System

Abstract

A system is provided.SOLUTION: A system comprising: means for detecting an incoming call from a customer; means for collecting voice data of the customer in real time and converting the voice data into text data using a voice recognition technique; means for analyzing the text data using a natural language processing technique and identifying inquiry content of the customer; means for generating an appropriate answer based on an analysis result; and means for collecting a reaction of the customer again and performing learning for improving accuracy of the system based on the reaction.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] Call centers face the following challenges: training new employees is time-consuming and costly, there is a wide variation in the level of service provided by operators, high turnover makes it difficult to secure a stable workforce, foreign language support is difficult, and the requirement to provide support 24 hours a day, 365 days a year makes operator shift management complicated. To solve these challenges, an efficient and consistent automated response system is needed. [Means for solving the problem]

[0005] The present invention aims to solve the above problems by the following means.

[0006] The system includes a means for detecting incoming calls from customers, a means for collecting customer voice data in real time and converting it into text data using voice recognition technology, a means for analyzing the text data using natural language processing technology to identify the content of the customer's inquiry, a means for generating an appropriate response based on the analysis results, a means for providing the generated response to the customer by voice, and a means for collecting customer responses again and learning to improve the accuracy of the system based on the results.Furthermore, the system is equipped with a means for analyzing customer emotions using voice analysis technology and adjusting responses based on the emotions, and is equipped with voice recognition technology and natural language processing technology that can handle multiple languages ​​and a means for providing automatic responses 24 hours a day, 365 days a year, thereby reducing the effort required for training new employees, eliminating differences in the response levels among operators, and enabling multilingual and 24-hour support.

[0007] A "customer" is a user who makes an inquiry to the call center.

[0008] "Incoming calls" refers to calls that come into a call center.

[0009] "Voice data" refers to data that records the voice spoken by a customer over the phone.

[0010] "Speech recognition technology" is a technology that analyzes voice data and converts it into corresponding text data.

[0011] "Text data" refers to data obtained by converting voice data into character information.

[0012] "Natural language processing technology" is a technology that analyzes text data and understands its meaning and intent.

[0013] The "analysis results" are the meaning and intent of the text data identified using natural language processing technology.

[0014] An "answer" is an appropriate response to a customer's inquiry.

[0015] "Speech synthesis technology" is a technology that converts generated text data into voice data.

[0016] "Customer response" refers to the response or behavior a customer exhibits in response to the answers provided.

[0017] "System accuracy" refers to the system's ability to accurately understand the query and provide an appropriate answer.

[0018] "Learning means" are means for analyzing customer responses and new data to improve knowledge and capabilities for the effective operation of the system.

[0019] "Emotion analysis technology" is a technology that reads emotions from customers' voices and text.

[0020] "Multilingual support" refers to the ability to process and respond to voice and text data in multiple languages.

[0021] "24 / 7 automatic response" is a system function that automatically responds to customer inquiries all day, every day of the year. [Brief explanation of the drawings]

[0022] [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

[0023] 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.

[0024] First, the terms used in the following description will be explained.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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."

[0030] [First embodiment]

[0031] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0032] 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.

[0033] 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).

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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.

[0038] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0039] 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.

[0040] 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.

[0041] 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.

[0042] 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."

[0043] This invention is a system for automating and streamlining customer service in a call center. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[0044] Overall system overview

[0045] 1. Detect incoming calls from customers

[0046] Server: Connects to the call center system and detects incoming calls in real time.

[0047] Device: Receives a call and starts recording.

[0048] User: Make a call and ask a question.

[0049] 2. Collection and analysis of audio data

[0050] Device: Records the customer's voice and sends the audio data to the server.

[0051] Server: Uses voice recognition technology to convert voice data into text data.

[0052] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[0053] 3. Identifying the inquiry and generating an answer

[0054] Server: Uses natural language processing technology to identify customer inquiries from text data.

[0055] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[0056] 4. Providing answers

[0057] Server: Uses speech synthesis technology to convert text data into voice data.

[0058] Terminal: Plays back the audio data and provides the answer to the customer.

[0059] 5. Feedback and learning

[0060] Terminal: Collects customer responses again and sends them to the server.

[0061] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[0062] As a specific processing example, consider the following scenario.

[0063] Example: Inquiry about product malfunction

[0064] 1. Receiving an incoming call

[0065] User: "My product isn't working and I'd like to check."

[0066] Server: Detects incoming calls and collects customer voice.

[0067] 2. Converting audio data into text

[0068] Terminal: Records what the customer says and sends the audio data to the server.

[0069] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[0070] 3. Query Analysis

[0071] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[0072] Server: Determines that the customer has a problem and searches the product manual for a solution.

[0073] 4. Answer Generation

[0074] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0075] Terminal: Speaks the generated answer to the customer.

[0076] 5. Analyzing and learning from customer responses

[0077] User: "The model number is ABC123. The serial number is XYZ890."

[0078] Terminal: Records customer responses and sends the audio data to the server.

[0079] Server: Analyzes the new voice data and retrieves product details.

[0080] Server: We will update our response with new information and provide further details as needed.

[0081] In this way, the server and terminals work together to create an automated response system that improves the efficiency of call center operations. This system is also multilingual, can operate 24 hours a day, 365 days a year, and is highly scalable.

[0082] The processing flow will be explained below.

[0083] Step 1:

[0084] User: Calls and says, "My product isn't working and I'd like to check it out."

[0085] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[0086] Terminal: Receives incoming calls and begins recording customer conversations.

[0087] Step 2:

[0088] Device: Sends recorded audio data to the server in real time.

[0089] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[0090] Server: Obtain the text data "My product is not working, so I would like to check it."

[0091] Step 3:

[0092] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[0093] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[0094] Server: Analyzes the customer's tone and language and determines the emotion is "confused."

[0095] Step 4:

[0096] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[0097] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0098] Step 5:

[0099] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0100] Terminal: Plays the generated voice data and provides the answer to the customer.

[0101] Step 6:

[0102] User: "The model number is ABC123. The serial number is XYZ890."

[0103] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0104] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0105] Step 7:

[0106] Server: Extract the product model number and serial number from the new text data.

[0107] Server: Uses the extracted information to retrieve product details from the database.

[0108] Server: Provides detailed fault diagnosis and updated response as needed.

[0109] Step 8:

[0110] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[0111] Terminal: Displays customer inquiry history and response content, and can be used to respond to future inquiries.

[0112] In this way, the server, terminal, and user work together to realize an automated voice response system that provides quick and accurate responses.

[0113] Example 1

[0114] 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."

[0115] In conventional call center operations, a large number of staff members have to deal with customers, requiring a great deal of effort and time, resulting in problems of inefficiency and human error. Furthermore, responding to customer inquiries in a timely and appropriate manner requires advanced skills and specialized knowledge, creating challenges in human resource development. To solve these problems and improve customer satisfaction, a more efficient and highly accurate automated response system is needed.

[0116] 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.

[0117] In this invention, the server includes a means for detecting incoming calls from customers, a means for collecting customer voice data in real time and converting it into text data using speech recognition technology, and a means for analyzing the text data using natural language processing technology to identify the content of the customer's inquiry. This enables an automated response that responds to customer inquiries in a timely and appropriate manner. Furthermore, by providing a means for recollecting customer responses and improving the accuracy of the system based on them, highly accurate responses based on the latest information are always possible. Furthermore, by providing a means for converting text data into speech data using speech synthesis technology and providing it to the customer audibly, more natural dialogue is realized.

[0118] "Means for detecting incoming calls" refers to a device or function that connects to a telephone system or network and recognizes incoming calls from customers in real time.

[0119] "Voice data" refers to information that has been recorded in digital form from a customer's voice.

[0120] "Speech recognition technology" is a technology for converting voice data into text data.

[0121] "Text data" is information in text format generated from voice data using voice recognition technology.

[0122] "Natural language processing technology" is a technology that analyzes text data, understands meaning and emotions, and generates appropriate responses.

[0123] "Means for identifying inquiry content" refers to the function of using natural language processing technology to identify specific customer requests or problems from text data.

[0124] An "answer generation means" is a device or function that generates an appropriate response or solution based on the identified inquiry.

[0125] "Speech synthesis technology" is a technology that converts text data into voice data.

[0126] The "means for providing to the customer" is a device or function for playing back the generated voice data to the customer.

[0127] A "means for recollecting customer responses" is a device or function that records new voice data that the customer gives in response to the answers provided and transmits it to the system.

[0128] "Learning means to improve system accuracy" refers to a function that analyzes collected customer response data and updates the model using machine learning technology to improve system performance.

[0129] A "means for establishing a call session" is a device or function that establishes a communication channel between a customer and a system, enabling two-way communication of voice data.

[0130] "Means for searching for relevant information from a database" refers to a search engine or function for quickly finding data related to the inquiry from stored information.

[0131] This invention is a system for automating and streamlining customer service in call center operations. The system is broadly composed of three elements: a server, a terminal, and a user. How each element specifically operates is described in detail below.

[0132] Hardware and software used

[0133] Server: A high-performance server is used to perform speech recognition, natural language processing, speech synthesis, data analysis, etc. Specifically, technologies such as Google Cloud Speech-to-Text, spaCy, Hugging Face's BERT model, ElasticSearch, Amazon Polly, and TensorFlow are used.

[0134] Terminal: A terminal equipped with a SIP (Session Initiation Protocol) client has the function of recording the customer's voice and sending it to the server.

[0135] User: The customer making the call.

[0136] Details of system processing

[0137] 1. Detecting incoming calls from customers and establishing call sessions

[0138] Server: Connects to the VoIP server to detect incoming calls from customers in real time and notify the terminal.

[0139] Device: Upon receiving the notification, it establishes a call session with the customer and begins recording audio.

[0140] User: Make a call and start an inquiry.

[0141] 2. Collecting and converting audio data into text

[0142] Terminal: Continuously streams the customer's voice to the server in real time.

[0143] Server: Sends the received audio data to the Google Cloud Speech-to-Text API to generate text data.

[0144] 3. Analyzing text data and identifying inquiry content

[0145] Server: Analyze the generated text data using the spaCy library to extract key phrases and sentiment information.

[0146] Server: Leverages Hugging Face's BERT model to identify specific queries based on key phrases.

[0147] 4. Answer Generation and Search

[0148] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals.

[0149] Server: Generates appropriate answers for customers based on the retrieved information.

[0150] 5. Conversion into audio data and provision

[0151] Server: The generated text data is converted into voice data using Amazon Polly.

[0152] Terminal: Plays the converted voice data and provides the answer to the customer.

[0153] 6. Collect and learn from customer responses

[0154] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server.

[0155] Server: Analyzes new audio data and uses TensorFlow to update the model to improve the accuracy of the system.

[0156] Specific examples

[0157] Below is a scenario for an inquiry about a product failure.

[0158] scenario

[0159] 1. Detecting incoming calls and establishing a call session

[0160] User: "My product isn't working, so I'd like to check it out."

[0161] Server: The VoIP server detects the incoming call and notifies the terminal.

[0162] Device: Establish a call session and start recording.

[0163] 2. Collecting and converting audio data into text

[0164] Terminal: Records what the customer says and sends the audio data to the server.

[0165] Server: Use Google Cloud Speech-to-Text to generate text data such as "My product isn't working, so I'd like to check it."

[0166] 3. Parsing the query and generating an answer

[0167] Server: Uses the spaCy library to extract the key phrase "product malfunction" from text data, and then uses the BERT model to identify the specific content.

[0168] Server: Use ElasticSearch to search the product manual for solutions related to the product failure. Based on the search results, generate a response such as "Please let us know the product model number and serial number so we can look into it."

[0169] 4. Conversion into audio data and provision

[0170] Server: Convert the response text into speech data using Amazon Polly.

[0171] Terminal: Plays back the audio data and provides the answer to the customer.

[0172] 5. Analyzing and learning from customer responses

[0173] User: "The model number is ABC123. The serial number is XYZ890."

[0174] Device: Record any additional comments and send them back to the server.

[0175] Server: Analyzes new audio data and retrieves product details, while simultaneously updating the model in TensorFlow to improve the system's accuracy.

[0176] Prompt Sentence Examples

[0177] "Generate a natural response for the customer who asks for the product's model number and serial number."

[0178] "When a customer calls to inquire about a malfunctioning product, generate the best response."

[0179] In this way, the server and terminals work together to create a highly efficient and accurate automatic response system for customer inquiries.

[0180] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0181] Step 1: Detecting an incoming call from a customer and establishing a call session

[0182] Server: Connects to the VoIP server and detects incoming calls from customers in real time. The input is the call signal from the customer, and the output is the call detection information.

[0183] Terminal: Upon receiving a notification from the server, it establishes a call session with the customer using a SIP client and starts recording the audio. The input is the notification from the server, and the output is the establishment of a call session and the start of recording.

[0184] User: Makes a call and initiates a query. The input is the user making the call, and the output is the initiation of the call.

[0185] Step 2: Collect and send audio data

[0186] Terminal: Collects voice data in real time during a call with a customer and streams the voice data to a server. The input is the voice signal, and the output is the voice data sent to the server.

[0187] Server: Sends the received voice data to the Google Cloud Speech-to-Text API and converts the voice data to text data. The input is voice data from the device, and the output is text data.

[0188] Step 3: Analyzing text data and identifying inquiry content

[0189] Server: The generated text data is analyzed using the spaCy library to extract key phrases and analyze customer sentiment. The input is text data, and the output is key phrases and sentiment information.

[0190] Server: Uses Hugging Face's BERT model to identify the specific query from the extracted key phrases. The input is the key phrase and the output is the query.

[0191] Step 4: Generate answers and find related information

[0192] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals. The input is the query and the output is the relevant information.

[0193] Server: Generates appropriate answers for customers based on the retrieved information. The input is the relevant information and the output is the answer text.

[0194] Step 5: Converting the response into audio data and providing it

[0195] Server: Use Amazon Polly to convert the generated answer text data into speech data. The input is the answer text and the output is speech data.

[0196] Terminal: Plays back the voice data received from the server and provides answers to the customer. The input is the voice data, and the output is the played voice.

[0197] Step 6: Collect and send customer responses

[0198] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server. The input is the customer's speech and the output is the recorded audio data.

[0199] User: Checks the answer and provides a follow-up question or additional information. The input is the answer provided and the output is the additional utterance.

[0200] Step 7: Analyze the data and train the system

[0201] Server: Analyzes new audio data received and uses machine learning algorithms to update the model to improve the accuracy of the system. The input is the new audio data and the output is the updated machine learning model.

[0202] (Application example 1)

[0203] 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."

[0204] Improving the quality and efficiency of customer service is a key challenge in modern brick-and-mortar stores. Providing prompt and appropriate responses to customers is particularly difficult during busy times, potentially leading to lower customer satisfaction. Another issue is the need for a large number of staff, which increases labor costs. Furthermore, providing accurate information in response to customer inquiries relies heavily on the knowledge and experience of staff. Given these circumstances, there is a need for systems that use voice recognition and natural language processing technologies to automate customer service in real time via smart devices, providing efficient, high-quality service.

[0205] 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.

[0206] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning based on the collected responses to improve the accuracy of the system, and means for collecting customer inquiries in real time via a smart device and generating and providing an appropriate answer, thereby enabling more efficient customer service and the provision of high-quality services.

[0207] A "server" is a central computer system that detects incoming calls from customers, collects voice data, analyzes it, and generates responses.

[0208] "Smart devices" are devices that have advanced functionality, primarily smart glasses and tablets, and interact through voice recognition and speech synthesis.

[0209] "Means for detecting incoming calls" refers to technology that recognizes when a call or voice message is received from a customer and uses that as a trigger to activate the system.

[0210] "Speech recognition technology" is a technology that converts voice data into text data.

[0211] "Natural language processing technology" is a technology that analyzes text data and understands customer intentions and emotions.

[0212] "Means for generating appropriate answers based on the analysis results" refers to technology that generates answers to customer questions based on the analyzed text data.

[0213] The "means for providing the generated answer to the customer by voice" is a technology for converting the generated text answer into voice data and transmitting it to the customer.

[0214] "Means of collecting customer responses again and learning based on that to improve the accuracy of the system" refers to a technology that collects customer responses and uses machine learning to improve the system's responses.

[0215] The system that realizes this application example is composed of the following programs and a combination of hardware and software.

[0216] Program Generation

[0217] The server contains programs that do the following:

[0218] 1. Detect incoming calls from customers.

[0219] 2. Collect customer voice data in real time and convert it into text data using voice recognition technology.

[0220] 3. Analyze the text data using natural language processing technology to identify the customer's inquiry.

[0221] 4. Generate appropriate answers based on the analysis results.

[0222] 5. Provide generated answers to customers via voice.

[0223] 6. Customer responses are collected again and the system learns from them to improve its accuracy.

[0224] 7. Collect customer inquiries in real time via smart devices and generate and provide appropriate answers.

[0225] Processing Description

[0226] Hardware / Software used

[0227] Hardware: Smart glasses (microphone, camera, speaker), server (high-performance processing unit)

[0228] Software: Python programming language, speech_recognition module, gTTS (Google Text-to-Speech) module, openai API

[0229] Data processing and calculation

[0230] 1. The server uses a speech recognition module to convert the customer's speech into text data in real time. Specifically, the server acquires the speech data using the speech_recognition module and outputs it as text data.

[0231] 2. The server uses natural language processing technology to analyze the converted text data. It uses the OpenAI API to analyze the text data and generate an appropriate answer. It uses the prompt sentence to input the data into the generative AI model to obtain the answer.

[0232] 3. The server converts the generated text response into voice data using the gTTS module, which allows the generated response to be output as voice and played through the smart glasses' speaker.

[0233] Adding specific examples

[0234] For example, consider the following scenario:

[0235] Customer: "Is this product waterproof?"

[0236] The camera in the smart glasses recognizes the customer and the microphone records their questions.

[0237] The server converts the voice data into text, such as "Is this product waterproof?"

[0238] The prompt sentence "Is this product waterproof?" was input into the generative AI model via the OpenAI API.

[0239] The server generates the response: "Yes, this product is IP67 waterproof."

[0240] gTTS converts text to speech and plays it through the smart glasses' speakers.

[0241] Example prompt:

[0242] Product feature questions: "Is this product waterproof?"

[0243] A: Yes, this product is IP67 rated waterproof.

[0244] In this way, it becomes possible to improve the efficiency of customer service and provide high-quality services in physical stores.

[0245] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0246] Step 1:

[0247] The server detects when a customer enters the store using the camera and microphone of the smart glasses. When the customer starts speaking, the smart glasses collect voice data. The input here is the customer's voice utterance, and the output is the recorded voice data.

[0248] Step 2:

[0249] The server converts the collected voice data into text data using the speech_recognition module. Specifically, it analyzes the voice data using speech recognition technology and generates text data such as "Is this product waterproof?". The input here is the recorded voice data, and the output is text data.

[0250] Step 3:

[0251] The server uses the OpenAI API to analyze the generated text data. A prompt sentence is constructed and input into the generative AI model to generate an appropriate answer. The input here is the text data and the prompt sentence, and the output is the generated answer text. An example of a prompt sentence is "Question about product features: Is this product waterproof?"

[0252] Step 4:

[0253] The server converts the generated answer text into voice data using the gTTS module. Specifically, it outputs the answer text as voice data using speech synthesis technology. The input here is the generated answer text, and the output is voice data.

[0254] Step 5:

[0255] The smart glasses at the terminal play the generated voice data and provide a response to the customer, where the input is the voice data and the output is the voice information provided to the customer.

[0256] Step 6:

[0257] The user responds to the server's answers and asks additional questions or comments as needed. Their responses are again recorded by the smart glasses' microphone. The input here is the customer's additional utterances, and the output is again the recorded voice data.

[0258] Step 7:

[0259] The server analyzes the re-collected voice data and performs learning to improve the accuracy of the system. Specifically, the customer response data and the analysis results are fed back to the machine learning model to improve future response accuracy. The input here is the re-collected voice data and the analysis results, and the output is a learning model with improved system accuracy.

[0260] 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.

[0261] This invention is a system for automating and streamlining customer service in call centers. In particular, by combining it with an emotion engine that recognizes customer emotions and adjusts responses based on those emotions, customer satisfaction can be further improved. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[0262] Overall system overview

[0263] 1. Detect incoming calls from customers

[0264] Server: Connects to the call center system and detects incoming calls in real time.

[0265] Device: Receives a call and starts recording.

[0266] User: Make a call and ask a question.

[0267] 2. Collection and analysis of audio data

[0268] Device: Records the customer's voice and sends the audio data to the server.

[0269] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[0270] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[0271] 3. Emotion analysis using an emotion engine

[0272] Server: Analyzes customer emotions and obtains emotional data using the emotion engine.

[0273] Server: Makes decisions to generate the best answer based on customer sentiment.

[0274] 4. Identifying the inquiry and generating an answer

[0275] Server: Uses a natural language processing (NLP) engine to identify customer inquiries from text data.

[0276] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[0277] Server: Dynamically adjust responses based on customer sentiment data.

[0278] 5. Providing answers

[0279] Server: Uses speech synthesis technology to convert text data into voice data.

[0280] Terminal: Plays the generated voice data and provides the answer to the customer.

[0281] 6. Feedback and Learning

[0282] Terminal: Collects customer responses again and sends them to the server.

[0283] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[0284] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[0285] As a specific processing example, consider the following scenario.

[0286] Example: Inquiry about product malfunction

[0287] 1. Receiving an incoming call

[0288] User: "My product isn't working and I'd like to check."

[0289] Server: Detects incoming calls and collects customer voice.

[0290] 2. Converting audio data into text

[0291] Terminal: Records what the customer says and sends the audio data to the server.

[0292] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[0293] 3. Emotion analysis

[0294] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions.

[0295] Server: Determines that the customer is in trouble and stores the emotion data.

[0296] 4. Query Analysis

[0297] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[0298] Server: Considers customer sentiment and searches for solutions in product manuals.

[0299] 5. Answer Generation

[0300] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0301] Servers: Use polite language that takes into account the customer's confused feelings.

[0302] 6. Providing answers

[0303] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0304] Terminal: Plays the generated voice data and provides the answer to the customer.

[0305] 7. Analyzing and learning from customer responses

[0306] User: "The model number is ABC123. The serial number is XYZ890."

[0307] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0308] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0309] Server: Extract the product model number and serial number from the new text data.

[0310] Server: Uses the extracted information to retrieve product details from the database.

[0311] Server: Provides detailed fault diagnosis and updated response as needed.

[0312] 8. Feedback and Data Storage

[0313] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[0314] Server: Emotional data is also stored and used for future improvements.

[0315] In this way, the server, terminals, and users work together to realize an automated voice response system that provides prompt and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[0316] The processing flow will be explained below.

[0317] Step 1:

[0318] User: Calls and says, "My product isn't working and I'd like to check it out."

[0319] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[0320] Terminal: Receives incoming calls and begins recording customer conversations.

[0321] Step 2:

[0322] Device: Sends recorded audio data to the server in real time.

[0323] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[0324] Server: Obtain the text data "My product is not working, so I would like to check it."

[0325] Step 3:

[0326] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[0327] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[0328] Server: Based on the text data and tone of voice, determine that the customer's emotion is "confused."

[0329] Step 4:

[0330] Server: Manages customer emotion data using the emotion engine.

[0331] Server: Makes decisions to generate the best answer based on the customer's confused feelings.

[0332] Step 5:

[0333] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[0334] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0335] Step 6:

[0336] Server: Dynamically adjusts responses based on customer sentiment determined by the sentiment engine.

[0337] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0338] Terminal: Plays the generated voice data and provides the answer to the customer.

[0339] Step 7:

[0340] User: "The model number is ABC123. The serial number is XYZ890."

[0341] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0342] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0343] Step 8:

[0344] Server: Extract the product model number and serial number from the new text data.

[0345] Server: Uses the extracted information to retrieve product details from the database.

[0346] Server: Provides detailed fault diagnosis and updated response as needed.

[0347] Step 9:

[0348] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[0349] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[0350] In this way, the server, terminals, and users work together to realize an automated voice response system that provides fast and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[0351] Example 2

[0352] 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."

[0353] Customer support at call centers relies heavily on human resources, and while they require fast and accurate responses, it is difficult to fully understand customer emotions and provide appropriate responses. In particular, a lack of real-time responses that reflect customer emotions can lead to a decline in customer satisfaction. Furthermore, existing systems are difficult to handle, as they also require multilingual support and 24 / 7 service provision.

[0354] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0355] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning to improve the accuracy of the system based on the analysis, means for analyzing customer emotions and dynamically adjusting answers based on the results, and a computer terminal and database with which the above means are linked via a communications network. This enables quick and accurate automatic responses based on an understanding of customer emotions and continuous improvement of the system.

[0356] "Means for detecting incoming calls from customers" refers to a system that detects incoming calls from customers in real time and notifies other systems of that information.

[0357] "Means of collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to a system that records the voices transmitted by customers and converts that data into text information using voice recognition technology.

[0358] "Means for analyzing text data using natural language processing technology and identifying the content of the customer's inquiry" refers to a system that analyzes text information using natural language processing technology and identifies the content of the customer's intended inquiry.

[0359] "Means for generating an appropriate response based on the analysis results" refers to a mechanism that automatically creates the most appropriate and accurate response based on the analyzed inquiry content.

[0360] The "means of providing the generated answer to the customer by voice" is a mechanism that converts the automatically generated answer into voice data using voice synthesis technology and provides it to the customer.

[0361] "Means of re-collecting customer responses and learning based on them to improve the accuracy of the system" refers to a mechanism that re-records responses and reactions from customers and uses that data to run a machine learning process to improve the performance of the system.

[0362] "Means for analyzing customer emotions and dynamically adjusting responses based on the results" refers to a system that analyzes emotions from customer voice and text data and dynamically changes the content of responses based on the results of that analysis.

[0363] "Computer terminals and databases linked via a communications network" refers to an infrastructure that enables different computer terminals and databases to share information and operate in cooperation with each other using the Internet or other communications networks.

[0364] This invention is a system for automating and streamlining customer service in a call center. This system is composed of three main elements: a server, a terminal, and a user, each of which operates in a specific way.

[0365] Overall system overview

[0366] 1. Detect incoming calls from customers

[0367] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls to specific numbers.

[0368] Terminal: Receives incoming calls and prepares to record the voice of the customer. Activates the recording function the moment it receives notification of an incoming call.

[0369] User: Calls to inquire.

[0370] 2. Collection and analysis of audio data

[0371] Device: Records the customer's voice and sends it to the server as audio data. A highly sensitive microphone is used to record the customer's speech with clear sound quality.

[0372] Server: Passes the received voice data to a voice recognition API (e.g., Google Speech-to-Text) and converts it into text data. Temporarily stores the voice data and sends a request to the voice recognition API.

[0373] Server: Analyzes customer sentiment and key phrases from text data and extracts important information.

[0374] 3. Emotion analysis using an emotion engine

[0375] Server: Analyzes customer emotions using an emotion engine (e.g., IBM Watson Tone Analyzer) and obtains emotion data. Inputs the text data and tone of voice data into the emotion engine.

[0376] Server: Determines the best response based on customer sentiment. Uses sentiment analysis results as a basis for customizing responses.

[0377] 4. Identifying the inquiry and generating an answer

[0378] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer's inquiry from the text data. Sends the text data to the NLP engine, which identifies the inquiry as "product malfunction."

[0379] Server: Searches product manuals and internal databases to generate appropriate answers. Automatically searches databases for relevant information and generates answers.

[0380] Server: Dynamically adjusts responses based on customer sentiment data. Uses polite language based on sentiment data.

[0381] 5. Providing answers

[0382] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert text data into speech data. Inputs the response text into the speech synthesis API to generate speech data.

[0383] Terminal: Plays the generated voice data and provides the answer to the customer. Receives the generated voice data and plays it back for the customer to hear.

[0384] 6. Feedback and Learning

[0385] Terminal: The customer's responses are collected again and sent to the server as audio data. Any additional answers or responses from the customer are also recorded and sent to the server.

[0386] Server: Analyzes customer responses and learns to improve the accuracy of the entire system. Analyzes new data and updates the system's machine learning algorithms.

[0387] Server: Stores emotion data based on the analysis results and uses it to improve future customer service responses. The updated database is used to respond to future customer inquiries.

[0388] Specific examples

[0389] Inquiries about product malfunctions

[0390] User: "My product isn't working and I'd like to check."

[0391] Server: Detects incoming calls and collects customer voice.

[0392] Terminal: Records what the customer says and sends the audio data to the server.

[0393] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[0394] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions. Detects confused emotions and stores the emotion data.

[0395] Server: Using natural language processing technology, identify the inquiry content as "product malfunction."

[0396] Server: Considers the customer's feelings and searches for a solution in the product manual. Generates a response such as "Please let us know your product model number and serial number so we can look into it."

[0397] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0398] Terminal: Plays the generated voice data and provides the answer to the customer.

[0399] User: Responds, "The model number is ABC123. The serial number is XYZ890."

[0400] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0401] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0402] Server: Extract the product model number and serial number from the new text data. Use the extracted information to retrieve detailed product information from the database.

[0403] Server: Provides detailed fault diagnosis and updated response as needed.

[0404] Example prompt sentence:

[0405] The system analyzes a user's statement, "My product isn't working, so I'd like to check it," and uses a sentiment analysis engine to identify the customer's emotions and generate an answer.

[0406] This allows the system to provide efficient and accurate automated responses while taking into account customer emotions.

[0407] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0408] Step 1:

[0409] Detecting incoming calls from customers

[0410] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls from customers.

[0411] Input: Incoming call signal from customer

[0412] Output: Incoming call notification signal

[0413] Specific operations: Monitors the phone number and receives an incoming call signal, then notifies the device of the incoming call signal.

[0414] Step 2:

[0415] Collecting audio data and starting recording

[0416] Terminal: Receives an incoming call and activates the recording function, recording what the customer says and preparing to collect voice data with clear sound quality using a high-sensitivity microphone.

[0417] Input: Incoming call notification signal

[0418] Output: Start recording

[0419] Specific behavior: When an incoming call notification is received, the recording function is activated and the customer's voice begins to be recorded.

[0420] Step 3:

[0421] Sending audio data to the server

[0422] Terminal: Sends recorded audio data to the server in real time.

[0423] Input: Recorded audio data

[0424] Output: Send audio data to the server

[0425] Specific operation: The recorded audio data is compressed and sent to a server via internet communication.

[0426] Step 4:

[0427] Converting audio data to text

[0428] Server: Passes the voice data to a speech recognition API (e.g., Google Speech-to-Text) and converts it to text data. Temporarily stores the voice data and sends a request to the speech recognition API.

[0429] Input: Audio data sent to the server

[0430] Output: Text data generated by speech recognition

[0431] Specific operation: Sends audio data to the API and retrieves the returned text data.

[0432] Step 5:

[0433] Sentiment analysis of text data

[0434] Server: Using an emotion engine (e.g., IBM Watson Tone Analyzer), analyze customer emotions based on text data and voice tone.

[0435] Input: Text data and voice tone data from speech recognition

[0436] Output: Emotion data

[0437] Specific operation: Text data and voice tone are input into the emotion engine to obtain the emotion analysis results.

[0438] Step 6:

[0439] Identifying the content of the inquiry

[0440] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer inquiry from the text data.

[0441] Input: Text data

[0442] Output: Identified inquiry

[0443] Specific behavior: Analyze what the customer says and identify the specific inquiry.

[0444] Step 7:

[0445] Generating the right answers

[0446] Server: Generates appropriate answers based on customer inquiry content and sentiment data. Searches product manuals and internal databases to generate answers.

[0447] Input: Identified query content and sentiment data

[0448] Output: Generated answer text

[0449] What it does: Searches a database, extracts the relevant information, and creates an answer.

[0450] Step 8:

[0451] Converting responses into audio data

[0452] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert the generated answer text into voice data.

[0453] Input: Generated answer text

[0454] Output: Audio data

[0455] Specific operation: The answer text is input into the speech synthesis API to generate voice data.

[0456] Step 9:

[0457] Providing answers

[0458] Terminal: Plays the generated voice data and provides the answer to the customer.

[0459] Input: Audio data

[0460] Output: A spoken response to the customer

[0461] Specific behavior: Plays back the audio data and provides the appropriate answer to the customer.

[0462] Step 10:

[0463] Collecting and sending customer responses

[0464] Terminal: Record the customer's response again and send it to the server.

[0465] Input: Customer response voice

[0466] Output: Sends response audio data to the server

[0467] What it does: Makes additional recordings and sends the collected data to a server.

[0468] Step 11:

[0469] Customer response analysis and system learning

[0470] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[0471] Input: Customer response voice data

[0472] Output: Training data for accuracy improvement

[0473] What it does: Analyzes response data and feeds it into the learning algorithm, continuously improving the system.

[0474] (Application example 2)

[0475] 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."

[0476] In autonomous vehicles, the lack of technology to adjust the in-car environment taking into account passenger emotions makes it difficult to improve passenger comfort and satisfaction. In addition, there is no system to provide immediate and appropriate guidance or responses when passengers are confused or inconvenienced, which can lead to a poor user experience.

[0477] The identification processing 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 detecting an incoming call from a customer, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer in voice, means for re-collecting customer responses and performing learning based on the collected responses to improve the accuracy of the system, means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology, means for adjusting the in-vehicle environment based on the voice input and the emotion analysis results, and means for presenting emotion-adjusted feedback using voice synthesis technology. This makes it possible to adjust the in-vehicle environment taking passengers' emotions into consideration.

[0478] "Means for detecting incoming calls from customers" refers to devices or functions that recognize phone calls or voice input from customers in real time and input that data into the system.

[0479] "Means for collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to devices or algorithms that record customer voice and convert it into text data using voice recognition technology.

[0480] "Means of analyzing text data using natural language processing technology and identifying the content of customer inquiries" refers to technology that analyzes text data using a natural language processing algorithm and clarifies the specific content of customer inquiries from the content.

[0481] "Means for generating an appropriate response based on the analysis results" refers to a function for creating an optimal response from a pre-programmed database or algorithm based on the analysis results of the inquiry content.

[0482] The "means for providing the generated answer to the customer by voice" refers to a device or technology that converts the generated text-based answer into voice data using voice synthesis technology and conveys it to the customer.

[0483] "Means for re-collecting customer responses and learning to improve the accuracy of the system based on that" refers to a self-learning function that re-collects and analyzes feedback from customers and uses that to improve the system's performance and response accuracy.

[0484] "Means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology" refers to a function for recording passenger voice data, determining emotions from the voice using emotion analysis technology, and adjusting the in-car entertainment system based on the determination result.

[0485] "Means for adjusting the in-car environment based on voice input and emotion analysis results" is a function for automatically adjusting the temperature, music, lighting, etc. inside the car based on the passenger's voice input and the results of emotion analysis.

[0486] The "means for presenting emotion-adjusted feedback using voice synthesis technology" is a function for expressing optimal feedback based on the analyzed emotions in voice using voice synthesis technology and providing it to the occupant.

[0487] This invention relates to an infotainment system for autonomous vehicles. The system collects and analyzes passenger voice data using emotion analysis technology, and automatically adjusts the operation of the infotainment system and the in-vehicle environment based on the results. This system uses the following hardware and software:

[0488] Hardware

[0489] 1. Infotainment system: An entertainment and information system installed in a vehicle.

[0490] 2. Microphone: A device for collecting the voices of the passengers.

[0491] 3. Speaker: A device for providing audio feedback.

[0492] software

[0493] 1. Speech recognition API: Converts passenger voice data into text data (e.g., Google Cloud Speech-to-Text).

[0494] 2. Sentiment analysis API: Analyzes text data and voice tone to identify emotions (e.g., IBM Watson Tone Analyzer).

[0495] 3. Natural language processing engine: processes the parsed text data and understands the passenger's intent (e.g., spaCy, NLTK).

[0496] 4. Text-to-speech API: Converts text-based responses into audio data (e.g., Amazon Polly).

[0497] System Operation

[0498] First, the passenger's voice data is collected through a microphone. The collected voice data is sent to a speech recognition API and converted into text data. This text data is then sent to an emotion analysis API to analyze the passenger's emotions. The analyzed emotion data is then analyzed by a natural language processing engine to determine the appropriate feedback action (for example, adjusting the temperature or changing the music).

[0499] Processing Description

[0500] 1. Voice data collection: The microphone collects the passenger's voice and sends it to the voice recognition API.

[0501] 2. Speech recognition: The speech recognition API converts the voice data into text data.

[0502] 3. Emotion analysis: Text data and voice tone are input into the emotion analysis API to analyze the passenger's emotions.

[0503] 4. Natural Language Processing: The natural language processing engine determines the appropriate response or action based on the analyzed emotional data.

[0504] 5. Environmental Adjustment and Feedback: Based on the determined responses and actions, the infotainment system adjusts the in-car environment and provides audible feedback of the adjustments to the occupants using a speech synthesis API.

[0505] Specific examples

[0506] For example, if a passenger says, "It's a little cold," the voice data is converted into text data saying "It's a little cold" by the speech recognition API. This text data and voice tone are analyzed by the emotion analysis API to determine that the passenger is feeling cold. Next, the natural language processing engine determines that the passenger is "feeling cold" and issues a command to raise the interior temperature by 2 degrees. The speech synthesis API is also used to provide feedback to the passenger, saying, "The interior temperature has been raised a little."

[0507] Prompt Sentence Examples

[0508] The server was told by voice, "It's a little cold."

[0509] The infotainment system analyzed the voice input text and sent it to the sentiment analysis API, which output the sentiment analysis result as "I feel cold."

[0510] The driver has been instructed to increase the interior temperature of the self-driving vehicle by 2 degrees.

[0511] "The interior temperature has been slightly increased," the voice synthesis API informed the occupants.

[0512] In this way, the various hardware and software components work together to provide passengers with a comfortable in-vehicle environment and appropriate feedback.

[0513] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0514] Step 1:

[0515] The user's voice data is collected through a microphone and sent to the device. Specifically, the infotainment system detects voice input and starts recording. The input is the user's voice, and the recorded voice file is generated as the output.

[0516] Step 2:

[0517] The device sends the collected voice data in real time to a voice recognition API, which converts the voice data into text data (e.g., Google Cloud Speech-to-Text). The input is an audio file, and the output is text data.

[0518] Step 3:

[0519] The server sends the text data and voice tone data to an emotion analysis API to analyze the customer's emotions. The emotion analysis API (e.g., IBM Watson Tone Analyzer) analyzes the text data and voice tone to generate emotion data. The input is text and voice tone, and the output is emotion data.

[0520] Step 4:

[0521] The server uses natural language processing technology to analyze the text data and emotion data to identify the customer's inquiry. The natural language processing engine (e.g., spaCy, NLTK) analyzes the text data and determines the appropriate response. The input is text data and emotion data, and the analysis results are generated as the output.

[0522] Step 5:

[0523] The server determines the appropriate response based on the analysis results and emotion data, and issues instructions to the infotainment system. For example, this could include raising the temperature inside the car by 2 degrees or changing the music. The inputs are the analysis results and emotion data, and the output is the setting data for the infotainment system.

[0524] Step 6:

[0525] The device adjusts the infotainment system according to instructions from the server, for example, adjusting the temperature or changing the music inside the car. The input is the setting data from the server, and the output is the adjusted in-car environment.

[0526] Step 7:

[0527] The server sends the generated settings to a speech synthesis API, which generates voice feedback. The speech synthesis API (e.g., Amazon Polly) converts text data into voice data. The input is the text data of the settings, and the output is voice feedback data.

[0528] Step 8:

[0529] The terminal plays back the generated voice feedback data and provides it to the user. This results in voice guidance such as "The interior temperature has been slightly increased." The voice feedback data is input, and voice guidance for the user is generated as output.

[0530] 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.

[0531] 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.

[0532] 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.

[0533] [Second embodiment]

[0534] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0535] 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.

[0536] 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).

[0537] 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.

[0538] 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.

[0539] 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).

[0540] 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.

[0541] 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.

[0542] 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.

[0543] 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.

[0544] 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.

[0545] 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."

[0546] This invention is a system for automating and streamlining customer service in a call center. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[0547] Overall system overview

[0548] 1. Detect incoming calls from customers

[0549] Server: Connects to the call center system and detects incoming calls in real time.

[0550] Device: Receives a call and starts recording.

[0551] User: Make a call and ask a question.

[0552] 2. Collection and analysis of audio data

[0553] Device: Records the customer's voice and sends the audio data to the server.

[0554] Server: Uses voice recognition technology to convert voice data into text data.

[0555] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[0556] 3. Identifying the inquiry and generating an answer

[0557] Server: Uses natural language processing technology to identify customer inquiries from text data.

[0558] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[0559] 4. Providing answers

[0560] Server: Uses speech synthesis technology to convert text data into voice data.

[0561] Terminal: Plays back the audio data and provides the answer to the customer.

[0562] 5. Feedback and learning

[0563] Terminal: Collects customer responses again and sends them to the server.

[0564] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[0565] As a specific processing example, consider the following scenario.

[0566] Example: Inquiry about product malfunction

[0567] 1. Receiving an incoming call

[0568] User: "My product isn't working and I'd like to check."

[0569] Server: Detects incoming calls and collects customer voice.

[0570] 2. Converting audio data into text

[0571] Terminal: Records what the customer says and sends the audio data to the server.

[0572] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[0573] 3. Query Analysis

[0574] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[0575] Server: Determines that the customer has a problem and searches the product manual for a solution.

[0576] 4. Answer Generation

[0577] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0578] Terminal: Speaks the generated answer to the customer.

[0579] 5. Analyzing and learning from customer responses

[0580] User: "The model number is ABC123. The serial number is XYZ890."

[0581] Terminal: Records customer responses and sends the audio data to the server.

[0582] Server: Analyzes the new voice data and retrieves product details.

[0583] Server: We will update our response with new information and provide further details as needed.

[0584] In this way, the server and terminals work together to create an automated response system that improves the efficiency of call center operations. This system is also multilingual, can operate 24 hours a day, 365 days a year, and is highly scalable.

[0585] The processing flow will be explained below.

[0586] Step 1:

[0587] User: Calls and says, "My product isn't working and I'd like to check it out."

[0588] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[0589] Terminal: Receives incoming calls and begins recording customer conversations.

[0590] Step 2:

[0591] Device: Sends recorded audio data to the server in real time.

[0592] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[0593] Server: Obtain the text data "My product is not working, so I would like to check it."

[0594] Step 3:

[0595] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[0596] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[0597] Server: Analyzes the customer's tone and language and determines the emotion is "confused."

[0598] Step 4:

[0599] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[0600] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0601] Step 5:

[0602] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0603] Terminal: Plays the generated voice data and provides the answer to the customer.

[0604] Step 6:

[0605] User: "The model number is ABC123. The serial number is XYZ890."

[0606] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0607] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0608] Step 7:

[0609] Server: Extract the product model number and serial number from the new text data.

[0610] Server: Uses the extracted information to retrieve product details from the database.

[0611] Server: Provides detailed fault diagnosis and updated response as needed.

[0612] Step 8:

[0613] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[0614] Terminal: Displays customer inquiry history and response content, and can be used to respond to future inquiries.

[0615] In this way, the server, terminal, and user work together to realize an automated voice response system that provides quick and accurate responses.

[0616] Example 1

[0617] 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."

[0618] In conventional call center operations, a large number of staff members have to deal with customers, requiring a great deal of effort and time, resulting in problems of inefficiency and human error. Furthermore, responding to customer inquiries in a timely and appropriate manner requires advanced skills and specialized knowledge, creating challenges in human resource development. To solve these problems and improve customer satisfaction, a more efficient and highly accurate automated response system is needed.

[0619] 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.

[0620] In this invention, the server includes a means for detecting incoming calls from customers, a means for collecting customer voice data in real time and converting it into text data using speech recognition technology, and a means for analyzing the text data using natural language processing technology to identify the content of the customer's inquiry. This enables an automated response that responds to customer inquiries in a timely and appropriate manner. Furthermore, by providing a means for recollecting customer responses and improving the accuracy of the system based on them, highly accurate responses based on the latest information are always possible. Furthermore, by providing a means for converting text data into speech data using speech synthesis technology and providing it to the customer audibly, more natural dialogue is realized.

[0621] "Means for detecting incoming calls" refers to a device or function that connects to a telephone system or network and recognizes incoming calls from customers in real time.

[0622] "Voice data" refers to information that has been recorded in digital form from a customer's voice.

[0623] "Speech recognition technology" is a technology for converting voice data into text data.

[0624] "Text data" is information in text format generated from voice data using voice recognition technology.

[0625] "Natural language processing technology" is a technology that analyzes text data, understands meaning and emotions, and generates appropriate responses.

[0626] "Means for identifying inquiry content" refers to the function of using natural language processing technology to identify specific customer requests or problems from text data.

[0627] An "answer generation means" is a device or function that generates an appropriate response or solution based on the identified inquiry.

[0628] "Speech synthesis technology" is a technology that converts text data into voice data.

[0629] The "means for providing to the customer" is a device or function for playing back the generated voice data to the customer.

[0630] A "means for recollecting customer responses" is a device or function that records new voice data that the customer gives in response to the answers provided and transmits it to the system.

[0631] "Learning means to improve system accuracy" refers to a function that analyzes collected customer response data and updates the model using machine learning technology to improve system performance.

[0632] A "means for establishing a call session" is a device or function that establishes a communication channel between a customer and a system, enabling two-way communication of voice data.

[0633] "Means for searching for relevant information from a database" refers to a search engine or function for quickly finding data related to the inquiry from stored information.

[0634] This invention is a system for automating and streamlining customer service in call center operations. The system is broadly composed of three elements: a server, a terminal, and a user. How each element specifically operates is described in detail below.

[0635] Hardware and software used

[0636] Server: A high-performance server is used to perform speech recognition, natural language processing, speech synthesis, data analysis, etc. Specifically, technologies such as Google Cloud Speech-to-Text, spaCy, Hugging Face's BERT model, ElasticSearch, Amazon Polly, and TensorFlow are used.

[0637] Terminal: A terminal equipped with a SIP (Session Initiation Protocol) client has the function of recording the customer's voice and sending it to the server.

[0638] User: The customer making the call.

[0639] Details of system processing

[0640] 1. Detecting incoming calls from customers and establishing call sessions

[0641] Server: Connects to the VoIP server to detect incoming calls from customers in real time and notify the terminal.

[0642] Device: Upon receiving the notification, it establishes a call session with the customer and begins recording audio.

[0643] User: Make a call and start an inquiry.

[0644] 2. Collecting and converting audio data into text

[0645] Terminal: Continuously streams the customer's voice to the server in real time.

[0646] Server: Sends the received audio data to the Google Cloud Speech-to-Text API to generate text data.

[0647] 3. Analyzing text data and identifying inquiry content

[0648] Server: Analyze the generated text data using the spaCy library to extract key phrases and sentiment information.

[0649] Server: Leverages Hugging Face's BERT model to identify specific queries based on key phrases.

[0650] 4. Answer Generation and Search

[0651] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals.

[0652] Server: Generates appropriate answers for customers based on the retrieved information.

[0653] 5. Conversion into audio data and provision

[0654] Server: The generated text data is converted into voice data using Amazon Polly.

[0655] Terminal: Plays the converted voice data and provides the answer to the customer.

[0656] 6. Collect and learn from customer responses

[0657] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server.

[0658] Server: Analyzes new audio data and uses TensorFlow to update the model to improve the accuracy of the system.

[0659] Specific examples

[0660] Below is a scenario for an inquiry about a product failure.

[0661] scenario

[0662] 1. Detecting incoming calls and establishing a call session

[0663] User: "My product isn't working, so I'd like to check it out."

[0664] Server: The VoIP server detects the incoming call and notifies the terminal.

[0665] Device: Establish a call session and start recording.

[0666] 2. Collecting and converting audio data into text

[0667] Terminal: Records what the customer says and sends the audio data to the server.

[0668] Server: Use Google Cloud Speech-to-Text to generate text data such as "My product isn't working, so I'd like to check it."

[0669] 3. Parsing the query and generating an answer

[0670] Server: Uses the spaCy library to extract the key phrase "product malfunction" from text data, and then uses the BERT model to identify the specific content.

[0671] Server: Use ElasticSearch to search the product manual for solutions related to the product failure. Based on the search results, generate a response such as "Please let us know the product model number and serial number so we can look into it."

[0672] 4. Conversion into audio data and provision

[0673] Server: Convert the response text into speech data using Amazon Polly.

[0674] Terminal: Plays back the audio data and provides the answer to the customer.

[0675] 5. Analyzing and learning from customer responses

[0676] User: "The model number is ABC123. The serial number is XYZ890."

[0677] Device: Record any additional comments and send them back to the server.

[0678] Server: Analyzes new audio data and retrieves product details, while simultaneously updating the model in TensorFlow to improve the system's accuracy.

[0679] Prompt Sentence Examples

[0680] "Generate a natural response for the customer who asks for the product's model number and serial number."

[0681] "When a customer calls to inquire about a malfunctioning product, generate the best response."

[0682] In this way, the server and terminals work together to create a highly efficient and accurate automatic response system for customer inquiries.

[0683] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0684] Step 1: Detecting an incoming call from a customer and establishing a call session

[0685] Server: Connects to the VoIP server and detects incoming calls from customers in real time. The input is the call signal from the customer, and the output is the call detection information.

[0686] Terminal: Upon receiving a notification from the server, it establishes a call session with the customer using a SIP client and starts recording the audio. The input is the notification from the server, and the output is the establishment of a call session and the start of recording.

[0687] User: Makes a call and initiates a query. The input is the user making the call, and the output is the initiation of the call.

[0688] Step 2: Collect and send audio data

[0689] Terminal: Collects voice data in real time during a call with a customer and streams the voice data to a server. The input is the voice signal, and the output is the voice data sent to the server.

[0690] Server: Sends the received voice data to the Google Cloud Speech-to-Text API and converts the voice data to text data. The input is voice data from the device, and the output is text data.

[0691] Step 3: Analyzing text data and identifying inquiry content

[0692] Server: The generated text data is analyzed using the spaCy library to extract key phrases and analyze customer sentiment. The input is text data, and the output is key phrases and sentiment information.

[0693] Server: Uses Hugging Face's BERT model to identify the specific query from the extracted key phrases. The input is the key phrase and the output is the query.

[0694] Step 4: Generate answers and find related information

[0695] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals. The input is the query and the output is the relevant information.

[0696] Server: Generates appropriate answers for customers based on the retrieved information. The input is the relevant information and the output is the answer text.

[0697] Step 5: Converting the response into audio data and providing it

[0698] Server: Use Amazon Polly to convert the generated answer text data into speech data. The input is the answer text and the output is speech data.

[0699] Terminal: Plays back the voice data received from the server and provides answers to the customer. The input is the voice data, and the output is the played voice.

[0700] Step 6: Collect and send customer responses

[0701] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server. The input is the customer's speech and the output is the recorded audio data.

[0702] User: Checks the answer and provides a follow-up question or additional information. The input is the answer provided and the output is the additional utterance.

[0703] Step 7: Analyze the data and train the system

[0704] Server: Analyzes new audio data received and uses machine learning algorithms to update the model to improve the accuracy of the system. The input is the new audio data and the output is the updated machine learning model.

[0705] (Application example 1)

[0706] 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."

[0707] Improving the quality and efficiency of customer service is a key challenge in modern brick-and-mortar stores. Providing prompt and appropriate responses to customers is particularly difficult during busy times, potentially leading to lower customer satisfaction. Another issue is the need for a large number of staff, which increases labor costs. Furthermore, providing accurate information in response to customer inquiries relies heavily on the knowledge and experience of staff. Given these circumstances, there is a need for systems that use voice recognition and natural language processing technologies to automate customer service in real time via smart devices, providing efficient, high-quality service.

[0708] 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.

[0709] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning based on the collected responses to improve the accuracy of the system, and means for collecting customer inquiries in real time via a smart device and generating and providing an appropriate answer, thereby enabling more efficient customer service and the provision of high-quality services.

[0710] A "server" is a central computer system that detects incoming calls from customers, collects voice data, analyzes it, and generates responses.

[0711] "Smart devices" are devices that have advanced functionality, primarily smart glasses and tablets, and interact through voice recognition and speech synthesis.

[0712] "Means for detecting incoming calls" refers to technology that recognizes when a call or voice message is received from a customer and uses that as a trigger to activate the system.

[0713] "Speech recognition technology" is a technology that converts voice data into text data.

[0714] "Natural language processing technology" is a technology that analyzes text data and understands customer intentions and emotions.

[0715] "Means for generating appropriate answers based on the analysis results" refers to technology that generates answers to customer questions based on the analyzed text data.

[0716] The "means for providing the generated answer to the customer by voice" is a technology for converting the generated text answer into voice data and transmitting it to the customer.

[0717] "Means of collecting customer responses again and learning based on that to improve the accuracy of the system" refers to a technology that collects customer responses and uses machine learning to improve the system's responses.

[0718] The system that realizes this application example is composed of the following programs and a combination of hardware and software.

[0719] Program Generation

[0720] The server contains programs that do the following:

[0721] 1. Detect incoming calls from customers.

[0722] 2. Collect customer voice data in real time and convert it into text data using voice recognition technology.

[0723] 3. Analyze the text data using natural language processing technology to identify the customer's inquiry.

[0724] 4. Generate appropriate answers based on the analysis results.

[0725] 5. Provide generated answers to customers via voice.

[0726] 6. Customer responses are collected again and the system learns from them to improve its accuracy.

[0727] 7. Collect customer inquiries in real time via smart devices and generate and provide appropriate answers.

[0728] Processing Description

[0729] Hardware / Software used

[0730] Hardware: Smart glasses (microphone, camera, speaker), server (high-performance processing unit)

[0731] Software: Python programming language, speech_recognition module, gTTS (Google Text-to-Speech) module, openai API

[0732] Data processing and calculation

[0733] 1. The server uses a speech recognition module to convert the customer's speech into text data in real time. Specifically, the server acquires the speech data using the speech_recognition module and outputs it as text data.

[0734] 2. The server uses natural language processing technology to analyze the converted text data. It uses the OpenAI API to analyze the text data and generate an appropriate answer. It uses the prompt sentence to input the data into the generative AI model to obtain the answer.

[0735] 3. The server converts the generated text response into voice data using the gTTS module, which allows the generated response to be output as voice and played through the smart glasses' speaker.

[0736] Adding specific examples

[0737] For example, consider the following scenario:

[0738] Customer: "Is this product waterproof?"

[0739] The camera in the smart glasses recognizes the customer and the microphone records their questions.

[0740] The server converts the voice data into text, such as "Is this product waterproof?"

[0741] The prompt sentence "Is this product waterproof?" was input into the generative AI model via the OpenAI API.

[0742] The server generates the response: "Yes, this product is IP67 waterproof."

[0743] gTTS converts text to speech and plays it through the smart glasses' speakers.

[0744] Example prompt:

[0745] Product feature questions: "Is this product waterproof?"

[0746] A: Yes, this product is IP67 rated waterproof.

[0747] In this way, it becomes possible to improve the efficiency of customer service and provide high-quality services in physical stores.

[0748] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0749] Step 1:

[0750] The server detects when a customer enters the store using the camera and microphone of the smart glasses. When the customer starts speaking, the smart glasses collect voice data. The input here is the customer's voice utterance, and the output is the recorded voice data.

[0751] Step 2:

[0752] The server converts the collected voice data into text data using the speech_recognition module. Specifically, it analyzes the voice data using speech recognition technology and generates text data such as "Is this product waterproof?". The input here is the recorded voice data, and the output is text data.

[0753] Step 3:

[0754] The server uses the OpenAI API to analyze the generated text data. A prompt sentence is constructed and input into the generative AI model to generate an appropriate answer. The input here is the text data and the prompt sentence, and the output is the generated answer text. An example of a prompt sentence is "Question about product features: Is this product waterproof?"

[0755] Step 4:

[0756] The server converts the generated answer text into voice data using the gTTS module. Specifically, it outputs the answer text as voice data using speech synthesis technology. The input here is the generated answer text, and the output is voice data.

[0757] Step 5:

[0758] The smart glasses at the terminal play the generated voice data and provide a response to the customer, where the input is the voice data and the output is the voice information provided to the customer.

[0759] Step 6:

[0760] The user responds to the server's answers and asks additional questions or comments as needed. Their responses are again recorded by the smart glasses' microphone. The input here is the customer's additional utterances, and the output is again the recorded voice data.

[0761] Step 7:

[0762] The server analyzes the re-collected voice data and performs learning to improve the accuracy of the system. Specifically, the customer response data and the analysis results are fed back to the machine learning model to improve future response accuracy. The input here is the re-collected voice data and the analysis results, and the output is a learning model with improved system accuracy.

[0763] 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.

[0764] This invention is a system for automating and streamlining customer service in call centers. In particular, by combining it with an emotion engine that recognizes customer emotions and adjusts responses based on those emotions, customer satisfaction can be further improved. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[0765] Overall system overview

[0766] 1. Detect incoming calls from customers

[0767] Server: Connects to the call center system and detects incoming calls in real time.

[0768] Device: Receives a call and starts recording.

[0769] User: Make a call and ask a question.

[0770] 2. Collection and analysis of audio data

[0771] Device: Records the customer's voice and sends the audio data to the server.

[0772] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[0773] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[0774] 3. Emotion analysis using an emotion engine

[0775] Server: Analyzes customer emotions and obtains emotional data using the emotion engine.

[0776] Server: Makes decisions to generate the best answer based on customer sentiment.

[0777] 4. Identifying the inquiry and generating an answer

[0778] Server: Uses a natural language processing (NLP) engine to identify customer inquiries from text data.

[0779] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[0780] Server: Dynamically adjust responses based on customer sentiment data.

[0781] 5. Providing answers

[0782] Server: Uses speech synthesis technology to convert text data into voice data.

[0783] Terminal: Plays the generated voice data and provides the answer to the customer.

[0784] 6. Feedback and Learning

[0785] Terminal: Collects customer responses again and sends them to the server.

[0786] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[0787] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[0788] As a specific processing example, consider the following scenario.

[0789] Example: Inquiry about product malfunction

[0790] 1. Receiving an incoming call

[0791] User: "My product isn't working and I'd like to check."

[0792] Server: Detects incoming calls and collects customer voice.

[0793] 2. Converting audio data into text

[0794] Terminal: Records what the customer says and sends the audio data to the server.

[0795] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[0796] 3. Emotion analysis

[0797] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions.

[0798] Server: Determines that the customer is in trouble and stores the emotion data.

[0799] 4. Query Analysis

[0800] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[0801] Server: Considers customer sentiment and searches for solutions in product manuals.

[0802] 5. Answer Generation

[0803] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0804] Servers: Use polite language that takes into account the customer's confused feelings.

[0805] 6. Providing answers

[0806] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0807] Terminal: Plays the generated voice data and provides the answer to the customer.

[0808] 7. Analyzing and learning from customer responses

[0809] User: "The model number is ABC123. The serial number is XYZ890."

[0810] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0811] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0812] Server: Extract the product model number and serial number from the new text data.

[0813] Server: Uses the extracted information to retrieve product details from the database.

[0814] Server: Provides detailed fault diagnosis and updated response as needed.

[0815] 8. Feedback and Data Storage

[0816] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[0817] Server: Emotional data is also stored and used for future improvements.

[0818] In this way, the server, terminals, and users work together to realize an automated voice response system that provides prompt and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[0819] The processing flow will be explained below.

[0820] Step 1:

[0821] User: Calls and says, "My product isn't working and I'd like to check it out."

[0822] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[0823] Terminal: Receives incoming calls and begins recording customer conversations.

[0824] Step 2:

[0825] Device: Sends recorded audio data to the server in real time.

[0826] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[0827] Server: Obtain the text data "My product is not working, so I would like to check it."

[0828] Step 3:

[0829] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[0830] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[0831] Server: Based on the text data and tone of voice, determine that the customer's emotion is "confused."

[0832] Step 4:

[0833] Server: Manages customer emotion data using the emotion engine.

[0834] Server: Makes decisions to generate the best answer based on the customer's confused feelings.

[0835] Step 5:

[0836] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[0837] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[0838] Step 6:

[0839] Server: Dynamically adjusts responses based on customer sentiment determined by the sentiment engine.

[0840] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0841] Terminal: Plays the generated voice data and provides the answer to the customer.

[0842] Step 7:

[0843] User: "The model number is ABC123. The serial number is XYZ890."

[0844] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0845] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0846] Step 8:

[0847] Server: Extract the product model number and serial number from the new text data.

[0848] Server: Uses the extracted information to retrieve product details from the database.

[0849] Server: Provides detailed fault diagnosis and updated response as needed.

[0850] Step 9:

[0851] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[0852] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[0853] In this way, the server, terminals, and users work together to realize an automated voice response system that provides fast and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[0854] Example 2

[0855] 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."

[0856] Customer support at call centers relies heavily on human resources, and while they require fast and accurate responses, it is difficult to fully understand customer emotions and provide appropriate responses. In particular, a lack of real-time responses that reflect customer emotions can lead to a decline in customer satisfaction. Furthermore, existing systems are difficult to handle, as they also require multilingual support and 24 / 7 service provision.

[0857] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0858] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning to improve the accuracy of the system based on the analysis, means for analyzing customer emotions and dynamically adjusting answers based on the results, and a computer terminal and database with which the above means are linked via a communications network. This enables quick and accurate automatic responses based on an understanding of customer emotions and continuous improvement of the system.

[0859] "Means for detecting incoming calls from customers" refers to a system that detects incoming calls from customers in real time and notifies other systems of that information.

[0860] "Means of collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to a system that records the voices transmitted by customers and converts that data into text information using voice recognition technology.

[0861] "Means for analyzing text data using natural language processing technology and identifying the content of the customer's inquiry" refers to a system that analyzes text information using natural language processing technology and identifies the content of the customer's intended inquiry.

[0862] "Means for generating an appropriate response based on the analysis results" refers to a mechanism that automatically creates the most appropriate and accurate response based on the analyzed inquiry content.

[0863] The "means of providing the generated answer to the customer by voice" is a mechanism that converts the automatically generated answer into voice data using voice synthesis technology and provides it to the customer.

[0864] "Means of re-collecting customer responses and learning based on them to improve the accuracy of the system" refers to a mechanism that re-records responses and reactions from customers and uses that data to run a machine learning process to improve the performance of the system.

[0865] "Means for analyzing customer emotions and dynamically adjusting responses based on the results" refers to a system that analyzes emotions from customer voice and text data and dynamically changes the content of responses based on the results of that analysis.

[0866] "Computer terminals and databases linked via a communications network" refers to an infrastructure that enables different computer terminals and databases to share information and operate in cooperation with each other using the Internet or other communications networks.

[0867] This invention is a system for automating and streamlining customer service in a call center. This system is composed of three main elements: a server, a terminal, and a user, each of which operates in a specific way.

[0868] Overall system overview

[0869] 1. Detect incoming calls from customers

[0870] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls to specific numbers.

[0871] Terminal: Receives incoming calls and prepares to record the voice of the customer. Activates the recording function the moment it receives notification of an incoming call.

[0872] User: Calls to inquire.

[0873] 2. Collection and analysis of audio data

[0874] Device: Records the customer's voice and sends it to the server as audio data. A highly sensitive microphone is used to record the customer's speech with clear sound quality.

[0875] Server: Passes the received voice data to a voice recognition API (e.g., Google Speech-to-Text) and converts it into text data. Temporarily stores the voice data and sends a request to the voice recognition API.

[0876] Server: Analyzes customer sentiment and key phrases from text data and extracts important information.

[0877] 3. Emotion analysis using an emotion engine

[0878] Server: Analyzes customer emotions using an emotion engine (e.g., IBM Watson Tone Analyzer) and obtains emotion data. Inputs the text data and tone of voice data into the emotion engine.

[0879] Server: Determines the best response based on customer sentiment. Uses sentiment analysis results as a basis for customizing responses.

[0880] 4. Identifying the inquiry and generating an answer

[0881] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer's inquiry from the text data. Sends the text data to the NLP engine, which identifies the inquiry as "product malfunction."

[0882] Server: Searches product manuals and internal databases to generate appropriate answers. Automatically searches databases for relevant information and generates answers.

[0883] Server: Dynamically adjusts responses based on customer sentiment data. Uses polite language based on sentiment data.

[0884] 5. Providing answers

[0885] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert text data into speech data. Inputs the response text into the speech synthesis API to generate speech data.

[0886] Terminal: Plays the generated voice data and provides the answer to the customer. Receives the generated voice data and plays it back for the customer to hear.

[0887] 6. Feedback and Learning

[0888] Terminal: The customer's responses are collected again and sent to the server as audio data. Any additional answers or responses from the customer are also recorded and sent to the server.

[0889] Server: Analyzes customer responses and learns to improve the accuracy of the entire system. Analyzes new data and updates the system's machine learning algorithms.

[0890] Server: Stores emotion data based on the analysis results and uses it to improve future customer service responses. The updated database is used to respond to future customer inquiries.

[0891] Specific examples

[0892] Inquiries about product malfunctions

[0893] User: "My product isn't working and I'd like to check."

[0894] Server: Detects incoming calls and collects customer voice.

[0895] Terminal: Records what the customer says and sends the audio data to the server.

[0896] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[0897] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions. Detects confused emotions and stores the emotion data.

[0898] Server: Using natural language processing technology, identify the inquiry content as "product malfunction."

[0899] Server: Considers the customer's feelings and searches for a solution in the product manual. Generates a response such as "Please let us know your product model number and serial number so we can look into it."

[0900] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[0901] Terminal: Plays the generated voice data and provides the answer to the customer.

[0902] User: Responds, "The model number is ABC123. The serial number is XYZ890."

[0903] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[0904] Server: Pass new voice data to the speech recognition API and convert it into text data.

[0905] Server: Extract the product model number and serial number from the new text data. Use the extracted information to retrieve detailed product information from the database.

[0906] Server: Provides detailed fault diagnosis and updated response as needed.

[0907] Example prompt sentence:

[0908] The system analyzes a user's statement, "My product isn't working, so I'd like to check it," and uses a sentiment analysis engine to identify the customer's emotions and generate an answer.

[0909] This allows the system to provide efficient and accurate automated responses while taking into account customer emotions.

[0910] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0911] Step 1:

[0912] Detecting incoming calls from customers

[0913] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls from customers.

[0914] Input: Incoming call signal from customer

[0915] Output: Incoming call notification signal

[0916] Specific operations: Monitors the phone number and receives an incoming call signal, then notifies the device of the incoming call signal.

[0917] Step 2:

[0918] Collecting audio data and starting recording

[0919] Terminal: Receives an incoming call and activates the recording function, recording what the customer says and preparing to collect voice data with clear sound quality using a high-sensitivity microphone.

[0920] Input: Incoming call notification signal

[0921] Output: Start recording

[0922] Specific behavior: When an incoming call notification is received, the recording function is activated and the customer's voice begins to be recorded.

[0923] Step 3:

[0924] Sending audio data to the server

[0925] Terminal: Sends recorded audio data to the server in real time.

[0926] Input: Recorded audio data

[0927] Output: Send audio data to the server

[0928] Specific operation: The recorded audio data is compressed and sent to a server via internet communication.

[0929] Step 4:

[0930] Converting audio data to text

[0931] Server: Passes the voice data to a speech recognition API (e.g., Google Speech-to-Text) and converts it to text data. Temporarily stores the voice data and sends a request to the speech recognition API.

[0932] Input: Audio data sent to the server

[0933] Output: Text data generated by speech recognition

[0934] Specific operation: Sends audio data to the API and retrieves the returned text data.

[0935] Step 5:

[0936] Sentiment analysis of text data

[0937] Server: Using an emotion engine (e.g., IBM Watson Tone Analyzer), analyze customer emotions based on text data and voice tone.

[0938] Input: Text data and voice tone data from speech recognition

[0939] Output: Emotion data

[0940] Specific operation: Text data and voice tone are input into the emotion engine to obtain the emotion analysis results.

[0941] Step 6:

[0942] Identifying the content of the inquiry

[0943] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer inquiry from the text data.

[0944] Input: Text data

[0945] Output: Identified inquiry

[0946] Specific behavior: Analyze what the customer says and identify the specific inquiry.

[0947] Step 7:

[0948] Generating the right answers

[0949] Server: Generates appropriate answers based on customer inquiry content and sentiment data. Searches product manuals and internal databases to generate answers.

[0950] Input: Identified query content and sentiment data

[0951] Output: Generated answer text

[0952] What it does: Searches a database, extracts the relevant information, and creates an answer.

[0953] Step 8:

[0954] Converting responses into audio data

[0955] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert the generated answer text into voice data.

[0956] Input: Generated answer text

[0957] Output: Audio data

[0958] Specific operation: The answer text is input into the speech synthesis API to generate voice data.

[0959] Step 9:

[0960] Providing answers

[0961] Terminal: Plays the generated voice data and provides the answer to the customer.

[0962] Input: Audio data

[0963] Output: A spoken response to the customer

[0964] Specific behavior: Plays back the audio data and provides the appropriate answer to the customer.

[0965] Step 10:

[0966] Collecting and sending customer responses

[0967] Terminal: Record the customer's response again and send it to the server.

[0968] Input: Customer response voice

[0969] Output: Sends response audio data to the server

[0970] What it does: Makes additional recordings and sends the collected data to a server.

[0971] Step 11:

[0972] Customer response analysis and system learning

[0973] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[0974] Input: Customer response voice data

[0975] Output: Training data for accuracy improvement

[0976] What it does: Analyzes response data and feeds it into the learning algorithm, continuously improving the system.

[0977] (Application example 2)

[0978] 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."

[0979] In autonomous vehicles, the lack of technology to adjust the in-car environment taking into account passenger emotions makes it difficult to improve passenger comfort and satisfaction. In addition, there is no system to provide immediate and appropriate guidance or responses when passengers are confused or inconvenienced, which can lead to a poor user experience.

[0980] The identification processing 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 detecting an incoming call from a customer, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer in voice, means for re-collecting customer responses and performing learning based on the collected responses to improve the accuracy of the system, means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology, means for adjusting the in-vehicle environment based on the voice input and the emotion analysis results, and means for presenting emotion-adjusted feedback using voice synthesis technology. This makes it possible to adjust the in-vehicle environment taking passengers' emotions into consideration.

[0981] "Means for detecting incoming calls from customers" refers to devices or functions that recognize phone calls or voice input from customers in real time and input that data into the system.

[0982] "Means for collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to devices or algorithms that record customer voice and convert it into text data using voice recognition technology.

[0983] "Means of analyzing text data using natural language processing technology and identifying the content of customer inquiries" refers to technology that analyzes text data using a natural language processing algorithm and clarifies the specific content of customer inquiries from the content.

[0984] "Means for generating an appropriate response based on the analysis results" refers to a function for creating an optimal response from a pre-programmed database or algorithm based on the analysis results of the inquiry content.

[0985] The "means for providing the generated answer to the customer by voice" refers to a device or technology that converts the generated text-based answer into voice data using voice synthesis technology and conveys it to the customer.

[0986] "Means for re-collecting customer responses and learning to improve the accuracy of the system based on that" refers to a self-learning function that re-collects and analyzes feedback from customers and uses that to improve the system's performance and response accuracy.

[0987] "Means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology" refers to a function for recording passenger voice data, determining emotions from the voice using emotion analysis technology, and adjusting the in-car entertainment system based on the determination result.

[0988] "Means for adjusting the in-car environment based on voice input and emotion analysis results" is a function for automatically adjusting the temperature, music, lighting, etc. inside the car based on the passenger's voice input and the results of emotion analysis.

[0989] The "means for presenting emotion-adjusted feedback using voice synthesis technology" is a function for expressing optimal feedback based on the analyzed emotions in voice using voice synthesis technology and providing it to the occupant.

[0990] This invention relates to an infotainment system for autonomous vehicles. The system collects and analyzes passenger voice data using emotion analysis technology, and automatically adjusts the operation of the infotainment system and the in-vehicle environment based on the results. This system uses the following hardware and software:

[0991] Hardware

[0992] 1. Infotainment system: An entertainment and information system installed in a vehicle.

[0993] 2. Microphone: A device for collecting the voices of the passengers.

[0994] 3. Speaker: A device for providing audio feedback.

[0995] software

[0996] 1. Speech recognition API: Converts passenger voice data into text data (e.g., Google Cloud Speech-to-Text).

[0997] 2. Sentiment analysis API: Analyzes text data and voice tone to identify emotions (e.g., IBM Watson Tone Analyzer).

[0998] 3. Natural language processing engine: processes the parsed text data and understands the passenger's intent (e.g., spaCy, NLTK).

[0999] 4. Text-to-speech API: Converts text-based responses into audio data (e.g., Amazon Polly).

[1000] System Operation

[1001] First, the passenger's voice data is collected through a microphone. The collected voice data is sent to a speech recognition API and converted into text data. This text data is then sent to an emotion analysis API to analyze the passenger's emotions. The analyzed emotion data is then analyzed by a natural language processing engine to determine the appropriate feedback action (for example, adjusting the temperature or changing the music).

[1002] Processing Description

[1003] 1. Voice data collection: The microphone collects the passenger's voice and sends it to the voice recognition API.

[1004] 2. Speech recognition: The speech recognition API converts the voice data into text data.

[1005] 3. Emotion analysis: Text data and voice tone are input into the emotion analysis API to analyze the passenger's emotions.

[1006] 4. Natural Language Processing: The natural language processing engine determines the appropriate response or action based on the analyzed emotional data.

[1007] 5. Environmental Adjustment and Feedback: Based on the determined responses and actions, the infotainment system adjusts the in-car environment and provides audible feedback of the adjustments to the occupants using a speech synthesis API.

[1008] Specific examples

[1009] For example, if a passenger says, "It's a little cold," the voice data is converted into text data saying "It's a little cold" by the speech recognition API. This text data and voice tone are analyzed by the emotion analysis API to determine that the passenger is feeling cold. Next, the natural language processing engine determines that the passenger is "feeling cold" and issues a command to raise the interior temperature by 2 degrees. The speech synthesis API is also used to provide feedback to the passenger, saying, "The interior temperature has been raised a little."

[1010] Prompt Sentence Examples

[1011] The server was told by voice, "It's a little cold."

[1012] The infotainment system analyzed the voice input text and sent it to the sentiment analysis API, which output the sentiment analysis result as "I feel cold."

[1013] The driver has been instructed to increase the interior temperature of the self-driving vehicle by 2 degrees.

[1014] "The interior temperature has been slightly increased," the voice synthesis API informed the occupants.

[1015] In this way, the various hardware and software components work together to provide passengers with a comfortable in-vehicle environment and appropriate feedback.

[1016] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1017] Step 1:

[1018] The user's voice data is collected through a microphone and sent to the device. Specifically, the infotainment system detects voice input and starts recording. The input is the user's voice, and the recorded voice file is generated as the output.

[1019] Step 2:

[1020] The device sends the collected voice data in real time to a voice recognition API, which converts the voice data into text data (e.g., Google Cloud Speech-to-Text). The input is an audio file, and the output is text data.

[1021] Step 3:

[1022] The server sends the text data and voice tone data to an emotion analysis API to analyze the customer's emotions. The emotion analysis API (e.g., IBM Watson Tone Analyzer) analyzes the text data and voice tone to generate emotion data. The input is text and voice tone, and the output is emotion data.

[1023] Step 4:

[1024] The server uses natural language processing technology to analyze the text data and emotion data to identify the customer's inquiry. The natural language processing engine (e.g., spaCy, NLTK) analyzes the text data and determines the appropriate response. The input is text data and emotion data, and the analysis results are generated as the output.

[1025] Step 5:

[1026] The server determines the appropriate response based on the analysis results and emotion data, and issues instructions to the infotainment system. For example, this could include raising the temperature inside the car by 2 degrees or changing the music. The inputs are the analysis results and emotion data, and the output is the setting data for the infotainment system.

[1027] Step 6:

[1028] The device adjusts the infotainment system according to instructions from the server, for example, adjusting the temperature or changing the music inside the car. The input is the setting data from the server, and the output is the adjusted in-car environment.

[1029] Step 7:

[1030] The server sends the generated settings to a speech synthesis API, which generates voice feedback. The speech synthesis API (e.g., Amazon Polly) converts text data into voice data. The input is the text data of the settings, and the output is voice feedback data.

[1031] Step 8:

[1032] The terminal plays back the generated voice feedback data and provides it to the user. This results in voice guidance such as "The interior temperature has been slightly increased." The voice feedback data is input, and voice guidance for the user is generated as output.

[1033] 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.

[1034] 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.

[1035] 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.

[1036] [Third embodiment]

[1037] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1038] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1039] 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).

[1040] 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.

[1041] 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.

[1042] 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).

[1043] 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.

[1044] 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.

[1045] 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.

[1046] 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.

[1047] 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.

[1048] 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."

[1049] This invention is a system for automating and streamlining customer service in a call center. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[1050] Overall system overview

[1051] 1. Detect incoming calls from customers

[1052] Server: Connects to the call center system and detects incoming calls in real time.

[1053] Device: Receives a call and starts recording.

[1054] User: Make a call and ask a question.

[1055] 2. Collection and analysis of audio data

[1056] Device: Records the customer's voice and sends the audio data to the server.

[1057] Server: Uses voice recognition technology to convert voice data into text data.

[1058] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[1059] 3. Identifying the inquiry and generating an answer

[1060] Server: Uses natural language processing technology to identify customer inquiries from text data.

[1061] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[1062] 4. Providing answers

[1063] Server: Uses speech synthesis technology to convert text data into voice data.

[1064] Terminal: Plays back the audio data and provides the answer to the customer.

[1065] 5. Feedback and learning

[1066] Terminal: Collects customer responses again and sends them to the server.

[1067] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[1068] As a specific processing example, consider the following scenario.

[1069] Example: Inquiry about product malfunction

[1070] 1. Receiving an incoming call

[1071] User: "My product isn't working and I'd like to check."

[1072] Server: Detects incoming calls and collects customer voice.

[1073] 2. Converting audio data into text

[1074] Terminal: Records what the customer says and sends the audio data to the server.

[1075] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[1076] 3. Query Analysis

[1077] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[1078] Server: Determines that the customer has a problem and searches the product manual for a solution.

[1079] 4. Answer Generation

[1080] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1081] Terminal: Speaks the generated answer to the customer.

[1082] 5. Analyzing and learning from customer responses

[1083] User: "The model number is ABC123. The serial number is XYZ890."

[1084] Terminal: Records customer responses and sends the audio data to the server.

[1085] Server: Analyzes the new voice data and retrieves product details.

[1086] Server: We will update our response with new information and provide further details as needed.

[1087] In this way, the server and terminals work together to create an automated response system that improves the efficiency of call center operations. This system is also multilingual, can operate 24 hours a day, 365 days a year, and is highly scalable.

[1088] The processing flow will be explained below.

[1089] Step 1:

[1090] User: Calls and says, "My product isn't working and I'd like to check it out."

[1091] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[1092] Terminal: Receives incoming calls and begins recording customer conversations.

[1093] Step 2:

[1094] Device: Sends recorded audio data to the server in real time.

[1095] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[1096] Server: Obtain the text data "My product is not working, so I would like to check it."

[1097] Step 3:

[1098] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[1099] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[1100] Server: Analyzes the customer's tone and language and determines the emotion is "confused."

[1101] Step 4:

[1102] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[1103] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1104] Step 5:

[1105] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1106] Terminal: Plays the generated voice data and provides the answer to the customer.

[1107] Step 6:

[1108] User: "The model number is ABC123. The serial number is XYZ890."

[1109] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1110] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1111] Step 7:

[1112] Server: Extract the product model number and serial number from the new text data.

[1113] Server: Uses the extracted information to retrieve product details from the database.

[1114] Server: Provides detailed fault diagnosis and updated response as needed.

[1115] Step 8:

[1116] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[1117] Terminal: Displays customer inquiry history and response content, and can be used to respond to future inquiries.

[1118] In this way, the server, terminal, and user work together to realize an automated voice response system that provides quick and accurate responses.

[1119] Example 1

[1120] 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."

[1121] In conventional call center operations, a large number of staff members have to deal with customers, requiring a great deal of effort and time, resulting in problems of inefficiency and human error. Furthermore, responding to customer inquiries in a timely and appropriate manner requires advanced skills and specialized knowledge, creating challenges in human resource development. To solve these problems and improve customer satisfaction, a more efficient and highly accurate automated response system is needed.

[1122] 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.

[1123] In this invention, the server includes a means for detecting incoming calls from customers, a means for collecting customer voice data in real time and converting it into text data using speech recognition technology, and a means for analyzing the text data using natural language processing technology to identify the content of the customer's inquiry. This enables an automated response that responds to customer inquiries in a timely and appropriate manner. Furthermore, by providing a means for recollecting customer responses and improving the accuracy of the system based on them, highly accurate responses based on the latest information are always possible. Furthermore, by providing a means for converting text data into speech data using speech synthesis technology and providing it to the customer audibly, more natural dialogue is realized.

[1124] "Means for detecting incoming calls" refers to a device or function that connects to a telephone system or network and recognizes incoming calls from customers in real time.

[1125] "Voice data" refers to information that has been recorded in digital form from a customer's voice.

[1126] "Speech recognition technology" is a technology for converting voice data into text data.

[1127] "Text data" is information in text format generated from voice data using voice recognition technology.

[1128] "Natural language processing technology" is a technology that analyzes text data, understands meaning and emotions, and generates appropriate responses.

[1129] "Means for identifying inquiry content" refers to the function of using natural language processing technology to identify specific customer requests or problems from text data.

[1130] An "answer generation means" is a device or function that generates an appropriate response or solution based on the identified inquiry.

[1131] "Speech synthesis technology" is a technology that converts text data into voice data.

[1132] The "means for providing to the customer" is a device or function for playing back the generated voice data to the customer.

[1133] A "means for recollecting customer responses" is a device or function that records new voice data that the customer gives in response to the answers provided and transmits it to the system.

[1134] "Learning means to improve system accuracy" refers to a function that analyzes collected customer response data and updates the model using machine learning technology to improve system performance.

[1135] A "means for establishing a call session" is a device or function that establishes a communication channel between a customer and a system, enabling two-way communication of voice data.

[1136] "Means for searching for relevant information from a database" refers to a search engine or function for quickly finding data related to the inquiry from stored information.

[1137] This invention is a system for automating and streamlining customer service in call center operations. The system is broadly composed of three elements: a server, a terminal, and a user. How each element specifically operates is described in detail below.

[1138] Hardware and software used

[1139] Server: A high-performance server is used to perform speech recognition, natural language processing, speech synthesis, data analysis, etc. Specifically, technologies such as Google Cloud Speech-to-Text, spaCy, Hugging Face's BERT model, ElasticSearch, Amazon Polly, and TensorFlow are used.

[1140] Terminal: A terminal equipped with a SIP (Session Initiation Protocol) client has the function of recording the customer's voice and sending it to the server.

[1141] User: The customer making the call.

[1142] Details of system processing

[1143] 1. Detecting incoming calls from customers and establishing call sessions

[1144] Server: Connects to the VoIP server to detect incoming calls from customers in real time and notify the terminal.

[1145] Device: Upon receiving the notification, it establishes a call session with the customer and begins recording audio.

[1146] User: Make a call and start an inquiry.

[1147] 2. Collecting and converting audio data into text

[1148] Terminal: Continuously streams the customer's voice to the server in real time.

[1149] Server: Sends the received audio data to the Google Cloud Speech-to-Text API to generate text data.

[1150] 3. Analyzing text data and identifying inquiry content

[1151] Server: Analyze the generated text data using the spaCy library to extract key phrases and sentiment information.

[1152] Server: Leverages Hugging Face's BERT model to identify specific queries based on key phrases.

[1153] 4. Answer Generation and Search

[1154] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals.

[1155] Server: Generates appropriate answers for customers based on the retrieved information.

[1156] 5. Conversion into audio data and provision

[1157] Server: The generated text data is converted into voice data using Amazon Polly.

[1158] Terminal: Plays the converted voice data and provides the answer to the customer.

[1159] 6. Collect and learn from customer responses

[1160] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server.

[1161] Server: Analyzes new audio data and uses TensorFlow to update the model to improve the accuracy of the system.

[1162] Specific examples

[1163] Below is a scenario for an inquiry about a product failure.

[1164] scenario

[1165] 1. Detecting incoming calls and establishing a call session

[1166] User: "My product isn't working, so I'd like to check it out."

[1167] Server: The VoIP server detects the incoming call and notifies the terminal.

[1168] Device: Establish a call session and start recording.

[1169] 2. Collecting and converting audio data into text

[1170] Terminal: Records what the customer says and sends the audio data to the server.

[1171] Server: Use Google Cloud Speech-to-Text to generate text data such as "My product isn't working, so I'd like to check it."

[1172] 3. Parsing the query and generating an answer

[1173] Server: Uses the spaCy library to extract the key phrase "product malfunction" from text data, and then uses the BERT model to identify the specific content.

[1174] Server: Use ElasticSearch to search the product manual for solutions related to the product failure. Based on the search results, generate a response such as "Please let us know the product model number and serial number so we can look into it."

[1175] 4. Conversion into audio data and provision

[1176] Server: Convert the response text into speech data using Amazon Polly.

[1177] Terminal: Plays back the audio data and provides the answer to the customer.

[1178] 5. Analyzing and learning from customer responses

[1179] User: "The model number is ABC123. The serial number is XYZ890."

[1180] Device: Record any additional comments and send them back to the server.

[1181] Server: Analyzes new audio data and retrieves product details, while simultaneously updating the model in TensorFlow to improve the system's accuracy.

[1182] Prompt Sentence Examples

[1183] "Generate a natural response for the customer who asks for the product's model number and serial number."

[1184] "When a customer calls to inquire about a malfunctioning product, generate the best response."

[1185] In this way, the server and terminals work together to create a highly efficient and accurate automatic response system for customer inquiries.

[1186] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1187] Step 1: Detecting an incoming call from a customer and establishing a call session

[1188] Server: Connects to the VoIP server and detects incoming calls from customers in real time. The input is the call signal from the customer, and the output is the call detection information.

[1189] Terminal: Upon receiving a notification from the server, it establishes a call session with the customer using a SIP client and starts recording the audio. The input is the notification from the server, and the output is the establishment of a call session and the start of recording.

[1190] User: Makes a call and initiates a query. The input is the user making the call, and the output is the initiation of the call.

[1191] Step 2: Collect and send audio data

[1192] Terminal: Collects voice data in real time during a call with a customer and streams the voice data to a server. The input is the voice signal, and the output is the voice data sent to the server.

[1193] Server: Sends the received voice data to the Google Cloud Speech-to-Text API and converts the voice data to text data. The input is voice data from the device, and the output is text data.

[1194] Step 3: Analyzing text data and identifying inquiry content

[1195] Server: The generated text data is analyzed using the spaCy library to extract key phrases and analyze customer sentiment. The input is text data, and the output is key phrases and sentiment information.

[1196] Server: Uses Hugging Face's BERT model to identify the specific query from the extracted key phrases. The input is the key phrase and the output is the query.

[1197] Step 4: Generate answers and find related information

[1198] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals. The input is the query and the output is the relevant information.

[1199] Server: Generates appropriate answers for customers based on the retrieved information. The input is the relevant information and the output is the answer text.

[1200] Step 5: Converting the response into audio data and providing it

[1201] Server: Use Amazon Polly to convert the generated answer text data into speech data. The input is the answer text and the output is speech data.

[1202] Terminal: Plays back the voice data received from the server and provides answers to the customer. The input is the voice data, and the output is the played voice.

[1203] Step 6: Collect and send customer responses

[1204] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server. The input is the customer's speech and the output is the recorded audio data.

[1205] User: Checks the answer and provides a follow-up question or additional information. The input is the answer provided and the output is the additional utterance.

[1206] Step 7: Analyze the data and train the system

[1207] Server: Analyzes new audio data received and uses machine learning algorithms to update the model to improve the accuracy of the system. The input is the new audio data and the output is the updated machine learning model.

[1208] (Application example 1)

[1209] 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."

[1210] Improving the quality and efficiency of customer service is a key challenge in modern brick-and-mortar stores. Providing prompt and appropriate responses to customers is particularly difficult during busy times, potentially leading to lower customer satisfaction. Another issue is the need for a large number of staff, which increases labor costs. Furthermore, providing accurate information in response to customer inquiries relies heavily on the knowledge and experience of staff. Given these circumstances, there is a need for systems that use voice recognition and natural language processing technologies to automate customer service in real time via smart devices, providing efficient, high-quality service.

[1211] 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.

[1212] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning based on the collected responses to improve the accuracy of the system, and means for collecting customer inquiries in real time via a smart device and generating and providing an appropriate answer, thereby enabling more efficient customer service and the provision of high-quality services.

[1213] A "server" is a central computer system that detects incoming calls from customers, collects voice data, analyzes it, and generates responses.

[1214] "Smart devices" are devices that have advanced functionality, primarily smart glasses and tablets, and interact through voice recognition and speech synthesis.

[1215] "Means for detecting incoming calls" refers to technology that recognizes when a call or voice message is received from a customer and uses that as a trigger to activate the system.

[1216] "Speech recognition technology" is a technology that converts voice data into text data.

[1217] "Natural language processing technology" is a technology that analyzes text data and understands customer intentions and emotions.

[1218] "Means for generating appropriate answers based on the analysis results" refers to technology that generates answers to customer questions based on the analyzed text data.

[1219] The "means for providing the generated answer to the customer by voice" is a technology for converting the generated text answer into voice data and transmitting it to the customer.

[1220] "Means of collecting customer responses again and learning based on that to improve the accuracy of the system" refers to a technology that collects customer responses and uses machine learning to improve the system's responses.

[1221] The system that realizes this application example is composed of the following programs and a combination of hardware and software.

[1222] Program Generation

[1223] The server contains programs that do the following:

[1224] 1. Detect incoming calls from customers.

[1225] 2. Collect customer voice data in real time and convert it into text data using voice recognition technology.

[1226] 3. Analyze the text data using natural language processing technology to identify the customer's inquiry.

[1227] 4. Generate appropriate answers based on the analysis results.

[1228] 5. Provide generated answers to customers via voice.

[1229] 6. Customer responses are collected again and the system learns from them to improve its accuracy.

[1230] 7. Collect customer inquiries in real time via smart devices and generate and provide appropriate answers.

[1231] Processing Description

[1232] Hardware / Software used

[1233] Hardware: Smart glasses (microphone, camera, speaker), server (high-performance processing unit)

[1234] Software: Python programming language, speech_recognition module, gTTS (Google Text-to-Speech) module, openai API

[1235] Data processing and calculation

[1236] 1. The server uses a speech recognition module to convert the customer's speech into text data in real time. Specifically, the server acquires the speech data using the speech_recognition module and outputs it as text data.

[1237] 2. The server uses natural language processing technology to analyze the converted text data. It uses the OpenAI API to analyze the text data and generate an appropriate answer. It uses the prompt sentence to input the data into the generative AI model to obtain the answer.

[1238] 3. The server converts the generated text response into voice data using the gTTS module, which allows the generated response to be output as voice and played through the smart glasses' speaker.

[1239] Adding specific examples

[1240] For example, consider the following scenario:

[1241] Customer: "Is this product waterproof?"

[1242] The camera in the smart glasses recognizes the customer and the microphone records their questions.

[1243] The server converts the voice data into text, such as "Is this product waterproof?"

[1244] The prompt sentence "Is this product waterproof?" was input into the generative AI model via the OpenAI API.

[1245] The server generates the response: "Yes, this product is IP67 waterproof."

[1246] gTTS converts text to speech and plays it through the smart glasses' speakers.

[1247] Example prompt:

[1248] Product feature questions: "Is this product waterproof?"

[1249] A: Yes, this product is IP67 rated waterproof.

[1250] In this way, it becomes possible to improve the efficiency of customer service and provide high-quality services in physical stores.

[1251] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1252] Step 1:

[1253] The server detects when a customer enters the store using the camera and microphone of the smart glasses. When the customer starts speaking, the smart glasses collect voice data. The input here is the customer's voice utterance, and the output is the recorded voice data.

[1254] Step 2:

[1255] The server converts the collected voice data into text data using the speech_recognition module. Specifically, it analyzes the voice data using speech recognition technology and generates text data such as "Is this product waterproof?". The input here is the recorded voice data, and the output is text data.

[1256] Step 3:

[1257] The server uses the OpenAI API to analyze the generated text data. A prompt sentence is constructed and input into the generative AI model to generate an appropriate answer. The input here is the text data and the prompt sentence, and the output is the generated answer text. An example of a prompt sentence is "Question about product features: Is this product waterproof?"

[1258] Step 4:

[1259] The server converts the generated answer text into voice data using the gTTS module. Specifically, it outputs the answer text as voice data using speech synthesis technology. The input here is the generated answer text, and the output is voice data.

[1260] Step 5:

[1261] The smart glasses at the terminal play the generated voice data and provide a response to the customer, where the input is the voice data and the output is the voice information provided to the customer.

[1262] Step 6:

[1263] The user responds to the server's answers and asks additional questions or comments as needed. Their responses are again recorded by the smart glasses' microphone. The input here is the customer's additional utterances, and the output is again the recorded voice data.

[1264] Step 7:

[1265] The server analyzes the re-collected voice data and performs learning to improve the accuracy of the system. Specifically, the customer response data and the analysis results are fed back to the machine learning model to improve future response accuracy. The input here is the re-collected voice data and the analysis results, and the output is a learning model with improved system accuracy.

[1266] 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.

[1267] This invention is a system for automating and streamlining customer service in call centers. In particular, by combining it with an emotion engine that recognizes customer emotions and adjusts responses based on those emotions, customer satisfaction can be further improved. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[1268] Overall system overview

[1269] 1. Detect incoming calls from customers

[1270] Server: Connects to the call center system and detects incoming calls in real time.

[1271] Device: Receives a call and starts recording.

[1272] User: Make a call and ask a question.

[1273] 2. Collection and analysis of audio data

[1274] Device: Records the customer's voice and sends the audio data to the server.

[1275] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[1276] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[1277] 3. Emotion analysis using an emotion engine

[1278] Server: Analyzes customer emotions and obtains emotional data using the emotion engine.

[1279] Server: Makes decisions to generate the best answer based on customer sentiment.

[1280] 4. Identifying the inquiry and generating an answer

[1281] Server: Uses a natural language processing (NLP) engine to identify customer inquiries from text data.

[1282] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[1283] Server: Dynamically adjust responses based on customer sentiment data.

[1284] 5. Providing answers

[1285] Server: Uses speech synthesis technology to convert text data into voice data.

[1286] Terminal: Plays the generated voice data and provides the answer to the customer.

[1287] 6. Feedback and Learning

[1288] Terminal: Collects customer responses again and sends them to the server.

[1289] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[1290] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[1291] As a specific processing example, consider the following scenario.

[1292] Example: Inquiry about product malfunction

[1293] 1. Receiving an incoming call

[1294] User: "My product isn't working and I'd like to check."

[1295] Server: Detects incoming calls and collects customer voice.

[1296] 2. Converting audio data into text

[1297] Terminal: Records what the customer says and sends the audio data to the server.

[1298] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[1299] 3. Emotion analysis

[1300] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions.

[1301] Server: Determines that the customer is in trouble and stores the emotion data.

[1302] 4. Query Analysis

[1303] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[1304] Server: Considers customer sentiment and searches for solutions in product manuals.

[1305] 5. Answer Generation

[1306] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1307] Servers: Use polite language that takes into account the customer's confused feelings.

[1308] 6. Providing answers

[1309] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1310] Terminal: Plays the generated voice data and provides the answer to the customer.

[1311] 7. Analyzing and learning from customer responses

[1312] User: "The model number is ABC123. The serial number is XYZ890."

[1313] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1314] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1315] Server: Extract the product model number and serial number from the new text data.

[1316] Server: Uses the extracted information to retrieve product details from the database.

[1317] Server: Provides detailed fault diagnosis and updated response as needed.

[1318] 8. Feedback and Data Storage

[1319] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[1320] Server: Emotional data is also stored and used for future improvements.

[1321] In this way, the server, terminals, and users work together to realize an automated voice response system that provides prompt and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[1322] The processing flow will be explained below.

[1323] Step 1:

[1324] User: Calls and says, "My product isn't working and I'd like to check it out."

[1325] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[1326] Terminal: Receives incoming calls and begins recording customer conversations.

[1327] Step 2:

[1328] Device: Sends recorded audio data to the server in real time.

[1329] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[1330] Server: Obtain the text data "My product is not working, so I would like to check it."

[1331] Step 3:

[1332] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[1333] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[1334] Server: Based on the text data and tone of voice, determine that the customer's emotion is "confused."

[1335] Step 4:

[1336] Server: Manages customer emotion data using the emotion engine.

[1337] Server: Makes decisions to generate the best answer based on the customer's confused feelings.

[1338] Step 5:

[1339] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[1340] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1341] Step 6:

[1342] Server: Dynamically adjusts responses based on customer sentiment determined by the sentiment engine.

[1343] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1344] Terminal: Plays the generated voice data and provides the answer to the customer.

[1345] Step 7:

[1346] User: "The model number is ABC123. The serial number is XYZ890."

[1347] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1348] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1349] Step 8:

[1350] Server: Extract the product model number and serial number from the new text data.

[1351] Server: Uses the extracted information to retrieve product details from the database.

[1352] Server: Provides detailed fault diagnosis and updated response as needed.

[1353] Step 9:

[1354] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[1355] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[1356] In this way, the server, terminals, and users work together to realize an automated voice response system that provides fast and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[1357] Example 2

[1358] 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."

[1359] Customer support at call centers relies heavily on human resources, and while they require fast and accurate responses, it is difficult to fully understand customer emotions and provide appropriate responses. In particular, a lack of real-time responses that reflect customer emotions can lead to a decline in customer satisfaction. Furthermore, existing systems are difficult to handle, as they also require multilingual support and 24 / 7 service provision.

[1360] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1361] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning to improve the accuracy of the system based on the analysis, means for analyzing customer emotions and dynamically adjusting answers based on the results, and a computer terminal and database with which the above means are linked via a communications network. This enables quick and accurate automatic responses based on an understanding of customer emotions and continuous improvement of the system.

[1362] "Means for detecting incoming calls from customers" refers to a system that detects incoming calls from customers in real time and notifies other systems of that information.

[1363] "Means of collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to a system that records the voices transmitted by customers and converts that data into text information using voice recognition technology.

[1364] "Means for analyzing text data using natural language processing technology and identifying the content of the customer's inquiry" refers to a system that analyzes text information using natural language processing technology and identifies the content of the customer's intended inquiry.

[1365] "Means for generating an appropriate response based on the analysis results" refers to a mechanism that automatically creates the most appropriate and accurate response based on the analyzed inquiry content.

[1366] The "means of providing the generated answer to the customer by voice" is a mechanism that converts the automatically generated answer into voice data using voice synthesis technology and provides it to the customer.

[1367] "Means of re-collecting customer responses and learning based on them to improve the accuracy of the system" refers to a mechanism that re-records responses and reactions from customers and uses that data to run a machine learning process to improve the performance of the system.

[1368] "Means for analyzing customer emotions and dynamically adjusting responses based on the results" refers to a system that analyzes emotions from customer voice and text data and dynamically changes the content of responses based on the results of that analysis.

[1369] "Computer terminals and databases linked via a communications network" refers to an infrastructure that enables different computer terminals and databases to share information and operate in cooperation with each other using the Internet or other communications networks.

[1370] This invention is a system for automating and streamlining customer service in a call center. This system is composed of three main elements: a server, a terminal, and a user, each of which operates in a specific way.

[1371] Overall system overview

[1372] 1. Detect incoming calls from customers

[1373] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls to specific numbers.

[1374] Terminal: Receives incoming calls and prepares to record the voice of the customer. Activates the recording function the moment it receives notification of an incoming call.

[1375] User: Calls to inquire.

[1376] 2. Collection and analysis of audio data

[1377] Device: Records the customer's voice and sends it to the server as audio data. A highly sensitive microphone is used to record the customer's speech with clear sound quality.

[1378] Server: Passes the received voice data to a voice recognition API (e.g., Google Speech-to-Text) and converts it into text data. Temporarily stores the voice data and sends a request to the voice recognition API.

[1379] Server: Analyzes customer sentiment and key phrases from text data and extracts important information.

[1380] 3. Emotion analysis using an emotion engine

[1381] Server: Analyzes customer emotions using an emotion engine (e.g., IBM Watson Tone Analyzer) and obtains emotion data. Inputs the text data and tone of voice data into the emotion engine.

[1382] Server: Determines the best response based on customer sentiment. Uses sentiment analysis results as a basis for customizing responses.

[1383] 4. Identifying the inquiry and generating an answer

[1384] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer's inquiry from the text data. Sends the text data to the NLP engine, which identifies the inquiry as "product malfunction."

[1385] Server: Searches product manuals and internal databases to generate appropriate answers. Automatically searches databases for relevant information and generates answers.

[1386] Server: Dynamically adjusts responses based on customer sentiment data. Uses polite language based on sentiment data.

[1387] 5. Providing answers

[1388] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert text data into speech data. Inputs the response text into the speech synthesis API to generate speech data.

[1389] Terminal: Plays the generated voice data and provides the answer to the customer. Receives the generated voice data and plays it back for the customer to hear.

[1390] 6. Feedback and Learning

[1391] Terminal: The customer's responses are collected again and sent to the server as audio data. Any additional answers or responses from the customer are also recorded and sent to the server.

[1392] Server: Analyzes customer responses and learns to improve the accuracy of the entire system. Analyzes new data and updates the system's machine learning algorithms.

[1393] Server: Stores emotion data based on the analysis results and uses it to improve future customer service responses. The updated database is used to respond to future customer inquiries.

[1394] Specific examples

[1395] Inquiries about product malfunctions

[1396] User: "My product isn't working and I'd like to check."

[1397] Server: Detects incoming calls and collects customer voice.

[1398] Terminal: Records what the customer says and sends the audio data to the server.

[1399] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[1400] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions. Detects confused emotions and stores the emotion data.

[1401] Server: Using natural language processing technology, identify the inquiry content as "product malfunction."

[1402] Server: Considers the customer's feelings and searches for a solution in the product manual. Generates a response such as "Please let us know your product model number and serial number so we can look into it."

[1403] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1404] Terminal: Plays the generated voice data and provides the answer to the customer.

[1405] User: Responds, "The model number is ABC123. The serial number is XYZ890."

[1406] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1407] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1408] Server: Extract the product model number and serial number from the new text data. Use the extracted information to retrieve detailed product information from the database.

[1409] Server: Provides detailed fault diagnosis and updated response as needed.

[1410] Example prompt sentence:

[1411] The system analyzes a user's statement, "My product isn't working, so I'd like to check it," and uses a sentiment analysis engine to identify the customer's emotions and generate an answer.

[1412] This allows the system to provide efficient and accurate automated responses while taking into account customer emotions.

[1413] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1414] Step 1:

[1415] Detecting incoming calls from customers

[1416] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls from customers.

[1417] Input: Incoming call signal from customer

[1418] Output: Incoming call notification signal

[1419] Specific operations: Monitors the phone number and receives an incoming call signal, then notifies the device of the incoming call signal.

[1420] Step 2:

[1421] Collecting audio data and starting recording

[1422] Terminal: Receives an incoming call and activates the recording function, recording what the customer says and preparing to collect voice data with clear sound quality using a high-sensitivity microphone.

[1423] Input: Incoming call notification signal

[1424] Output: Start recording

[1425] Specific behavior: When an incoming call notification is received, the recording function is activated and the customer's voice begins to be recorded.

[1426] Step 3:

[1427] Sending audio data to the server

[1428] Terminal: Sends recorded audio data to the server in real time.

[1429] Input: Recorded audio data

[1430] Output: Send audio data to the server

[1431] Specific operation: The recorded audio data is compressed and sent to a server via internet communication.

[1432] Step 4:

[1433] Converting audio data to text

[1434] Server: Passes the voice data to a speech recognition API (e.g., Google Speech-to-Text) and converts it to text data. Temporarily stores the voice data and sends a request to the speech recognition API.

[1435] Input: Audio data sent to the server

[1436] Output: Text data generated by speech recognition

[1437] Specific operation: Sends audio data to the API and retrieves the returned text data.

[1438] Step 5:

[1439] Sentiment analysis of text data

[1440] Server: Using an emotion engine (e.g., IBM Watson Tone Analyzer), analyze customer emotions based on text data and voice tone.

[1441] Input: Text data and voice tone data from speech recognition

[1442] Output: Emotion data

[1443] Specific operation: Text data and voice tone are input into the emotion engine to obtain the emotion analysis results.

[1444] Step 6:

[1445] Identifying the content of the inquiry

[1446] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer inquiry from the text data.

[1447] Input: Text data

[1448] Output: Identified inquiry

[1449] Specific behavior: Analyze what the customer says and identify the specific inquiry.

[1450] Step 7:

[1451] Generating the right answers

[1452] Server: Generates appropriate answers based on customer inquiry content and sentiment data. Searches product manuals and internal databases to generate answers.

[1453] Input: Identified query content and sentiment data

[1454] Output: Generated answer text

[1455] What it does: Searches a database, extracts the relevant information, and creates an answer.

[1456] Step 8:

[1457] Converting responses into audio data

[1458] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert the generated answer text into voice data.

[1459] Input: Generated answer text

[1460] Output: Audio data

[1461] Specific operation: The answer text is input into the speech synthesis API to generate voice data.

[1462] Step 9:

[1463] Providing answers

[1464] Terminal: Plays the generated voice data and provides the answer to the customer.

[1465] Input: Audio data

[1466] Output: A spoken response to the customer

[1467] Specific behavior: Plays back the audio data and provides the appropriate answer to the customer.

[1468] Step 10:

[1469] Collecting and sending customer responses

[1470] Terminal: Record the customer's response again and send it to the server.

[1471] Input: Customer response voice

[1472] Output: Sends response audio data to the server

[1473] What it does: Makes additional recordings and sends the collected data to a server.

[1474] Step 11:

[1475] Customer response analysis and system learning

[1476] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[1477] Input: Customer response voice data

[1478] Output: Training data for accuracy improvement

[1479] What it does: Analyzes response data and feeds it into the learning algorithm, continuously improving the system.

[1480] (Application example 2)

[1481] 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."

[1482] In autonomous vehicles, the lack of technology to adjust the in-car environment taking into account passenger emotions makes it difficult to improve passenger comfort and satisfaction. In addition, there is no system to provide immediate and appropriate guidance or responses when passengers are confused or inconvenienced, which can lead to a poor user experience.

[1483] The identification processing 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 detecting an incoming call from a customer, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer in voice, means for re-collecting customer responses and performing learning based on the collected responses to improve the accuracy of the system, means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology, means for adjusting the in-vehicle environment based on the voice input and the emotion analysis results, and means for presenting emotion-adjusted feedback using voice synthesis technology. This makes it possible to adjust the in-vehicle environment taking passengers' emotions into consideration.

[1484] "Means for detecting incoming calls from customers" refers to devices or functions that recognize phone calls or voice input from customers in real time and input that data into the system.

[1485] "Means for collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to devices or algorithms that record customer voice and convert it into text data using voice recognition technology.

[1486] "Means of analyzing text data using natural language processing technology and identifying the content of customer inquiries" refers to technology that analyzes text data using a natural language processing algorithm and clarifies the specific content of customer inquiries from the content.

[1487] "Means for generating an appropriate response based on the analysis results" refers to a function for creating an optimal response from a pre-programmed database or algorithm based on the analysis results of the inquiry content.

[1488] The "means for providing the generated answer to the customer by voice" refers to a device or technology that converts the generated text-based answer into voice data using voice synthesis technology and conveys it to the customer.

[1489] "Means for re-collecting customer responses and learning to improve the accuracy of the system based on that" refers to a self-learning function that re-collects and analyzes feedback from customers and uses that to improve the system's performance and response accuracy.

[1490] "Means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology" refers to a function for recording passenger voice data, determining emotions from the voice using emotion analysis technology, and adjusting the in-car entertainment system based on the determination result.

[1491] "Means for adjusting the in-car environment based on voice input and emotion analysis results" is a function for automatically adjusting the temperature, music, lighting, etc. inside the car based on the passenger's voice input and the results of emotion analysis.

[1492] The "means for presenting emotion-adjusted feedback using voice synthesis technology" is a function for expressing optimal feedback based on the analyzed emotions in voice using voice synthesis technology and providing it to the occupant.

[1493] This invention relates to an infotainment system for autonomous vehicles. The system collects and analyzes passenger voice data using emotion analysis technology, and automatically adjusts the operation of the infotainment system and the in-vehicle environment based on the results. This system uses the following hardware and software:

[1494] Hardware

[1495] 1. Infotainment system: An entertainment and information system installed in a vehicle.

[1496] 2. Microphone: A device for collecting the voices of the passengers.

[1497] 3. Speaker: A device for providing audio feedback.

[1498] software

[1499] 1. Speech recognition API: Converts passenger voice data into text data (e.g., Google Cloud Speech-to-Text).

[1500] 2. Sentiment analysis API: Analyzes text data and voice tone to identify emotions (e.g., IBM Watson Tone Analyzer).

[1501] 3. Natural language processing engine: processes the parsed text data and understands the passenger's intent (e.g., spaCy, NLTK).

[1502] 4. Text-to-speech API: Converts text-based responses into audio data (e.g., Amazon Polly).

[1503] System Operation

[1504] First, the passenger's voice data is collected through a microphone. The collected voice data is sent to a speech recognition API and converted into text data. This text data is then sent to an emotion analysis API to analyze the passenger's emotions. The analyzed emotion data is then analyzed by a natural language processing engine to determine the appropriate feedback action (for example, adjusting the temperature or changing the music).

[1505] Processing Description

[1506] 1. Voice data collection: The microphone collects the passenger's voice and sends it to the voice recognition API.

[1507] 2. Speech recognition: The speech recognition API converts the voice data into text data.

[1508] 3. Emotion analysis: Text data and voice tone are input into the emotion analysis API to analyze the passenger's emotions.

[1509] 4. Natural Language Processing: The natural language processing engine determines the appropriate response or action based on the analyzed emotional data.

[1510] 5. Environmental Adjustment and Feedback: Based on the determined responses and actions, the infotainment system adjusts the in-car environment and provides audible feedback of the adjustments to the occupants using a speech synthesis API.

[1511] Specific examples

[1512] For example, if a passenger says, "It's a little cold," the voice data is converted into text data saying "It's a little cold" by the speech recognition API. This text data and voice tone are analyzed by the emotion analysis API to determine that the passenger is feeling cold. Next, the natural language processing engine determines that the passenger is "feeling cold" and issues a command to raise the interior temperature by 2 degrees. The speech synthesis API is also used to provide feedback to the passenger, saying, "The interior temperature has been raised a little."

[1513] Prompt Sentence Examples

[1514] The server was told by voice, "It's a little cold."

[1515] The infotainment system analyzed the voice input text and sent it to the sentiment analysis API, which output the sentiment analysis result as "I feel cold."

[1516] The driver has been instructed to increase the interior temperature of the self-driving vehicle by 2 degrees.

[1517] "The interior temperature has been slightly increased," the voice synthesis API informed the occupants.

[1518] In this way, the various hardware and software components work together to provide passengers with a comfortable in-vehicle environment and appropriate feedback.

[1519] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1520] Step 1:

[1521] The user's voice data is collected through a microphone and sent to the device. Specifically, the infotainment system detects voice input and starts recording. The input is the user's voice, and the recorded voice file is generated as the output.

[1522] Step 2:

[1523] The device sends the collected voice data in real time to a voice recognition API, which converts the voice data into text data (e.g., Google Cloud Speech-to-Text). The input is an audio file, and the output is text data.

[1524] Step 3:

[1525] The server sends the text data and voice tone data to an emotion analysis API to analyze the customer's emotions. The emotion analysis API (e.g., IBM Watson Tone Analyzer) analyzes the text data and voice tone to generate emotion data. The input is text and voice tone, and the output is emotion data.

[1526] Step 4:

[1527] The server uses natural language processing technology to analyze the text data and emotion data to identify the customer's inquiry. The natural language processing engine (e.g., spaCy, NLTK) analyzes the text data and determines the appropriate response. The input is text data and emotion data, and the analysis results are generated as the output.

[1528] Step 5:

[1529] The server determines the appropriate response based on the analysis results and emotion data, and issues instructions to the infotainment system. For example, this could include raising the temperature inside the car by 2 degrees or changing the music. The inputs are the analysis results and emotion data, and the output is the setting data for the infotainment system.

[1530] Step 6:

[1531] The device adjusts the infotainment system according to instructions from the server, for example, adjusting the temperature or changing the music inside the car. The input is the setting data from the server, and the output is the adjusted in-car environment.

[1532] Step 7:

[1533] The server sends the generated settings to a speech synthesis API, which generates voice feedback. The speech synthesis API (e.g., Amazon Polly) converts text data into voice data. The input is the text data of the settings, and the output is voice feedback data.

[1534] Step 8:

[1535] The terminal plays back the generated voice feedback data and provides it to the user. This results in voice guidance such as "The interior temperature has been slightly increased." The voice feedback data is input, and voice guidance for the user is generated as output.

[1536] 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.

[1537] 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.

[1538] 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.

[1539] [Fourth embodiment]

[1540] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1541] 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.

[1542] 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).

[1543] 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.

[1544] 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.

[1545] 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).

[1546] 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.

[1547] 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.

[1548] 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.

[1549] 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.

[1550] 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.

[1551] 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.

[1552] 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."

[1553] This invention is a system for automating and streamlining customer service in a call center. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[1554] Overall system overview

[1555] 1. Detect incoming calls from customers

[1556] Server: Connects to the call center system and detects incoming calls in real time.

[1557] Device: Receives a call and starts recording.

[1558] User: Make a call and ask a question.

[1559] 2. Collection and analysis of audio data

[1560] Device: Records the customer's voice and sends the audio data to the server.

[1561] Server: Uses voice recognition technology to convert voice data into text data.

[1562] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[1563] 3. Identifying the inquiry and generating an answer

[1564] Server: Uses natural language processing technology to identify customer inquiries from text data.

[1565] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[1566] 4. Providing answers

[1567] Server: Uses speech synthesis technology to convert text data into voice data.

[1568] Terminal: Plays back the audio data and provides the answer to the customer.

[1569] 5. Feedback and learning

[1570] Terminal: Collects customer responses again and sends them to the server.

[1571] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[1572] As a specific processing example, consider the following scenario.

[1573] Example: Inquiry about product malfunction

[1574] 1. Receiving an incoming call

[1575] User: "My product isn't working and I'd like to check."

[1576] Server: Detects incoming calls and collects customer voice.

[1577] 2. Converting audio data into text

[1578] Terminal: Records what the customer says and sends the audio data to the server.

[1579] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[1580] 3. Query Analysis

[1581] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[1582] Server: Determines that the customer has a problem and searches the product manual for a solution.

[1583] 4. Answer Generation

[1584] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1585] Terminal: Speaks the generated answer to the customer.

[1586] 5. Analyzing and learning from customer responses

[1587] User: "The model number is ABC123. The serial number is XYZ890."

[1588] Terminal: Records customer responses and sends the audio data to the server.

[1589] Server: Analyzes the new voice data and retrieves product details.

[1590] Server: We will update our response with new information and provide further details as needed.

[1591] In this way, the server and terminals work together to create an automated response system that improves the efficiency of call center operations. This system is also multilingual, can operate 24 hours a day, 365 days a year, and is highly scalable.

[1592] The processing flow will be explained below.

[1593] Step 1:

[1594] User: Calls and says, "My product isn't working and I'd like to check it out."

[1595] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[1596] Terminal: Receives incoming calls and begins recording customer conversations.

[1597] Step 2:

[1598] Device: Sends recorded audio data to the server in real time.

[1599] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[1600] Server: Obtain the text data "My product is not working, so I would like to check it."

[1601] Step 3:

[1602] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[1603] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[1604] Server: Analyzes the customer's tone and language and determines the emotion is "confused."

[1605] Step 4:

[1606] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[1607] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1608] Step 5:

[1609] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1610] Terminal: Plays the generated voice data and provides the answer to the customer.

[1611] Step 6:

[1612] User: "The model number is ABC123. The serial number is XYZ890."

[1613] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1614] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1615] Step 7:

[1616] Server: Extract the product model number and serial number from the new text data.

[1617] Server: Uses the extracted information to retrieve product details from the database.

[1618] Server: Provides detailed fault diagnosis and updated response as needed.

[1619] Step 8:

[1620] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[1621] Terminal: Displays customer inquiry history and response content, and can be used to respond to future inquiries.

[1622] In this way, the server, terminal, and user work together to realize an automated voice response system that provides quick and accurate responses.

[1623] Example 1

[1624] 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."

[1625] In conventional call center operations, a large number of staff members have to deal with customers, requiring a great deal of effort and time, resulting in problems of inefficiency and human error. Furthermore, responding to customer inquiries in a timely and appropriate manner requires advanced skills and specialized knowledge, creating challenges in human resource development. To solve these problems and improve customer satisfaction, a more efficient and highly accurate automated response system is needed.

[1626] 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.

[1627] In this invention, the server includes a means for detecting incoming calls from customers, a means for collecting customer voice data in real time and converting it into text data using speech recognition technology, and a means for analyzing the text data using natural language processing technology to identify the content of the customer's inquiry. This enables an automated response that responds to customer inquiries in a timely and appropriate manner. Furthermore, by providing a means for recollecting customer responses and improving the accuracy of the system based on them, highly accurate responses based on the latest information are always possible. Furthermore, by providing a means for converting text data into speech data using speech synthesis technology and providing it to the customer audibly, more natural dialogue is realized.

[1628] "Means for detecting incoming calls" refers to a device or function that connects to a telephone system or network and recognizes incoming calls from customers in real time.

[1629] "Voice data" refers to information that has been recorded in digital form from a customer's voice.

[1630] "Speech recognition technology" is a technology for converting voice data into text data.

[1631] "Text data" is information in text format generated from voice data using voice recognition technology.

[1632] "Natural language processing technology" is a technology that analyzes text data, understands meaning and emotions, and generates appropriate responses.

[1633] "Means for identifying inquiry content" refers to the function of using natural language processing technology to identify specific customer requests or problems from text data.

[1634] An "answer generation means" is a device or function that generates an appropriate response or solution based on the identified inquiry.

[1635] "Speech synthesis technology" is a technology that converts text data into voice data.

[1636] The "means for providing to the customer" is a device or function for playing back the generated voice data to the customer.

[1637] A "means for recollecting customer responses" is a device or function that records new voice data that the customer gives in response to the answers provided and transmits it to the system.

[1638] "Learning means to improve system accuracy" refers to a function that analyzes collected customer response data and updates the model using machine learning technology to improve system performance.

[1639] A "means for establishing a call session" is a device or function that establishes a communication channel between a customer and a system, enabling two-way communication of voice data.

[1640] "Means for searching for relevant information from a database" refers to a search engine or function for quickly finding data related to the inquiry from stored information.

[1641] This invention is a system for automating and streamlining customer service in call center operations. The system is broadly composed of three elements: a server, a terminal, and a user. How each element specifically operates is described in detail below.

[1642] Hardware and software used

[1643] Server: A high-performance server is used to perform speech recognition, natural language processing, speech synthesis, data analysis, etc. Specifically, technologies such as Google Cloud Speech-to-Text, spaCy, Hugging Face's BERT model, ElasticSearch, Amazon Polly, and TensorFlow are used.

[1644] Terminal: A terminal equipped with a SIP (Session Initiation Protocol) client has the function of recording the customer's voice and sending it to the server.

[1645] User: The customer making the call.

[1646] Details of system processing

[1647] 1. Detecting incoming calls from customers and establishing call sessions

[1648] Server: Connects to the VoIP server to detect incoming calls from customers in real time and notify the terminal.

[1649] Device: Upon receiving the notification, it establishes a call session with the customer and begins recording audio.

[1650] User: Make a call and start an inquiry.

[1651] 2. Collecting and converting audio data into text

[1652] Terminal: Continuously streams the customer's voice to the server in real time.

[1653] Server: Sends the received audio data to the Google Cloud Speech-to-Text API to generate text data.

[1654] 3. Analyzing text data and identifying inquiry content

[1655] Server: Analyze the generated text data using the spaCy library to extract key phrases and sentiment information.

[1656] Server: Leverages Hugging Face's BERT model to identify specific queries based on key phrases.

[1657] 4. Answer Generation and Search

[1658] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals.

[1659] Server: Generates appropriate answers for customers based on the retrieved information.

[1660] 5. Conversion into audio data and provision

[1661] Server: The generated text data is converted into voice data using Amazon Polly.

[1662] Terminal: Plays the converted voice data and provides the answer to the customer.

[1663] 6. Collect and learn from customer responses

[1664] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server.

[1665] Server: Analyzes new audio data and uses TensorFlow to update the model to improve the accuracy of the system.

[1666] Specific examples

[1667] Below is a scenario for an inquiry about a product failure.

[1668] scenario

[1669] 1. Detecting incoming calls and establishing a call session

[1670] User: "My product isn't working, so I'd like to check it out."

[1671] Server: The VoIP server detects the incoming call and notifies the terminal.

[1672] Device: Establish a call session and start recording.

[1673] 2. Collecting and converting audio data into text

[1674] Terminal: Records what the customer says and sends the audio data to the server.

[1675] Server: Use Google Cloud Speech-to-Text to generate text data such as "My product isn't working, so I'd like to check it."

[1676] 3. Parsing the query and generating an answer

[1677] Server: Uses the spaCy library to extract the key phrase "product malfunction" from text data, and then uses the BERT model to identify the specific content.

[1678] Server: Use ElasticSearch to search the product manual for solutions related to the product failure. Based on the search results, generate a response such as "Please let us know the product model number and serial number so we can look into it."

[1679] 4. Conversion into audio data and provision

[1680] Server: Convert the response text into speech data using Amazon Polly.

[1681] Terminal: Plays back the audio data and provides the answer to the customer.

[1682] 5. Analyzing and learning from customer responses

[1683] User: "The model number is ABC123. The serial number is XYZ890."

[1684] Device: Record any additional comments and send them back to the server.

[1685] Server: Analyzes new audio data and retrieves product details, while simultaneously updating the model in TensorFlow to improve the system's accuracy.

[1686] Prompt Sentence Examples

[1687] "Generate a natural response for the customer who asks for the product's model number and serial number."

[1688] "When a customer calls to inquire about a malfunctioning product, generate the best response."

[1689] In this way, the server and terminals work together to create a highly efficient and accurate automatic response system for customer inquiries.

[1690] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1691] Step 1: Detecting an incoming call from a customer and establishing a call session

[1692] Server: Connects to the VoIP server and detects incoming calls from customers in real time. The input is the call signal from the customer, and the output is the call detection information.

[1693] Terminal: Upon receiving a notification from the server, it establishes a call session with the customer using a SIP client and starts recording the audio. The input is the notification from the server, and the output is the establishment of a call session and the start of recording.

[1694] User: Makes a call and initiates a query. The input is the user making the call, and the output is the initiation of the call.

[1695] Step 2: Collect and send audio data

[1696] Terminal: Collects voice data in real time during a call with a customer and streams the voice data to a server. The input is the voice signal, and the output is the voice data sent to the server.

[1697] Server: Sends the received voice data to the Google Cloud Speech-to-Text API and converts the voice data to text data. The input is voice data from the device, and the output is text data.

[1698] Step 3: Analyzing text data and identifying inquiry content

[1699] Server: The generated text data is analyzed using the spaCy library to extract key phrases and analyze customer sentiment. The input is text data, and the output is key phrases and sentiment information.

[1700] Server: Uses Hugging Face's BERT model to identify the specific query from the extracted key phrases. The input is the key phrase and the output is the query.

[1701] Step 4: Generate answers and find related information

[1702] Server: Based on the identified query, ElasticSearch is used to search for relevant information from product manuals and proposals. The input is the query and the output is the relevant information.

[1703] Server: Generates appropriate answers for customers based on the retrieved information. The input is the relevant information and the output is the answer text.

[1704] Step 5: Converting the response into audio data and providing it

[1705] Server: Use Amazon Polly to convert the generated answer text data into speech data. The input is the answer text and the output is speech data.

[1706] Terminal: Plays back the voice data received from the server and provides answers to the customer. The input is the voice data, and the output is the played voice.

[1707] Step 6: Collect and send customer responses

[1708] Terminal: Records the customer's additional comments and reactions again and sends the audio data to the server. The input is the customer's speech and the output is the recorded audio data.

[1709] User: Checks the answer and provides a follow-up question or additional information. The input is the answer provided and the output is the additional utterance.

[1710] Step 7: Analyze the data and train the system

[1711] Server: Analyzes new audio data received and uses machine learning algorithms to update the model to improve the accuracy of the system. The input is the new audio data and the output is the updated machine learning model.

[1712] (Application example 1)

[1713] 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."

[1714] Improving the quality and efficiency of customer service is a key challenge in modern brick-and-mortar stores. Providing prompt and appropriate responses to customers is particularly difficult during busy times, potentially leading to lower customer satisfaction. Another issue is the need for a large number of staff, which increases labor costs. Furthermore, providing accurate information in response to customer inquiries relies heavily on the knowledge and experience of staff. Given these circumstances, there is a need for systems that use voice recognition and natural language processing technologies to automate customer service in real time via smart devices, providing efficient, high-quality service.

[1715] 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.

[1716] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning based on the collected responses to improve the accuracy of the system, and means for collecting customer inquiries in real time via a smart device and generating and providing an appropriate answer, thereby enabling more efficient customer service and the provision of high-quality services.

[1717] A "server" is a central computer system that detects incoming calls from customers, collects voice data, analyzes it, and generates responses.

[1718] "Smart devices" are devices that have advanced functionality, primarily smart glasses and tablets, and interact through voice recognition and speech synthesis.

[1719] "Means for detecting incoming calls" refers to technology that recognizes when a call or voice message is received from a customer and uses that as a trigger to activate the system.

[1720] "Speech recognition technology" is a technology that converts voice data into text data.

[1721] "Natural language processing technology" is a technology that analyzes text data and understands customer intentions and emotions.

[1722] "Means for generating appropriate answers based on the analysis results" refers to technology that generates answers to customer questions based on the analyzed text data.

[1723] The "means for providing the generated answer to the customer by voice" is a technology for converting the generated text answer into voice data and transmitting it to the customer.

[1724] "Means of collecting customer responses again and learning based on that to improve the accuracy of the system" refers to a technology that collects customer responses and uses machine learning to improve the system's responses.

[1725] The system that realizes this application example is composed of the following programs and a combination of hardware and software.

[1726] Program Generation

[1727] The server contains programs that do the following:

[1728] 1. Detect incoming calls from customers.

[1729] 2. Collect customer voice data in real time and convert it into text data using voice recognition technology.

[1730] 3. Analyze the text data using natural language processing technology to identify the customer's inquiry.

[1731] 4. Generate appropriate answers based on the analysis results.

[1732] 5. Provide generated answers to customers via voice.

[1733] 6. Customer responses are collected again and the system learns from them to improve its accuracy.

[1734] 7. Collect customer inquiries in real time via smart devices and generate and provide appropriate answers.

[1735] Processing Description

[1736] Hardware / Software used

[1737] Hardware: Smart glasses (microphone, camera, speaker), server (high-performance processing unit)

[1738] Software: Python programming language, speech_recognition module, gTTS (Google Text-to-Speech) module, openai API

[1739] Data processing and calculation

[1740] 1. The server uses a speech recognition module to convert the customer's speech into text data in real time. Specifically, the server acquires the speech data using the speech_recognition module and outputs it as text data.

[1741] 2. The server uses natural language processing technology to analyze the converted text data. It uses the OpenAI API to analyze the text data and generate an appropriate answer. It uses the prompt sentence to input the data into the generative AI model to obtain the answer.

[1742] 3. The server converts the generated text response into voice data using the gTTS module, which allows the generated response to be output as voice and played through the smart glasses' speaker.

[1743] Adding specific examples

[1744] For example, consider the following scenario:

[1745] Customer: "Is this product waterproof?"

[1746] The camera in the smart glasses recognizes the customer and the microphone records their questions.

[1747] The server converts the voice data into text, such as "Is this product waterproof?"

[1748] The prompt sentence "Is this product waterproof?" was input into the generative AI model via the OpenAI API.

[1749] The server generates the response: "Yes, this product is IP67 waterproof."

[1750] gTTS converts text to speech and plays it through the smart glasses' speakers.

[1751] Example prompt:

[1752] Product feature questions: "Is this product waterproof?"

[1753] A: Yes, this product is IP67 rated waterproof.

[1754] In this way, it becomes possible to improve the efficiency of customer service and provide high-quality services in physical stores.

[1755] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1756] Step 1:

[1757] The server detects when a customer enters the store using the camera and microphone of the smart glasses. When the customer starts speaking, the smart glasses collect voice data. The input here is the customer's voice utterance, and the output is the recorded voice data.

[1758] Step 2:

[1759] The server converts the collected voice data into text data using the speech_recognition module. Specifically, it analyzes the voice data using speech recognition technology and generates text data such as "Is this product waterproof?". The input here is the recorded voice data, and the output is text data.

[1760] Step 3:

[1761] The server uses the OpenAI API to analyze the generated text data. A prompt sentence is constructed and input into the generative AI model to generate an appropriate answer. The input here is the text data and the prompt sentence, and the output is the generated answer text. An example of a prompt sentence is "Question about product features: Is this product waterproof?"

[1762] Step 4:

[1763] The server converts the generated answer text into voice data using the gTTS module. Specifically, it outputs the answer text as voice data using speech synthesis technology. The input here is the generated answer text, and the output is voice data.

[1764] Step 5:

[1765] The smart glasses at the terminal play the generated voice data and provide a response to the customer, where the input is the voice data and the output is the voice information provided to the customer.

[1766] Step 6:

[1767] The user responds to the server's answers and asks additional questions or comments as needed. Their responses are again recorded by the smart glasses' microphone. The input here is the customer's additional utterances, and the output is again the recorded voice data.

[1768] Step 7:

[1769] The server analyzes the re-collected voice data and performs learning to improve the accuracy of the system. Specifically, the customer response data and the analysis results are fed back to the machine learning model to improve future response accuracy. The input here is the re-collected voice data and the analysis results, and the output is a learning model with improved system accuracy.

[1770] 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.

[1771] This invention is a system for automating and streamlining customer service in call centers. In particular, by combining it with an emotion engine that recognizes customer emotions and adjusts responses based on those emotions, customer satisfaction can be further improved. This system consists of three main elements: a server, a terminal, and a user. The specific operation of each element is explained below.

[1772] Overall system overview

[1773] 1. Detect incoming calls from customers

[1774] Server: Connects to the call center system and detects incoming calls in real time.

[1775] Device: Receives a call and starts recording.

[1776] User: Make a call and ask a question.

[1777] 2. Collection and analysis of audio data

[1778] Device: Records the customer's voice and sends the audio data to the server.

[1779] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[1780] Server: Analyzes and extracts customer sentiment and key phrases from text data.

[1781] 3. Emotion analysis using an emotion engine

[1782] Server: Analyzes customer emotions and obtains emotional data using the emotion engine.

[1783] Server: Makes decisions to generate the best answer based on customer sentiment.

[1784] 4. Identifying the inquiry and generating an answer

[1785] Server: Uses a natural language processing (NLP) engine to identify customer inquiries from text data.

[1786] Server: Searches for relevant information in product manuals and proposals and generates appropriate answers.

[1787] Server: Dynamically adjust responses based on customer sentiment data.

[1788] 5. Providing answers

[1789] Server: Uses speech synthesis technology to convert text data into voice data.

[1790] Terminal: Plays the generated voice data and provides the answer to the customer.

[1791] 6. Feedback and Learning

[1792] Terminal: Collects customer responses again and sends them to the server.

[1793] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[1794] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[1795] As a specific processing example, consider the following scenario.

[1796] Example: Inquiry about product malfunction

[1797] 1. Receiving an incoming call

[1798] User: "My product isn't working and I'd like to check."

[1799] Server: Detects incoming calls and collects customer voice.

[1800] 2. Converting audio data into text

[1801] Terminal: Records what the customer says and sends the audio data to the server.

[1802] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[1803] 3. Emotion analysis

[1804] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions.

[1805] Server: Determines that the customer is in trouble and stores the emotion data.

[1806] 4. Query Analysis

[1807] Server: Using natural language processing technology, the server analyzes the text data and identifies the inquiry as "product malfunction."

[1808] Server: Considers customer sentiment and searches for solutions in product manuals.

[1809] 5. Answer Generation

[1810] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1811] Servers: Use polite language that takes into account the customer's confused feelings.

[1812] 6. Providing answers

[1813] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1814] Terminal: Plays the generated voice data and provides the answer to the customer.

[1815] 7. Analyzing and learning from customer responses

[1816] User: "The model number is ABC123. The serial number is XYZ890."

[1817] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1818] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1819] Server: Extract the product model number and serial number from the new text data.

[1820] Server: Uses the extracted information to retrieve product details from the database.

[1821] Server: Provides detailed fault diagnosis and updated response as needed.

[1822] 8. Feedback and Data Storage

[1823] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[1824] Server: Emotional data is also stored and used for future improvements.

[1825] In this way, the server, terminals, and users work together to realize an automated voice response system that provides prompt and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[1826] The processing flow will be explained below.

[1827] Step 1:

[1828] User: Calls and says, "My product isn't working and I'd like to check it out."

[1829] Server: Detects incoming calls to the call center system and confirms that a new call has been received.

[1830] Terminal: Receives incoming calls and begins recording customer conversations.

[1831] Step 2:

[1832] Device: Sends recorded audio data to the server in real time.

[1833] Server: Passes the received voice data to the voice recognition API and converts it into text data.

[1834] Server: Obtain the text data "My product is not working, so I would like to check it."

[1835] Step 3:

[1836] Server: The acquired text data is passed to a natural language processing (NLP) engine for analysis.

[1837] Server: As a result of the analysis, the "inquiry regarding product malfunction" is identified.

[1838] Server: Based on the text data and tone of voice, determine that the customer's emotion is "confused."

[1839] Step 4:

[1840] Server: Manages customer emotion data using the emotion engine.

[1841] Server: Makes decisions to generate the best answer based on the customer's confused feelings.

[1842] Step 5:

[1843] Server: Searches a database of product manuals and proposals to retrieve information related to "product malfunction."

[1844] Server: Generate a response saying, "Please let us know your product model number and serial number so we can look into it."

[1845] Step 6:

[1846] Server: Dynamically adjusts responses based on customer sentiment determined by the sentiment engine.

[1847] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1848] Terminal: Plays the generated voice data and provides the answer to the customer.

[1849] Step 7:

[1850] User: "The model number is ABC123. The serial number is XYZ890."

[1851] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1852] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1853] Step 8:

[1854] Server: Extract the product model number and serial number from the new text data.

[1855] Server: Uses the extracted information to retrieve product details from the database.

[1856] Server: Provides detailed fault diagnosis and updated response as needed.

[1857] Step 9:

[1858] Server: Accumulates data on customer responses, inquiries, and answers provided, and uses this data as learning data to improve the accuracy of the system.

[1859] Server: Based on the analysis results, emotional data is saved and used to improve future response to inquiries.

[1860] In this way, the server, terminals, and users work together to realize an automated voice response system that provides fast and accurate responses.The use of an emotion engine is a feature of the present invention, as it allows for understanding customer emotions and providing more personalized service.

[1861] Example 2

[1862] 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."

[1863] Customer support at call centers relies heavily on human resources, and while they require fast and accurate responses, it is difficult to fully understand customer emotions and provide appropriate responses. In particular, a lack of real-time responses that reflect customer emotions can lead to a decline in customer satisfaction. Furthermore, existing systems are difficult to handle, as they also require multilingual support and 24 / 7 service provision.

[1864] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1865] In this invention, the server includes means for detecting incoming calls from customers, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the content of the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer by voice, means for re-collecting customer responses and learning to improve the accuracy of the system based on the analysis, means for analyzing customer emotions and dynamically adjusting answers based on the results, and a computer terminal and database with which the above means are linked via a communications network. This enables quick and accurate automatic responses based on an understanding of customer emotions and continuous improvement of the system.

[1866] "Means for detecting incoming calls from customers" refers to a system that detects incoming calls from customers in real time and notifies other systems of that information.

[1867] "Means of collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to a system that records the voices transmitted by customers and converts that data into text information using voice recognition technology.

[1868] "Means for analyzing text data using natural language processing technology and identifying the content of the customer's inquiry" refers to a system that analyzes text information using natural language processing technology and identifies the content of the customer's intended inquiry.

[1869] "Means for generating an appropriate response based on the analysis results" refers to a mechanism that automatically creates the most appropriate and accurate response based on the analyzed inquiry content.

[1870] The "means of providing the generated answer to the customer by voice" is a mechanism that converts the automatically generated answer into voice data using voice synthesis technology and provides it to the customer.

[1871] "Means of re-collecting customer responses and learning based on them to improve the accuracy of the system" refers to a mechanism that re-records responses and reactions from customers and uses that data to run a machine learning process to improve the performance of the system.

[1872] "Means for analyzing customer emotions and dynamically adjusting responses based on the results" refers to a system that analyzes emotions from customer voice and text data and dynamically changes the content of responses based on the results of that analysis.

[1873] "Computer terminals and databases linked via a communications network" refers to an infrastructure that enables different computer terminals and databases to share information and operate in cooperation with each other using the Internet or other communications networks.

[1874] This invention is a system for automating and streamlining customer service in a call center. This system is composed of three main elements: a server, a terminal, and a user, each of which operates in a specific way.

[1875] Overall system overview

[1876] 1. Detect incoming calls from customers

[1877] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls to specific numbers.

[1878] Terminal: Receives incoming calls and prepares to record the voice of the customer. Activates the recording function the moment it receives notification of an incoming call.

[1879] User: Calls to inquire.

[1880] 2. Collection and analysis of audio data

[1881] Device: Records the customer's voice and sends it to the server as audio data. A highly sensitive microphone is used to record the customer's speech with clear sound quality.

[1882] Server: Passes the received voice data to a voice recognition API (e.g., Google Speech-to-Text) and converts it into text data. Temporarily stores the voice data and sends a request to the voice recognition API.

[1883] Server: Analyzes customer sentiment and key phrases from text data and extracts important information.

[1884] 3. Emotion analysis using an emotion engine

[1885] Server: Analyzes customer emotions using an emotion engine (e.g., IBM Watson Tone Analyzer) and obtains emotion data. Inputs the text data and tone of voice data into the emotion engine.

[1886] Server: Determines the best response based on customer sentiment. Uses sentiment analysis results as a basis for customizing responses.

[1887] 4. Identifying the inquiry and generating an answer

[1888] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer's inquiry from the text data. Sends the text data to the NLP engine, which identifies the inquiry as "product malfunction."

[1889] Server: Searches product manuals and internal databases to generate appropriate answers. Automatically searches databases for relevant information and generates answers.

[1890] Server: Dynamically adjusts responses based on customer sentiment data. Uses polite language based on sentiment data.

[1891] 5. Providing answers

[1892] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert text data into speech data. Inputs the response text into the speech synthesis API to generate speech data.

[1893] Terminal: Plays the generated voice data and provides the answer to the customer. Receives the generated voice data and plays it back for the customer to hear.

[1894] 6. Feedback and Learning

[1895] Terminal: The customer's responses are collected again and sent to the server as audio data. Any additional answers or responses from the customer are also recorded and sent to the server.

[1896] Server: Analyzes customer responses and learns to improve the accuracy of the entire system. Analyzes new data and updates the system's machine learning algorithms.

[1897] Server: Stores emotion data based on the analysis results and uses it to improve future customer service responses. The updated database is used to respond to future customer inquiries.

[1898] Specific examples

[1899] Inquiries about product malfunctions

[1900] User: "My product isn't working and I'd like to check."

[1901] Server: Detects incoming calls and collects customer voice.

[1902] Terminal: Records what the customer says and sends the audio data to the server.

[1903] Server: Using voice recognition technology, generate text data such as "My product isn't working, so I'd like to check it."

[1904] Server: Inputs text data and voice tone into the emotion engine to analyze customer emotions. Detects confused emotions and stores the emotion data.

[1905] Server: Using natural language processing technology, identify the inquiry content as "product malfunction."

[1906] Server: Considers the customer's feelings and searches for a solution in the product manual. Generates a response such as "Please let us know your product model number and serial number so we can look into it."

[1907] Server: Pass the answer text to the speech synthesis API and convert it into voice data.

[1908] Terminal: Plays the generated voice data and provides the answer to the customer.

[1909] User: Responds, "The model number is ABC123. The serial number is XYZ890."

[1910] Terminal: Any additional responses from the customer are also recorded and sent to the server.

[1911] Server: Pass new voice data to the speech recognition API and convert it into text data.

[1912] Server: Extract the product model number and serial number from the new text data. Use the extracted information to retrieve detailed product information from the database.

[1913] Server: Provides detailed fault diagnosis and updated response as needed.

[1914] Example prompt sentence:

[1915] The system analyzes a user's statement, "My product isn't working, so I'd like to check it," and uses a sentiment analysis engine to identify the customer's emotions and generate an answer.

[1916] This allows the system to provide efficient and accurate automated responses while taking into account customer emotions.

[1917] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1918] Step 1:

[1919] Detecting incoming calls from customers

[1920] Server: Connects to the call center system and detects incoming calls in real time. Works with the IP telephone system to monitor incoming calls from customers.

[1921] Input: Incoming call signal from customer

[1922] Output: Incoming call notification signal

[1923] Specific operations: Monitors the phone number and receives an incoming call signal, then notifies the device of the incoming call signal.

[1924] Step 2:

[1925] Collecting audio data and starting recording

[1926] Terminal: Receives an incoming call and activates the recording function, recording what the customer says and preparing to collect voice data with clear sound quality using a high-sensitivity microphone.

[1927] Input: Incoming call notification signal

[1928] Output: Start recording

[1929] Specific behavior: When an incoming call notification is received, the recording function is activated and the customer's voice begins to be recorded.

[1930] Step 3:

[1931] Sending audio data to the server

[1932] Terminal: Sends recorded audio data to the server in real time.

[1933] Input: Recorded audio data

[1934] Output: Send audio data to the server

[1935] Specific operation: The recorded audio data is compressed and sent to a server via internet communication.

[1936] Step 4:

[1937] Converting audio data to text

[1938] Server: Passes the voice data to a speech recognition API (e.g., Google Speech-to-Text) and converts it to text data. Temporarily stores the voice data and sends a request to the speech recognition API.

[1939] Input: Audio data sent to the server

[1940] Output: Text data generated by speech recognition

[1941] Specific operation: Sends audio data to the API and retrieves the returned text data.

[1942] Step 5:

[1943] Sentiment analysis of text data

[1944] Server: Using an emotion engine (e.g., IBM Watson Tone Analyzer), analyze customer emotions based on text data and voice tone.

[1945] Input: Text data and voice tone data from speech recognition

[1946] Output: Emotion data

[1947] Specific operation: Text data and voice tone are input into the emotion engine to obtain the emotion analysis results.

[1948] Step 6:

[1949] Identifying the content of the inquiry

[1950] Server: Uses a natural language processing (NLP) engine (e.g., OpenAI GPT-3) to identify the customer inquiry from the text data.

[1951] Input: Text data

[1952] Output: Identified inquiry

[1953] Specific behavior: Analyze what the customer says and identify the specific inquiry.

[1954] Step 7:

[1955] Generating the right answers

[1956] Server: Generates appropriate answers based on customer inquiry content and sentiment data. Searches product manuals and internal databases to generate answers.

[1957] Input: Identified query content and sentiment data

[1958] Output: Generated answer text

[1959] What it does: Searches a database, extracts the relevant information, and creates an answer.

[1960] Step 8:

[1961] Converting responses into audio data

[1962] Server: Uses speech synthesis technology (e.g., Amazon Polly) to convert the generated answer text into voice data.

[1963] Input: Generated answer text

[1964] Output: Audio data

[1965] Specific operation: The answer text is input into the speech synthesis API to generate voice data.

[1966] Step 9:

[1967] Providing answers

[1968] Terminal: Plays the generated voice data and provides the answer to the customer.

[1969] Input: Audio data

[1970] Output: A spoken response to the customer

[1971] Specific behavior: Plays back the audio data and provides the appropriate answer to the customer.

[1972] Step 10:

[1973] Collecting and sending customer responses

[1974] Terminal: Record the customer's response again and send it to the server.

[1975] Input: Customer response voice

[1976] Output: Sends response audio data to the server

[1977] What it does: Makes additional recordings and sends the collected data to a server.

[1978] Step 11:

[1979] Customer response analysis and system learning

[1980] Server: Analyzes customer responses and learns to improve the accuracy of the system.

[1981] Input: Customer response voice data

[1982] Output: Training data for accuracy improvement

[1983] What it does: Analyzes response data and feeds it into the learning algorithm, continuously improving the system.

[1984] (Application example 2)

[1985] 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."

[1986] In autonomous vehicles, the lack of technology to adjust the in-car environment taking into account passenger emotions makes it difficult to improve passenger comfort and satisfaction. In addition, there is no system to provide immediate and appropriate guidance or responses when passengers are confused or inconvenienced, which can lead to a poor user experience.

[1987] The identification processing 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 detecting an incoming call from a customer, means for collecting customer voice data in real time and converting it into text data using voice recognition technology, means for analyzing the text data using natural language processing technology and identifying the customer's inquiry, means for generating an appropriate answer based on the analysis results, means for providing the generated answer to the customer in voice, means for re-collecting customer responses and performing learning based on the collected responses to improve the accuracy of the system, means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology, means for adjusting the in-vehicle environment based on the voice input and the emotion analysis results, and means for presenting emotion-adjusted feedback using voice synthesis technology. This makes it possible to adjust the in-vehicle environment taking passengers' emotions into consideration.

[1988] "Means for detecting incoming calls from customers" refers to devices or functions that recognize phone calls or voice input from customers in real time and input that data into the system.

[1989] "Means for collecting customer voice data in real time and converting it into text data using voice recognition technology" refers to devices or algorithms that record customer voice and convert it into text data using voice recognition technology.

[1990] "Means of analyzing text data using natural language processing technology and identifying the content of customer inquiries" refers to technology that analyzes text data using a natural language processing algorithm and clarifies the specific content of customer inquiries from the content.

[1991] "Means for generating an appropriate response based on the analysis results" refers to a function for creating an optimal response from a pre-programmed database or algorithm based on the analysis results of the inquiry content.

[1992] The "means for providing the generated answer to the customer by voice" refers to a device or technology that converts the generated text-based answer into voice data using voice synthesis technology and conveys it to the customer.

[1993] "Means for re-collecting customer responses and learning to improve the accuracy of the system based on that" refers to a self-learning function that re-collects and analyzes feedback from customers and uses that to improve the system's performance and response accuracy.

[1994] "Means for collecting passenger voice data and adjusting the operation of the infotainment system based on the passenger's emotions using emotion analysis technology" refers to a function for recording passenger voice data, determining emotions from the voice using emotion analysis technology, and adjusting the in-car entertainment system based on the determination result.

[1995] "Means for adjusting the in-car environment based on voice input and emotion analysis results" is a function for automatically adjusting the temperature, music, lighting, etc. inside the car based on the passenger's voice input and the results of emotion analysis.

[1996] The "means for presenting emotion-adjusted feedback using voice synthesis technology" is a function for expressing optimal feedback based on the analyzed emotions in voice using voice synthesis technology and providing it to the occupant.

[1997] This invention relates to an infotainment system for autonomous vehicles. The system collects and analyzes passenger voice data using emotion analysis technology, and automatically adjusts the operation of the infotainment system and the in-vehicle environment based on the results. This system uses the following hardware and software:

[1998] Hardware

[1999] 1. Infotainment system: An entertainment and information system installed in a vehicle.

[2000] 2. Microphone: A device for collecting the voices of the passengers.

[2001] 3. Speaker: A device for providing audio feedback.

[2002] software

[2003] 1. Speech recognition API: Converts passenger voice data into text data (e.g., Google Cloud Speech-to-Text).

[2004] 2. Sentiment analysis API: Analyzes text data and voice tone to identify emotions (e.g., IBM Watson Tone Analyzer).

[2005] 3. Natural language processing engine: processes the parsed text data and understands the passenger's intent (e.g., spaCy, NLTK).

[2006] 4. Text-to-speech API: Converts text-based responses into audio data (e.g., Amazon Polly).

[2007] System Operation

[2008] First, the passenger's voice data is collected through a microphone. The collected voice data is sent to a speech recognition API and converted into text data. This text data is then sent to an emotion analysis API to analyze the passenger's emotions. The analyzed emotion data is then analyzed by a natural language processing engine to determine the appropriate feedback action (for example, adjusting the temperature or changing the music).

[2009] Processing Description

[2010] 1. Voice data collection: The microphone collects the passenger's voice and sends it to the voice recognition API.

[2011] 2. Speech recognition: The speech recognition API converts the voice data into text data.

[2012] 3. Emotion analysis: Text data and voice tone are input into the emotion analysis API to analyze the passenger's emotions.

[2013] 4. Natural Language Processing: The natural language processing engine determines the appropriate response or action based on the analyzed emotional data.

[2014] 5. Environmental Adjustment and Feedback: Based on the determined responses and actions, the infotainment system adjusts the in-car environment and provides audible feedback of the adjustments to the occupants using a speech synthesis API.

[2015] Specific examples

[2016] For example, if a passenger says, "It's a little cold," the voice data is converted into text data saying "It's a little cold" by the speech recognition API. This text data and voice tone are analyzed by the emotion analysis API to determine that the passenger is feeling cold. Next, the natural language processing engine determines that the passenger is "feeling cold" and issues a command to raise the interior temperature by 2 degrees. The speech synthesis API is also used to provide feedback to the passenger, saying, "The interior temperature has been raised a little."

[2017] Prompt Sentence Examples

[2018] The server was told by voice, "It's a little cold."

[2019] The infotainment system analyzed the voice input text and sent it to the sentiment analysis API, which output the sentiment analysis result as "I feel cold."

[2020] The driver has been instructed to increase the interior temperature of the self-driving vehicle by 2 degrees.

[2021] "The interior temperature has been slightly increased," the voice synthesis API informed the occupants.

[2022] In this way, the various hardware and software components work together to provide passengers with a comfortable in-vehicle environment and appropriate feedback.

[2023] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2024] Step 1:

[2025] The user's voice data is collected through a microphone and sent to the device. Specifically, the infotainment system detects voice input and starts recording. The input is the user's voice, and the recorded voice file is generated as the output.

[2026] Step 2:

[2027] The device sends the collected voice data in real time to a voice recognition API, which converts the voice data into text data (e.g., Google Cloud Speech-to-Text). The input is an audio file, and the output is text data.

[2028] Step 3:

[2029] The server sends the text data and voice tone data to an emotion analysis API to analyze the customer's emotions. The emotion analysis API (e.g., IBM Watson Tone Analyzer) analyzes the text data and voice tone to generate emotion data. The input is text and voice tone, and the output is emotion data.

[2030] Step 4:

[2031] The server uses natural language processing technology to analyze the text data and emotion data to identify the customer's inquiry. The natural language processing engine (e.g., spaCy, NLTK) analyzes the text data and determines the appropriate response. The input is text data and emotion data, and the analysis results are generated as the output.

[2032] Step 5:

[2033] The server determines the appropriate response based on the analysis results and emotion data, and issues instructions to the infotainment system. For example, this could include raising the temperature inside the car by 2 degrees or changing the music. The inputs are the analysis results and emotion data, and the output is the setting data for the infotainment system.

[2034] Step 6:

[2035] The device adjusts the infotainment system according to instructions from the server, for example, adjusting the temperature or changing the music inside the car. The input is the setting data from the server, and the output is the adjusted in-car environment.

[2036] Step 7:

[2037] The server sends the generated settings to a speech synthesis API, which generates voice feedback. The speech synthesis API (e.g., Amazon Polly) converts text data into voice data. The input is the text data of the settings, and the output is voice feedback data.

[2038] Step 8:

[2039] The terminal plays back the generated voice feedback data and provides it to the user. This results in voice guidance such as "The interior temperature has been slightly increased." The voice feedback data is input, and voice guidance for the user is generated as output.

[2040] 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.

[2041] 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.

[2042] 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.

[2043] 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.

[2044] 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.

[2045] 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.

[2046] 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).

[2047] 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.

[2048] 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."

[2049] 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.

[2050] 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).

[2051] 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.

[2052] 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.

[2053] 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.

[2054] 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.

[2055] 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.

[2056] 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.

[2057] 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.

[2058] 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.

[2059] 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.

[2060] 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.

[2061] The following is further disclosed regarding the above embodiment.

[2062] (Claim 1)

[2063] means for detecting an incoming call from a customer;

[2064] A means for collecting customer voice data in real time and converting it into text data using voice recognition technology;

[2065] A means for analyzing text data using natural language processing technology to identify the content of customer inquiries;

[2066] A means for generating an appropriate response based on the analysis results;

[2067] a means for providing the generated answer to the customer by voice;

[2068] It will be a means of collecting customer responses again and learning from them to improve the accuracy of the system.

[2069] A system including:

[2070] (Claim 2)

[2071] 10. The system of claim 1, further comprising means for analyzing a customer's emotions using voice analysis technology and adjusting responses based on the emotions.

[2072] (Claim 3)

[2073] 2. The system according to claim 1, further comprising means for providing automated response 24 hours a day, 365 days a year, and equipped with speech recognition technology and natural language processing technology that can handle multiple languages.

[2074] "Example 1"

[2075] (Claim 1)

[2076] means for detecting an incoming call from a customer;

[2077] A means for collecting customer voice data in real time and converting it into text data using voice recognition technology;

[2078] A means for analyzing text data using natural language processing technology to identify the content of customer inquiries;

[2079] A means for generating an appropriate response based on the analysis results;

[2080] a means for providing the generated answer to the customer by voice;

[2081] A means of collecting customer responses again and learning based on them to improve the accuracy of the system;

[2082] a means for establishing a call session and recording the customer's voice and transmitting the recording to a server;

[2083] a means for searching multiple databases for relevant information;

[2084] A means for converting the acquired text data into voice data, and

[2085] A system including:

[2086] (Claim 2)

[2087] 10. The system of claim 1, further comprising means for analyzing a customer's emotions using voice analysis technology and adjusting responses based on the emotions.

[2088] (Claim 3)

[2089] 2. The system according to claim 1, further comprising means for providing automated response 24 hours a day, 365 days a year, and equipped with speech recognition technology and natural language processing technology that can handle multiple languages.

[2090] "Application Example 1"

[2091] (Claim 1)

[2092] means for detecting an incoming call from a customer;

[2093] A means for collecting customer voice data in real time and converting it into text data using voice recognition technology;

[2094] A means for analyzing text data using natural language processing technology to identify the content of customer inquiries;

[2095] A means for generating an appropriate response based on the analysis results;

[2096] a means for providing the generated answer to the customer by voice;

[2097] It will be a means of collecting customer responses again and learning from them to improve the accuracy of the system.

[2098] A means of collecting customer inquiries in real time via smart devices and generating and providing appropriate answers;

[2099] A system including:

[2100] (Claim 2)

[2101] 10. The system of claim 1, further comprising means for analyzing a customer's emotions using voice analysis technology and adjusting responses based on the emotions.

[2102] (Claim 3)

[2103] Equipped with voice recognition technology and natural language processing technology that can handle multiple languages, it provides automated responses 24 hours a day, 365 days a year.

[2104] 10. The system of claim 1, further comprising means for interacting with the customer using a smart device.

[2105] "Example 2: Combining Emotion Engines"

[2106] (Claim 1)

[2107] means for detecting an incoming call from a customer;

[2108] A means for collecting customer voice data in real time and converting it into text data using voice recognition technology;

[2109] A means for analyzing text data using natural language processing technology to identify the content of customer inquiries;

[2110] A means for generating an appropriate response based on the analysis results;

[2111] a means for providing the generated answer to the customer by voice;

[2112] A means of collecting customer responses again and learning based on them to improve the accuracy of the system;

[2113] A means of analyzing customer sentiment and dynamically adjusting responses based on the results;

[2114] The system includes a computer terminal and a database that are linked together via a communications network.

[2115] (Claim 2)

[2116] 10. The system of claim 1, further comprising means for analyzing a customer's emotions using voice analysis technology and adjusting responses based on the emotions.

[2117] (Claim 3)

[2118] 2. The system according to claim 1, further comprising means for providing automated response 24 hours a day, 365 days a year, and equipped with speech recognition technology and natural language processing technology that can handle multiple languages.

[2119] "Application example 2 when combining emotion engines"

[2120] (Claim 1)

[2121] means for detecting an incoming call from a customer;

[2122] A means for collecting customer voice data in real time and converting it into text data using voice recognition technology;

[2123] A means for analyzing text data using natural language processing technology to identify the content of customer inquiries;

[2124] A means for generating an appropriate response based on the analysis results;

[2125] a means for providing the generated answer to the customer by voice;

[2126] A means of collecting customer responses again and learning based on them to improve the accuracy of the system;

[2127] means for collecting voice data from occupants and adjusting the operation of the infotainment system based on the emotions of the occupants using emotion analysis techniques;

[2128] means for adjusting the in-vehicle environment based on voice input and emotion analysis results;

[2129] a means for presenting emotion-adjusted feedback using speech synthesis technology;

[2130] A system including:

[2131] (Claim 2)

[2132] 10. The system of claim 1, further comprising means for analyzing a customer's emotions using voice analysis technology and adjusting responses based on the emotions.

[2133] (Claim 3)

[2134] 2. The system according to claim 1, further comprising means for providing automated response 24 hours a day, 365 days a year, and equipped with speech recognition technology and natural language processing technology that can handle multiple languages. [Explanation of symbols]

[2135] 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 detecting an incoming call from a customer; A means for collecting customer voice data in real time and converting it into text data using voice recognition technology; A means for analyzing text data using natural language processing technology to identify the content of customer inquiries; A means for generating an appropriate response based on the analysis results; a means for providing the generated answer to the customer by voice; It will be a means of collecting customer responses again and learning from them to improve the accuracy of the system. A system including:

2. 10. The system of claim 1, further comprising means for analyzing a customer's emotions using voice analysis technology and adjusting responses based on the emotions.

3. 2. The system according to claim 1, further comprising means for providing an automated response 24 hours a day, 365 days a year, and including a voice recognition technology and a natural language processing technology that can handle multiple languages.

Images