yen," which it then sends to the user. The appropriateness of this response is then further improved based on the user's reaction and feedback.

[0467] This invention aims to improve the quality of customer service by providing efficient support for sales personnel through these processes.

[0468] The following describes the processing flow.

[0469] Step 1:

[0470] A user submits an inquiry via chat, voice assistant, email, or social media. This initiates an information request to the customer support system.

[0471] Step 2:

[0472] The server receives the query. Since the received information comes from multiple communication methods, the server converts each piece of data into a standardized format and prepares it for analysis.

[0473] Step 3:

[0474] The server uses a generated natural language processing model to analyze the query. Through this analysis, the server identifies the user's intent and determines the appropriate response category.

[0475] Step 4:

[0476] Based on the identified intent, the server accesses FAQ databases, customer information systems, and external APIs to collect necessary information in order to obtain relevant data.

[0477] Step 5:

[0478] Based on the information collected by the server, a response is generated in natural language. In this process, the generative AI summarizes the key points of the information and creates a clear answer.

[0479] Step 6:

[0480] The server sends the generated response to the user via the communication method used for the initial inquiry. This is done according to the method the user used, such as chat or email.

[0481] Step 7:

[0482] The user receives and reviews the response. The user can ask additional questions or provide feedback on the response as needed.

[0483] Step 8:

[0484] The server receives and analyzes user feedback. Based on this analysis, the effectiveness of the response is evaluated, and the training database of the generative AI model is updated, contributing to future improvements in responses.

[0485] (Example 1)

[0486] 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."

[0487] There is a need to improve the work efficiency of sales staff, ensure uninterrupted customer support even when the assigned representative is absent, and maintain customer satisfaction. Furthermore, it is necessary to continuously improve the quality of responses to user inquiries and enhance the accuracy of those inquiries. To address these challenges, there is a lack of efficient and flexible customer support systems.

[0488] 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.

[0489] In this invention, the server includes means for acquiring queries via multiple communication technologies and converting them into a standardized data format, means for analyzing query information and identifying intent using a natural language processing device, and means for collecting necessary information from relevant information sources based on the identified intent. This makes it possible to provide efficient and flexible customer service and continuously improve the quality of responses.

[0490] "Inquiry" refers to requests for information or support submitted by users.

[0491] "Communication technology" refers to the means and methods for sending and receiving data, and includes chat, voice, email, and social networking services (SNS).

[0492] A "standardized data format" refers to a method of representing data based on conventions used to unify data in different formats.

[0493] A "natural language processing system" refers to the technology and devices that enable computers to understand and analyze human language.

[0494] "Identifying intent" refers to the act of analyzing a user's inquiry to determine its main point and purpose.

[0495] "Information source" refers to a database, system, or external service accessed to retrieve the necessary information.

[0496] A "generation device" refers to a technology or system used to create a response in natural language based on collected information.

[0497] "Response" refers to the feedback or reply that a user gives to the information or service they receive.

[0498] This invention provides a system in which a server plays a central role in receiving user inquiries through multiple communication technologies. This system utilizes communication methods such as chat, voice assistants, email, and social networking services (SNS), and the server converts the inquiries received through these means into a standardized data format. For example, formats such as XML or JSON are used for conversion, and appropriate libraries are employed for data formatting.

[0499] The server, acting as a generative AI model, uses a natural language processing unit (NLP) to analyze and identify the user's intent. Specifically, existing technologies such as "GPT" and "BERT" are applied. Based on the analysis results, the server collects necessary information from sources related to the identified intent. This information collection process includes access to database management systems (DBMS) and external APIs.

[0500] Based on the information collected, the server uses a generator to produce a response in natural language. This generation process uses prompts such as, "Create an appropriate response to the user's inquiry." The generated response is then sent to the user via the original communication technology.

[0501] For example, if a user asks in a chat, "Please tell me the latest product price," the server analyzes the message and identifies the intent as a "request for product information." The server then queries sources containing price information, generates a response such as "The latest product price is XX yen," and sends it back to the user's chat.

[0502] Through the process described above, this invention aims to efficiently support sales personnel and improve the quality of customer service.

[0503] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0504] Step 1:

[0505] Users submit inquiries via chat, voice assistant, email, or social media. The input is the user's inquiry message, which is then captured by the system. The server receives messages from these different communication technologies and converts the inquiries into a unified data format. For example, upon receiving a text message, the server converts it to JSON format for standardization. The output is formatted and parseable data.

[0506] Step 2:

[0507] The server analyzes standardized data using a generative AI model to identify the user's intent. The input is standardized query data. The server uses a natural language processing unit to analyze the query content and apply a generative AI model to identify the intent. In this process, tokenization and grammatical analysis are performed, and the output is the result with the intent identified.

[0508] Step 3:

[0509] Based on the identified intent, the server queries relevant information sources to collect the necessary information. The input is the result of the identified intent. The server accesses database management systems and external APIs to retrieve the relevant information. For example, if the request is for price information, it sends an SQL query to the product database to retrieve the corresponding price information. The output of this process is the collected relevant data.

[0510] Step 4:

[0511] The server uses a generator to produce a natural language response based on the collected information. The input is the collected information data. The generating AI model is given a prompt sentence, instructing it to "generate an appropriate response based on this information." The output is a response message in the form of text, which is then formatted in language that is easy for the user to understand.

[0512] Step 5:

[0513] The server sends the generated response to the user via the original communication method. The input is the generated response message. The server sends the response back to the user using a chat API, mail server, or SNS platform. As output, the user receives the corresponding message, reviews it, and takes appropriate action.

[0514] Step 6:

[0515] The server collects user feedback, analyzes it, and evaluates the effectiveness of its responses. The input is user feedback. The server uses feedback analysis tools to analyze the feedback and identify areas for improvement. This process yields an evaluation as output, which can be used to improve the generated AI model.

[0516] (Application Example 1)

[0517] 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."

[0518] In today's business environment, there is a demand for reduced workload for sales staff and more efficient customer service. However, conventional systems often experience delays in service when sales staff are absent, leading to decreased customer satisfaction and reduced operational efficiency. Furthermore, it is difficult to improve the system using feedback, making it challenging to adequately improve the quality of service. This invention aims to solve these problems and achieve more effective customer service and system improvement.

[0519] 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.

[0520] In this invention, the server includes means for receiving inquiries via multiple communication technologies and converting them into a unified format, means for analyzing the content of the inquiries and identifying their intent using the generated natural language processing technology, and means for providing a function to immediately respond in natural language to inquiries made by users through voice and visual devices. This enables stable customer support even when sales staff are absent and allows for continuous improvement through the use of feedback.

[0521] An "inquiry" is a request from a user for information or support to resolve a problem.

[0522] "Communication technology" refers to the technical means of transmitting information through different channels.

[0523] A "unified format" is a transformation of information from different formats into a consistent state, intended to facilitate analysis.

[0524] "Natural language processing technology" refers to technologies for analyzing, understanding, and generating human language.

[0525] "Intention" refers to the purpose or request that the user is trying to obtain through their inquiry.

[0526] "Information resources" is a general term for databases and external information sources that are referenced in order to provide answers.

[0527] "Responding in natural language" refers to a response generated in a format that is intuitively understandable to the user.

[0528] "Audio and visual devices" are hardware devices that allow users to exchange information through sound and vision.

[0529] "Feedback" refers to the responses, evaluations, and opinions that users provide regarding services.

[0530] "Training" is the process of repeatedly learning by utilizing evaluation results to improve the performance of a system or model.

[0531] This invention realizes a system that receives inquiries using various communication technologies and generates responses using natural language processing technology. The following describes specific embodiments of the invention.

[0532] The server receives inquiries from users via devices such as smartphones and smart glasses. These inquiries are converted into a standardized format and analyzed using natural language processing techniques. Based on the analysis, the server identifies the user's intent and collects necessary data from relevant information resources. This system utilizes generative AI models, particularly OpenAI, to generate natural language responses that are intuitively understandable to the user.

[0533] The generated response is sent to the user via the original communication technology. For example, if a user makes a voice inquiry saying, "I want to check my usage details for this month," the server understands the inquiry, immediately collects the information to display the latest usage details, and presents that information on the smart device.

[0534] User feedback verifies the quality of the system's responses and is used as training data to improve the model. This mechanism allows the accuracy and reliability of responses to improve over time.

[0535] As a concrete example, we will use the following example prompt:

[0536] "Generate an appropriate response when a user requests their monthly billing statement."

[0537] This system makes it possible to maintain consistent customer service and improve the user experience even when sales staff are absent.

[0538] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0539] Step 1:

[0540] The user sends an inquiry via their device. The inquiry can be sent as voice, text, or visual input. The device forwards this to the server, providing initial data for processing.

[0541] Step 2:

[0542] The server prepares to parse the received query. First, it converts the query into a unified format, even if it contains data in different formats. This conversion ensures data consistency and facilitates subsequent processing. The input is the raw query data, and the output is data in a unified format.

[0543] Step 3:

[0544] The server analyzes the query data transformed using a generative AI model. During the analysis, the generative AI model utilizes natural language processing to understand the subject and purpose of the query. The input is data in a unified format, and the output is information indicating the intent of the query. For example, if a user makes a query about pricing, the related intent is identified.

[0545] Step 4:

[0546] Based on the identified intent, the server queries relevant information resources. In this step, the server retrieves data related to the query from databases, external APIs, etc. The input is the query intent, and the output is data containing the necessary information. For example, the most recent billing information is retrieved for a billing inquiry intent.

[0547] Step 5:

[0548] Based on the collected information, the server generates a response in natural language. It utilizes a generative AI model to construct answers that are intuitively understandable to the user. The input is the collected information, and the output is a natural language response. Specifically, it is provided in the form of "Your latest usage details are XX."

[0549] Step 6:

[0550] The server sends the generated response to the user via the original communication method. The terminal displays or speaks this response to the user, conveying the information. The input is the generated response statement, and the output is the information provided to the user.

[0551] Step 7:

[0552] Users provide feedback on the response. This feedback is collected and analyzed by the server. This information is used to improve the model, aiming to increase the accuracy of future responses. The input is user feedback, and the output is data intended for improvement.

[0553] 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.

[0554] This invention aims to improve communication with customers by combining emotion recognition with an automated response system for inquiries. This system centrally processes inquiries received through multiple communication channels, identifies the user's emotions using an emotion engine, and generates a response based on those emotions.

[0555] First, users submit inquiries through any communication method they choose, such as chat, voice assistants, email, or social media. The server receives these inquiries and standardizes the data format. This integrates data from different channels, making it easier to analyze.

[0556] Next, the server uses a natural language processing model to analyze the query and identify the user's intent. Furthermore, it uses an emotion engine to analyze the emotions contained in the text and voice. Emotions can, for example, identify whether the user is angry, confused, or satisfied.

[0557] Based on identified intentions and emotions, the server accesses FAQ databases, customer information systems, and external services to collect relevant information. This information is then comprehensively evaluated in light of newly acquired emotional information.

[0558] The server then generates an optimized response in natural language based on the collected information and sentiment analysis. The response is tailored to the user's emotions and adjusted as needed. For example, an angry user will be provided with a formal response emphasizing quick problem resolution.

[0559] The generated response is sent back to the original communication method via the server and then sent to the user. This allows the user to receive the response quickly.

[0560] Furthermore, user responses and feedback are analyzed by the server, and the effectiveness of the responses is evaluated. This feedback and sentiment analysis results are added to the training data of the generative model and used to improve the sentiment relevance of future responses.

[0561] As a concrete example, consider a scenario where a user sends a message via chat saying, "My ordered item hasn't arrived yet, and I'm very unhappy." In response to this inquiry, the server generates a response that demonstrates understanding and prompt action, such as, "We understand you're distressed that your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." This process aims to improve customer satisfaction by providing a response that emotionally satisfies the user.

[0562] The following describes the processing flow.

[0563] Step 1:

[0564] Users send inquiries using various communication methods (chat, voice assistant, email, social media, etc.). These messages may contain emotional content.

[0565] Step 2:

[0566] The server receives the query. It converts the received data into a standardized format and prepares it for data analysis. This ensures consistent processing of data from different input channels.

[0567] Step 3:

[0568] The server uses a generated natural language processing (NLP) model to analyze the query and identify the user's intent. This helps it understand what kind of information needs to be returned.

[0569] Step 4:

[0570] The server activates the emotion engine and identifies the user's emotions from the analyzed text data. In this step, the emotion engine determines whether the user is angry, excited, calm, etc.

[0571] Step 5:

[0572] The server collects relevant information based on the identified intent and sentiment. This information is obtained by the server using FAQ databases, customer information systems, and external APIs.

[0573] Step 6:

[0574] The server generates responses in natural language based on collected information and sentiment data. These responses are sensitive to the user's emotions and are adjusted as needed. For example, if the user is angry, the responses may emphasize solutions.

[0575] Step 7:

[0576] The server sends the generated response to the user via the original communication method. This allows the user to receive service and support quickly.

[0577] Step 8:

[0578] The system sends feedback to the server regarding the response received by the user. This feedback includes the effectiveness of the response and the user's emotional reaction.

[0579] Step 9:

[0580] The server receives feedback, analyzes the user's response, and evaluates its effectiveness. Based on this evaluation, the server continuously learns and improves its emotion engine and generative AI models to improve the emotional relevance of future responses.

[0581] (Example 2)

[0582] 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."

[0583] The diversification of information and communication technologies has made it difficult to maintain effective communication with users. Accurately understanding user intent and providing responses that consider appropriate emotions is essential for improving customer satisfaction. However, current technology has a problem in that it is difficult to effectively utilize sentiment analysis in generating such responses.

[0584] 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.

[0585] In this invention, the server includes means for receiving queries via multiple information transmission means and converting them into a unified format; means for analyzing the content of the queries and identifying their purpose using generated natural language processing techniques; and means for generating responses in emotionally sensitive natural language based on the collected information and user sentiment analysis. This makes it possible to provide timely and appropriate responses that take user sentiment into account, even for queries transmitted through diverse information channels.

[0586] An "inquiry" is a message or request that a user sends seeking some kind of information or assistance.

[0587] "Information transmission means" refers to a variety of methods and platforms for users to communicate with a server, such as chat, voice assistants, email, and social networking services (SNS).

[0588] A "standardized format" is a data format that has been standardized to make it easier to analyze data received from different means of information transmission.

[0589] "Natural language processing technology" is a technology that allows computers to understand and analyze human language, and is used to grasp the user's intent.

[0590] "Purpose" refers to the objective or intention that a user expects when making an inquiry.

[0591] "Sentiment analysis" is a technology that identifies a user's emotional state from the words and expressions contained in their inquiries.

[0592] An "information source" is a database or external system that is referenced to obtain an appropriate answer in response to a user's inquiry.

[0593] "Generating responses in natural language" refers to the process by which a computer creates an answer to a user's inquiry using language that is understandable to humans.

[0594] "Response" refers to feedback from users, including their impressions and suggestions for improvement, regarding the response provided.

[0595] This invention relates to a system in which a server centrally receives and processes inquiries made by a user through multiple information transmission means, and generates responses using sentiment analysis. This system includes a server and a user terminal as hardware, and incorporates natural language processing technology and a sentiment analysis engine as software.

[0596] Users submit inquiries using various means of information transmission, such as chat, voice assistants, email, and social media. User devices have access to these information transmission platforms and the ability to send inquiries to the server. Upon receiving an inquiry, the server first converts the data into a standardized format. This standardizes data from different communication methods, ensuring consistency in analysis.

[0597] Next, the server analyzes the query using natural language processing techniques to identify the user's purpose. Various natural language processing platforms, such as GPT and BERT, which are generative AI models, are used here. In parallel, an emotion analysis engine (e.g., an emotion recognition API) analyzes the user's emotional state from the text information and identifies emotions such as anger, anxiety, and joy.

[0598] Based on this information, the server accesses relevant information sources (FAQ databases and external services) to collect the necessary information. Considering the collected information and the results of sentiment analysis, the server generates a natural language response that takes emotions into account. In this process, a generative AI model is used, and input in the form of text, such as prompts, is used. For example, a prompt might read: "Generate an appropriate response based on the user's emotions. Inquiry: 'I am dissatisfied because the product I ordered has not yet arrived.'"

[0599] The generated response is returned to the user's original information transmission method via the server and then sent back to the user. The user is expected to receive this response and experience greater satisfaction due to the prompt and emotionally appropriate response. This allows users to use the service with confidence and improves the quality of customer service provided by the company.

[0600] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0601] Step 1:

[0602] Users use their devices to submit inquiries through various means of communication, such as chat, voice assistants, email, and social media. Input is raw text or voice data. These inquiries are then forwarded to the server as output.

[0603] Step 2:

[0604] The server converts the query data received from the server from a format specific to each information transmission method to a standardized format. The input is query data in a heterogeneous format, and the output is a standardized data format. This process allows subsequent analysis to proceed linearly.

[0605] Step 3:

[0606] The server analyzes data in a standardized format using natural language processing techniques. The input is standardized text data, and the output consists of keywords and intents identified as the user's purpose. This analysis utilizes AI models (e.g., natural language processing platforms) to process the data and identify the user's intent.

[0607] Step 4:

[0608] Simultaneously, the server uses an emotion analysis engine to identify the user's emotions from the text. The input is standardized text, and the output is a set of emotional state labels (e.g., anger, anxiety, joy, etc.). This allows the user's psychological state to be recognized.

[0609] Step 5:

[0610] Based on identified intent and sentiment information, the server accesses relevant sources to collect the necessary information. The input is the user's intent and the type of information associated with it, while the output is a specific set of information to respond to the user's query. Databases and external APIs are utilized here.

[0611] Step 6:

[0612] Using the collected information and the results of sentiment analysis, the server generates a response in natural language through a generative AI model. The input is an information set and sentiment labels, and the output is a sentiment-sensitive natural language response. A prompt is used, and the AI ​​model generates the text.

[0613] Step 7:

[0614] The generated response is sent via the server to the user's original information transmission means. The input is the generated response, and the output is the response displayed or played back as audio on the user's terminal. Specifically, the response is provided to the user visually or audibly.

[0615] Step 8:

[0616] The server evaluates the effectiveness of its responses by receiving user responses or feedback. The input is user feedback data, and the output is evaluation information regarding the effectiveness of the responses. This data is used as feedback in the learning process of the generative AI model to improve the quality of the responses.

[0617] (Application Example 2)

[0618] 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."

[0619] Traditional automated response systems provide a uniform response to user inquiries, making it difficult to offer appropriate responses that reflect the user's emotions. Furthermore, the lack of emotionally responsive responses hindered user satisfaction. As a result, there was a need to improve the quality of communication and customer satisfaction.

[0620] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving queries via multiple communication means and converting them into a standardized format, means for analyzing the content of the query and identifying the intent using a generated natural language processing model, and means for identifying the user's emotions using an emotion analysis engine. This makes it possible to take the user's emotions into consideration and generate and transmit an emotion-appropriate response.

[0621] Point down

[0622] An "inquiry" refers to information such as questions or requests that users submit through communication methods such as chat, voice assistants, email, and social media.

[0623] "Communication methods" refer to the channels and technologies used to send or receive information, and specifically include chat, voice assistants, email, and social networking services (SNS).

[0624] A "standardized format" is the process of converting data received from different channels into a unified format that facilitates analysis.

[0625] A "natural language processing model" is a collection of algorithms and technologies for understanding and appropriately analyzing human language.

[0626] "Intent" refers to the purpose or request that is understood by identifying what the user is seeking in their inquiry.

[0627] A "sentiment analysis engine" is a collection of algorithms and technologies that identify a user's emotions from text or audio.

[0628] "Information source" refers to a database or system that is queried to collect relevant information based on the user's intent.

[0629] "Feedback" refers to evaluations and opinions that users provide regarding responses, and this information is used to improve the system's performance.

[0630] A "generative model" is a set of algorithms and techniques trained to create the best possible response to a user's inquiry.

[0631] The system implementing this invention consists of a server and multiple terminals working together. Users send inquiries via terminals using communication methods such as chat, voice assistants, email, and social networking services. The server receives these inquiries and first converts them into a standardized format. This unifies data from different channels and makes analysis easier.

[0632] Next, the server uses a natural language processing model to analyze the content of the query and identify the user's intent. Simultaneously, it uses an emotion analysis engine to identify the user's emotions from the text and voice. This emotion analysis determines the user's psychological state.

[0633] Subsequently, based on the identified intent and sentiment, the server collects relevant information from various sources. These sources include FAQ databases, customer information systems, and external services. This makes it possible to obtain accurate and relevant information in response to user inquiries.

[0634] Based on the collected information and emotional data, the server uses a generative AI model to generate a natural language-optimized response. This response takes the user's emotions into consideration and is adjusted as needed. For example, if the user is emotionally dissatisfied, the response may be designed to emphasize rapid problem resolution.

[0635] Finally, the server sends the generated response to the user via the original communication method. This allows the user to receive a quick and appropriate response. The user's feedback is also analyzed by the server to evaluate the effectiveness of the response. This feedback and sentiment analysis results are used as training data for the generative model to improve the sentiment relevance of future responses.

[0636] For example, if a user enters via their device, "My ordered item hasn't arrived yet, and I'm very unhappy," the server will generate a response such as, "It seems you're having trouble because your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." In this way, it becomes possible to provide appropriate responses tailored to the user's requests.

[0637] A concrete example of a prompt message for a generative AI model would be: "The user seems confused. They want to know more details about product delivery, so please generate a reassuring explanation."

[0638] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0639] Point down

[0640] Step 1:

[0641] The terminal receives inquiries from users. Users send inquiries using chat, voice assistants, email, social media, etc. The input at this time is the user's inquiry information, and the output is the data sent directly to the server. Specifically, the terminal sends data to the server when it receives an inquiry.

[0642] Step 2:

[0643] The server standardizes the received queries. The input is query data sent from the terminal, and the output is in a standardized data format. As part of the data processing, it converts queries in different formats into a unified format. Specifically, this involves unifying character encodings and removing unnecessary tags.

[0644] Step 3:

[0645] The server analyzes queries using a natural language processing model. The input is standardized query data, and the output is the result of identifying the user's intent. The data calculation involves analyzing the query text to identify what information and responses are needed. Specific operations include keyword extraction and intent classification.

[0646] Step 4:

[0647] The server identifies the user's emotions using an emotion analysis engine. The input is the analyzed query data, and the output is the user's emotional state (e.g., anger, confusion, satisfaction). As a data calculation, it analyzes emotions from text or audio to determine the priority and style of response. Specifically, it evaluates emotion-related words and expressions within the text.

[0648] Step 5:

[0649] The server collects relevant information from sources based on identified intentions and sentiments. The input is intention and sentiment information, and the output is a collection of relevant information. As a data operation, it accesses sources and picks out the information necessary to resolve the user's inquiry. Specific actions include accessing FAQ databases and calling external APIs.

[0650] Step 6:

[0651] The server generates responses in natural language based on collected information and emotional data. The input is relevant information and user emotions, and the output is the generated response message. A generative AI model is used to select expressions appropriate to the user's psychological state. Specifically, it synthesizes response sentences with appropriate wording and tone.

[0652] Step 7:

[0653] The server sends the generated response to the terminal via the original communication method. The input is the generated response message, and the output is the response that reaches the user. Specifically, the information is encoded and returned to the original channel.

[0654] Step 8:

[0655] The user can receive a response and send feedback back to the server. The server receives this feedback and evaluates the effectiveness of the response. The input is user feedback information, and the output is the feedback analysis result. As a data calculation, problems and areas for improvement are analyzed from the feedback and recorded as training data. Specifically, positive and negative analysis of the feedback is performed.

[0656] 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.

[0657] 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.

[0658] 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.

[0659] [Fourth Embodiment]

[0660] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0661] 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.

[0662] 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).

[0663] 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.

[0664] 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.

[0665] 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).

[0666] 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.

[0667] 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.

[0668] 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.

[0669] 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.

[0670] 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.

[0671] 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.

[0672] 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".

[0673] This invention utilizes a multi-functional customer service system to improve the work efficiency of sales staff and promote the use of paid leave. This system can receive user inquiries through multiple communication channels, process them effectively, and return responses.

[0674] First, the user submits an inquiry via chat, voice assistant, email, or social media. The server receives these inquiries and converts them into a standardized format. This conversion makes it easy for the server to parse data from different channels.

[0675] Next, the server uses the generated natural language processing model to analyze the query and identify the user's intent. Based on this analysis, the server categorizes the query into relevant categories and determines what information needs to be retrieved.

[0676] The server queries FAQ databases, customer information systems, and external APIs to gather information related to the identified intent. Based on this information, the server generates a response in natural language and ensures that its content is clear to the user.

[0677] The generated response is sent from the server to the user using the original communication method. The user receiving and reviewing this response facilitates problem resolution and further support.

[0678] Furthermore, the server evaluates user feedback and verifies the effectiveness of its responses. This improves the quality of responses to future inquiries. The collected feedback is used to improve the generative AI model, thus increasing the overall response accuracy of the system.

[0679] As a concrete example, consider a case where a user asks "What is the latest product price?" via chat. In this case, the server analyzes the inquiry and determines it is a "request for product information." It retrieves information from the FAQ database, which includes pricing information, and generates a response in natural language such as "The latest product price is XX yen," which it then sends to the user. The appropriateness of this response is then further improved based on the user's reaction and feedback.

[0680] This invention aims to improve the quality of customer service by providing efficient support for sales personnel through these processes.

[0681] The following describes the processing flow.

[0682] Step 1:

[0683] A user submits an inquiry via chat, voice assistant, email, or social media. This initiates an information request to the customer support system.

[0684] Step 2:

[0685] The server receives the query. Since the received information comes from multiple communication methods, the server converts each piece of data into a standardized format and prepares it for analysis.

[0686] Step 3:

[0687] The server uses a generated natural language processing model to analyze the query. Through this analysis, the server identifies the user's intent and determines the appropriate response category.

[0688] Step 4:

[0689] Based on the identified intent, the server accesses FAQ databases, customer information systems, and external APIs to collect necessary information in order to obtain relevant data.

[0690] Step 5:

[0691] Based on the information collected by the server, a response is generated in natural language. In this process, the generative AI summarizes the key points of the information and creates a clear answer.

[0692] Step 6:

[0693] The server sends the generated response to the user via the communication method used for the initial inquiry. This is done according to the method the user used, such as chat or email.

[0694] Step 7:

[0695] The user receives and reviews the response. The user can ask additional questions or provide feedback on the response as needed.

[0696] Step 8:

[0697] The server receives and analyzes user feedback. Based on this analysis, the effectiveness of the response is evaluated, and the training database of the generative AI model is updated, contributing to future improvements in responses.

[0698] (Example 1)

[0699] 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".

[0700] There is a need to improve the work efficiency of sales staff, ensure uninterrupted customer support even when the assigned representative is absent, and maintain customer satisfaction. Furthermore, it is necessary to continuously improve the quality of responses to user inquiries and enhance the accuracy of those inquiries. To address these challenges, there is a lack of efficient and flexible customer support systems.

[0701] 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.

[0702] In this invention, the server includes means for acquiring queries via multiple communication technologies and converting them into a standardized data format, means for analyzing query information and identifying intent using a natural language processing device, and means for collecting necessary information from relevant information sources based on the identified intent. This makes it possible to provide efficient and flexible customer service and continuously improve the quality of responses.

[0703] "Inquiry" refers to requests for information or support submitted by users.

[0704] "Communication technology" refers to the means and methods for sending and receiving data, and includes chat, voice, email, and social networking services (SNS).

[0705] A "standardized data format" refers to a method of representing data based on conventions used to unify data in different formats.

[0706] A "natural language processing system" refers to the technology and devices that enable computers to understand and analyze human language.

[0707] "Identifying intent" refers to the act of analyzing a user's inquiry to determine its main point and purpose.

[0708] "Information source" refers to a database, system, or external service accessed to retrieve the necessary information.

[0709] A "generation device" refers to a technology or system used to create a response in natural language based on collected information.

[0710] "Response" refers to the feedback or reply that a user gives to the information or service they receive.

[0711] This invention provides a system in which a server plays a central role in receiving user inquiries through multiple communication technologies. This system utilizes communication methods such as chat, voice assistants, email, and social networking services (SNS), and the server converts the inquiries received through these means into a standardized data format. For example, formats such as XML or JSON are used for conversion, and appropriate libraries are employed for data formatting.

[0712] The server, acting as a generative AI model, uses a natural language processing unit (NLP) to analyze and identify the user's intent. Specifically, existing technologies such as "GPT" and "BERT" are applied. Based on the analysis results, the server collects necessary information from sources related to the identified intent. This information collection process includes access to database management systems (DBMS) and external APIs.

[0713] Based on the information collected, the server uses a generator to produce a response in natural language. This generation process uses prompts such as, "Create an appropriate response to the user's inquiry." The generated response is then sent to the user via the original communication technology.

[0714] For example, if a user asks in a chat, "Please tell me the latest product price," the server analyzes the message and identifies the intent as a "request for product information." The server then queries sources containing price information, generates a response such as "The latest product price is XX yen," and sends it back to the user's chat.

[0715] Through the process described above, this invention aims to efficiently support sales personnel and improve the quality of customer service.

[0716] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0717] Step 1:

[0718] Users submit inquiries via chat, voice assistant, email, or social media. The input is the user's inquiry message, which is then captured by the system. The server receives messages from these different communication technologies and converts the inquiries into a unified data format. For example, upon receiving a text message, the server converts it to JSON format for standardization. The output is formatted and parseable data.

[0719] Step 2:

[0720] The server analyzes standardized data using a generative AI model to identify the user's intent. The input is standardized query data. The server uses a natural language processing unit to analyze the query content and apply a generative AI model to identify the intent. In this process, tokenization and grammatical analysis are performed, and the output is the result with the intent identified.

[0721] Step 3:

[0722] Based on the identified intent, the server queries relevant information sources to collect the necessary information. The input is the result of the identified intent. The server accesses database management systems and external APIs to retrieve the relevant information. For example, if the request is for price information, it sends an SQL query to the product database to retrieve the corresponding price information. The output of this process is the collected relevant data.

[0723] Step 4:

[0724] The server uses a generator to produce a natural language response based on the collected information. The input is the collected information data. The generating AI model is given a prompt sentence, instructing it to "generate an appropriate response based on this information." The output is a response message in the form of text, which is then formatted in language that is easy for the user to understand.

[0725] Step 5:

[0726] The server sends the generated response to the user via the original communication method. The input is the generated response message. The server sends the response back to the user using a chat API, mail server, or SNS platform. As output, the user receives the corresponding message, reviews it, and takes appropriate action.

[0727] Step 6:

[0728] The server collects user feedback, analyzes it, and evaluates the effectiveness of its responses. The input is user feedback. The server uses feedback analysis tools to analyze the feedback and identify areas for improvement. This process yields an evaluation as output, which can be used to improve the generated AI model.

[0729] (Application Example 1)

[0730] 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".

[0731] In today's business environment, there is a demand for reduced workload for sales staff and more efficient customer service. However, conventional systems often experience delays in service when sales staff are absent, leading to decreased customer satisfaction and reduced operational efficiency. Furthermore, it is difficult to improve the system using feedback, making it challenging to adequately improve the quality of service. This invention aims to solve these problems and achieve more effective customer service and system improvement.

[0732] 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.

[0733] In this invention, the server includes means for receiving inquiries via multiple communication technologies and converting them into a unified format, means for analyzing the content of the inquiries and identifying their intent using the generated natural language processing technology, and means for providing a function to immediately respond in natural language to inquiries made by users through voice and visual devices. This enables stable customer support even when sales staff are absent and allows for continuous improvement through the use of feedback.

[0734] An "inquiry" is a request from a user for information or support to resolve a problem.

[0735] "Communication technology" refers to the technical means of transmitting information through different channels.

[0736] A "unified format" is a transformation of information from different formats into a consistent state, intended to facilitate analysis.

[0737] "Natural language processing technology" refers to technologies for analyzing, understanding, and generating human language.

[0738] "Intention" refers to the purpose or request that the user is trying to obtain through their inquiry.

[0739] "Information resources" is a general term for databases and external information sources that are referenced in order to provide answers.

[0740] "Responding in natural language" refers to a response generated in a format that is intuitively understandable to the user.

[0741] "Audio and visual devices" are hardware devices that allow users to exchange information through sound and vision.

[0742] "Feedback" refers to the responses, evaluations, and opinions that users provide regarding services.

[0743] "Training" is the process of repeatedly learning by utilizing evaluation results to improve the performance of a system or model.

[0744] This invention realizes a system that receives inquiries using various communication technologies and generates responses using natural language processing technology. The following describes specific embodiments of the invention.

[0745] The server receives inquiries from users via devices such as smartphones and smart glasses. These inquiries are converted into a standardized format and analyzed using natural language processing techniques. Based on the analysis, the server identifies the user's intent and collects necessary data from relevant information resources. This system utilizes generative AI models, particularly OpenAI, to generate natural language responses that are intuitively understandable to the user.

[0746] The generated response is sent to the user via the original communication technology. For example, if a user makes a voice inquiry saying, "I want to check my usage details for this month," the server understands the inquiry, immediately collects the information to display the latest usage details, and presents that information on the smart device.

[0747] User feedback verifies the quality of the system's responses and is used as training data to improve the model. This mechanism allows the accuracy and reliability of responses to improve over time.

[0748] As a concrete example, we will use the following example prompt:

[0749] "Generate an appropriate response when a user requests their monthly billing statement."

[0750] This system makes it possible to maintain consistent customer service and improve the user experience even when sales staff are absent.

[0751] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0752] Step 1:

[0753] The user sends an inquiry via their device. The inquiry can be sent as voice, text, or visual input. The device forwards this to the server, providing initial data for processing.

[0754] Step 2:

[0755] The server prepares to parse the received query. First, it converts the query into a unified format, even if it contains data in different formats. This conversion ensures data consistency and facilitates subsequent processing. The input is the raw query data, and the output is data in a unified format.

[0756] Step 3:

[0757] The server analyzes the query data transformed using a generative AI model. During the analysis, the generative AI model utilizes natural language processing to understand the subject and purpose of the query. The input is data in a unified format, and the output is information indicating the intent of the query. For example, if a user makes a query about pricing, the related intent is identified.

[0758] Step 4:

[0759] Based on the identified intent, the server queries relevant information resources. In this step, the server retrieves data related to the query from databases, external APIs, etc. The input is the query intent, and the output is data containing the necessary information. For example, the most recent billing information is retrieved for a billing inquiry intent.

[0760] Step 5:

[0761] Based on the collected information, the server generates a response in natural language. It utilizes a generative AI model to construct answers that are intuitively understandable to the user. The input is the collected information, and the output is a natural language response. Specifically, it is provided in the form of "Your latest usage details are XX."

[0762] Step 6:

[0763] The server sends the generated response to the user via the original communication method. The terminal displays or speaks this response to the user, conveying the information. The input is the generated response statement, and the output is the information provided to the user.

[0764] Step 7:

[0765] Users provide feedback on the response. This feedback is collected and analyzed by the server. This information is used to improve the model, aiming to increase the accuracy of future responses. The input is user feedback, and the output is data intended for improvement.

[0766] 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.

[0767] This invention aims to improve communication with customers by combining emotion recognition with an automated response system for inquiries. This system centrally processes inquiries received through multiple communication channels, identifies the user's emotions using an emotion engine, and generates a response based on those emotions.

[0768] First, users submit inquiries through any communication method they choose, such as chat, voice assistants, email, or social media. The server receives these inquiries and standardizes the data format. This integrates data from different channels, making it easier to analyze.

[0769] Next, the server uses a natural language processing model to analyze the query and identify the user's intent. Furthermore, it uses an emotion engine to analyze the emotions contained in the text and voice. Emotions can, for example, identify whether the user is angry, confused, or satisfied.

[0770] Based on identified intentions and emotions, the server accesses FAQ databases, customer information systems, and external services to collect relevant information. This information is then comprehensively evaluated in light of newly acquired emotional information.

[0771] The server then generates an optimized response in natural language based on the collected information and sentiment analysis. The response is tailored to the user's emotions and adjusted as needed. For example, an angry user will be provided with a formal response emphasizing quick problem resolution.

[0772] The generated response is sent back to the original communication method via the server and then sent to the user. This allows the user to receive the response quickly.

[0773] Furthermore, user responses and feedback are analyzed by the server, and the effectiveness of the responses is evaluated. This feedback and sentiment analysis results are added to the training data of the generative model and used to improve the sentiment relevance of future responses.

[0774] As a concrete example, consider a scenario where a user sends a message via chat saying, "My ordered item hasn't arrived yet, and I'm very unhappy." In response to this inquiry, the server generates a response that demonstrates understanding and prompt action, such as, "We understand you're distressed that your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." This process aims to improve customer satisfaction by providing a response that emotionally satisfies the user.

[0775] The following describes the processing flow.

[0776] Step 1:

[0777] Users send inquiries using various communication methods (chat, voice assistant, email, social media, etc.). These messages may contain emotional content.

[0778] Step 2:

[0779] The server receives the query. It converts the received data into a standardized format and prepares it for data analysis. This ensures consistent processing of data from different input channels.

[0780] Step 3:

[0781] The server uses a generated natural language processing (NLP) model to analyze the query and identify the user's intent. This helps it understand what kind of information needs to be returned.

[0782] Step 4:

[0783] The server activates the emotion engine and identifies the user's emotions from the analyzed text data. In this step, the emotion engine determines whether the user is angry, excited, calm, etc.

[0784] Step 5:

[0785] The server collects relevant information based on the identified intent and sentiment. This information is obtained by the server using FAQ databases, customer information systems, and external APIs.

[0786] Step 6:

[0787] The server generates responses in natural language based on collected information and sentiment data. These responses are sensitive to the user's emotions and are adjusted as needed. For example, if the user is angry, the responses may emphasize solutions.

[0788] Step 7:

[0789] The server sends the generated response to the user via the original communication method. This allows the user to receive service and support quickly.

[0790] Step 8:

[0791] The system sends feedback to the server regarding the response received by the user. This feedback includes the effectiveness of the response and the user's emotional reaction.

[0792] Step 9:

[0793] The server receives feedback, analyzes the user's response, and evaluates its effectiveness. Based on this evaluation, the server continuously learns and improves its emotion engine and generative AI models to improve the emotional relevance of future responses.

[0794] (Example 2)

[0795] 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".

[0796] The diversification of information and communication technologies has made it difficult to maintain effective communication with users. Accurately understanding user intent and providing responses that consider appropriate emotions is essential for improving customer satisfaction. However, current technology has a problem in that it is difficult to effectively utilize sentiment analysis in generating such responses.

[0797] 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.

[0798] In this invention, the server includes means for receiving queries via multiple information transmission means and converting them into a unified format; means for analyzing the content of the queries and identifying their purpose using generated natural language processing techniques; and means for generating responses in emotionally sensitive natural language based on the collected information and user sentiment analysis. This makes it possible to provide timely and appropriate responses that take user sentiment into account, even for queries transmitted through diverse information channels.

[0799] An "inquiry" is a message or request that a user sends seeking some kind of information or assistance.

[0800] "Information transmission means" refers to a variety of methods and platforms for users to communicate with a server, such as chat, voice assistants, email, and social networking services (SNS).

[0801] A "standardized format" is a data format that has been standardized to make it easier to analyze data received from different means of information transmission.

[0802] "Natural language processing technology" is a technology that allows computers to understand and analyze human language, and is used to grasp the user's intent.

[0803] "Purpose" refers to the objective or intention that a user expects when making an inquiry.

[0804] "Sentiment analysis" is a technology that identifies a user's emotional state from the words and expressions contained in their inquiries.

[0805] An "information source" is a database or external system that is referenced to obtain an appropriate answer in response to a user's inquiry.

[0806] "Generating responses in natural language" refers to the process by which a computer creates an answer to a user's inquiry using language that is understandable to humans.

[0807] "Response" refers to feedback from users, including their impressions and suggestions for improvement, regarding the response provided.

[0808] This invention relates to a system in which a server centrally receives and processes inquiries made by a user through multiple information transmission means, and generates responses using sentiment analysis. This system includes a server and a user terminal as hardware, and incorporates natural language processing technology and a sentiment analysis engine as software.

[0809] Users submit inquiries using various means of information transmission, such as chat, voice assistants, email, and social media. User devices have access to these information transmission platforms and the ability to send inquiries to the server. Upon receiving an inquiry, the server first converts the data into a standardized format. This standardizes data from different communication methods, ensuring consistency in analysis.

[0810] Next, the server analyzes the query using natural language processing techniques to identify the user's purpose. Various natural language processing platforms, such as GPT and BERT, which are generative AI models, are used here. In parallel, an emotion analysis engine (e.g., an emotion recognition API) analyzes the user's emotional state from the text information and identifies emotions such as anger, anxiety, and joy.

[0811] Based on this information, the server accesses relevant information sources (FAQ databases and external services) to collect the necessary information. Considering the collected information and the results of sentiment analysis, the server generates a natural language response that takes emotions into account. In this process, a generative AI model is used, and input in the form of text, such as prompts, is used. For example, a prompt might read: "Generate an appropriate response based on the user's emotions. Inquiry: 'I am dissatisfied because the product I ordered has not yet arrived.'"

[0812] The generated response is returned to the user's original information transmission method via the server and then sent back to the user. The user is expected to receive this response and experience greater satisfaction due to the prompt and emotionally appropriate response. This allows users to use the service with confidence and improves the quality of customer service provided by the company.

[0813] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0814] Step 1:

[0815] Users use their devices to submit inquiries through various means of communication, such as chat, voice assistants, email, and social media. Input is raw text or voice data. These inquiries are then forwarded to the server as output.

[0816] Step 2:

[0817] The server converts the query data received from the server from a format specific to each information transmission method to a standardized format. The input is query data in a heterogeneous format, and the output is a standardized data format. This process allows subsequent analysis to proceed linearly.

[0818] Step 3:

[0819] The server analyzes data in a standardized format using natural language processing techniques. The input is standardized text data, and the output consists of keywords and intents identified as the user's purpose. This analysis utilizes AI models (e.g., natural language processing platforms) to process the data and identify the user's intent.

[0820] Step 4:

[0821] Simultaneously, the server uses an emotion analysis engine to identify the user's emotions from the text. The input is standardized text, and the output is a set of emotional state labels (e.g., anger, anxiety, joy, etc.). This allows the user's psychological state to be recognized.

[0822] Step 5:

[0823] Based on identified intent and sentiment information, the server accesses relevant sources to collect the necessary information. The input is the user's intent and the type of information associated with it, while the output is a specific set of information to respond to the user's query. Databases and external APIs are utilized here.

[0824] Step 6:

[0825] Using the collected information and the results of sentiment analysis, the server generates a response in natural language through a generative AI model. The input is an information set and sentiment labels, and the output is a sentiment-sensitive natural language response. A prompt is used, and the AI ​​model generates the text.

[0826] Step 7:

[0827] The generated response is sent via the server to the user's original information transmission means. The input is the generated response, and the output is the response displayed or played back as audio on the user's terminal. Specifically, the response is provided to the user visually or audibly.

[0828] Step 8:

[0829] The server evaluates the effectiveness of its responses by receiving user responses or feedback. The input is user feedback data, and the output is evaluation information regarding the effectiveness of the responses. This data is used as feedback in the learning process of the generative AI model to improve the quality of the responses.

[0830] (Application Example 2)

[0831] 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".

[0832] Traditional automated response systems provide a uniform response to user inquiries, making it difficult to offer appropriate responses that reflect the user's emotions. Furthermore, the lack of emotionally responsive responses hindered user satisfaction. As a result, there was a need to improve the quality of communication and customer satisfaction.

[0833] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving queries via multiple communication means and converting them into a standardized format, means for analyzing the content of the query and identifying the intent using a generated natural language processing model, and means for identifying the user's emotions using an emotion analysis engine. This makes it possible to take the user's emotions into consideration and generate and transmit an emotion-appropriate response.

[0834] Point down

[0835] An "inquiry" refers to information such as questions or requests that users submit through communication methods such as chat, voice assistants, email, and social media.

[0836] "Communication methods" refer to the channels and technologies used to send or receive information, and specifically include chat, voice assistants, email, and social networking services (SNS).

[0837] A "standardized format" is the process of converting data received from different channels into a unified format that facilitates analysis.

[0838] A "natural language processing model" is a collection of algorithms and technologies for understanding and appropriately analyzing human language.

[0839] "Intent" refers to the purpose or request that is understood by identifying what the user is seeking in their inquiry.

[0840] A "sentiment analysis engine" is a collection of algorithms and technologies that identify a user's emotions from text or audio.

[0841] "Information source" refers to a database or system that is queried to collect relevant information based on the user's intent.

[0842] "Feedback" refers to evaluations and opinions that users provide regarding responses, and this information is used to improve the system's performance.

[0843] A "generative model" is a set of algorithms and techniques trained to create the best possible response to a user's inquiry.

[0844] The system implementing this invention consists of a server and multiple terminals working together. Users send inquiries via terminals using communication methods such as chat, voice assistants, email, and social networking services. The server receives these inquiries and first converts them into a standardized format. This unifies data from different channels and makes analysis easier.

[0845] Next, the server uses a natural language processing model to analyze the content of the query and identify the user's intent. Simultaneously, it uses an emotion analysis engine to identify the user's emotions from the text and voice. This emotion analysis determines the user's psychological state.

[0846] Subsequently, based on the identified intent and sentiment, the server collects relevant information from various sources. These sources include FAQ databases, customer information systems, and external services. This makes it possible to obtain accurate and relevant information in response to user inquiries.

[0847] Based on the collected information and emotional data, the server uses a generative AI model to generate a natural language-optimized response. This response takes the user's emotions into consideration and is adjusted as needed. For example, if the user is emotionally dissatisfied, the response may be designed to emphasize rapid problem resolution.

[0848] Finally, the server sends the generated response to the user via the original communication method. This allows the user to receive a quick and appropriate response. The user's feedback is also analyzed by the server to evaluate the effectiveness of the response. This feedback and sentiment analysis results are used as training data for the generative model to improve the sentiment relevance of future responses.

[0849] For example, if a user enters via their device, "My ordered item hasn't arrived yet, and I'm very unhappy," the server will generate a response such as, "It seems you're having trouble because your item hasn't arrived. We'll check the situation immediately and do our best to resolve it as quickly as possible." In this way, it becomes possible to provide appropriate responses tailored to the user's requests.

[0850] A concrete example of a prompt message for a generative AI model would be: "The user seems confused. They want to know more details about product delivery, so please generate a reassuring explanation."

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

[0852] Point down

[0853] Step 1:

[0854] The terminal receives inquiries from users. Users send inquiries using chat, voice assistants, email, social media, etc. The input at this time is the user's inquiry information, and the output is the data sent directly to the server. Specifically, the terminal sends data to the server when it receives an inquiry.

[0855] Step 2:

[0856] The server standardizes the received queries. The input is query data sent from the terminal, and the output is in a standardized data format. As part of the data processing, it converts queries in different formats into a unified format. Specifically, this involves unifying character encodings and removing unnecessary tags.

[0857] Step 3:

[0858] The server analyzes queries using a natural language processing model. The input is standardized query data, and the output is the result of identifying the user's intent. The data calculation involves analyzing the query text to identify what information and responses are needed. Specific operations include keyword extraction and intent classification.

[0859] Step 4:

[0860] The server identifies the user's emotions using an emotion analysis engine. The input is the analyzed query data, and the output is the user's emotional state (e.g., anger, confusion, satisfaction). As a data calculation, it analyzes emotions from text or audio to determine the priority and style of response. Specifically, it evaluates emotion-related words and expressions within the text.

[0861] Step 5:

[0862] The server collects relevant information from sources based on identified intentions and sentiments. The input is intention and sentiment information, and the output is a collection of relevant information. As a data operation, it accesses sources and picks out the information necessary to resolve the user's inquiry. Specific actions include accessing FAQ databases and calling external APIs.

[0863] Step 6:

[0864] The server generates responses in natural language based on collected information and emotional data. The input is relevant information and user emotions, and the output is the generated response message. A generative AI model is used to select expressions appropriate to the user's psychological state. Specifically, it synthesizes response sentences with appropriate wording and tone.

[0865] Step 7:

[0866] The server sends the generated response to the terminal via the original communication method. The input is the generated response message, and the output is the response that reaches the user. Specifically, the information is encoded and returned to the original channel.

[0867] Step 8:

[0868] The user can receive a response and send feedback back to the server. The server receives this feedback and evaluates the effectiveness of the response. The input is user feedback information, and the output is the feedback analysis result. As a data calculation, problems and areas for improvement are analyzed from the feedback and recorded as training data. Specifically, positive and negative analysis of the feedback is performed.

[0869] 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.

[0870] 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.

[0871] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0872] 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.

[0873] 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.

[0874] 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.

[0875] 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.

[0876] 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.

[0877] 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."

[0878] 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.

[0879] 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.

[0880] 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.

[0881] 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.

[0882] 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.

[0883] 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.

[0884] 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.

[0885] 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.

[0886] 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.

[0887] 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.

[0888] 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.

[0889] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0890] The following is further disclosed regarding the embodiments described above.

[0891] (Claim 1)

[0892] A means for receiving inquiries via multiple communication methods and converting them into a standardized format,

[0893] A means of analyzing the content of a query and identifying its intent using a generated natural language processing model,

[0894] Means for collecting necessary information from relevant sources based on identified intentions,

[0895] A means of generating a response in natural language based on the collected information,

[0896] A means for transmitting the generated response via the original communication means,

[0897] A means of analyzing user feedback and evaluating the effectiveness of responses,

[0898] A system that includes means to improve the quality of responses by learning a generative model based on evaluation results.

[0899] (Claim 2)

[0900] The system according to claim 1, which provides a function to maintain user support even when sales personnel are absent.

[0901] (Claim 3)

[0902] The system according to claim 1, which updates the database of information sources based on user feedback and improves the accuracy of future queries.

[0903] "Example 1"

[0904] (Claim 1)

[0905] A means for obtaining queries via multiple communication technologies and converting them into a standardized data format,

[0906] A means for analyzing query information and identifying intent using a natural language processing device,

[0907] Means for collecting necessary information from relevant sources based on identified intent,

[0908] A means of generating a response in natural language based on the collected information,

[0909] A means for transmitting the generated response via the original communication technology,

[0910] A means of analyzing user responses and evaluating the effectiveness of those responses,

[0911] A system that includes means for adjusting the generating device based on evaluation results to improve the quality of the response.

[0912] (Claim 2)

[0913] The system according to claim 1, which provides a function to maintain user support even when the person in charge of operations is absent.

[0914] (Claim 3)

[0915] The system according to claim 1, which updates the data set of information sources based on user feedback and improves the accuracy of future queries.

[0916] "Application Example 1"

[0917] (Claim 1)

[0918] A means of receiving inquiries via multiple communication technologies and converting them into a unified format,

[0919] A means of analyzing the content of a query and identifying its intent using generated natural language processing techniques,

[0920] Means for obtaining necessary information from relevant information resources based on identified intentions,

[0921] A means of generating a response in natural language based on acquired information,

[0922] A means for transmitting the generated response via the original communication technology,

[0923] A means of analyzing user feedback and evaluating the effectiveness of responses,

[0924] A means to train the generation technology based on the evaluation results and improve the quality of the response,

[0925] A system that includes means for providing a function to instantly respond in natural language to inquiries made by users through voice or visual devices.

[0926] (Claim 2)

[0927] The system according to claim 1, which provides a function to maintain customer service even when sales staff are absent.

[0928] (Claim 3)

[0929] The system according to claim 1, which updates the data collection of information resources based on user feedback and improves the accuracy of future queries.

[0930] "Example 2 of combining an emotion engine"

[0931] (Claim 1)

[0932] A means of receiving inquiries via multiple information transmission methods and converting them into a standardized format,

[0933] A means of analyzing the content of a query and identifying its purpose using generated natural language processing techniques,

[0934] Means for collecting necessary information from relevant sources based on identified purposes,

[0935] A means of generating responses in emotionally sensitive natural language based on collected information and user sentiment analysis,

[0936] A means for transmitting the generated response via the original information transmission means,

[0937] A means of analyzing user responses and verifying the effectiveness of the response,

[0938] A system that includes means to improve response quality by learning generation techniques based on verification results.

[0939] (Claim 2)

[0940] The system according to claim 1, which provides a function to maintain user support even when the person performing the job is absent.

[0941] (Claim 3)

[0942] The system according to claim 1, which updates the data aggregation device of the information source based on user feedback and improves the accuracy of future inquiries.

[0943] "Application example 2 when combining with an emotional engine"

[0944] Point down

[0945] (Claim 1)

[0946] A means for receiving inquiries via multiple communication methods and converting them into a standardized format,

[0947] A means of analyzing the content of a query and identifying its intent using a generated natural language processing model,

[0948] A means of identifying a user's emotions using an emotion analysis engine,

[0949] Means for gathering necessary information from relevant sources based on identified intentions and feelings,

[0950] A means of generating emotionally sensitive responses in natural language based on collected information,

[0951] A means for transmitting the generated response via the original communication means,

[0952] A means of analyzing user feedback and evaluating the effectiveness of responses,

[0953] A system that includes means to improve the quality of responses by learning a generative model based on evaluation results.

[0954] (Claim 2)

[0955] The system according to claim 1, which maintains user support and provides customer-centric responses based on emotion, even when sales personnel are absent.

[0956] (Claim 3)

[0957] The system according to claim 1, which updates the information source database based on user feedback and improves the accuracy and sentiment relevance of future inquiries through sentiment analysis. [Explanation of symbols]

[0958] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for receiving inquiries via multiple communication methods and converting them into a standardized format, A means of analyzing the content of a query and identifying its intent using a generated natural language processing model, Means for collecting necessary information from relevant sources based on identified intentions, A means of generating a response in natural language based on the collected information, A means for transmitting the generated response via the original communication means, A means of analyzing user feedback and evaluating the effectiveness of responses, A system that includes means to improve the quality of responses by learning a generative model based on evaluation results.

2. The system according to claim 1, which provides a function to maintain user support even when sales personnel are absent.

3. The system according to claim 1, which updates the database of information sources based on user feedback and improves the accuracy of future queries.

Images