system

Abstract

Provide a system. [Solution] means for receiving text data; means for preprocessing the received text data; means for analyzing the preprocessed text data to extract query intent and entities; a means for classifying the query based on the extracted intent and entities; means for generating an appropriate response to the classified query; means for returning the generated response to the user; A system including:

Description

[Technical Field]

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

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

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

[0004] In current call centers, the increasing number of inquiries and complaints are handled manually, resulting in problems such as slow response times and reduced operational efficiency. Furthermore, as inquiries become more diverse and complex, the amount of work required for human representatives to respond to them increases, raising concerns about a decline in customer satisfaction. To solve these problems, a system is needed that can efficiently classify inquiries, generate automatic responses, and determine escalation. [Means for solving the problem]

[0005] In order to solve the above problems, the present invention provides the following means: A call center system including a means for receiving text data, a means for preprocessing the received text data, a means for analyzing the preprocessed text data to extract the intent and entities of the inquiry, a means for classifying the inquiry based on the extracted intent and entities, a means for generating an appropriate response to the classified inquiry, and a means for returning the generated response to the user. Furthermore, by including a means for automatically determining escalation based on the content of the inquiry and escalating the inquiry to a human representative as necessary, and a means for generating the response to be returned to the user based on a template and an FAQ database, the system aims to improve the efficiency of inquiry responses and customer satisfaction.

[0006] "Text data" is data consisting of a series of character strings including the content of an inquiry sent by a user.

[0007] The "receiving means" is a device or program that provides a function for the server to receive the query text entered by the user.

[0008] The "preprocessing means" is a device or program that performs processing to remove unnecessary characters and spaces from the received text data and prepare it in a format suitable for analysis.

[0009] The "analyzing means" is a device or program that analyzes the preprocessed text data using a natural language processing engine and extracts the intent and entities of the query.

[0010] "Intention" refers to the meaning or purpose of a user's inquiry, and is a concept that represents a specific request, such as "I want to return an item" or "I want to ask a question about a product."

[0011] An "entity" refers to a specific object or keyword contained in text, and is a concept that represents concrete information such as a product name or date and time.

[0012] A "classifying means" is a device or program that categorizes query content into different predefined categories based on analyzed intent and entities.

[0013] The "means for generating a response" is a device or program for generating an appropriate response sentence in response to a classified inquiry.

[0014] A "template" is a model that defines the basic structure and format for generating a response.

[0015] An "FAQ database" is a database that stores frequently asked questions and their answers in advance, and is used to provide appropriate answers according to the content of inquiries.

[0016] The "means for sending back" is a device or program that provides the function of sending the generated response to the user.

[0017] The "means for determining escalation" is a device or program that automatically determines whether the response should be handed over to a human agent, depending on the complexity and urgency of the inquiry.

[0018] "Escalation" is the process of transferring an inquiry to a human representative with greater expertise and response capabilities when automated processing is difficult to handle. [Brief explanation of the drawings]

[0019] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4]FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0020] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0021] First, the terms used in the following description will be explained.

[0022] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0023] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0024] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0025] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0026] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0027] [First embodiment]

[0028] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0029] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0030] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0031] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0032] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0033] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0034] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0035] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0036] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0037] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0038] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0039] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0040] The present invention is a system for improving the efficiency of inquiry operations at a call center, and is mainly responsible for receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, and returning responses. Specific embodiments are described below.

[0041] System configuration

[0042] This system is implemented mainly by three entities: a server, a terminal, and a user.

[0043] System Operation

[0044] 1. User submits an inquiry

[0045] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[0046] 2. The server receives the query

[0047] The server receives the query text sent by the user. At this time, the text data is received in the form included in the request body.

[0048] 3. The server preprocesses the text

[0049] The server performs preprocessing on the received text data. Preprocessing includes deleting unnecessary characters and spaces and improving grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is reformatted to "Product ABC is broken. Please tell me how to return it."

[0050] 4. The server analyzes the data using an NLP engine.

[0051] The server passes the preprocessed text data to a natural language processing (NLP) engine for text analysis. This analysis extracts the query intent and entities (keywords). For example, the intent "return" and the entities "product ABC" and "break" are extracted.

[0052] 5. The server classifies the query

[0053] The server classifies the query into a predefined category based on the parsed intent and entities, for example, into the "returns" category.

[0054] 6. The server decides on a course of action

[0055] The server determines the appropriate response to the classified inquiry, for example, whether to generate an automated response based on a template or escalate the inquiry to a human agent.

[0056] 7. The server generates an automatic response

[0057] The server uses response templates and an FAQ database to generate an appropriate response to the user's inquiry, such as "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[0058] 8. Server determines escalation

[0059] The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative.

[0060] 9. The server returns the answer

[0061] The server returns the generated response to the user via the original communication channel (chat box, email, etc.) through which the user made the inquiry.

[0062] 10. The user receives the answer

[0063] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[0064] Specific examples

[0065] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "Returns" category and generates an answer using an appropriate template. The generated answer is, "You can return product ABC by clicking the link below. Please see the return policy page for details." and is sent back to the user. The user can then receive this answer and proceed with the return process.

[0066] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[0067] The processing flow will be explained below.

[0068] Step 1:

[0069] The user submits a query.

[0070] The user uses the call center's chat box or email form to enter the inquiry in text format and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[0071] Step 2:

[0072] The server receives the query.

[0073] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed through the API operated by the server.

[0074] Step 3:

[0075] The server preprocesses the text.

[0076] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0077] Step 4:

[0078] The server analyzes the text using an NLP engine.

[0079] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[0080] Step 5:

[0081] The server classifies the query.

[0082] The server classifies the query into a predefined category based on the extracted intent and entities, for example, into the "returns" category.

[0083] Step 6:

[0084] The server decides the response policy.

[0085] The server determines the appropriate course of action for the classified query, possibly generating an automatic response or determining whether escalation is necessary. For example, a simple query can be handled with an automatic response.

[0086] Step 7:

[0087] The server generates an automatic response.

[0088] The server generates an appropriate answer using a response template or FAQ database. During this generation process, it replaces placeholders in the template with actual data. For example, it might generate a response like, "You can return product ABC by clicking the link below. For more information, please see our returns policy page."

[0089] Step 8:

[0090] The server decides on the escalation.

[0091] The server automatically determines the need for escalation based on the content and classification of the inquiry, and transfers particularly complex inquiries or urgent complaints to a human representative.

[0092] Step 9:

[0093] The server sends the response back to the user.

[0094] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0095] Step 10:

[0096] The user receives the response.

[0097] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[0098] Example 1

[0099] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0100] While call center operations require a rapid and accurate response to a huge volume of inquiries, the speed and efficiency of response was insufficient with conventional systems, resulting in a decline in customer satisfaction. Furthermore, automated response systems were unable to handle complex inquiries or urgent complaints, and lacked the functionality to escalate to a human representative at the appropriate time.

[0101] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0102] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for determining an appropriate response to the classified inquiry, means for generating an automatic response based on the response, and means for returning the generated response to the user. This enables a prompt and accurate response to the inquiry, and automatically determines whether escalation is necessary and transfers the case to a human representative, thereby improving overall business efficiency and customer satisfaction.

[0103] "Text data" refers to the inquiry content sent by the user in the form of a character string.

[0104] "Means for receiving" refers to the function of importing text data sent by a user into the server.

[0105] "Preprocessing" refers to the ability to format received text data into a format that is easy to parse, including removing unnecessary characters and spaces, and improving grammar and syntax.

[0106] "Means for analyzing" refers to the function of analyzing preprocessed text data using a natural language processing (NLP) engine to extract query intent and entities.

[0107] "Inquiry intent" refers to the purpose or request contained in the user's inquiry, such as "returning a product" or "technical support."

[0108] An "entity" refers to an important keyword or unit of information contained in a user's inquiry. For example, a specific item such as "product ABC" or "breaks."

[0109] "Means for query classification" refers to the ability to classify query content into predefined categories based on extracted intent and entities.

[0110] "Means for determining response" refers to the function that determines the appropriate response to a classified inquiry, including the process for determining whether to generate an automated response or escalate.

[0111] "Means for generating automatic responses" refers to a function that generates appropriate responses to user inquiries using templates and an FAQ database.

[0112] "Means for sending back" refers to the function of sending back the generated response to the user through the communication channel used by the user (chat box, email, etc.).

[0113] The present invention is a system for improving the efficiency of call center inquiry operations, and is primarily responsible for receiving text data, preprocessing, analyzing, classifying, determining countermeasures, generating automatic responses, determining escalation, and returning responses.

[0114] System configuration

[0115] This system is implemented mainly by three entities: a server, a terminal, and a user. Specific hardware used includes a server and a user terminal (e.g., a PC or smartphone). Software used includes a natural language processing engine (e.g., SpaCy, BERT), a query management system (e.g., Zendesk), and a database (e.g., MySQL (registered trademark)).

[0116] System Operation

[0117] The server receives the text data sent by the user, preprocesses it, and analyzes it. Based on the analysis results, it classifies the inquiry and determines the appropriate response, which may include generating an automated response or escalating the case to a human agent. Finally, the generated response is sent back to the user.

[0118] Examples of hardware and software

[0119] 1. Hardware:

[0120] Server: Processes and stores query data.

[0121] User terminal: Sends queries and receives responses.

[0122] 2. Software:

[0123] Natural language processing engines: Analyze text data (e.g., SpaCy, BERT).

[0124] Enquiry management system: Manages inquiry data and sends responses (e.g. Zendesk).

[0125] Database: Stores query and response data (e.g. MySQL).

[0126] Specific examples of operation

[0127] For example, if a user sends an inquiry saying "Product ABC is broken. I would like to return it. Please let me know how to do this," the inquiry will be processed as follows:

[0128] 1. The server receives this query text and retrieves the text data from the request body.

[0129] 2. The server uses regular expressions to remove unnecessary characters and spaces and preprocess the text into a format suitable for natural language processing. Specifically, it converts the text into the format "Product ABC is broken. Please tell me how to return it."

[0130] 3. The preprocessed text is passed to a natural language processing engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted.

[0131] 4. The server categorizes the inquiry based on the analysis results, for example, into the "returned goods" category.

[0132] 5. The server determines the appropriate response to the classified inquiry, for example, generating an auto-response based on a template or escalating the inquiry.

[0133] 6. An appropriate response is generated and sent back to the user, for example, "You can return product ABC by following the link below. Please see our returns policy page for more information."

[0134] Examples of prompt statements

[0135] 1. "Please preprocess this text data to remove unnecessary characters and spaces: Product ABC is broken. I would like to return it. What should I do?"

[0136] 2. "Please parse this preprocessed text data and extract the query intent and entities: Product ABC is broken. Please tell me how to return it."

[0137] 3. Generate an appropriate response based on the extracted intent and entity: Intent 'Return', Entity 'Product ABC', 'Broken'

[0138] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[0139] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0140] Step 1:

[0141] The user sends an inquiry. Using the call center's chat box or email form, the user sends a text inquiry from their terminal, such as "Product ABC is broken. I would like to return it. Please tell me what to do." Input: Inquiry text sent from the user's terminal. Output: Inquiry data in text format.

[0142] Step 2:

[0143] The server receives the query. The server receives the query text sent by the user from the request body. Input: Query text sent by the user. Output: Received text data.

[0144] Step 3:

[0145] The server preprocesses the text. The server uses regular expressions to remove unnecessary characters and spaces from the received text data and converts it into a format suitable for natural language processing. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted into "Product ABC is broken. Please tell me how to return it." Input: Received text data. Output: Preprocessed text data.

[0146] Step 4:

[0147] The server analyzes the data using an NLP engine. The server passes the preprocessed text data to a natural language processing (NLP) engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted. Input: Preprocessed text data. Output: Extracted intent and entities.

[0148] Step 5:

[0149] The server classifies the inquiry. Based on the parsed intent and entities, the server classifies the inquiry into a pre-defined category (e.g., "returns," "technical support," "product information"). For example, it is classified into the "returns" category. Input: Extracted intent and entities. Output: Classified inquiry category.

[0150] Step 6:

[0151] The server determines the appropriate response for the classified inquiry category. Specifically, it determines whether to generate an automatic response based on a template or escalate to a human agent. Input: Classified inquiry category. Output: Determined response (generating an automatic response or escalation).

[0152] Step 7:

[0153] The server generates an automatic response. The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. For example, it generates a response such as, "You can return product ABC by clicking the link below. Please see the return policy page for details." Input: The determined response. Output: The generated automatic response.

[0154] Step 8:

[0155] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative. Input: Inquiry content. Output: Escalation decision.

[0156] Step 9:

[0157] The server returns an answer. The server returns the generated response to the user. This is done through the original communication channel through which the user made the query (chat box, email, etc.). Input: The generated automated response. Output: The response returned to the user.

[0158] Step 10:

[0159] The user receives the answer. The user checks the response sent from the server on their own device. For example, a message is displayed in the chat box or an email is received. Input: The response sent from the server. Output: The response checked on the user device.

[0160] (Application example 1)

[0161] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0162] Conventional systems for responding to inquiries at call centers and monitoring centers have made it difficult to respond quickly and appropriately, and responses are often delayed, especially in urgent situations. This has led to a decline in customer satisfaction and increased risk of serious problems. Furthermore, there was no system in place to efficiently process inquiries using real-time voice input.

[0163] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0164] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for generating an appropriate response to the classified inquiry, means for returning the generated response to the user, means for converting voice input into text data in real time, and means for determining urgency and performing automatic escalation. This enables prompt and appropriate processing of inquiries, and enables prompt responses, especially in emergencies.

[0165] The "means for receiving text data" is a device or program that has the function of allowing the server to receive a text-formatted inquiry sent from a user.

[0166] The "means for preprocessing received text data" refers to a device or program that executes a process to remove unnecessary characters and spaces from the received text data and to improve grammar and syntax.

[0167] "Means for analyzing preprocessed text data to extract query intent and entities" refers to a device or program that has the function of analyzing preprocessed text data using natural language processing (NLP) technology and extracting the main intent of a user's query and important keywords (entities).

[0168] The "means for classifying inquiries based on the extracted intent and entities" refers to a device or program that has the function of classifying the content of inquiries into pre-defined categories based on the intent and entities extracted through analysis.

[0169] The "means for generating appropriate responses to classified inquiries" refers to a device or program that has the function of automatically generating appropriate answers based on the classified inquiry content using templates and an FAQ database.

[0170] The "means for returning the generated response to the user" refers to a device or program that has the function of returning the automatically generated response to the user, and performs the return via a communication channel.

[0171] The "means for converting voice input into text data in real time" refers to a device or program that has the function of recognizing a user's voice input in real time and converting it into text data.

[0172] "Means for determining urgency and automatically escalating" refers to a device or program that has the function of automatically determining the urgency of an inquiry and, if necessary, transferring the inquiry to a human representative.

[0173] The system for implementing this invention is composed of three main entities: a server, a terminal, and a user. To understand the specific processing procedure, the operation of this system will be explained.

[0174] Server Features

[0175] The server has the following functions:

[0176] 1. Function to receive text data:

[0177] The server receives the inquiry sent by the user in text format. This function also includes a function to convert voice data into text data in real time.

[0178] 2. Text data preprocessing function:

[0179] Removes unnecessary characters and spaces from received text data and improves grammar and syntax, making it easier to analyze.

[0180] 3. Text data analysis function:

[0181] The preprocessed text data is analyzed using natural language processing (NLP) technology to extract the user's query intent and important keywords (entities). Specifically, open source NLP engines such as spaCy and TextBlob are used.

[0182] 4. Inquiry classification function:

[0183] Based on the extracted intent and entities, the inquiry is classified into pre-defined categories, and an appropriate response policy is then determined based on the pre-defined categories.

[0184] 5. Appropriate response generation:

[0185] For classified inquiries, an automatic response is generated using templates and an FAQ database. This automatically generated response is returned to the user in real time, allowing for a quick response.

[0186] 6. Urgency assessment and automatic escalation:

[0187] The system automatically determines the urgency of the inquiry and escalates it to a human representative if necessary, preventing serious problems.

[0188] Device Features

[0189] The devices used by security personnel and those on-site to receive inquiries are smartphones, smart glasses, head-mounted displays, etc. These devices are equipped with a voice input function, converting voice into text in real time and sending it to a server.

[0190] User Roles

[0191] When a problem occurs on-site, the user, as a security officer, sends a query to the server by voice input. For example, the following prompt sentences can be used:

[0192] "I've spotted a suspicious person. What should I do?"

[0193] "We just discovered a fire at our facility. What should we do?"

[0194] "I found a suspicious package. Do you need urgent action?"

[0195] Hardware and software used

[0196] This system uses the following hardware and software:

[0197] Hardware: Servers (with powerful CPUs and large amounts of memory), smartphones, smart glasses, head-mounted displays

[0198] Software: Python, spaCy (NLP engine), TextBlob (NLP engine), open source libraries for speech recognition (e.g., Google® Speech-to-Text)

[0199] This enables immediate response on-site and automatic escalation in emergencies.

[0200] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0201] Step 1:

[0202] The user inputs voice data. For example, they might ask, "I spotted a suspicious person. What should I do?" This is sent to the server in real time using the voice input function of smart glasses or a head-mounted display. The input here is voice data, and the output is data converted into text format.

[0203] Step 2:

[0204] The server converts the received voice data into text data. An open source library for speech recognition (e.g., Google Speech-to-Text) is used to convert the voice data into text data. This process results in the voice data being converted into text data.

[0205] Step 3:

[0206] The server preprocesses the text data. It removes unnecessary characters and spaces from the received text data and improves grammar and syntax. This makes it easier to analyze. The input of this step is the converted text data, and the output is the preprocessed text data.

[0207] Step 4:

[0208] The server analyzes the preprocessed text data. Using a natural language processing (NLP) engine (e.g., spaCy), it analyzes the text data and extracts the user's query intent and important keywords (entities). The input is the preprocessed text data, and the output is the extracted intent and entities.

[0209] Step 5:

[0210] The server classifies queries based on intent and entities. Based on the extracted intent and entities, the query content is classified into predefined categories. The input is the extracted intent and entities, and the output is the classified query.

[0211] Step 6:

[0212] The server generates an appropriate response to the classified query. It uses templates and an FAQ database to automatically generate responses. For example, the response generated is "Contact the police immediately and monitor the scene from a safe location." The input is the classified query and the output is the generated response.

[0213] Step 7:

[0214] The server determines the urgency of the inquiry and automatically escalates it. If the inquiry is deemed urgent, it is escalated to a human agent. The input is the generated response and the inquiry itself, and the output is the decision to escalate or not.

[0215] Step 8:

[0216] The server sends the generated response back to the user. The server sends the response back to the user through the original communication channel (e.g., smart glasses or a head-mounted display). The input is the generated response and the output is the response sent back to the user.

[0217] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0218] The present invention is a system for improving the efficiency of call center inquiry operations, and mainly includes receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, returning responses, and an emotion engine that analyzes user emotions. Specific embodiments are described below.

[0219] System configuration

[0220] This system is implemented mainly by three entities: the server, the terminal, and the user. In addition, by incorporating an emotion engine for emotion analysis, more advanced inquiry response can be achieved.

[0221] System Operation

[0222] 1. User submits an inquiry

[0223] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[0224] 2. The server receives the query

[0225] The server receives the query text sent by the user and takes it as the request body.

[0226] 3. The server preprocesses the text

[0227] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0228] 4. The server analyzes the text using an NLP engine

[0229] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[0230] 5. The server analyzes emotions using an emotion engine

[0231] The server also sends the parsed text to an emotion engine, which analyzes the user's emotion (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords in the text. At this stage, the emotion may be analyzed as "anger."

[0232] 6. The server classifies the query

[0233] The server classifies the query into a predefined category based on the extracted intent and entities, as well as the analyzed sentiment, e.g., into the "returns" category.

[0234] 7. The server decides on a course of action

[0235] The server determines the appropriate response policy for the classified query. If the user's emotion is strong, it adjusts the tone of the response. For example, if the user is angry, it generates a calmer and more polite response.

[0236] 8. The server generates an automatic response

[0237] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During this generation process, placeholders in the template are replaced with actual data. For example, the server might generate a response such as, "You can return product ABC by clicking the link below. For more information, please see our return policy page."

[0238] 9. Server determines escalation

[0239] The server automatically determines whether a call needs to be escalated based on the content of the call and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, the server will hand over to a human agent. This decision is made when anger or strong dissatisfaction is detected.

[0240] 10. The server sends the response back to the user

[0241] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0242] 11. The user receives a response

[0243] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[0244] Specific examples

[0245] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "return" category, and the emotion engine further analyzes that the user's emotion is "anger." The system generates a response using an appropriate template and returns it to the user with the following response: "You can return product ABC by clicking the link below. Please see the return policy page for details." The user can then accept this response and proceed with the return process.

[0246] This embodiment not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[0247] The processing flow will be explained below.

[0248] Step 1:

[0249] The user submits a query.

[0250] The user enters the inquiry in text format using the call center's chat box or email form and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[0251] Step 2:

[0252] The server receives the query.

[0253] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed using a protocol such as an HTTP request or socket communication.

[0254] Step 3:

[0255] The server preprocesses the text.

[0256] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters from the text and grammatically arranges it. For example, it converts "Product ABC is broken. I would like to return it. Please tell me what to do." into "Product ABC is broken. Please tell me how to return it."

[0257] Step 4:

[0258] The server analyzes the text using an NLP engine.

[0259] The server sends the preprocessed text data to a natural language processing (NLP) engine for text analysis. The NLP engine breaks the text down into tokens and extracts the user's intent and important information (entities). For example, "intent = return" and "entity = product ABC breaks" are extracted.

[0260] Step 5:

[0261] The server analyzes emotions using an emotion engine.

[0262] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions. The emotion engine identifies emotions such as positive, negative, and neutral from the text. For example, "anger" may be detected from the user's text.

[0263] Step 6:

[0264] The server classifies the query.

[0265] The server classifies the query into a predefined category based on the extracted intent and entities, and the analyzed sentiment, e.g., into the "returns" category.

[0266] Step 7:

[0267] The server decides the response policy.

[0268] The server determines the appropriate response policy for the classified inquiry. If the user's emotion is strong, the tone of the response is adjusted. For example, if the user is angry, the server generates a polite and calm response.

[0269] Step 8:

[0270] The server generates an automatic response.

[0271] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During the generation process, specific information is filled in the placeholders in the template. For example, it generates a response such as, "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[0272] Step 9:

[0273] The server decides on the escalation.

[0274] The server automatically determines whether or not to escalate the inquiry based on the content of the inquiry and the analyzed sentiment. If the user's sentiment is particularly strong or the inquiry is complex, the server will hand it over to a human agent. This decision is also based on the results of sentiment analysis.

[0275] Step 10:

[0276] The server sends the response back to the user.

[0277] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0278] Step 11:

[0279] The user receives the response.

[0280] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[0281] Example 2

[0282] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0283] Call centers are required to respond to many inquiries quickly and appropriately, but typical manual responses are inefficient and prone to user dissatisfaction. Responses that ignore the user's feelings can also lower customer satisfaction. Furthermore, if appropriate escalation procedures are not carried out, there is a risk that important issues will be overlooked. To solve these issues, a system is needed that can automatically analyze the content of inquiries and respond in a way that takes the user's feelings into account.

[0284] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the query, means for classifying the query based on the extracted intent and entity, means for classifying the query based on the analyzed emotion, means for determining a response policy based on the classified query, means for generating an appropriate response based on the response policy, and means for returning the generated response to the user. This enables an efficient response to the query that takes emotion into consideration.

[0285] The "means for receiving text data" is a device or function for receiving a text-formatted inquiry sent by a user.

[0286] The "means for preprocessing received text data" refers to a device or function for removing unnecessary spaces and special characters from the received text data and for adjusting the grammar and syntax.

[0287] The "means for analyzing preprocessed text data to extract query intent and entities" is a device or function for automatically analyzing and extracting user intent and important information from preprocessed text data.

[0288] The "means for classifying queries based on extracted intent and entities" refers to a device or function for classifying queries into predefined categories based on the intent and entities extracted by the analysis.

[0289] A "means for classifying queries based on analyzed emotions" is a device or function for classifying queries according to their emotional state based on emotion analysis.

[0290] The "means for determining a response policy based on the classified inquiry" is a device or function for determining an appropriate response policy based on the classified inquiry.

[0291] The "means for generating an appropriate response based on a response policy" is a device or function for automatically generating an appropriate response based on a determined response policy.

[0292] A "means for returning a generated response to a user" is a device or function for returning a generated response to a user through the original communication channel.

[0293] This invention is a system for improving the efficiency of call center inquiry operations, and mainly includes text data reception, preprocessing, analysis, classification, response generation, automatic escalation, response return, and an emotion engine that analyzes user emotions. This system is implemented mainly by three entities: a server, a terminal, and a user.

[0294] System configuration

[0295] 1. Receiving text data

[0296] The server receives the text data (inquiry content) sent by the user. The hardware used includes a regular web server, mail server, or chatbot server.

[0297] 2. Preprocessing of text data

[0298] The server preprocesses the received text data, including removing unnecessary spaces and special characters, grammar reformatting, etc. The software used includes natural language processing libraries (e.g., NLTK and SpaCy).

[0299] 3. Text Data Analysis

[0300] The preprocessed text data is then analyzed by an NLP engine, which uses services such as Google Cloud NLP or Amazon Comprehend to extract user intent and important information (entities).

[0301] 4. Emotion analysis

[0302] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions using services such as IBM Watson® Tone Analyzer, which identifies emotions (e.g., joy, anger, etc.) from the tone and keywords in the text.

[0303] 5. Categorizing inquiries

[0304] The server classifies queries into predefined categories based on the extracted intent and entities, and the analyzed sentiment, using rule-based classification algorithms and machine learning models.

[0305] 6. Deciding on a response policy

[0306] The server determines the appropriate response policy for each classified inquiry. In particular, it adjusts the tone of the response according to the user's emotions. If the user's emotions are strong, it generates a calmer and more polite response.

[0307] 7. Generate Auto-Responses

[0308] The server generates an appropriate answer using response templates and an FAQ database. During the generation process, it fills in the placeholders in the template with actual data. For example, it might generate an answer like, "You can return product ABC by following the link below. For more information, please see our returns policy page."

[0309] 8. Escalation Decision

[0310] The server automatically determines the need for escalation based on the query content and analyzed sentiment, specifically handing over to a human agent if the user's sentiment exceeds a certain threshold.

[0311] 9. Returning Responses

[0312] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0313] Specific examples

[0314] As a concrete example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me what to do." When the system receives this query, it preprocesses the text and analyzes it using an NLP engine. As a result of the analysis, the system classifies the query into the "return" category and further analyzes the user's emotion as "anger" using an emotion engine. Using an appropriate template, the system generates a response that reads, "You can return product ABC by clicking the link below. For details, please see the return policy page." and sends it back to the user. The user can then accept this response and proceed with the return process.

[0315] The present invention not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[0316] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0317] Step 1:

[0318] The user submits an inquiry. The user enters the inquiry details in the chat box or email form and presses the send button. For example, a prompt message such as "Product ABC is broken. I would like to return it. Please tell me what to do." is sent.

[0319] Step 2:

[0320] The server receives the query. The server receives the user's input and takes it as the request body. The input is the user's query text, and the output is this text data. The query content and meta information (user ID, timestamp, etc.) are saved in a database.

[0321] Step 3:

[0322] The server preprocesses the text data. It removes unnecessary spaces and special characters from the received text data and arranges the grammatical structure. For example, "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is the received text, and the output is the preprocessed text data. A natural language processing library (for example, NLTK or SpaCy) is used for processing.

[0323] Step 4:

[0324] The server analyzes the preprocessed text and extracts the query intent and entities. The server then sends the preprocessed text data to an NLP engine, which extracts the intent and entities. The input is the preprocessed text, and the output is the extracted intent and entities. Natural language processing services such as Google Cloud NLP and Amazon Comprehend are used. The extracted results are "intent = return" and "entity = product ABC breaks."

[0325] Step 5:

[0326] The server analyzes emotions from the extracted data. The server sends the analysis results to an emotion engine to analyze the user's emotions. For example, IBM Watson Tone Analyzer is used. The input is text containing the extracted intent and entities, and the output is the emotion analysis result (e.g., "anger").

[0327] Step 6:

[0328] The server classifies the query. The server classifies the query into the appropriate category based on the extracted intent, entities, and sentiment. The input is the parsed text data, and the output is the query category (e.g., the "returns" category). The processing uses rule-based classification algorithms or machine learning models.

[0329] Step 7:

[0330] The server determines the response policy based on the classified inquiry. In particular, if the user's emotion is strong, the tone of the response is adjusted. The input is the inquiry category and the emotion analysis result, and the output is the response policy.

[0331] Step 8:

[0332] The server generates an automatic response. Based on the response policy, the server generates an appropriate answer using response templates and an FAQ database. The input is the response policy, and the output is the generated response text. For example, it generates the text "You can return product ABC by clicking the link below. Please see the return policy page for details."

[0333] Step 9:

[0334] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry and the analyzed sentiment. The input is the inquiry data and sentiment analysis results, and the output is the escalation decision. In particular, if the user's sentiment is strong or the problem is complex, the server will hand over to a human agent.

[0335] Step 10:

[0336] The server sends the generated response back to the user The server sends the generated response back to the user through the original communication channel (chat box or email) The input is the generated response text, the output is the response sent back to the user

[0337] Step 11:

[0338] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The output is the response text displayed on the user's screen.

[0339] (Application example 2)

[0340] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0341] Modern online shopping sites are receiving an increasing number of inquiries from users, requiring efficient and accurate responses. However, conventional systems struggle to receive and preprocess text data, extract intent and entities, analyze emotions, and automatically escalate inquiries, making it difficult to improve customer satisfaction. Furthermore, there is a lack of a way to generate appropriate responses based on user emotions, limiting the efficiency of inquiry response and customer satisfaction. Therefore, a solution to this issue is needed.

[0342] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intention and events of the inquiry, means for analyzing the user's emotions, means for classifying the inquiry based on the extracted intention and events and the analyzed emotions, means for generating an appropriate response to the classified inquiry, and means for returning the generated response to the user. This makes it possible to improve the efficiency of inquiry operations and provide an appropriate response according to the user's emotions.

[0343] "Text data" is information in the form of text that a user transmits over a communication channel.

[0344] "Preprocessing" is a process that removes unnecessary spaces and special characters from the received text data and corrects the grammar and syntax.

[0345] "Intention" refers to the request or purpose that a user wants to convey through text data.

[0346] An "event" is specific information or an object extracted from text data.

[0347] "Emotions" are psychological states contained in the user's text data, and include joy, anger, anxiety, and the like.

[0348] "Query classification" refers to dividing the query content into specific categories based on the analyzed intent, events, and sentiment.

[0349] "Response generation" is the process of generating appropriate responses or information to classified inquiries.

[0350] "Returning a response" is the process of retransmitting the generated response to the user's communication channel.

[0351] "Automatic escalation" refers to handing over an inquiry to a human agent as needed based on the content of the inquiry and analyzed sentiment.

[0352] A "template" is a pre-prepared response template that can be customized for a specific situation.

[0353] The "FAQ Data Collection" is a database that compiles frequently asked questions and their answers.

[0354] The present invention provides a system for improving the efficiency of inquiry operations on mail-order sites and enabling appropriate responses according to the user's emotions. Specific embodiments for carrying out the invention will be described below.

[0355] System configuration

[0356] The system has three main components: a server, a terminal, and a user. The server plays a central role in processing everything from receiving text data to generating responses. The terminal provides a communication channel with the user, and the user makes inquiries through this terminal.

[0357] Hardware and Software

[0358] The server includes the following main software modules:

[0359] 1. NLP service: A software module that performs natural language processing (NLP) and analyzes the user's query text to extract intent and events.

[0360] 2. Sentiment Analysis Engine: A software module for analyzing user emotions from text data and identifying emotions such as anger, anxiety, and joy.

[0361] 3. FAQ data collection: A database containing pre-prepared response templates for many user inquiries.

[0362] Data processing and calculation

[0363] 1. Preprocessing of text data: The server performs preprocessing by formatting the text data received from the user and removing unnecessary spaces and special characters. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0364] 2. NLP analysis: The preprocessed text data is sent to the NLP service to extract intent and events. The NLP service analyzes information such as "intent = return" and "event = product ABC breaks."

[0365] 3. Sentiment Analysis: The text data is sent to a sentiment analysis engine to identify the user's emotions. The sentiment analysis engine extracts emotions such as anger, anxiety, or joy from the text. At this stage, the emotion may be analyzed as "anger."

[0366] 4. Enquiry Classification: Based on the extracted intent, events, and sentiment, the inquiry is classified into a specific category, e.g., “return” category.

[0367] 5. Response Generation: An appropriate response is generated for the classified inquiry based on the FAQ data set. If the user is angry, the tone of the response is adjusted and the placeholders in the template are replaced with actual data. For example, a response might be generated that reads, "You can return product ABC by clicking the link below. Please see the return policy page for details."

[0368] 6. Automatic Escalation: Automatically escalate queries based on query content and analyzed sentiment. If the user's sentiment is strong or exceeds a certain threshold, the query will be escalated to a human agent.

[0369] 7. Return Response: The generated response is returned to the user via the original communication channel (e.g., chat box or email).

[0370] Specific examples

[0371] When a user sends a query such as "Product ABC is broken. I would like to return it. Please tell me what to do," the server receives this query and performs preprocessing. It then uses NLP analysis to extract "Intention = Return" and "Event = Product ABC breaks," and uses sentiment analysis to identify the user's emotion as "Anger." Using an appropriate template, it generates a response saying, "You can return Product ABC by clicking the link below. Please see the return policy page for details," and sends it back to the user.

[0372] Example prompts to input to the generative AI model

[0373] "A user submits a query: (Product ABC is broken. I would like to return it. What should I do?) Preprocess this text, perform NLP analysis and sentiment analysis, and generate an optimal response. Please also consider the emotional tone of the response."

[0374] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0375] Step 1:

[0376] The user sends an inquiry. The user inputs and sends a text-format inquiry via the terminal. An example of such input is, "Product ABC is broken. I would like to return it. Please tell me what to do." The input text data is sent to the server.

[0377] Step 2:

[0378] The server receives the query. The server receives the text data sent by the user and takes it as the request body. The input data is left as is, and the output is raw text data that is subject to preprocessing.

[0379] Step 3:

[0380] The server preprocesses the text. It removes unnecessary spaces and special characters from the input text data and adjusts the grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is raw text data, and the output is preprocessed, clean text data.

[0381] Step 4:

[0382] The server analyzes the text using an NLP engine. The preprocessed text is sent to a natural language processing (NLP) engine, which analyzes the text. The NLP engine extracts the user's intent and important information (entities) from the text. From the input preprocessed text data, "Intent = Return" and "Event = Product ABC breaks" are extracted. The output is the analyzed intent and event.

[0383] Step 5:

[0384] The server analyzes emotions using an emotion engine. The analyzed text data is also sent to the emotion engine, which analyzes the user's emotions (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords of the text. The input is the analyzed text data, and the output is the identified emotion. At this stage, the emotion may be analyzed as "anger."

[0385] Step 6:

[0386] The server classifies the query. Based on the extracted intent and events, as well as the analyzed sentiment, the query is classified into a pre-defined category. For example, it is classified into the "return" category. The input is the intent, event, and sentiment data, and the output is a specific category.

[0387] Step 7:

[0388] The server determines the appropriate response policy and generates a response. It determines the appropriate response policy for the classified inquiry. If the user is angry, it adjusts the tone of the response and generates an appropriate answer to the user's inquiry using response templates and FAQ data. For example, it generates a response that reads, "You can return product ABC using the link below. Please see the return policy page for details." The input is the category and emotion data, and the output is the response sent back to the user.

[0389] Step 8:

[0390] The server makes the decision on escalation. It automatically determines the need for escalation based on the content of the inquiry and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, it will hand over to a human agent. The input is emotion data, and the output is instructions or execution of escalation.

[0391] Step 9:

[0392] The server sends the response back to the user. The generated response is sent back to the user via the original communication channel (chat box or email). The input is the generated response data, and the output is the message sent to the user.

[0393] Step 10:

[0394] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The input is the message sent from the server, and the output is the response displayed to the user.

[0395] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0396] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0397] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0398] [Second embodiment]

[0399] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0400] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0401] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0402] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0403] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0404] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0405] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0406] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0407] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0408] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0409] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0410] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0411] The present invention is a system for improving the efficiency of inquiry operations at a call center, and is mainly responsible for receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, and returning responses. Specific embodiments are described below.

[0412] System configuration

[0413] This system is implemented mainly by three entities: a server, a terminal, and a user.

[0414] System Operation

[0415] 1. User submits an inquiry

[0416] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[0417] 2. The server receives the query

[0418] The server receives the query text sent by the user. At this time, the text data is received in the form included in the request body.

[0419] 3. The server preprocesses the text

[0420] The server performs preprocessing on the received text data. Preprocessing includes deleting unnecessary characters and spaces and improving grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is reformatted to "Product ABC is broken. Please tell me how to return it."

[0421] 4. The server analyzes the data using an NLP engine.

[0422] The server passes the preprocessed text data to a natural language processing (NLP) engine for text analysis. This analysis extracts the query intent and entities (keywords). For example, the intent "return" and the entities "product ABC" and "break" are extracted.

[0423] 5. The server classifies the query

[0424] The server classifies the query into a predefined category based on the parsed intent and entities, for example, into the "returns" category.

[0425] 6. The server decides on a course of action

[0426] The server determines the appropriate response to the classified inquiry, for example, whether to generate an automated response based on a template or escalate the inquiry to a human agent.

[0427] 7. The server generates an automatic response

[0428] The server uses response templates and an FAQ database to generate an appropriate response to the user's inquiry, such as "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[0429] 8. Server determines escalation

[0430] The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative.

[0431] 9. The server returns the answer

[0432] The server returns the generated response to the user via the original communication channel (chat box, email, etc.) through which the user made the inquiry.

[0433] 10. The user receives the answer

[0434] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[0435] Specific examples

[0436] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "Returns" category and generates an answer using an appropriate template. The generated answer is, "You can return product ABC by clicking the link below. Please see the return policy page for details." and is sent back to the user. The user can then receive this answer and proceed with the return process.

[0437] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[0438] The processing flow will be explained below.

[0439] Step 1:

[0440] The user submits a query.

[0441] The user uses the call center's chat box or email form to enter the inquiry in text format and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[0442] Step 2:

[0443] The server receives the query.

[0444] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed through the API operated by the server.

[0445] Step 3:

[0446] The server preprocesses the text.

[0447] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0448] Step 4:

[0449] The server analyzes the text using an NLP engine.

[0450] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[0451] Step 5:

[0452] The server classifies the query.

[0453] The server classifies the query into a predefined category based on the extracted intent and entities, for example, into the "returns" category.

[0454] Step 6:

[0455] The server decides the response policy.

[0456] The server determines the appropriate course of action for the classified query, possibly generating an automatic response or determining whether escalation is necessary. For example, a simple query can be handled with an automatic response.

[0457] Step 7:

[0458] The server generates an automatic response.

[0459] The server generates an appropriate answer using a response template or FAQ database. During this generation process, it replaces placeholders in the template with actual data. For example, it might generate a response like, "You can return product ABC by clicking the link below. For more information, please see our returns policy page."

[0460] Step 8:

[0461] The server decides on the escalation.

[0462] The server automatically determines the need for escalation based on the content and classification of the inquiry, and transfers particularly complex inquiries or urgent complaints to a human representative.

[0463] Step 9:

[0464] The server sends the response back to the user.

[0465] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0466] Step 10:

[0467] The user receives the response.

[0468] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[0469] Example 1

[0470] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0471] While call center operations require a rapid and accurate response to a huge volume of inquiries, the speed and efficiency of response was insufficient with conventional systems, resulting in a decline in customer satisfaction. Furthermore, automated response systems were unable to handle complex inquiries or urgent complaints, and lacked the functionality to escalate to a human representative at the appropriate time.

[0472] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0473] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for determining an appropriate response to the classified inquiry, means for generating an automatic response based on the response, and means for returning the generated response to the user. This enables a prompt and accurate response to the inquiry, and automatically determines whether escalation is necessary and transfers the case to a human representative, thereby improving overall business efficiency and customer satisfaction.

[0474] "Text data" refers to the inquiry content sent by the user in the form of a character string.

[0475] "Means for receiving" refers to the function of importing text data sent by a user into the server.

[0476] "Preprocessing" refers to the ability to format received text data into a format that is easy to parse, including removing unnecessary characters and spaces, and improving grammar and syntax.

[0477] "Means for analyzing" refers to the function of analyzing preprocessed text data using a natural language processing (NLP) engine to extract query intent and entities.

[0478] "Inquiry intent" refers to the purpose or request contained in the user's inquiry, such as "returning a product" or "technical support."

[0479] An "entity" refers to an important keyword or unit of information contained in a user's inquiry. For example, a specific item such as "product ABC" or "breaks."

[0480] "Means for query classification" refers to the ability to classify query content into predefined categories based on extracted intent and entities.

[0481] "Means for determining response" refers to the function that determines the appropriate response to a classified inquiry, including the process for determining whether to generate an automated response or escalate.

[0482] "Means for generating automatic responses" refers to a function that generates appropriate responses to user inquiries using templates and an FAQ database.

[0483] "Means for sending back" refers to the function of sending back the generated response to the user through the communication channel used by the user (chat box, email, etc.).

[0484] The present invention is a system for improving the efficiency of call center inquiry operations, and is primarily responsible for receiving text data, preprocessing, analyzing, classifying, determining countermeasures, generating automatic responses, determining escalation, and returning responses.

[0485] System configuration

[0486] This system is implemented mainly by three entities: a server, a terminal, and a user. The specific hardware used includes a server and a user terminal (e.g., a PC or smartphone). The software used includes a natural language processing engine (e.g., SpaCy, BERT), a query management system (e.g., Zendesk), and a database (e.g., MySQL).

[0487] System Operation

[0488] The server receives the text data sent by the user, preprocesses it, and analyzes it. Based on the analysis results, it classifies the inquiry and determines the appropriate response, which may include generating an automated response or escalating the case to a human agent. Finally, the generated response is sent back to the user.

[0489] Examples of hardware and software

[0490] 1. Hardware:

[0491] Server: Processes and stores query data.

[0492] User terminal: Sends queries and receives responses.

[0493] 2. Software:

[0494] Natural language processing engines: Analyze text data (e.g., SpaCy, BERT).

[0495] Enquiry management system: Manages inquiry data and sends responses (e.g. Zendesk).

[0496] Database: Stores query and response data (e.g. MySQL).

[0497] Specific examples of operation

[0498] For example, if a user sends an inquiry saying "Product ABC is broken. I would like to return it. Please let me know how to do this," the inquiry will be processed as follows:

[0499] 1. The server receives this query text and retrieves the text data from the request body.

[0500] 2. The server uses regular expressions to remove unnecessary characters and spaces and preprocess the text into a format suitable for natural language processing. Specifically, it converts the text into the format "Product ABC is broken. Please tell me how to return it."

[0501] 3. The preprocessed text is passed to a natural language processing engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted.

[0502] 4. The server categorizes the inquiry based on the analysis results, for example, into the "returned goods" category.

[0503] 5. The server determines the appropriate response to the classified inquiry, for example, generating an auto-response based on a template or escalating the inquiry.

[0504] 6. An appropriate response is generated and sent back to the user, for example, "You can return product ABC by following the link below. Please see our returns policy page for more information."

[0505] Examples of prompt statements

[0506] 1. "Please preprocess this text data to remove unnecessary characters and spaces: Product ABC is broken. I would like to return it. What should I do?"

[0507] 2. "Please parse this preprocessed text data and extract the query intent and entities: Product ABC is broken. Please tell me how to return it."

[0508] 3. Generate an appropriate response based on the extracted intent and entity: Intent 'Return', Entity 'Product ABC', 'Broken'

[0509] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[0510] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0511] Step 1:

[0512] The user sends an inquiry. Using the call center's chat box or email form, the user sends a text inquiry from their terminal, such as "Product ABC is broken. I would like to return it. Please tell me what to do." Input: Inquiry text sent from the user's terminal. Output: Inquiry data in text format.

[0513] Step 2:

[0514] The server receives the query. The server receives the query text sent by the user from the request body. Input: Query text sent by the user. Output: Received text data.

[0515] Step 3:

[0516] The server preprocesses the text. The server uses regular expressions to remove unnecessary characters and spaces from the received text data and converts it into a format suitable for natural language processing. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted into "Product ABC is broken. Please tell me how to return it." Input: Received text data. Output: Preprocessed text data.

[0517] Step 4:

[0518] The server analyzes the data using an NLP engine. The server passes the preprocessed text data to a natural language processing (NLP) engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted. Input: Preprocessed text data. Output: Extracted intent and entities.

[0519] Step 5:

[0520] The server classifies the inquiry. Based on the parsed intent and entities, the server classifies the inquiry into a pre-defined category (e.g., "returns," "technical support," "product information"). For example, it is classified into the "returns" category. Input: Extracted intent and entities. Output: Classified inquiry category.

[0521] Step 6:

[0522] The server determines the appropriate response for the classified inquiry category. Specifically, it determines whether to generate an automatic response based on a template or escalate to a human agent. Input: Classified inquiry category. Output: Determined response (generating an automatic response or escalation).

[0523] Step 7:

[0524] The server generates an automatic response. The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. For example, it generates a response such as, "You can return product ABC by clicking the link below. Please see the return policy page for details." Input: The determined response. Output: The generated automatic response.

[0525] Step 8:

[0526] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative. Input: Inquiry content. Output: Escalation decision.

[0527] Step 9:

[0528] The server returns an answer. The server returns the generated response to the user. This is done through the original communication channel through which the user made the query (chat box, email, etc.). Input: The generated automated response. Output: The response returned to the user.

[0529] Step 10:

[0530] The user receives the answer. The user checks the response sent from the server on their own device. For example, a message is displayed in the chat box or an email is received. Input: The response sent from the server. Output: The response checked on the user device.

[0531] (Application example 1)

[0532] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0533] Conventional systems for responding to inquiries at call centers and monitoring centers have made it difficult to respond quickly and appropriately, and responses are often delayed, especially in urgent situations. This has led to a decline in customer satisfaction and increased risk of serious problems. Furthermore, there was no system in place to efficiently process inquiries using real-time voice input.

[0534] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0535] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for generating an appropriate response to the classified inquiry, means for returning the generated response to the user, means for converting voice input into text data in real time, and means for determining urgency and performing automatic escalation. This enables prompt and appropriate processing of inquiries, and enables prompt responses, especially in emergencies.

[0536] The "means for receiving text data" is a device or program that has the function of allowing the server to receive a text-formatted inquiry sent from a user.

[0537] The "means for preprocessing received text data" refers to a device or program that executes a process to remove unnecessary characters and spaces from the received text data and to improve grammar and syntax.

[0538] "Means for analyzing preprocessed text data to extract query intent and entities" refers to a device or program that has the function of analyzing preprocessed text data using natural language processing (NLP) technology and extracting the main intent of a user's query and important keywords (entities).

[0539] The "means for classifying inquiries based on the extracted intent and entities" refers to a device or program that has the function of classifying the content of inquiries into pre-defined categories based on the intent and entities extracted through analysis.

[0540] The "means for generating appropriate responses to classified inquiries" refers to a device or program that has the function of automatically generating appropriate answers based on the classified inquiry content using templates and an FAQ database.

[0541] The "means for returning the generated response to the user" refers to a device or program that has the function of returning the automatically generated response to the user, and performs the return via a communication channel.

[0542] The "means for converting voice input into text data in real time" refers to a device or program that has the function of recognizing a user's voice input in real time and converting it into text data.

[0543] "Means for determining urgency and automatically escalating" refers to a device or program that has the function of automatically determining the urgency of an inquiry and, if necessary, transferring the inquiry to a human representative.

[0544] The system for implementing this invention is composed of three main entities: a server, a terminal, and a user. To understand the specific processing procedure, the operation of this system will be explained.

[0545] Server Features

[0546] The server has the following functions:

[0547] 1. Function to receive text data:

[0548] The server receives the inquiry sent by the user in text format. This function also includes a function to convert voice data into text data in real time.

[0549] 2. Text data preprocessing function:

[0550] Removes unnecessary characters and spaces from received text data and improves grammar and syntax, making it easier to analyze.

[0551] 3. Text data analysis function:

[0552] The preprocessed text data is analyzed using natural language processing (NLP) technology to extract the user's query intent and important keywords (entities). Specifically, open source NLP engines such as spaCy and TextBlob are used.

[0553] 4. Inquiry classification function:

[0554] Based on the extracted intent and entities, the inquiry is classified into pre-defined categories, and an appropriate response policy is then determined based on the pre-defined categories.

[0555] 5. Appropriate response generation:

[0556] For classified inquiries, an automatic response is generated using templates and an FAQ database. This automatically generated response is returned to the user in real time, allowing for a quick response.

[0557] 6. Urgency assessment and automatic escalation:

[0558] The system automatically determines the urgency of the inquiry and escalates it to a human representative if necessary, preventing serious problems.

[0559] Device Features

[0560] The devices used by security personnel and those on-site to receive inquiries are smartphones, smart glasses, head-mounted displays, etc. These devices are equipped with a voice input function, converting voice into text in real time and sending it to a server.

[0561] User Roles

[0562] When a problem occurs on-site, the user, as a security officer, sends a query to the server by voice input. For example, the following prompt sentences can be used:

[0563] "I've spotted a suspicious person. What should I do?"

[0564] "We just discovered a fire at our facility. What should we do?"

[0565] "I found a suspicious package. Do you need urgent action?"

[0566] Hardware and software used

[0567] This system uses the following hardware and software:

[0568] Hardware: Servers (with powerful CPUs and large amounts of memory), smartphones, smart glasses, head-mounted displays

[0569] Software: Python, spaCy (NLP engine), TextBlob (NLP engine), open source libraries for speech recognition (e.g., Google Speech-to-Text)

[0570] This enables immediate response on-site and automatic escalation in emergencies.

[0571] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0572] Step 1:

[0573] The user inputs voice data. For example, they might ask, "I spotted a suspicious person. What should I do?" This is sent to the server in real time using the voice input function of smart glasses or a head-mounted display. The input here is voice data, and the output is data converted into text format.

[0574] Step 2:

[0575] The server converts the received voice data into text data. An open source library for speech recognition (e.g., Google Speech-to-Text) is used to convert the voice data into text data. This process results in the voice data being converted into text data.

[0576] Step 3:

[0577] The server preprocesses the text data. It removes unnecessary characters and spaces from the received text data and improves grammar and syntax. This makes it easier to analyze. The input of this step is the converted text data, and the output is the preprocessed text data.

[0578] Step 4:

[0579] The server analyzes the preprocessed text data. Using a natural language processing (NLP) engine (e.g., spaCy), it analyzes the text data and extracts the user's query intent and important keywords (entities). The input is the preprocessed text data, and the output is the extracted intent and entities.

[0580] Step 5:

[0581] The server classifies queries based on intent and entities. Based on the extracted intent and entities, the query content is classified into predefined categories. The input is the extracted intent and entities, and the output is the classified query.

[0582] Step 6:

[0583] The server generates an appropriate response to the classified query. It uses templates and an FAQ database to automatically generate responses. For example, the response generated is "Contact the police immediately and monitor the scene from a safe location." The input is the classified query and the output is the generated response.

[0584] Step 7:

[0585] The server determines the urgency of the inquiry and automatically escalates it. If the inquiry is deemed urgent, it is escalated to a human agent. The input is the generated response and the inquiry itself, and the output is the decision to escalate or not.

[0586] Step 8:

[0587] The server sends the generated response back to the user. The server sends the response back to the user through the original communication channel (e.g., smart glasses or a head-mounted display). The input is the generated response and the output is the response sent back to the user.

[0588] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0589] The present invention is a system for improving the efficiency of call center inquiry operations, and mainly includes receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, returning responses, and an emotion engine that analyzes user emotions. Specific embodiments are described below.

[0590] System configuration

[0591] This system is implemented mainly by three entities: the server, the terminal, and the user. In addition, by incorporating an emotion engine for emotion analysis, more advanced inquiry response can be achieved.

[0592] System Operation

[0593] 1. User submits an inquiry

[0594] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[0595] 2. The server receives the query

[0596] The server receives the query text sent by the user and takes it as the request body.

[0597] 3. The server preprocesses the text

[0598] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0599] 4. The server analyzes the text using an NLP engine

[0600] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[0601] 5. The server analyzes emotions using an emotion engine

[0602] The server also sends the parsed text to an emotion engine, which analyzes the user's emotion (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords in the text. At this stage, the emotion may be analyzed as "anger."

[0603] 6. The server classifies the query

[0604] The server classifies the query into a predefined category based on the extracted intent and entities, as well as the analyzed sentiment, e.g., into the "returns" category.

[0605] 7. The server decides on a course of action

[0606] The server determines the appropriate response policy for the classified query. If the user's emotion is strong, it adjusts the tone of the response. For example, if the user is angry, it generates a calmer and more polite response.

[0607] 8. The server generates an automatic response

[0608] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During this generation process, placeholders in the template are replaced with actual data. For example, the server might generate a response such as, "You can return product ABC by clicking the link below. For more information, please see our return policy page."

[0609] 9. Server determines escalation

[0610] The server automatically determines whether a call needs to be escalated based on the content of the call and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, the server will hand over to a human agent. This decision is made when anger or strong dissatisfaction is detected.

[0611] 10. The server sends the response back to the user

[0612] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0613] 11. The user receives a response

[0614] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[0615] Specific examples

[0616] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "return" category, and the emotion engine further analyzes that the user's emotion is "anger." The system generates a response using an appropriate template and returns it to the user with the following response: "You can return product ABC by clicking the link below. Please see the return policy page for details." The user can then accept this response and proceed with the return process.

[0617] This embodiment not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[0618] The processing flow will be explained below.

[0619] Step 1:

[0620] The user submits a query.

[0621] The user enters the inquiry in text format using the call center's chat box or email form and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[0622] Step 2:

[0623] The server receives the query.

[0624] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed using a protocol such as an HTTP request or socket communication.

[0625] Step 3:

[0626] The server preprocesses the text.

[0627] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters from the text and grammatically arranges it. For example, it converts "Product ABC is broken. I would like to return it. Please tell me what to do." into "Product ABC is broken. Please tell me how to return it."

[0628] Step 4:

[0629] The server analyzes the text using an NLP engine.

[0630] The server sends the preprocessed text data to a natural language processing (NLP) engine for text analysis. The NLP engine breaks the text down into tokens and extracts the user's intent and important information (entities). For example, "intent = return" and "entity = product ABC breaks" are extracted.

[0631] Step 5:

[0632] The server analyzes emotions using an emotion engine.

[0633] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions. The emotion engine identifies emotions such as positive, negative, and neutral from the text. For example, "anger" may be detected from the user's text.

[0634] Step 6:

[0635] The server classifies the query.

[0636] The server classifies the query into a predefined category based on the extracted intent and entities, and the analyzed sentiment, e.g., into the "returns" category.

[0637] Step 7:

[0638] The server decides the response policy.

[0639] The server determines the appropriate response policy for the classified inquiry. If the user's emotion is strong, the tone of the response is adjusted. For example, if the user is angry, the server generates a polite and calm response.

[0640] Step 8:

[0641] The server generates an automatic response.

[0642] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During the generation process, specific information is filled in the placeholders in the template. For example, it generates a response such as, "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[0643] Step 9:

[0644] The server decides on the escalation.

[0645] The server automatically determines whether or not to escalate the inquiry based on the content of the inquiry and the analyzed sentiment. If the user's sentiment is particularly strong or the inquiry is complex, the server will hand it over to a human agent. This decision is also based on the results of sentiment analysis.

[0646] Step 10:

[0647] The server sends the response back to the user.

[0648] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0649] Step 11:

[0650] The user receives the response.

[0651] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[0652] Example 2

[0653] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0654] Call centers are required to respond to many inquiries quickly and appropriately, but typical manual responses are inefficient and prone to user dissatisfaction. Responses that ignore the user's feelings can also lower customer satisfaction. Furthermore, if appropriate escalation procedures are not carried out, there is a risk that important issues will be overlooked. To solve these issues, a system is needed that can automatically analyze the content of inquiries and respond in a way that takes the user's feelings into account.

[0655] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the query, means for classifying the query based on the extracted intent and entity, means for classifying the query based on the analyzed emotion, means for determining a response policy based on the classified query, means for generating an appropriate response based on the response policy, and means for returning the generated response to the user. This enables an efficient response to the query that takes emotion into consideration.

[0656] The "means for receiving text data" is a device or function for receiving a text-formatted inquiry sent by a user.

[0657] The "means for preprocessing received text data" refers to a device or function for removing unnecessary spaces and special characters from the received text data and for adjusting the grammar and syntax.

[0658] The "means for analyzing preprocessed text data to extract query intent and entities" is a device or function for automatically analyzing and extracting user intent and important information from preprocessed text data.

[0659] The "means for classifying queries based on extracted intent and entities" refers to a device or function for classifying queries into predefined categories based on the intent and entities extracted by the analysis.

[0660] A "means for classifying queries based on analyzed emotions" is a device or function for classifying queries according to their emotional state based on emotion analysis.

[0661] The "means for determining a response policy based on the classified inquiry" is a device or function for determining an appropriate response policy based on the classified inquiry.

[0662] The "means for generating an appropriate response based on a response policy" is a device or function for automatically generating an appropriate response based on a determined response policy.

[0663] A "means for returning a generated response to a user" is a device or function for returning a generated response to a user through the original communication channel.

[0664] This invention is a system for improving the efficiency of call center inquiry operations, and mainly includes text data reception, preprocessing, analysis, classification, response generation, automatic escalation, response return, and an emotion engine that analyzes user emotions. This system is implemented mainly by three entities: a server, a terminal, and a user.

[0665] System configuration

[0666] 1. Receiving text data

[0667] The server receives the text data (inquiry content) sent by the user. The hardware used includes a regular web server, mail server, or chatbot server.

[0668] 2. Preprocessing of text data

[0669] The server preprocesses the received text data, including removing unnecessary spaces and special characters, grammar reformatting, etc. The software used includes natural language processing libraries (e.g., NLTK and SpaCy).

[0670] 3. Text Data Analysis

[0671] The preprocessed text data is then analyzed by an NLP engine, which uses services such as Google Cloud NLP or Amazon Comprehend to extract user intent and important information (entities).

[0672] 4. Emotion analysis

[0673] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions using services such as IBM Watson Tone Analyzer, which identifies emotions (e.g., joy, anger, etc.) from the tone and keywords in the text.

[0674] 5. Categorizing inquiries

[0675] The server classifies queries into predefined categories based on the extracted intent and entities, and the analyzed sentiment, using rule-based classification algorithms and machine learning models.

[0676] 6. Deciding on a response policy

[0677] The server determines the appropriate response policy for each classified inquiry. In particular, it adjusts the tone of the response according to the user's emotions. If the user's emotions are strong, it generates a calmer and more polite response.

[0678] 7. Generate Auto-Responses

[0679] The server generates an appropriate answer using response templates and an FAQ database. During the generation process, it fills in the placeholders in the template with actual data. For example, it might generate an answer like, "You can return product ABC by following the link below. For more information, please see our returns policy page."

[0680] 8. Escalation Decision

[0681] The server automatically determines the need for escalation based on the query content and analyzed sentiment, specifically handing over to a human agent if the user's sentiment exceeds a certain threshold.

[0682] 9. Returning Responses

[0683] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0684] Specific examples

[0685] As a concrete example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me what to do." When the system receives this query, it preprocesses the text and analyzes it using an NLP engine. As a result of the analysis, the system classifies the query into the "return" category and further analyzes the user's emotion as "anger" using an emotion engine. Using an appropriate template, the system generates a response that reads, "You can return product ABC by clicking the link below. For details, please see the return policy page." and sends it back to the user. The user can then accept this response and proceed with the return process.

[0686] The present invention not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[0687] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0688] Step 1:

[0689] The user submits an inquiry. The user enters the inquiry details in the chat box or email form and presses the send button. For example, a prompt message such as "Product ABC is broken. I would like to return it. Please tell me what to do." is sent.

[0690] Step 2:

[0691] The server receives the query. The server receives the user's input and takes it as the request body. The input is the user's query text, and the output is this text data. The query content and meta information (user ID, timestamp, etc.) are saved in a database.

[0692] Step 3:

[0693] The server preprocesses the text data. It removes unnecessary spaces and special characters from the received text data and arranges the grammatical structure. For example, "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is the received text, and the output is the preprocessed text data. A natural language processing library (for example, NLTK or SpaCy) is used for processing.

[0694] Step 4:

[0695] The server analyzes the preprocessed text and extracts the query intent and entities. The server then sends the preprocessed text data to an NLP engine, which extracts the intent and entities. The input is the preprocessed text, and the output is the extracted intent and entities. Natural language processing services such as Google Cloud NLP and Amazon Comprehend are used. The extracted results are "intent = return" and "entity = product ABC breaks."

[0696] Step 5:

[0697] The server analyzes emotions from the extracted data. The server sends the analysis results to an emotion engine to analyze the user's emotions. For example, IBM Watson Tone Analyzer is used. The input is text containing the extracted intent and entities, and the output is the emotion analysis result (e.g., "anger").

[0698] Step 6:

[0699] The server classifies the query. The server classifies the query into the appropriate category based on the extracted intent, entities, and sentiment. The input is the parsed text data, and the output is the query category (e.g., the "returns" category). The processing uses rule-based classification algorithms or machine learning models.

[0700] Step 7:

[0701] The server determines the response policy based on the classified inquiry. In particular, if the user's emotion is strong, the tone of the response is adjusted. The input is the inquiry category and the emotion analysis result, and the output is the response policy.

[0702] Step 8:

[0703] The server generates an automatic response. Based on the response policy, the server generates an appropriate answer using response templates and an FAQ database. The input is the response policy, and the output is the generated response text. For example, it generates the text "You can return product ABC by clicking the link below. Please see the return policy page for details."

[0704] Step 9:

[0705] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry and the analyzed sentiment. The input is the inquiry data and sentiment analysis results, and the output is the escalation decision. In particular, if the user's sentiment is strong or the problem is complex, the server will hand over to a human agent.

[0706] Step 10:

[0707] The server sends the generated response back to the user The server sends the generated response back to the user through the original communication channel (chat box or email) The input is the generated response text, the output is the response sent back to the user

[0708] Step 11:

[0709] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The output is the response text displayed on the user's screen.

[0710] (Application example 2)

[0711] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0712] Modern online shopping sites are receiving an increasing number of inquiries from users, requiring efficient and accurate responses. However, conventional systems struggle to receive and preprocess text data, extract intent and entities, analyze emotions, and automatically escalate inquiries, making it difficult to improve customer satisfaction. Furthermore, there is a lack of a way to generate appropriate responses based on user emotions, limiting the efficiency of inquiry response and customer satisfaction. Therefore, a solution to this issue is needed.

[0713] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intention and events of the inquiry, means for analyzing the user's emotions, means for classifying the inquiry based on the extracted intention and events and the analyzed emotions, means for generating an appropriate response to the classified inquiry, and means for returning the generated response to the user. This makes it possible to improve the efficiency of inquiry operations and provide an appropriate response according to the user's emotions.

[0714] "Text data" is information in the form of text that a user transmits over a communication channel.

[0715] "Preprocessing" is a process that removes unnecessary spaces and special characters from the received text data and corrects the grammar and syntax.

[0716] "Intention" refers to the request or purpose that a user wants to convey through text data.

[0717] An "event" is specific information or an object extracted from text data.

[0718] "Emotions" are psychological states contained in the user's text data, and include joy, anger, anxiety, and the like.

[0719] "Query classification" refers to dividing the query content into specific categories based on the analyzed intent, events, and sentiment.

[0720] "Response generation" is the process of generating appropriate responses or information to classified inquiries.

[0721] "Returning a response" is the process of retransmitting the generated response to the user's communication channel.

[0722] "Automatic escalation" refers to handing over an inquiry to a human agent as needed based on the content of the inquiry and analyzed sentiment.

[0723] A "template" is a pre-prepared response template that can be customized for a specific situation.

[0724] The "FAQ Data Collection" is a database that compiles frequently asked questions and their answers.

[0725] The present invention provides a system for improving the efficiency of inquiry operations on mail-order sites and enabling appropriate responses according to the user's emotions. Specific embodiments for carrying out the invention will be described below.

[0726] System configuration

[0727] The system has three main components: a server, a terminal, and a user. The server plays a central role in processing everything from receiving text data to generating responses. The terminal provides a communication channel with the user, and the user makes inquiries through this terminal.

[0728] Hardware and Software

[0729] The server includes the following main software modules:

[0730] 1. NLP service: A software module that performs natural language processing (NLP) and analyzes the user's query text to extract intent and events.

[0731] 2. Sentiment Analysis Engine: A software module for analyzing user emotions from text data and identifying emotions such as anger, anxiety, and joy.

[0732] 3. FAQ data collection: A database containing pre-prepared response templates for many user inquiries.

[0733] Data processing and calculation

[0734] 1. Preprocessing of text data: The server performs preprocessing by formatting the text data received from the user and removing unnecessary spaces and special characters. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0735] 2. NLP analysis: The preprocessed text data is sent to the NLP service to extract intent and events. The NLP service analyzes information such as "intent = return" and "event = product ABC breaks."

[0736] 3. Sentiment Analysis: The text data is sent to a sentiment analysis engine to identify the user's emotions. The sentiment analysis engine extracts emotions such as anger, anxiety, or joy from the text. At this stage, the emotion may be analyzed as "anger."

[0737] 4. Enquiry Classification: Based on the extracted intent, events, and sentiment, the inquiry is classified into a specific category, e.g., “return” category.

[0738] 5. Response Generation: An appropriate response is generated for the classified inquiry based on the FAQ data set. If the user is angry, the tone of the response is adjusted and the placeholders in the template are replaced with actual data. For example, a response might be generated that reads, "You can return product ABC by clicking the link below. Please see the return policy page for details."

[0739] 6. Automatic Escalation: Automatically escalate queries based on query content and analyzed sentiment. If the user's sentiment is strong or exceeds a certain threshold, the query will be escalated to a human agent.

[0740] 7. Return Response: The generated response is returned to the user via the original communication channel (e.g., chat box or email).

[0741] Specific examples

[0742] When a user sends a query such as "Product ABC is broken. I would like to return it. Please tell me what to do," the server receives this query and performs preprocessing. It then uses NLP analysis to extract "Intention = Return" and "Event = Product ABC breaks," and uses sentiment analysis to identify the user's emotion as "Anger." Using an appropriate template, it generates a response saying, "You can return Product ABC by clicking the link below. Please see the return policy page for details," and sends it back to the user.

[0743] Example prompts to input to the generative AI model

[0744] "A user submits a query: (Product ABC is broken. I would like to return it. What should I do?) Preprocess this text, perform NLP analysis and sentiment analysis, and generate an optimal response. Please also consider the emotional tone of the response."

[0745] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0746] Step 1:

[0747] The user sends an inquiry. The user inputs and sends a text-format inquiry via the terminal. An example of such input is, "Product ABC is broken. I would like to return it. Please tell me what to do." The input text data is sent to the server.

[0748] Step 2:

[0749] The server receives the query. The server receives the text data sent by the user and takes it as the request body. The input data is left as is, and the output is raw text data that is subject to preprocessing.

[0750] Step 3:

[0751] The server preprocesses the text. It removes unnecessary spaces and special characters from the input text data and adjusts the grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is raw text data, and the output is preprocessed, clean text data.

[0752] Step 4:

[0753] The server analyzes the text using an NLP engine. The preprocessed text is sent to a natural language processing (NLP) engine, which analyzes the text. The NLP engine extracts the user's intent and important information (entities) from the text. From the input preprocessed text data, "Intent = Return" and "Event = Product ABC breaks" are extracted. The output is the analyzed intent and event.

[0754] Step 5:

[0755] The server analyzes emotions using an emotion engine. The analyzed text data is also sent to the emotion engine, which analyzes the user's emotions (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords of the text. The input is the analyzed text data, and the output is the identified emotion. At this stage, the emotion may be analyzed as "anger."

[0756] Step 6:

[0757] The server classifies the query. Based on the extracted intent and events, as well as the analyzed sentiment, the query is classified into a pre-defined category. For example, it is classified into the "return" category. The input is the intent, event, and sentiment data, and the output is a specific category.

[0758] Step 7:

[0759] The server determines the appropriate response policy and generates a response. It determines the appropriate response policy for the classified inquiry. If the user is angry, it adjusts the tone of the response and generates an appropriate answer to the user's inquiry using response templates and FAQ data. For example, it generates a response that reads, "You can return product ABC using the link below. Please see the return policy page for details." The input is the category and emotion data, and the output is the response sent back to the user.

[0760] Step 8:

[0761] The server makes the decision on escalation. It automatically determines the need for escalation based on the content of the inquiry and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, it will hand over to a human agent. The input is emotion data, and the output is instructions or execution of escalation.

[0762] Step 9:

[0763] The server sends the response back to the user. The generated response is sent back to the user via the original communication channel (chat box or email). The input is the generated response data, and the output is the message sent to the user.

[0764] Step 10:

[0765] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The input is the message sent from the server, and the output is the response displayed to the user.

[0766] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0767] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0768] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0769] [Third embodiment]

[0770] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0771] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0772] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0773] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0774] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0775] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0776] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0777] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0778] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0779] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0780] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0781] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0782] The present invention is a system for improving the efficiency of inquiry operations at a call center, and is mainly responsible for receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, and returning responses. Specific embodiments are described below.

[0783] System configuration

[0784] This system is implemented mainly by three entities: a server, a terminal, and a user.

[0785] System Operation

[0786] 1. User submits an inquiry

[0787] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[0788] 2. The server receives the query

[0789] The server receives the query text sent by the user. At this time, the text data is received in the form included in the request body.

[0790] 3. The server preprocesses the text

[0791] The server performs preprocessing on the received text data. Preprocessing includes deleting unnecessary characters and spaces and improving grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is reformatted to "Product ABC is broken. Please tell me how to return it."

[0792] 4. The server analyzes the data using an NLP engine.

[0793] The server passes the preprocessed text data to a natural language processing (NLP) engine for text analysis. This analysis extracts the query intent and entities (keywords). For example, the intent "return" and the entities "product ABC" and "break" are extracted.

[0794] 5. The server classifies the query

[0795] The server classifies the query into a predefined category based on the parsed intent and entities, for example, into the "returns" category.

[0796] 6. The server decides on a course of action

[0797] The server determines the appropriate response to the classified inquiry, for example, whether to generate an automated response based on a template or escalate the inquiry to a human agent.

[0798] 7. The server generates an automatic response

[0799] The server uses response templates and an FAQ database to generate an appropriate response to the user's inquiry, such as "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[0800] 8. Server determines escalation

[0801] The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative.

[0802] 9. The server returns the answer

[0803] The server returns the generated response to the user via the original communication channel (chat box, email, etc.) through which the user made the inquiry.

[0804] 10. The user receives the answer

[0805] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[0806] Specific examples

[0807] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "Returns" category and generates an answer using an appropriate template. The generated answer is, "You can return product ABC by clicking the link below. Please see the return policy page for details." and is sent back to the user. The user can then receive this answer and proceed with the return process.

[0808] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[0809] The processing flow will be explained below.

[0810] Step 1:

[0811] The user submits a query.

[0812] The user uses the call center's chat box or email form to enter the inquiry in text format and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[0813] Step 2:

[0814] The server receives the query.

[0815] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed through the API operated by the server.

[0816] Step 3:

[0817] The server preprocesses the text.

[0818] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0819] Step 4:

[0820] The server analyzes the text using an NLP engine.

[0821] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[0822] Step 5:

[0823] The server classifies the query.

[0824] The server classifies the query into a predefined category based on the extracted intent and entities, for example, into the "returns" category.

[0825] Step 6:

[0826] The server decides the response policy.

[0827] The server determines the appropriate course of action for the classified query, possibly generating an automatic response or determining whether escalation is necessary. For example, a simple query can be handled with an automatic response.

[0828] Step 7:

[0829] The server generates an automatic response.

[0830] The server generates an appropriate answer using a response template or FAQ database. During this generation process, it replaces placeholders in the template with actual data. For example, it might generate a response like, "You can return product ABC by clicking the link below. For more information, please see our returns policy page."

[0831] Step 8:

[0832] The server decides on the escalation.

[0833] The server automatically determines the need for escalation based on the content and classification of the inquiry, and transfers particularly complex inquiries or urgent complaints to a human representative.

[0834] Step 9:

[0835] The server sends the response back to the user.

[0836] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0837] Step 10:

[0838] The user receives the response.

[0839] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[0840] Example 1

[0841] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0842] While call center operations require a rapid and accurate response to a huge volume of inquiries, the speed and efficiency of response was insufficient with conventional systems, resulting in a decline in customer satisfaction. Furthermore, automated response systems were unable to handle complex inquiries or urgent complaints, and lacked the functionality to escalate to a human representative at the appropriate time.

[0843] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0844] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for determining an appropriate response to the classified inquiry, means for generating an automatic response based on the response, and means for returning the generated response to the user. This enables a prompt and accurate response to the inquiry, and automatically determines whether escalation is necessary and transfers the case to a human representative, thereby improving overall business efficiency and customer satisfaction.

[0845] "Text data" refers to the inquiry content sent by the user in the form of a character string.

[0846] "Means for receiving" refers to the function of importing text data sent by a user into the server.

[0847] "Preprocessing" refers to the ability to format received text data into a format that is easy to parse, including removing unnecessary characters and spaces, and improving grammar and syntax.

[0848] "Means for analyzing" refers to the function of analyzing preprocessed text data using a natural language processing (NLP) engine to extract query intent and entities.

[0849] "Inquiry intent" refers to the purpose or request contained in the user's inquiry, such as "returning a product" or "technical support."

[0850] An "entity" refers to an important keyword or unit of information contained in a user's inquiry. For example, a specific item such as "product ABC" or "breaks."

[0851] "Means for query classification" refers to the ability to classify query content into predefined categories based on extracted intent and entities.

[0852] "Means for determining response" refers to the function that determines the appropriate response to a classified inquiry, including the process for determining whether to generate an automated response or escalate.

[0853] "Means for generating automatic responses" refers to a function that generates appropriate responses to user inquiries using templates and an FAQ database.

[0854] "Means for sending back" refers to the function of sending back the generated response to the user through the communication channel used by the user (chat box, email, etc.).

[0855] The present invention is a system for improving the efficiency of call center inquiry operations, and is primarily responsible for receiving text data, preprocessing, analyzing, classifying, determining countermeasures, generating automatic responses, determining escalation, and returning responses.

[0856] System configuration

[0857] This system is implemented mainly by three entities: a server, a terminal, and a user. The specific hardware used includes a server and a user terminal (e.g., a PC or smartphone). The software used includes a natural language processing engine (e.g., SpaCy, BERT), a query management system (e.g., Zendesk), and a database (e.g., MySQL).

[0858] System Operation

[0859] The server receives the text data sent by the user, preprocesses it, and analyzes it. Based on the analysis results, it classifies the inquiry and determines the appropriate response, which may include generating an automated response or escalating the case to a human agent. Finally, the generated response is sent back to the user.

[0860] Examples of hardware and software

[0861] 1. Hardware:

[0862] Server: Processes and stores query data.

[0863] User terminal: Sends queries and receives responses.

[0864] 2. Software:

[0865] Natural language processing engines: Analyze text data (e.g., SpaCy, BERT).

[0866] Enquiry management system: Manages inquiry data and sends responses (e.g. Zendesk).

[0867] Database: Stores query and response data (e.g. MySQL).

[0868] Specific examples of operation

[0869] For example, if a user sends an inquiry saying "Product ABC is broken. I would like to return it. Please let me know how to do this," the inquiry will be processed as follows:

[0870] 1. The server receives this query text and retrieves the text data from the request body.

[0871] 2. The server uses regular expressions to remove unnecessary characters and spaces and preprocess the text into a format suitable for natural language processing. Specifically, it converts the text into the format "Product ABC is broken. Please tell me how to return it."

[0872] 3. The preprocessed text is passed to a natural language processing engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted.

[0873] 4. The server categorizes the inquiry based on the analysis results, for example, into the "returned goods" category.

[0874] 5. The server determines the appropriate response to the classified inquiry, for example, generating an auto-response based on a template or escalating the inquiry.

[0875] 6. An appropriate response is generated and sent back to the user, for example, "You can return product ABC by following the link below. Please see our returns policy page for more information."

[0876] Examples of prompt statements

[0877] 1. "Please preprocess this text data to remove unnecessary characters and spaces: Product ABC is broken. I would like to return it. What should I do?"

[0878] 2. "Please parse this preprocessed text data and extract the query intent and entities: Product ABC is broken. Please tell me how to return it."

[0879] 3. Generate an appropriate response based on the extracted intent and entity: Intent 'Return', Entity 'Product ABC', 'Broken'

[0880] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[0881] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0882] Step 1:

[0883] The user sends an inquiry. Using the call center's chat box or email form, the user sends a text inquiry from their terminal, such as "Product ABC is broken. I would like to return it. Please tell me what to do." Input: Inquiry text sent from the user's terminal. Output: Inquiry data in text format.

[0884] Step 2:

[0885] The server receives the query. The server receives the query text sent by the user from the request body. Input: Query text sent by the user. Output: Received text data.

[0886] Step 3:

[0887] The server preprocesses the text. The server uses regular expressions to remove unnecessary characters and spaces from the received text data and converts it into a format suitable for natural language processing. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted into "Product ABC is broken. Please tell me how to return it." Input: Received text data. Output: Preprocessed text data.

[0888] Step 4:

[0889] The server analyzes the data using an NLP engine. The server passes the preprocessed text data to a natural language processing (NLP) engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted. Input: Preprocessed text data. Output: Extracted intent and entities.

[0890] Step 5:

[0891] The server classifies the inquiry. Based on the parsed intent and entities, the server classifies the inquiry into a pre-defined category (e.g., "returns," "technical support," "product information"). For example, it is classified into the "returns" category. Input: Extracted intent and entities. Output: Classified inquiry category.

[0892] Step 6:

[0893] The server determines the appropriate response for the classified inquiry category. Specifically, it determines whether to generate an automatic response based on a template or escalate to a human agent. Input: Classified inquiry category. Output: Determined response (generating an automatic response or escalation).

[0894] Step 7:

[0895] The server generates an automatic response. The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. For example, it generates a response such as, "You can return product ABC by clicking the link below. Please see the return policy page for details." Input: The determined response. Output: The generated automatic response.

[0896] Step 8:

[0897] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative. Input: Inquiry content. Output: Escalation decision.

[0898] Step 9:

[0899] The server returns an answer. The server returns the generated response to the user. This is done through the original communication channel through which the user made the query (chat box, email, etc.). Input: The generated automated response. Output: The response returned to the user.

[0900] Step 10:

[0901] The user receives the answer. The user checks the response sent from the server on their own device. For example, a message is displayed in the chat box or an email is received. Input: The response sent from the server. Output: The response checked on the user device.

[0902] (Application example 1)

[0903] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0904] Conventional systems for responding to inquiries at call centers and monitoring centers have made it difficult to respond quickly and appropriately, and responses are often delayed, especially in urgent situations. This has led to a decline in customer satisfaction and increased risk of serious problems. Furthermore, there was no system in place to efficiently process inquiries using real-time voice input.

[0905] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0906] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for generating an appropriate response to the classified inquiry, means for returning the generated response to the user, means for converting voice input into text data in real time, and means for determining urgency and performing automatic escalation. This enables prompt and appropriate processing of inquiries, and enables prompt responses, especially in emergencies.

[0907] The "means for receiving text data" is a device or program that has the function of allowing the server to receive a text-formatted inquiry sent from a user.

[0908] The "means for preprocessing received text data" refers to a device or program that executes a process to remove unnecessary characters and spaces from the received text data and to improve grammar and syntax.

[0909] "Means for analyzing preprocessed text data to extract query intent and entities" refers to a device or program that has the function of analyzing preprocessed text data using natural language processing (NLP) technology and extracting the main intent of a user's query and important keywords (entities).

[0910] The "means for classifying inquiries based on the extracted intent and entities" refers to a device or program that has the function of classifying the content of inquiries into pre-defined categories based on the intent and entities extracted through analysis.

[0911] The "means for generating appropriate responses to classified inquiries" refers to a device or program that has the function of automatically generating appropriate answers based on the classified inquiry content using templates and an FAQ database.

[0912] The "means for returning the generated response to the user" refers to a device or program that has the function of returning the automatically generated response to the user, and performs the return via a communication channel.

[0913] The "means for converting voice input into text data in real time" refers to a device or program that has the function of recognizing a user's voice input in real time and converting it into text data.

[0914] "Means for determining urgency and automatically escalating" refers to a device or program that has the function of automatically determining the urgency of an inquiry and, if necessary, transferring the inquiry to a human representative.

[0915] The system for implementing this invention is composed of three main entities: a server, a terminal, and a user. To understand the specific processing procedure, the operation of this system will be explained.

[0916] Server Features

[0917] The server has the following functions:

[0918] 1. Function to receive text data:

[0919] The server receives the inquiry sent by the user in text format. This function also includes a function to convert voice data into text data in real time.

[0920] 2. Text data preprocessing function:

[0921] Removes unnecessary characters and spaces from received text data and improves grammar and syntax, making it easier to analyze.

[0922] 3. Text data analysis function:

[0923] The preprocessed text data is analyzed using natural language processing (NLP) technology to extract the user's query intent and important keywords (entities). Specifically, open source NLP engines such as spaCy and TextBlob are used.

[0924] 4. Inquiry classification function:

[0925] Based on the extracted intent and entities, the inquiry is classified into pre-defined categories, and an appropriate response policy is then determined based on the pre-defined categories.

[0926] 5. Appropriate response generation:

[0927] For classified inquiries, an automatic response is generated using templates and an FAQ database. This automatically generated response is returned to the user in real time, allowing for a quick response.

[0928] 6. Urgency assessment and automatic escalation:

[0929] The system automatically determines the urgency of the inquiry and escalates it to a human representative if necessary, preventing serious problems.

[0930] Device Features

[0931] The devices used by security personnel and those on-site to receive inquiries are smartphones, smart glasses, head-mounted displays, etc. These devices are equipped with a voice input function, converting voice into text in real time and sending it to a server.

[0932] User Roles

[0933] When a problem occurs on-site, the user, as a security officer, sends a query to the server by voice input. For example, the following prompt sentences can be used:

[0934] "I've spotted a suspicious person. What should I do?"

[0935] "We just discovered a fire at our facility. What should we do?"

[0936] "I found a suspicious package. Do you need urgent action?"

[0937] Hardware and software used

[0938] This system uses the following hardware and software:

[0939] Hardware: Servers (with powerful CPUs and large amounts of memory), smartphones, smart glasses, head-mounted displays

[0940] Software: Python, spaCy (NLP engine), TextBlob (NLP engine), open source libraries for speech recognition (e.g., Google Speech-to-Text)

[0941] This enables immediate response on-site and automatic escalation in emergencies.

[0942] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0943] Step 1:

[0944] The user inputs voice data. For example, they might ask, "I spotted a suspicious person. What should I do?" This is sent to the server in real time using the voice input function of smart glasses or a head-mounted display. The input here is voice data, and the output is data converted into text format.

[0945] Step 2:

[0946] The server converts the received voice data into text data. An open source library for speech recognition (e.g., Google Speech-to-Text) is used to convert the voice data into text data. This process results in the voice data being converted into text data.

[0947] Step 3:

[0948] The server preprocesses the text data. It removes unnecessary characters and spaces from the received text data and improves grammar and syntax. This makes it easier to analyze. The input of this step is the converted text data, and the output is the preprocessed text data.

[0949] Step 4:

[0950] The server analyzes the preprocessed text data. Using a natural language processing (NLP) engine (e.g., spaCy), it analyzes the text data and extracts the user's query intent and important keywords (entities). The input is the preprocessed text data, and the output is the extracted intent and entities.

[0951] Step 5:

[0952] The server classifies queries based on intent and entities. Based on the extracted intent and entities, the query content is classified into predefined categories. The input is the extracted intent and entities, and the output is the classified query.

[0953] Step 6:

[0954] The server generates an appropriate response to the classified query. It uses templates and an FAQ database to automatically generate responses. For example, the response generated is "Contact the police immediately and monitor the scene from a safe location." The input is the classified query and the output is the generated response.

[0955] Step 7:

[0956] The server determines the urgency of the inquiry and automatically escalates it. If the inquiry is deemed urgent, it is escalated to a human agent. The input is the generated response and the inquiry itself, and the output is the decision to escalate or not.

[0957] Step 8:

[0958] The server sends the generated response back to the user. The server sends the response back to the user through the original communication channel (e.g., smart glasses or a head-mounted display). The input is the generated response and the output is the response sent back to the user.

[0959] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0960] The present invention is a system for improving the efficiency of call center inquiry operations, and mainly includes receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, returning responses, and an emotion engine that analyzes user emotions. Specific embodiments are described below.

[0961] System configuration

[0962] This system is implemented mainly by three entities: the server, the terminal, and the user. In addition, by incorporating an emotion engine for emotion analysis, more advanced inquiry response can be achieved.

[0963] System Operation

[0964] 1. User submits an inquiry

[0965] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[0966] 2. The server receives the query

[0967] The server receives the query text sent by the user and takes it as the request body.

[0968] 3. The server preprocesses the text

[0969] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[0970] 4. The server analyzes the text using an NLP engine

[0971] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[0972] 5. The server analyzes emotions using an emotion engine

[0973] The server also sends the parsed text to an emotion engine, which analyzes the user's emotion (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords in the text. At this stage, the emotion may be analyzed as "anger."

[0974] 6. The server classifies the query

[0975] The server classifies the query into a predefined category based on the extracted intent and entities, as well as the analyzed sentiment, e.g., into the "returns" category.

[0976] 7. The server decides on a course of action

[0977] The server determines the appropriate response policy for the classified query. If the user's emotion is strong, it adjusts the tone of the response. For example, if the user is angry, it generates a calmer and more polite response.

[0978] 8. The server generates an automatic response

[0979] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During this generation process, placeholders in the template are replaced with actual data. For example, the server might generate a response such as, "You can return product ABC by clicking the link below. For more information, please see our return policy page."

[0980] 9. Server determines escalation

[0981] The server automatically determines whether a call needs to be escalated based on the content of the call and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, the server will hand over to a human agent. This decision is made when anger or strong dissatisfaction is detected.

[0982] 10. The server sends the response back to the user

[0983] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[0984] 11. The user receives a response

[0985] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[0986] Specific examples

[0987] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "return" category, and the emotion engine further analyzes that the user's emotion is "anger." The system generates a response using an appropriate template and returns it to the user with the following response: "You can return product ABC by clicking the link below. Please see the return policy page for details." The user can then accept this response and proceed with the return process.

[0988] This embodiment not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[0989] The processing flow will be explained below.

[0990] Step 1:

[0991] The user submits a query.

[0992] The user enters the inquiry in text format using the call center's chat box or email form and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[0993] Step 2:

[0994] The server receives the query.

[0995] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed using a protocol such as an HTTP request or socket communication.

[0996] Step 3:

[0997] The server preprocesses the text.

[0998] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters from the text and grammatically arranges it. For example, it converts "Product ABC is broken. I would like to return it. Please tell me what to do." into "Product ABC is broken. Please tell me how to return it."

[0999] Step 4:

[1000] The server analyzes the text using an NLP engine.

[1001] The server sends the preprocessed text data to a natural language processing (NLP) engine for text analysis. The NLP engine breaks the text down into tokens and extracts the user's intent and important information (entities). For example, "intent = return" and "entity = product ABC breaks" are extracted.

[1002] Step 5:

[1003] The server analyzes emotions using an emotion engine.

[1004] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions. The emotion engine identifies emotions such as positive, negative, and neutral from the text. For example, "anger" may be detected from the user's text.

[1005] Step 6:

[1006] The server classifies the query.

[1007] The server classifies the query into a predefined category based on the extracted intent and entities, and the analyzed sentiment, e.g., into the "returns" category.

[1008] Step 7:

[1009] The server decides the response policy.

[1010] The server determines the appropriate response policy for the classified inquiry. If the user's emotion is strong, the tone of the response is adjusted. For example, if the user is angry, the server generates a polite and calm response.

[1011] Step 8:

[1012] The server generates an automatic response.

[1013] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During the generation process, specific information is filled in the placeholders in the template. For example, it generates a response such as, "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[1014] Step 9:

[1015] The server decides on the escalation.

[1016] The server automatically determines whether or not to escalate the inquiry based on the content of the inquiry and the analyzed sentiment. If the user's sentiment is particularly strong or the inquiry is complex, the server will hand it over to a human agent. This decision is also based on the results of sentiment analysis.

[1017] Step 10:

[1018] The server sends the response back to the user.

[1019] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[1020] Step 11:

[1021] The user receives the response.

[1022] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[1023] Example 2

[1024] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1025] Call centers are required to respond to many inquiries quickly and appropriately, but typical manual responses are inefficient and prone to user dissatisfaction. Responses that ignore the user's feelings can also lower customer satisfaction. Furthermore, if appropriate escalation procedures are not carried out, there is a risk that important issues will be overlooked. To solve these issues, a system is needed that can automatically analyze the content of inquiries and respond in a way that takes the user's feelings into account.

[1026] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the query, means for classifying the query based on the extracted intent and entity, means for classifying the query based on the analyzed emotion, means for determining a response policy based on the classified query, means for generating an appropriate response based on the response policy, and means for returning the generated response to the user. This enables an efficient response to the query that takes emotion into consideration.

[1027] The "means for receiving text data" is a device or function for receiving a text-formatted inquiry sent by a user.

[1028] The "means for preprocessing received text data" refers to a device or function for removing unnecessary spaces and special characters from the received text data and for adjusting the grammar and syntax.

[1029] The "means for analyzing preprocessed text data to extract query intent and entities" is a device or function for automatically analyzing and extracting user intent and important information from preprocessed text data.

[1030] The "means for classifying queries based on extracted intent and entities" refers to a device or function for classifying queries into predefined categories based on the intent and entities extracted by the analysis.

[1031] A "means for classifying queries based on analyzed emotions" is a device or function for classifying queries according to their emotional state based on emotion analysis.

[1032] The "means for determining a response policy based on the classified inquiry" is a device or function for determining an appropriate response policy based on the classified inquiry.

[1033] The "means for generating an appropriate response based on a response policy" is a device or function for automatically generating an appropriate response based on a determined response policy.

[1034] A "means for returning a generated response to a user" is a device or function for returning a generated response to a user through the original communication channel.

[1035] This invention is a system for improving the efficiency of call center inquiry operations, and mainly includes text data reception, preprocessing, analysis, classification, response generation, automatic escalation, response return, and an emotion engine that analyzes user emotions. This system is implemented mainly by three entities: a server, a terminal, and a user.

[1036] System configuration

[1037] 1. Receiving text data

[1038] The server receives the text data (inquiry content) sent by the user. The hardware used includes a regular web server, mail server, or chatbot server.

[1039] 2. Preprocessing of text data

[1040] The server preprocesses the received text data, including removing unnecessary spaces and special characters, grammar reformatting, etc. The software used includes natural language processing libraries (e.g., NLTK and SpaCy).

[1041] 3. Text Data Analysis

[1042] The preprocessed text data is then analyzed by an NLP engine, which uses services such as Google Cloud NLP or Amazon Comprehend to extract user intent and important information (entities).

[1043] 4. Emotion analysis

[1044] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions using services such as IBM Watson Tone Analyzer, which identifies emotions (e.g., joy, anger, etc.) from the tone and keywords in the text.

[1045] 5. Categorizing inquiries

[1046] The server classifies queries into predefined categories based on the extracted intent and entities, and the analyzed sentiment, using rule-based classification algorithms and machine learning models.

[1047] 6. Deciding on a response policy

[1048] The server determines the appropriate response policy for each classified inquiry. In particular, it adjusts the tone of the response according to the user's emotions. If the user's emotions are strong, it generates a calmer and more polite response.

[1049] 7. Generate Auto-Responses

[1050] The server generates an appropriate answer using response templates and an FAQ database. During the generation process, it fills in the placeholders in the template with actual data. For example, it might generate an answer like, "You can return product ABC by following the link below. For more information, please see our returns policy page."

[1051] 8. Escalation Decision

[1052] The server automatically determines the need for escalation based on the query content and analyzed sentiment, specifically handing over to a human agent if the user's sentiment exceeds a certain threshold.

[1053] 9. Returning Responses

[1054] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[1055] Specific examples

[1056] As a concrete example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me what to do." When the system receives this query, it preprocesses the text and analyzes it using an NLP engine. As a result of the analysis, the system classifies the query into the "return" category and further analyzes the user's emotion as "anger" using an emotion engine. Using an appropriate template, the system generates a response that reads, "You can return product ABC by clicking the link below. For details, please see the return policy page." and sends it back to the user. The user can then accept this response and proceed with the return process.

[1057] The present invention not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[1058] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1059] Step 1:

[1060] The user submits an inquiry. The user enters the inquiry details in the chat box or email form and presses the send button. For example, a prompt message such as "Product ABC is broken. I would like to return it. Please tell me what to do." is sent.

[1061] Step 2:

[1062] The server receives the query. The server receives the user's input and takes it as the request body. The input is the user's query text, and the output is this text data. The query content and meta information (user ID, timestamp, etc.) are saved in a database.

[1063] Step 3:

[1064] The server preprocesses the text data. It removes unnecessary spaces and special characters from the received text data and arranges the grammatical structure. For example, "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is the received text, and the output is the preprocessed text data. A natural language processing library (for example, NLTK or SpaCy) is used for processing.

[1065] Step 4:

[1066] The server analyzes the preprocessed text and extracts the query intent and entities. The server then sends the preprocessed text data to an NLP engine, which extracts the intent and entities. The input is the preprocessed text, and the output is the extracted intent and entities. Natural language processing services such as Google Cloud NLP and Amazon Comprehend are used. The extracted results are "intent = return" and "entity = product ABC breaks."

[1067] Step 5:

[1068] The server analyzes emotions from the extracted data. The server sends the analysis results to an emotion engine to analyze the user's emotions. For example, IBM Watson Tone Analyzer is used. The input is text containing the extracted intent and entities, and the output is the emotion analysis result (e.g., "anger").

[1069] Step 6:

[1070] The server classifies the query. The server classifies the query into the appropriate category based on the extracted intent, entities, and sentiment. The input is the parsed text data, and the output is the query category (e.g., the "returns" category). The processing uses rule-based classification algorithms or machine learning models.

[1071] Step 7:

[1072] The server determines the response policy based on the classified inquiry. In particular, if the user's emotion is strong, the tone of the response is adjusted. The input is the inquiry category and the emotion analysis result, and the output is the response policy.

[1073] Step 8:

[1074] The server generates an automatic response. Based on the response policy, the server generates an appropriate answer using response templates and an FAQ database. The input is the response policy, and the output is the generated response text. For example, it generates the text "You can return product ABC by clicking the link below. Please see the return policy page for details."

[1075] Step 9:

[1076] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry and the analyzed sentiment. The input is the inquiry data and sentiment analysis results, and the output is the escalation decision. In particular, if the user's sentiment is strong or the problem is complex, the server will hand over to a human agent.

[1077] Step 10:

[1078] The server sends the generated response back to the user The server sends the generated response back to the user through the original communication channel (chat box or email) The input is the generated response text, the output is the response sent back to the user

[1079] Step 11:

[1080] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The output is the response text displayed on the user's screen.

[1081] (Application example 2)

[1082] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1083] Modern online shopping sites are receiving an increasing number of inquiries from users, requiring efficient and accurate responses. However, conventional systems struggle to receive and preprocess text data, extract intent and entities, analyze emotions, and automatically escalate inquiries, making it difficult to improve customer satisfaction. Furthermore, there is a lack of a way to generate appropriate responses based on user emotions, limiting the efficiency of inquiry response and customer satisfaction. Therefore, a solution to this issue is needed.

[1084] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intention and events of the inquiry, means for analyzing the user's emotions, means for classifying the inquiry based on the extracted intention and events and the analyzed emotions, means for generating an appropriate response to the classified inquiry, and means for returning the generated response to the user. This makes it possible to improve the efficiency of inquiry operations and provide an appropriate response according to the user's emotions.

[1085] "Text data" is information in the form of text that a user transmits over a communication channel.

[1086] "Preprocessing" is a process that removes unnecessary spaces and special characters from the received text data and corrects the grammar and syntax.

[1087] "Intention" refers to the request or purpose that a user wants to convey through text data.

[1088] An "event" is specific information or an object extracted from text data.

[1089] "Emotions" are psychological states contained in the user's text data, and include joy, anger, anxiety, and the like.

[1090] "Query classification" refers to dividing the query content into specific categories based on the analyzed intent, events, and sentiment.

[1091] "Response generation" is the process of generating appropriate responses or information to classified inquiries.

[1092] "Returning a response" is the process of retransmitting the generated response to the user's communication channel.

[1093] "Automatic escalation" refers to handing over an inquiry to a human agent as needed based on the content of the inquiry and analyzed sentiment.

[1094] A "template" is a pre-prepared response template that can be customized for a specific situation.

[1095] The "FAQ Data Collection" is a database that compiles frequently asked questions and their answers.

[1096] The present invention provides a system for improving the efficiency of inquiry operations on mail-order sites and enabling appropriate responses according to the user's emotions. Specific embodiments for carrying out the invention will be described below.

[1097] System configuration

[1098] The system has three main components: a server, a terminal, and a user. The server plays a central role in processing everything from receiving text data to generating responses. The terminal provides a communication channel with the user, and the user makes inquiries through this terminal.

[1099] Hardware and Software

[1100] The server includes the following main software modules:

[1101] 1. NLP service: A software module that performs natural language processing (NLP) and analyzes the user's query text to extract intent and events.

[1102] 2. Sentiment Analysis Engine: A software module for analyzing user emotions from text data and identifying emotions such as anger, anxiety, and joy.

[1103] 3. FAQ data collection: A database containing pre-prepared response templates for many user inquiries.

[1104] Data processing and calculation

[1105] 1. Preprocessing of text data: The server performs preprocessing by formatting the text data received from the user and removing unnecessary spaces and special characters. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[1106] 2. NLP analysis: The preprocessed text data is sent to the NLP service to extract intent and events. The NLP service analyzes information such as "intent = return" and "event = product ABC breaks."

[1107] 3. Sentiment Analysis: The text data is sent to a sentiment analysis engine to identify the user's emotions. The sentiment analysis engine extracts emotions such as anger, anxiety, or joy from the text. At this stage, the emotion may be analyzed as "anger."

[1108] 4. Enquiry Classification: Based on the extracted intent, events, and sentiment, the inquiry is classified into a specific category, e.g., “return” category.

[1109] 5. Response Generation: An appropriate response is generated for the classified inquiry based on the FAQ data set. If the user is angry, the tone of the response is adjusted and the placeholders in the template are replaced with actual data. For example, a response might be generated that reads, "You can return product ABC by clicking the link below. Please see the return policy page for details."

[1110] 6. Automatic Escalation: Automatically escalate queries based on query content and analyzed sentiment. If the user's sentiment is strong or exceeds a certain threshold, the query will be escalated to a human agent.

[1111] 7. Return Response: The generated response is returned to the user via the original communication channel (e.g., chat box or email).

[1112] Specific examples

[1113] When a user sends a query such as "Product ABC is broken. I would like to return it. Please tell me what to do," the server receives this query and performs preprocessing. It then uses NLP analysis to extract "Intention = Return" and "Event = Product ABC breaks," and uses sentiment analysis to identify the user's emotion as "Anger." Using an appropriate template, it generates a response saying, "You can return Product ABC by clicking the link below. Please see the return policy page for details," and sends it back to the user.

[1114] Example prompts to input to the generative AI model

[1115] "A user submits a query: (Product ABC is broken. I would like to return it. What should I do?) Preprocess this text, perform NLP analysis and sentiment analysis, and generate an optimal response. Please also consider the emotional tone of the response."

[1116] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1117] Step 1:

[1118] The user sends an inquiry. The user inputs and sends a text-format inquiry via the terminal. An example of such input is, "Product ABC is broken. I would like to return it. Please tell me what to do." The input text data is sent to the server.

[1119] Step 2:

[1120] The server receives the query. The server receives the text data sent by the user and takes it as the request body. The input data is left as is, and the output is raw text data that is subject to preprocessing.

[1121] Step 3:

[1122] The server preprocesses the text. It removes unnecessary spaces and special characters from the input text data and adjusts the grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is raw text data, and the output is preprocessed, clean text data.

[1123] Step 4:

[1124] The server analyzes the text using an NLP engine. The preprocessed text is sent to a natural language processing (NLP) engine, which analyzes the text. The NLP engine extracts the user's intent and important information (entities) from the text. From the input preprocessed text data, "Intent = Return" and "Event = Product ABC breaks" are extracted. The output is the analyzed intent and event.

[1125] Step 5:

[1126] The server analyzes emotions using an emotion engine. The analyzed text data is also sent to the emotion engine, which analyzes the user's emotions (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords of the text. The input is the analyzed text data, and the output is the identified emotion. At this stage, the emotion may be analyzed as "anger."

[1127] Step 6:

[1128] The server classifies the query. Based on the extracted intent and events, as well as the analyzed sentiment, the query is classified into a pre-defined category. For example, it is classified into the "return" category. The input is the intent, event, and sentiment data, and the output is a specific category.

[1129] Step 7:

[1130] The server determines the appropriate response policy and generates a response. It determines the appropriate response policy for the classified inquiry. If the user is angry, it adjusts the tone of the response and generates an appropriate answer to the user's inquiry using response templates and FAQ data. For example, it generates a response that reads, "You can return product ABC using the link below. Please see the return policy page for details." The input is the category and emotion data, and the output is the response sent back to the user.

[1131] Step 8:

[1132] The server makes the decision on escalation. It automatically determines the need for escalation based on the content of the inquiry and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, it will hand over to a human agent. The input is emotion data, and the output is instructions or execution of escalation.

[1133] Step 9:

[1134] The server sends the response back to the user. The generated response is sent back to the user via the original communication channel (chat box or email). The input is the generated response data, and the output is the message sent to the user.

[1135] Step 10:

[1136] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The input is the message sent from the server, and the output is the response displayed to the user.

[1137] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1138] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1139] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1140] [Fourth embodiment]

[1141] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1142] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1143] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1144] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1145] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1146] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1147] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1148] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1149] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1150] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1151] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1152] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1153] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1154] The present invention is a system for improving the efficiency of inquiry operations at a call center, and is mainly responsible for receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, and returning responses. Specific embodiments are described below.

[1155] System configuration

[1156] This system is implemented mainly by three entities: a server, a terminal, and a user.

[1157] System Operation

[1158] 1. User submits an inquiry

[1159] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[1160] 2. The server receives the query

[1161] The server receives the query text sent by the user. At this time, the text data is received in the form included in the request body.

[1162] 3. The server preprocesses the text

[1163] The server performs preprocessing on the received text data. Preprocessing includes deleting unnecessary characters and spaces and improving grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is reformatted to "Product ABC is broken. Please tell me how to return it."

[1164] 4. The server analyzes the data using an NLP engine.

[1165] The server passes the preprocessed text data to a natural language processing (NLP) engine for text analysis. This analysis extracts the query intent and entities (keywords). For example, the intent "return" and the entities "product ABC" and "break" are extracted.

[1166] 5. The server classifies the query

[1167] The server classifies the query into a predefined category based on the parsed intent and entities, for example, into the "returns" category.

[1168] 6. The server decides on a course of action

[1169] The server determines the appropriate response to the classified inquiry, for example, whether to generate an automated response based on a template or escalate the inquiry to a human agent.

[1170] 7. The server generates an automatic response

[1171] The server uses response templates and an FAQ database to generate an appropriate response to the user's inquiry, such as "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[1172] 8. Server determines escalation

[1173] The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative.

[1174] 9. The server returns the answer

[1175] The server returns the generated response to the user via the original communication channel (chat box, email, etc.) through which the user made the inquiry.

[1176] 10. The user receives the answer

[1177] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[1178] Specific examples

[1179] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "Returns" category and generates an answer using an appropriate template. The generated answer is, "You can return product ABC by clicking the link below. Please see the return policy page for details." and is sent back to the user. The user can then receive this answer and proceed with the return process.

[1180] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[1181] The processing flow will be explained below.

[1182] Step 1:

[1183] The user submits a query.

[1184] The user uses the call center's chat box or email form to enter the inquiry in text format and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[1185] Step 2:

[1186] The server receives the query.

[1187] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed through the API operated by the server.

[1188] Step 3:

[1189] The server preprocesses the text.

[1190] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[1191] Step 4:

[1192] The server analyzes the text using an NLP engine.

[1193] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[1194] Step 5:

[1195] The server classifies the query.

[1196] The server classifies the query into a predefined category based on the extracted intent and entities, for example, into the "returns" category.

[1197] Step 6:

[1198] The server decides the response policy.

[1199] The server determines the appropriate course of action for the classified query, possibly generating an automatic response or determining whether escalation is necessary. For example, a simple query can be handled with an automatic response.

[1200] Step 7:

[1201] The server generates an automatic response.

[1202] The server generates an appropriate answer using a response template or FAQ database. During this generation process, it replaces placeholders in the template with actual data. For example, it might generate a response like, "You can return product ABC by clicking the link below. For more information, please see our returns policy page."

[1203] Step 8:

[1204] The server decides on the escalation.

[1205] The server automatically determines the need for escalation based on the content and classification of the inquiry, and transfers particularly complex inquiries or urgent complaints to a human representative.

[1206] Step 9:

[1207] The server sends the response back to the user.

[1208] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[1209] Step 10:

[1210] The user receives the response.

[1211] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[1212] Example 1

[1213] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1214] While call center operations require a rapid and accurate response to a huge volume of inquiries, the speed and efficiency of response was insufficient with conventional systems, resulting in a decline in customer satisfaction. Furthermore, automated response systems were unable to handle complex inquiries or urgent complaints, and lacked the functionality to escalate to a human representative at the appropriate time.

[1215] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1216] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for determining an appropriate response to the classified inquiry, means for generating an automatic response based on the response, and means for returning the generated response to the user. This enables a prompt and accurate response to the inquiry, and automatically determines whether escalation is necessary and transfers the case to a human representative, thereby improving overall business efficiency and customer satisfaction.

[1217] "Text data" refers to the inquiry content sent by the user in the form of a character string.

[1218] "Means for receiving" refers to the function of importing text data sent by a user into the server.

[1219] "Preprocessing" refers to the ability to format received text data into a format that is easy to parse, including removing unnecessary characters and spaces, and improving grammar and syntax.

[1220] "Means for analyzing" refers to the function of analyzing preprocessed text data using a natural language processing (NLP) engine to extract query intent and entities.

[1221] "Inquiry intent" refers to the purpose or request contained in the user's inquiry, such as "returning a product" or "technical support."

[1222] An "entity" refers to an important keyword or unit of information contained in a user's inquiry. For example, a specific item such as "product ABC" or "breaks."

[1223] "Means for query classification" refers to the ability to classify query content into predefined categories based on extracted intent and entities.

[1224] "Means for determining response" refers to the function that determines the appropriate response to a classified inquiry, including the process for determining whether to generate an automated response or escalate.

[1225] "Means for generating automatic responses" refers to a function that generates appropriate responses to user inquiries using templates and an FAQ database.

[1226] "Means for sending back" refers to the function of sending back the generated response to the user through the communication channel used by the user (chat box, email, etc.).

[1227] The present invention is a system for improving the efficiency of call center inquiry operations, and is primarily responsible for receiving text data, preprocessing, analyzing, classifying, determining countermeasures, generating automatic responses, determining escalation, and returning responses.

[1228] System configuration

[1229] This system is implemented mainly by three entities: a server, a terminal, and a user. The specific hardware used includes a server and a user terminal (e.g., a PC or smartphone). The software used includes a natural language processing engine (e.g., SpaCy, BERT), a query management system (e.g., Zendesk), and a database (e.g., MySQL).

[1230] System Operation

[1231] The server receives the text data sent by the user, preprocesses it, and analyzes it. Based on the analysis results, it classifies the inquiry and determines the appropriate response, which may include generating an automated response or escalating the case to a human agent. Finally, the generated response is sent back to the user.

[1232] Examples of hardware and software

[1233] 1. Hardware:

[1234] Server: Processes and stores query data.

[1235] User terminal: Sends queries and receives responses.

[1236] 2. Software:

[1237] Natural language processing engines: Analyze text data (e.g., SpaCy, BERT).

[1238] Enquiry management system: Manages inquiry data and sends responses (e.g. Zendesk).

[1239] Database: Stores query and response data (e.g. MySQL).

[1240] Specific examples of operation

[1241] For example, if a user sends an inquiry saying "Product ABC is broken. I would like to return it. Please let me know how to do this," the inquiry will be processed as follows:

[1242] 1. The server receives this query text and retrieves the text data from the request body.

[1243] 2. The server uses regular expressions to remove unnecessary characters and spaces and preprocess the text into a format suitable for natural language processing. Specifically, it converts the text into the format "Product ABC is broken. Please tell me how to return it."

[1244] 3. The preprocessed text is passed to a natural language processing engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted.

[1245] 4. The server categorizes the inquiry based on the analysis results, for example, into the "returned goods" category.

[1246] 5. The server determines the appropriate response to the classified inquiry, for example, generating an auto-response based on a template or escalating the inquiry.

[1247] 6. An appropriate response is generated and sent back to the user, for example, "You can return product ABC by following the link below. Please see our returns policy page for more information."

[1248] Examples of prompt statements

[1249] 1. "Please preprocess this text data to remove unnecessary characters and spaces: Product ABC is broken. I would like to return it. What should I do?"

[1250] 2. "Please parse this preprocessed text data and extract the query intent and entities: Product ABC is broken. Please tell me how to return it."

[1251] 3. Generate an appropriate response based on the extracted intent and entity: Intent 'Return', Entity 'Product ABC', 'Broken'

[1252] This embodiment is expected to significantly improve the efficiency of call center inquiry operations, shorten response times, and improve customer satisfaction.

[1253] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1254] Step 1:

[1255] The user sends an inquiry. Using the call center's chat box or email form, the user sends a text inquiry from their terminal, such as "Product ABC is broken. I would like to return it. Please tell me what to do." Input: Inquiry text sent from the user's terminal. Output: Inquiry data in text format.

[1256] Step 2:

[1257] The server receives the query. The server receives the query text sent by the user from the request body. Input: Query text sent by the user. Output: Received text data.

[1258] Step 3:

[1259] The server preprocesses the text. The server uses regular expressions to remove unnecessary characters and spaces from the received text data and converts it into a format suitable for natural language processing. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted into "Product ABC is broken. Please tell me how to return it." Input: Received text data. Output: Preprocessed text data.

[1260] Step 4:

[1261] The server analyzes the data using an NLP engine. The server passes the preprocessed text data to a natural language processing (NLP) engine (e.g., SpaCy, BERT) to extract the query intent and entities (important keywords). For example, the intent "return" and entities "product ABC" and "break" are extracted. Input: Preprocessed text data. Output: Extracted intent and entities.

[1262] Step 5:

[1263] The server classifies the inquiry. Based on the parsed intent and entities, the server classifies the inquiry into a pre-defined category (e.g., "returns," "technical support," "product information"). For example, it is classified into the "returns" category. Input: Extracted intent and entities. Output: Classified inquiry category.

[1264] Step 6:

[1265] The server determines the appropriate response for the classified inquiry category. Specifically, it determines whether to generate an automatic response based on a template or escalate to a human agent. Input: Classified inquiry category. Output: Determined response (generating an automatic response or escalation).

[1266] Step 7:

[1267] The server generates an automatic response. The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. For example, it generates a response such as, "You can return product ABC by clicking the link below. Please see the return policy page for details." Input: The determined response. Output: The generated automatic response.

[1268] Step 8:

[1269] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry, and transfers complex inquiries or urgent complaints to a human representative. Input: Inquiry content. Output: Escalation decision.

[1270] Step 9:

[1271] The server returns an answer. The server returns the generated response to the user. This is done through the original communication channel through which the user made the query (chat box, email, etc.). Input: The generated automated response. Output: The response returned to the user.

[1272] Step 10:

[1273] The user receives the answer. The user checks the response sent from the server on their own device. For example, a message is displayed in the chat box or an email is received. Input: The response sent from the server. Output: The response checked on the user device.

[1274] (Application example 1)

[1275] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1276] Conventional systems for responding to inquiries at call centers and monitoring centers have made it difficult to respond quickly and appropriately, and responses are often delayed, especially in urgent situations. This has led to a decline in customer satisfaction and increased risk of serious problems. Furthermore, there was no system in place to efficiently process inquiries using real-time voice input.

[1277] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1278] In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the inquiry, means for classifying the inquiry based on the extracted intent and entity, means for generating an appropriate response to the classified inquiry, means for returning the generated response to the user, means for converting voice input into text data in real time, and means for determining urgency and performing automatic escalation. This enables prompt and appropriate processing of inquiries, and enables prompt responses, especially in emergencies.

[1279] The "means for receiving text data" is a device or program that has the function of allowing the server to receive a text-formatted inquiry sent from a user.

[1280] The "means for preprocessing received text data" refers to a device or program that executes a process to remove unnecessary characters and spaces from the received text data and to improve grammar and syntax.

[1281] "Means for analyzing preprocessed text data to extract query intent and entities" refers to a device or program that has the function of analyzing preprocessed text data using natural language processing (NLP) technology and extracting the main intent of a user's query and important keywords (entities).

[1282] The "means for classifying inquiries based on the extracted intent and entities" refers to a device or program that has the function of classifying the content of inquiries into pre-defined categories based on the intent and entities extracted through analysis.

[1283] The "means for generating appropriate responses to classified inquiries" refers to a device or program that has the function of automatically generating appropriate answers based on the classified inquiry content using templates and an FAQ database.

[1284] The "means for returning the generated response to the user" refers to a device or program that has the function of returning the automatically generated response to the user, and performs the return via a communication channel.

[1285] The "means for converting voice input into text data in real time" refers to a device or program that has the function of recognizing a user's voice input in real time and converting it into text data.

[1286] "Means for determining urgency and automatically escalating" refers to a device or program that has the function of automatically determining the urgency of an inquiry and, if necessary, transferring the inquiry to a human representative.

[1287] The system for implementing this invention is composed of three main entities: a server, a terminal, and a user. To understand the specific processing procedure, the operation of this system will be explained.

[1288] Server Features

[1289] The server has the following functions:

[1290] 1. Function to receive text data:

[1291] The server receives the inquiry sent by the user in text format. This function also includes a function to convert voice data into text data in real time.

[1292] 2. Text data preprocessing function:

[1293] Removes unnecessary characters and spaces from received text data and improves grammar and syntax, making it easier to analyze.

[1294] 3. Text data analysis function:

[1295] The preprocessed text data is analyzed using natural language processing (NLP) technology to extract the user's query intent and important keywords (entities). Specifically, open source NLP engines such as spaCy and TextBlob are used.

[1296] 4. Inquiry classification function:

[1297] Based on the extracted intent and entities, the inquiry is classified into pre-defined categories, and an appropriate response policy is then determined based on the pre-defined categories.

[1298] 5. Appropriate response generation:

[1299] For classified inquiries, an automatic response is generated using templates and an FAQ database. This automatically generated response is returned to the user in real time, allowing for a quick response.

[1300] 6. Urgency assessment and automatic escalation:

[1301] The system automatically determines the urgency of the inquiry and escalates it to a human representative if necessary, preventing serious problems.

[1302] Device Features

[1303] The devices used by security personnel and those on-site to receive inquiries are smartphones, smart glasses, head-mounted displays, etc. These devices are equipped with a voice input function, converting voice into text in real time and sending it to a server.

[1304] User Roles

[1305] When a problem occurs on-site, the user, as a security officer, sends a query to the server by voice input. For example, the following prompt sentences can be used:

[1306] "I've spotted a suspicious person. What should I do?"

[1307] "We just discovered a fire at our facility. What should we do?"

[1308] "I found a suspicious package. Do you need urgent action?"

[1309] Hardware and software used

[1310] This system uses the following hardware and software:

[1311] Hardware: Servers (with powerful CPUs and large amounts of memory), smartphones, smart glasses, head-mounted displays

[1312] Software: Python, spaCy (NLP engine), TextBlob (NLP engine), open source libraries for speech recognition (e.g., Google Speech-to-Text)

[1313] This enables immediate response on-site and automatic escalation in emergencies.

[1314] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1315] Step 1:

[1316] The user inputs voice data. For example, they might ask, "I spotted a suspicious person. What should I do?" This is sent to the server in real time using the voice input function of smart glasses or a head-mounted display. The input here is voice data, and the output is data converted into text format.

[1317] Step 2:

[1318] The server converts the received voice data into text data. An open source library for speech recognition (e.g., Google Speech-to-Text) is used to convert the voice data into text data. This process results in the voice data being converted into text data.

[1319] Step 3:

[1320] The server preprocesses the text data. It removes unnecessary characters and spaces from the received text data and improves grammar and syntax. This makes it easier to analyze. The input of this step is the converted text data, and the output is the preprocessed text data.

[1321] Step 4:

[1322] The server analyzes the preprocessed text data. Using a natural language processing (NLP) engine (e.g., spaCy), it analyzes the text data and extracts the user's query intent and important keywords (entities). The input is the preprocessed text data, and the output is the extracted intent and entities.

[1323] Step 5:

[1324] The server classifies queries based on intent and entities. Based on the extracted intent and entities, the query content is classified into predefined categories. The input is the extracted intent and entities, and the output is the classified query.

[1325] Step 6:

[1326] The server generates an appropriate response to the classified query. It uses templates and an FAQ database to automatically generate responses. For example, the response generated is "Contact the police immediately and monitor the scene from a safe location." The input is the classified query and the output is the generated response.

[1327] Step 7:

[1328] The server determines the urgency of the inquiry and automatically escalates it. If the inquiry is deemed urgent, it is escalated to a human agent. The input is the generated response and the inquiry itself, and the output is the decision to escalate or not.

[1329] Step 8:

[1330] The server sends the generated response back to the user. The server sends the response back to the user through the original communication channel (e.g., smart glasses or a head-mounted display). The input is the generated response and the output is the response sent back to the user.

[1331] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1332] The present invention is a system for improving the efficiency of call center inquiry operations, and mainly includes receiving text data, preprocessing, analyzing, classifying, generating responses, automatically escalating, returning responses, and an emotion engine that analyzes user emotions. Specific embodiments are described below.

[1333] System configuration

[1334] This system is implemented mainly by three entities: the server, the terminal, and the user. In addition, by incorporating an emotion engine for emotion analysis, more advanced inquiry response can be achieved.

[1335] System Operation

[1336] 1. User submits an inquiry

[1337] A user inputs and sends a text inquiry using a chat box or email form at a call center. For example, the user sends a text message such as, "Product ABC is broken. I would like to return it. Please tell me how to do this."

[1338] 2. The server receives the query

[1339] The server receives the query text sent by the user and takes it as the request body.

[1340] 3. The server preprocesses the text

[1341] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters and corrects grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[1342] 4. The server analyzes the text using an NLP engine

[1343] The server sends the preprocessed text to a natural language processing (NLP) engine for text analysis. The NLP engine extracts the user's intent and important information (entities) from the text. For example, "intent = return" and "entity = product ABC break" are extracted.

[1344] 5. The server analyzes emotions using an emotion engine

[1345] The server also sends the parsed text to an emotion engine, which analyzes the user's emotion (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords in the text. At this stage, the emotion may be analyzed as "anger."

[1346] 6. The server classifies the query

[1347] The server classifies the query into a predefined category based on the extracted intent and entities, as well as the analyzed sentiment, e.g., into the "returns" category.

[1348] 7. The server decides on a course of action

[1349] The server determines the appropriate response policy for the classified query. If the user's emotion is strong, it adjusts the tone of the response. For example, if the user is angry, it generates a calmer and more polite response.

[1350] 8. The server generates an automatic response

[1351] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During this generation process, placeholders in the template are replaced with actual data. For example, the server might generate a response such as, "You can return product ABC by clicking the link below. For more information, please see our return policy page."

[1352] 9. Server determines escalation

[1353] The server automatically determines whether a call needs to be escalated based on the content of the call and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, the server will hand over to a human agent. This decision is made when anger or strong dissatisfaction is detected.

[1354] 10. The server sends the response back to the user

[1355] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[1356] 11. The user receives a response

[1357] The user checks the response sent from the server on their own device, for example, a message is displayed in a chat box or an email is received.

[1358] Specific examples

[1359] For example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me how to do this." The system receives this query, preprocesses the text, and then analyzes it. Based on the analysis results, the system classifies it into the "return" category, and the emotion engine further analyzes that the user's emotion is "anger." The system generates a response using an appropriate template and returns it to the user with the following response: "You can return product ABC by clicking the link below. Please see the return policy page for details." The user can then accept this response and proceed with the return process.

[1360] This embodiment not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[1361] The processing flow will be explained below.

[1362] Step 1:

[1363] The user submits a query.

[1364] The user enters the inquiry in text format using the call center's chat box or email form and presses the send button. For example, the user might send an inquiry such as, "Product ABC is broken. I would like to return it. Please tell me what to do."

[1365] Step 2:

[1366] The server receives the query.

[1367] The server receives the text data sent by the user and takes it as the request body. This receiving operation is performed using a protocol such as an HTTP request or socket communication.

[1368] Step 3:

[1369] The server preprocesses the text.

[1370] The server performs preprocessing on the received text data. Specifically, it removes unnecessary spaces and special characters from the text and grammatically arranges it. For example, it converts "Product ABC is broken. I would like to return it. Please tell me what to do." into "Product ABC is broken. Please tell me how to return it."

[1371] Step 4:

[1372] The server analyzes the text using an NLP engine.

[1373] The server sends the preprocessed text data to a natural language processing (NLP) engine for text analysis. The NLP engine breaks the text down into tokens and extracts the user's intent and important information (entities). For example, "intent = return" and "entity = product ABC breaks" are extracted.

[1374] Step 5:

[1375] The server analyzes emotions using an emotion engine.

[1376] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions. The emotion engine identifies emotions such as positive, negative, and neutral from the text. For example, "anger" may be detected from the user's text.

[1377] Step 6:

[1378] The server classifies the query.

[1379] The server classifies the query into a predefined category based on the extracted intent and entities, and the analyzed sentiment, e.g., into the "returns" category.

[1380] Step 7:

[1381] The server decides the response policy.

[1382] The server determines the appropriate response policy for the classified inquiry. If the user's emotion is strong, the tone of the response is adjusted. For example, if the user is angry, the server generates a polite and calm response.

[1383] Step 8:

[1384] The server generates an automatic response.

[1385] The server uses response templates and an FAQ database to generate an appropriate answer to the user's inquiry. During the generation process, specific information is filled in the placeholders in the template. For example, it generates a response such as, "You can return product ABC by clicking the link below. For more information, please see the return policy page."

[1386] Step 9:

[1387] The server decides on the escalation.

[1388] The server automatically determines whether or not to escalate the inquiry based on the content of the inquiry and the analyzed sentiment. If the user's sentiment is particularly strong or the inquiry is complex, the server will hand it over to a human agent. This decision is also based on the results of sentiment analysis.

[1389] Step 10:

[1390] The server sends the response back to the user.

[1391] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[1392] Step 11:

[1393] The user receives the response.

[1394] The user checks the response sent from the server on his / her own terminal, for example, a message is displayed in a chat box or an email is received.

[1395] Example 2

[1396] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1397] Call centers are required to respond to many inquiries quickly and appropriately, but typical manual responses are inefficient and prone to user dissatisfaction. Responses that ignore the user's feelings can also lower customer satisfaction. Furthermore, if appropriate escalation procedures are not carried out, there is a risk that important issues will be overlooked. To solve these issues, a system is needed that can automatically analyze the content of inquiries and respond in a way that takes the user's feelings into account.

[1398] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intent and entity of the query, means for classifying the query based on the extracted intent and entity, means for classifying the query based on the analyzed emotion, means for determining a response policy based on the classified query, means for generating an appropriate response based on the response policy, and means for returning the generated response to the user. This enables an efficient response to the query that takes emotion into consideration.

[1399] The "means for receiving text data" is a device or function for receiving a text-formatted inquiry sent by a user.

[1400] The "means for preprocessing received text data" refers to a device or function for removing unnecessary spaces and special characters from the received text data and for adjusting the grammar and syntax.

[1401] The "means for analyzing preprocessed text data to extract query intent and entities" is a device or function for automatically analyzing and extracting user intent and important information from preprocessed text data.

[1402] The "means for classifying queries based on extracted intent and entities" refers to a device or function for classifying queries into predefined categories based on the intent and entities extracted by the analysis.

[1403] A "means for classifying queries based on analyzed emotions" is a device or function for classifying queries according to their emotional state based on emotion analysis.

[1404] The "means for determining a response policy based on the classified inquiry" is a device or function for determining an appropriate response policy based on the classified inquiry.

[1405] The "means for generating an appropriate response based on a response policy" is a device or function for automatically generating an appropriate response based on a determined response policy.

[1406] A "means for returning a generated response to a user" is a device or function for returning a generated response to a user through the original communication channel.

[1407] This invention is a system for improving the efficiency of call center inquiry operations, and mainly includes text data reception, preprocessing, analysis, classification, response generation, automatic escalation, response return, and an emotion engine that analyzes user emotions. This system is implemented mainly by three entities: a server, a terminal, and a user.

[1408] System configuration

[1409] 1. Receiving text data

[1410] The server receives the text data (inquiry content) sent by the user. The hardware used includes a regular web server, mail server, or chatbot server.

[1411] 2. Preprocessing of text data

[1412] The server preprocesses the received text data, including removing unnecessary spaces and special characters, grammar reformatting, etc. The software used includes natural language processing libraries (e.g., NLTK and SpaCy).

[1413] 3. Text Data Analysis

[1414] The preprocessed text data is then analyzed by an NLP engine, which uses services such as Google Cloud NLP or Amazon Comprehend to extract user intent and important information (entities).

[1415] 4. Emotion analysis

[1416] The server sends the parsed text data to an emotion engine, which analyzes the user's emotions using services such as IBM Watson Tone Analyzer, which identifies emotions (e.g., joy, anger, etc.) from the tone and keywords in the text.

[1417] 5. Categorizing inquiries

[1418] The server classifies queries into predefined categories based on the extracted intent and entities, and the analyzed sentiment, using rule-based classification algorithms and machine learning models.

[1419] 6. Deciding on a response policy

[1420] The server determines the appropriate response policy for each classified inquiry. In particular, it adjusts the tone of the response according to the user's emotions. If the user's emotions are strong, it generates a calmer and more polite response.

[1421] 7. Generate Auto-Responses

[1422] The server generates an appropriate answer using response templates and an FAQ database. During the generation process, it fills in the placeholders in the template with actual data. For example, it might generate an answer like, "You can return product ABC by following the link below. For more information, please see our returns policy page."

[1423] 8. Escalation Decision

[1424] The server automatically determines the need for escalation based on the query content and analyzed sentiment, specifically handing over to a human agent if the user's sentiment exceeds a certain threshold.

[1425] 9. Returning Responses

[1426] The server generates a response and sends it back to the user via the original communication channel (chat box or email).

[1427] Specific examples

[1428] As a concrete example, consider the case where a user sends a query such as, "Product ABC is broken. I would like to return it. Please tell me what to do." When the system receives this query, it preprocesses the text and analyzes it using an NLP engine. As a result of the analysis, the system classifies the query into the "return" category and further analyzes the user's emotion as "anger" using an emotion engine. Using an appropriate template, the system generates a response that reads, "You can return product ABC by clicking the link below. For details, please see the return policy page." and sends it back to the user. The user can then accept this response and proceed with the return process.

[1429] The present invention not only significantly improves the efficiency of call center inquiry operations, but also enables responses that take into consideration the feelings of users, which is expected to reduce response times and improve customer satisfaction.

[1430] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1431] Step 1:

[1432] The user submits an inquiry. The user enters the inquiry details in the chat box or email form and presses the send button. For example, a prompt message such as "Product ABC is broken. I would like to return it. Please tell me what to do." is sent.

[1433] Step 2:

[1434] The server receives the query. The server receives the user's input and takes it as the request body. The input is the user's query text, and the output is this text data. The query content and meta information (user ID, timestamp, etc.) are saved in a database.

[1435] Step 3:

[1436] The server preprocesses the text data. It removes unnecessary spaces and special characters from the received text data and arranges the grammatical structure. For example, "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is the received text, and the output is the preprocessed text data. A natural language processing library (for example, NLTK or SpaCy) is used for processing.

[1437] Step 4:

[1438] The server analyzes the preprocessed text and extracts the query intent and entities. The server then sends the preprocessed text data to an NLP engine, which extracts the intent and entities. The input is the preprocessed text, and the output is the extracted intent and entities. Natural language processing services such as Google Cloud NLP and Amazon Comprehend are used. The extracted results are "intent = return" and "entity = product ABC breaks."

[1439] Step 5:

[1440] The server analyzes emotions from the extracted data. The server sends the analysis results to an emotion engine to analyze the user's emotions. For example, IBM Watson Tone Analyzer is used. The input is text containing the extracted intent and entities, and the output is the emotion analysis result (e.g., "anger").

[1441] Step 6:

[1442] The server classifies the query. The server classifies the query into the appropriate category based on the extracted intent, entities, and sentiment. The input is the parsed text data, and the output is the query category (e.g., the "returns" category). The processing uses rule-based classification algorithms or machine learning models.

[1443] Step 7:

[1444] The server determines the response policy based on the classified inquiry. In particular, if the user's emotion is strong, the tone of the response is adjusted. The input is the inquiry category and the emotion analysis result, and the output is the response policy.

[1445] Step 8:

[1446] The server generates an automatic response. Based on the response policy, the server generates an appropriate answer using response templates and an FAQ database. The input is the response policy, and the output is the generated response text. For example, it generates the text "You can return product ABC by clicking the link below. Please see the return policy page for details."

[1447] Step 9:

[1448] The server decides whether to escalate. The server automatically determines the need for escalation based on the content of the inquiry and the analyzed sentiment. The input is the inquiry data and sentiment analysis results, and the output is the escalation decision. In particular, if the user's sentiment is strong or the problem is complex, the server will hand over to a human agent.

[1449] Step 10:

[1450] The server sends the generated response back to the user The server sends the generated response back to the user through the original communication channel (chat box or email) The input is the generated response text, the output is the response sent back to the user

[1451] Step 11:

[1452] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The output is the response text displayed on the user's screen.

[1453] (Application example 2)

[1454] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1455] Modern online shopping sites are receiving an increasing number of inquiries from users, requiring efficient and accurate responses. However, conventional systems struggle to receive and preprocess text data, extract intent and entities, analyze emotions, and automatically escalate inquiries, making it difficult to improve customer satisfaction. Furthermore, there is a lack of a way to generate appropriate responses based on user emotions, limiting the efficiency of inquiry response and customer satisfaction. Therefore, a solution to this issue is needed.

[1456] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving text data, means for preprocessing the received text data, means for analyzing the preprocessed text data to extract the intention and events of the inquiry, means for analyzing the user's emotions, means for classifying the inquiry based on the extracted intention and events and the analyzed emotions, means for generating an appropriate response to the classified inquiry, and means for returning the generated response to the user. This makes it possible to improve the efficiency of inquiry operations and provide an appropriate response according to the user's emotions.

[1457] "Text data" is information in the form of text that a user transmits over a communication channel.

[1458] "Preprocessing" is a process that removes unnecessary spaces and special characters from the received text data and corrects the grammar and syntax.

[1459] "Intention" refers to the request or purpose that a user wants to convey through text data.

[1460] An "event" is specific information or an object extracted from text data.

[1461] "Emotions" are psychological states contained in the user's text data, and include joy, anger, anxiety, and the like.

[1462] "Query classification" refers to dividing the query content into specific categories based on the analyzed intent, events, and sentiment.

[1463] "Response generation" is the process of generating appropriate responses or information to classified inquiries.

[1464] "Returning a response" is the process of retransmitting the generated response to the user's communication channel.

[1465] "Automatic escalation" refers to handing over an inquiry to a human agent as needed based on the content of the inquiry and analyzed sentiment.

[1466] A "template" is a pre-prepared response template that can be customized for a specific situation.

[1467] The "FAQ Data Collection" is a database that compiles frequently asked questions and their answers.

[1468] The present invention provides a system for improving the efficiency of inquiry operations on mail-order sites and enabling appropriate responses according to the user's emotions. Specific embodiments for carrying out the invention will be described below.

[1469] System configuration

[1470] The system has three main components: a server, a terminal, and a user. The server plays a central role in processing everything from receiving text data to generating responses. The terminal provides a communication channel with the user, and the user makes inquiries through this terminal.

[1471] Hardware and Software

[1472] The server includes the following main software modules:

[1473] 1. NLP service: A software module that performs natural language processing (NLP) and analyzes the user's query text to extract intent and events.

[1474] 2. Sentiment Analysis Engine: A software module for analyzing user emotions from text data and identifying emotions such as anger, anxiety, and joy.

[1475] 3. FAQ data collection: A database containing pre-prepared response templates for many user inquiries.

[1476] Data processing and calculation

[1477] 1. Preprocessing of text data: The server performs preprocessing by formatting the text data received from the user and removing unnecessary spaces and special characters. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it."

[1478] 2. NLP analysis: The preprocessed text data is sent to the NLP service to extract intent and events. The NLP service analyzes information such as "intent = return" and "event = product ABC breaks."

[1479] 3. Sentiment Analysis: The text data is sent to a sentiment analysis engine to identify the user's emotions. The sentiment analysis engine extracts emotions such as anger, anxiety, or joy from the text. At this stage, the emotion may be analyzed as "anger."

[1480] 4. Enquiry Classification: Based on the extracted intent, events, and sentiment, the inquiry is classified into a specific category, e.g., “return” category.

[1481] 5. Response Generation: An appropriate response is generated for the classified inquiry based on the FAQ data set. If the user is angry, the tone of the response is adjusted and the placeholders in the template are replaced with actual data. For example, a response might be generated that reads, "You can return product ABC by clicking the link below. Please see the return policy page for details."

[1482] 6. Automatic Escalation: Automatically escalate queries based on query content and analyzed sentiment. If the user's sentiment is strong or exceeds a certain threshold, the query will be escalated to a human agent.

[1483] 7. Return Response: The generated response is returned to the user via the original communication channel (e.g., chat box or email).

[1484] Specific examples

[1485] When a user sends a query such as "Product ABC is broken. I would like to return it. Please tell me what to do," the server receives this query and performs preprocessing. It then uses NLP analysis to extract "Intention = Return" and "Event = Product ABC breaks," and uses sentiment analysis to identify the user's emotion as "Anger." Using an appropriate template, it generates a response saying, "You can return Product ABC by clicking the link below. Please see the return policy page for details," and sends it back to the user.

[1486] Example prompts to input to the generative AI model

[1487] "A user submits a query: (Product ABC is broken. I would like to return it. What should I do?) Preprocess this text, perform NLP analysis and sentiment analysis, and generate an optimal response. Please also consider the emotional tone of the response."

[1488] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1489] Step 1:

[1490] The user sends an inquiry. The user inputs and sends a text-format inquiry via the terminal. An example of such input is, "Product ABC is broken. I would like to return it. Please tell me what to do." The input text data is sent to the server.

[1491] Step 2:

[1492] The server receives the query. The server receives the text data sent by the user and takes it as the request body. The input data is left as is, and the output is raw text data that is subject to preprocessing.

[1493] Step 3:

[1494] The server preprocesses the text. It removes unnecessary spaces and special characters from the input text data and adjusts the grammar and syntax. For example, the text "Product ABC is broken. I would like to return it. Please tell me what to do." is converted to "Product ABC is broken. Please tell me how to return it." The input is raw text data, and the output is preprocessed, clean text data.

[1495] Step 4:

[1496] The server analyzes the text using an NLP engine. The preprocessed text is sent to a natural language processing (NLP) engine, which analyzes the text. The NLP engine extracts the user's intent and important information (entities) from the text. From the input preprocessed text data, "Intent = Return" and "Event = Product ABC breaks" are extracted. The output is the analyzed intent and event.

[1497] Step 5:

[1498] The server analyzes emotions using an emotion engine. The analyzed text data is also sent to the emotion engine, which analyzes the user's emotions (e.g., joy, anger, anxiety, etc.). The emotion engine identifies emotions from the tone and keywords of the text. The input is the analyzed text data, and the output is the identified emotion. At this stage, the emotion may be analyzed as "anger."

[1499] Step 6:

[1500] The server classifies the query. Based on the extracted intent and events, as well as the analyzed sentiment, the query is classified into a pre-defined category. For example, it is classified into the "return" category. The input is the intent, event, and sentiment data, and the output is a specific category.

[1501] Step 7:

[1502] The server determines the appropriate response policy and generates a response. It determines the appropriate response policy for the classified inquiry. If the user is angry, it adjusts the tone of the response and generates an appropriate answer to the user's inquiry using response templates and FAQ data. For example, it generates a response that reads, "You can return product ABC using the link below. Please see the return policy page for details." The input is the category and emotion data, and the output is the response sent back to the user.

[1503] Step 8:

[1504] The server makes the decision on escalation. It automatically determines the need for escalation based on the content of the inquiry and the analyzed emotions. In particular, if the user's emotions exceed a certain threshold, it will hand over to a human agent. The input is emotion data, and the output is instructions or execution of escalation.

[1505] Step 9:

[1506] The server sends the response back to the user. The generated response is sent back to the user via the original communication channel (chat box or email). The input is the generated response data, and the output is the message sent to the user.

[1507] Step 10:

[1508] The user receives a response. The user sees the response sent from the server on their device. For example, a message appears in a chat box or an email is received. The input is the message sent from the server, and the output is the response displayed to the user.

[1509] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1510] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1511] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1512] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1513] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1514] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1515] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1516] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1517] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1518] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1519] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1520] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1521] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1522] 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.

[1523] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1524] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1525] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1526] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1527] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1528] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1529] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1530] The following is further disclosed regarding the above embodiment.

[1531] (Claim 1)

[1532] means for receiving text data;

[1533] means for preprocessing the received text data;

[1534] means for analyzing the preprocessed text data to extract query intent and entities;

[1535] a means for classifying the query based on the extracted intent and entities;

[1536] means for generating an appropriate response to the classified query;

[1537] means for returning the generated response to the user;

[1538] A system including:

[1539] (Claim 2)

[1540] 2. The system according to claim 1, further comprising means for automatically determining escalation based on the content of the inquiry and escalating the inquiry to a human representative as necessary.

[1541] (Claim 3)

[1542] 10. The system of claim 1, wherein the response sent back to the user is generated based on a template and an FAQ database.

[1543] "Example 1"

[1544] (Claim 1)

[1545] means for receiving text data;

[1546] means for preprocessing the received text data;

[1547] means for analyzing the preprocessed text data to extract query intent and entities;

[1548] a means for classifying the query based on the extracted intent and entities;

[1549] a means for determining an appropriate response to the classified inquiry;

[1550] means for generating an automated response based on the countermeasure;

[1551] means for returning the generated response to the user;

[1552] A system including:

[1553] (Claim 2)

[1554] 2. The system according to claim 1, further comprising means for automatically determining escalation based on the content of the inquiry and escalating the inquiry to a human representative as necessary.

[1555] (Claim 3)

[1556] 10. The system of claim 1, wherein the response sent back to the user is generated based on a template and an FAQ database.

[1557] "Application Example 1"

[1558] (Claim 1)

[1559] means for receiving text data;

[1560] means for preprocessing the received text data;

[1561] means for analyzing the preprocessed text data to extract query intent and entities;

[1562] a means for classifying the query based on the extracted intent and entities;

[1563] means for generating an appropriate response to the classified query;

[1564] means for returning the generated response to the user;

[1565] a means for converting voice input into text data in real time;

[1566] A means of determining urgency and automatic escalation;

[1567] A system including:

[1568] (Claim 2)

[1569] 2. The system according to claim 1, further comprising means for automatically determining escalation based on the content of the inquiry and escalating the inquiry to a human representative as necessary.

[1570] (Claim 3)

[1571] 10. The system of claim 1, wherein the response sent back to the user is generated based on a template and an FAQ database.

[1572] "Example 2: Combining Emotion Engines"

[1573] (Claim 1)

[1574] means for receiving text data;

[1575] means for preprocessing the received text data;

[1576] means for analyzing the preprocessed text data to extract query intent and entities;

[1577] a means for classifying the query based on the extracted intent and entities;

[1578] means for classifying the query based on the analyzed sentiment;

[1579] A means for determining a response policy based on the classified inquiries;

[1580] means for generating an appropriate response based on a response policy;

[1581] means for returning the generated response to the user;

[1582] A system including:

[1583] (Claim 2)

[1584] 10. The system of claim 1, further comprising means for automatically determining escalation based on the content of the inquiry and the analyzed sentiment, and escalating to a human agent as necessary.

[1585] (Claim 3)

[1586] 10. The system of claim 1, wherein the response sent back to the user is generated based on a template and an FAQ database.

[1587] "Application example 2 when combining emotion engines"

[1588] (Claim 1)

[1589] means for receiving text data;

[1590] means for preprocessing the received text data;

[1591] means for analyzing the preprocessed text data to extract query intent and events;

[1592] means for analyzing user emotions;

[1593] means for classifying queries based on the extracted intents and events and the analyzed sentiment;

[1594] means for generating an appropriate response to the classified query;

[1595] means for returning the generated response to the user;

[1596] A system including:

[1597] (Claim 2)

[1598] 10. The system of claim 1, further comprising means for automatically determining escalation based on the content of the inquiry and the analyzed sentiment, and escalating to a human agent as necessary.

[1599] (Claim 3)

[1600] 2. The system of claim 1, wherein the response sent back to the user is generated based on a template and a collection of FAQ data. [Explanation of symbols]

[1601] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for receiving text data; means for preprocessing the received text data; means for analyzing the preprocessed text data to extract query intent and entities; a means for classifying the query based on the extracted intent and entities; means for generating an appropriate response to the classified query; means for returning the generated response to the user; A system including:

2. 2. The system according to claim 1, further comprising means for automatically determining escalation depending on the content of an inquiry, and escalating the inquiry to a human representative as necessary.

3. 2. The system of claim 1, wherein the response sent back to the user is generated based on a template and an FAQ database.

Images