system

Abstract

We provide the system. [Solution] A means for saving the received text, A means of analyzing saved texts and performing sentiment analysis, A means of evaluating the intensity of emotions based on the results of emotion analysis, A means of sending a notification to the person in charge when strong emotions are detected, A means of generating appropriate text, A means of providing the generated draft to the person in charge, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Mails and messages from customers often contain emotions. In particular, strong emotions such as dissatisfaction and anger require prompt and appropriate responses. However, it is difficult for the person in charge to immediately check all mails and read the emotions to consider the optimal countermeasures in terms of time and labor. As a result, customer satisfaction may decrease. The purpose of the present invention is to improve customer satisfaction by providing a system that can respond promptly and appropriately to emotional mails and messages from customers.

Means for Solving the Problems

[0005] The system according to the present invention includes means for saving received text, means for analyzing the saved text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of the sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating an appropriate draft, and means for providing the generated draft to the person in charge. This makes it possible to respond quickly and appropriately to strong customer emotions, thereby improving customer satisfaction.

[0006] "Means for saving received text" refers to devices or software that record emails and messages received by a user.

[0007] "Means for analyzing saved text and performing sentiment analysis" refers to programs or devices that use natural language processing to automatically determine the emotional state of saved text.

[0008] "Means for evaluating the intensity of emotions based on the results of emotion analysis" refers to algorithms or devices that quantify or classify the intensity of emotions based on data obtained through emotion analysis.

[0009] "Means of sending notifications to the person in charge when strong emotions are detected" refers to a notification system or device that informs the person in charge who needs to take action when the detected emotions exceed a predetermined threshold.

[0010] "Means for generating appropriate sentences" refers to generative AI or programs that automatically create appropriate response sentences based on detected emotions.

[0011] "Means of providing the generated draft text to the person in charge" refers to a system or interface for communicating the generated corresponding text to the person in charge. [Brief explanation of the drawing]

[0012] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of the data processing device and smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14] This is a sequence diagram showing the processing flow of the data processing system in Application Example 2, which combines an emotion engine. [Modes for carrying out the invention]

[0013] Hereinafter, an example of an embodiment of the system relating to the technology of this disclosure will be described with reference to the attached drawings.

[0014] First, the terms used in the following description will be explained.

[0015] In the following embodiments, the labeled processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0016] In the following embodiments, the labeled RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0017] In the following embodiments, the labeled storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, and the like.

[0018] In the following embodiments, the labeled communication I / F (Interface) is an interface including a communication processor and 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), or Bluetooth (registered trademark), and the like.

[0019] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0020] [First Embodiment]

[0021] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0022] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0023] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0024] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0025] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0026] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0027] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0028] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0029] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0030] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0031] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0032] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0033] This invention relates to a system that inputs customer emails and messages into a generative AI and performs sentiment analysis. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has the function of quickly notifying the relevant personnel of this information and automatically generating and providing appropriate drafts of messages to help them manage their emotions.

[0034] Specifically, the following processes are performed.

[0035] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[0036] Next, the server prepares the stored email content for input into the generative AI. Specifically, it extracts the text data from the email and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[0037] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[0038] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[0039] At the same time, the server instructs the generative AI to generate a draft message that suppresses strong emotions. The generative AI generates a draft message that includes appropriate language and solutions, and sends it back to the server. For example, it might say, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0040] The server sends the generated draft to the employee's terminal, who then uses that draft to respond to the customer. This allows for a quick and appropriate response to strong customer emotions.

[0041] By using this system, staff can handle customer inquiries efficiently, which is expected to lead to improved customer satisfaction. A concrete example is the ability to respond quickly to complaint emails. If a customer sends an email saying, "The delivery is delayed and I'm very frustrated!", the system can quickly respond using a generated message such as, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," thereby calming the customer's anger.

[0042] The following describes the processing flow.

[0043] Step 1:

[0044] Users receive emails and messages from customers.

[0045] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[0046] Step 2:

[0047] The server saves received emails to the database.

[0048] Specifically, the server records information such as the email body, sender information, and the date and time of receipt in a database.

[0049] Step 3:

[0050] The server reads emails stored in the database and prepares them for input into the generative AI.

[0051] Specifically, the server extracts the text data from the email and converts it into a format for analysis by generative AI.

[0052] Step 4:

[0053] The server converts the email content and sends it to a generation AI.

[0054] Specifically, the server sends email data to the generative AI via an API.

[0055] Step 5:

[0056] Generative AI analyzes the content of emails and performs sentiment analysis.

[0057] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[0058] Step 6:

[0059] The server receives the emotion score and evaluates the intensity of the emotion.

[0060] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[0061] Step 7:

[0062] The server sends a notification to the person in charge when strong emotions are detected.

[0063] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[0064] Step 8:

[0065] The device displays a notification to the person in charge.

[0066] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[0067] Step 9:

[0068] The server requests the generative AI to generate appropriate text.

[0069] Specifically, the server specifies the requirements for the emotionally restrained text and sends them again to the generative AI via the API.

[0070] Step 10:

[0071] The generative AI generates text based on the specified conditions.

[0072] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0073] Step 11:

[0074] The server provides the generated draft to the person in charge.

[0075] Specifically, the server generates a draft document, which is then sent to the person in charge, allowing them to view it on their terminal.

[0076] Step 12:

[0077] The terminal displays the draft document to the person in charge.

[0078] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[0079] Step 13:

[0080] The person in charge will use the generated draft as a reference and reply to the customer.

[0081] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[0082] This makes it possible to respond quickly and effectively to customers' strong emotions.

[0083] (Example 1)

[0084] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0085] In customer communication, a challenge is responding quickly and appropriately to messages that evoke strong emotions. Conventional systems have been unable to respond quickly to such messages, potentially resulting in decreased customer satisfaction. In contrast, this invention aims to improve the efficiency and quality of customer service by automatically analyzing emotions and providing appropriate responses.

[0086] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0087] In this invention, the server includes means for storing received communications, means for analyzing the stored communications and performing sentiment analysis, means for evaluating the intensity of emotions based on the results of the sentiment analysis, means for sending notifications when the intensity of emotions exceeds a pre-set threshold, means for generating selected example sentences, and means for providing the generated example sentences to the relevant personnel. This makes it possible to quickly detect messages from customers with strong emotions and automatically generate and provide appropriate countermeasures.

[0088] "Received communications" refers to messages and data received by the target system or device.

[0089] "Means of preservation" refers to devices or programs for recording received communications in a database or storage medium.

[0090] "Means of analysis and sentiment analysis" refer to algorithms and software that process stored communications and identify emotions from their content.

[0091] A "means for evaluating the intensity of emotions" is a mechanism that determines, either numerically or qualitatively, how strong an emotion is, based on the results of an emotion analysis.

[0092] "Means of sending notifications" refers to communication methods and protocols used to transmit warnings and information to relevant personnel in real time, based on evaluation results and other factors.

[0093] "Means for generating selected example sentences" refers to artificial intelligence or programs that automatically create appropriate sentences to correspond to specific emotions.

[0094] "Means of providing" refers to means of sending or displaying generated example sentences, etc., for the purpose of showing them to the person in charge.

[0095] This invention is a system for analyzing customer communications and providing appropriate responses. This system stores received communications, performs sentiment analysis, evaluates the intensity of emotions, and generates appropriate example sentences. The following hardware and software are used to implement these functions.

[0096] System Configuration

[0097] This system consists of a user's terminal, a server, and a generative AI model.

[0098] User's device:

[0099] Receive communications from customers using email client software (e.g., Outlook, Gmail).

[0100] server:

[0101] The system uses a mail server (e.g., Postfix, Sendmail) to receive communications and saves them to a database server (e.g., MySQL®, PostgreSQL).

[0102] Use a data transformation library (e.g., Pandas) to convert the data into a format suitable for sentiment analysis.

[0103] A notification system (e.g., Slack API, email notification system) is used to notify the responsible person when a sentiment score exceeding a threshold is detected.

[0104] Generative AI models:

[0105] We analyze communication content using natural language processing libraries (e.g., NLTK, SpaCy) and perform sentiment analysis.

[0106] The analysis results are returned to the server as an emotion score.

[0107] Based on the prompt, it generates example sentences that include appropriate wording and solutions.

[0108] Processing Overview

[0109] 1. Check for communications received by the user (e.g., emails such as "I am very upset about the product defect").

[0110] 2. The server stores the emails on the database server.

[0111] 3. The server uses the Python Pandas library to convert the email text data into a format that can be processed by the generative AI.

[0112] 4. The generative AI uses a natural language processing library to detect strong emotions from the context of the email. This detection result is returned to the server as a numerical emotion score.

[0113] 5. The server compares the sentiment score to a threshold, and if it exceeds the threshold, it sends a real-time notification to the person in charge using the notification system.

[0114] 6. The server sends prompt text to the generative AI, instructing it to generate example sentences to suppress emotions.

[0115] Example prompt: For the sentence "Customer email: I am very upset about the product defect," please generate an example sentence that includes an apology and a solution.

[0116] 7. The generative AI generates appropriate example sentences and sends them back to the server.

[0117] 8. The server sends the generated example sentences to the employee's terminal and displays the draft sentences on the notification system.

[0118] 9. The person in charge uses email client software to reply to the customer based on example sentences provided by the generative AI.

[0119] Specific example

[0120] This is an example of how to handle a customer who sends an email saying, "The delivery is delayed and I'm extremely frustrated!" The user who receives this email checks it using their email client software, and the server saves its contents to a database server. The server then converts the email text into a format that can be processed by a generative AI and sends it to the generative AI. The generative AI performs sentiment analysis, detects a strong emotion such as "extremely frustrated," and sends a sentiment score back to the server.

[0121] The server evaluates the sentiment score and sends a notification to the employee's terminal if it exceeds the threshold. At the same time, it sends a prompt to the generative AI, instructing it to generate example sentences. The generative AI generates an example sentence, for example, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," and sends it back to the server. The server sends the generated example sentence to the employee's terminal, and the employee uses email client software to reply to the customer based on that example sentence.

[0122] This will enable us to respond quickly and appropriately to customers' strong emotions, which is expected to improve customer satisfaction.

[0123] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0124] Step 1:

[0125] The user receives communications from the customer.

[0126] (Specific actions) The user's email client software (e.g., Outlook, Gmail) checks for emails from customers (e.g., "I am very upset about the product defect").

[0127] (Input) Email from customer

[0128] (Output) Received email content

[0129] Step 2:

[0130] The server saves the received communications to the database.

[0131] (Specific operation) The mail server saves received emails to a database server (e.g., MySQL, PostgreSQL) and records the email content.

[0132] (Input) Received email content

[0133] (Output) Email data stored in the database

[0134] Step 3:

[0135] The server performs a format conversion so that the email content can be input into the generation AI.

[0136] (Specific operation) The server uses the Python Pandas library to convert the saved email text data into a format that can be processed by the generative AI (e.g., JSON format).

[0137] (Input) Email data stored in the database

[0138] (Output) Data converted into a format that can be processed by a generative AI.

[0139] Step 4:

[0140] A generative AI analyzes the formatted email data and performs sentiment analysis.

[0141] (Specific operation) A generative AI uses a natural language processing library (e.g., SpaCy, NLTK) to detect a strong emotion such as "very angry" from the context of the email and quantifies the emotion score (e.g., anger 0.8).

[0142] (Input) Data converted into a format that can be processed by a generative AI.

[0143] (Output) Emotion score (Example: Anger 0.8)

[0144] Step 5:

[0145] The server evaluates the sentiment score and notifies the person in charge if it exceeds the threshold.

[0146] (Specific operation) The server compares the emotion score to a threshold (e.g., anger 0.7), and if the threshold is exceeded, it sends a real-time notification to the person in charge's terminal using a notification system (e.g., Slack API).

[0147] (Input) Emotion score

[0148] (Output) Notification sent to the responsible person's terminal

[0149] Step 6:

[0150] The server instructs the generative AI to generate text that suppresses emotions.

[0151] (Specific operation) The server sends a prompt message to the generation AI: "Customer email: Please generate a draft email that includes an apology and a solution regarding a customer's extremely upset about a product defect."

[0152] (Input) Sentiment score, prompt text

[0153] (Output) Instructions for text generation sent to the generation system AI.

[0154] Step 7:

[0155] A generative AI generates the text.

[0156] (Specific operation) The generative AI creates an appropriate message (e.g., "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible.") and sends it back to the server.

[0157] (Input) Instructions for generating text draft

[0158] (Output) Generated text

[0159] Step 8:

[0160] The server sends the generated draft to the user's terminal.

[0161] (Specific operation) The server sends the draft document to the person in charge's PC or tablet, and the document is displayed in the notification system (e.g., MICROSOFT® TEAMS®, ​​Slack).

[0162] (Input) Generated text

[0163] (Output) Draft sent to the terminal of the person in charge

[0164] Step 9:

[0165] The person in charge will use the generated draft as a reference to reply to the customer.

[0166] (Specific Action) The person in charge uses email client software to reply, based on a template provided by the generation AI, with the message: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0167] (Input) Generated text

[0168] (Output) Reply email sent to the customer

[0169] (Application Example 1)

[0170] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0171] Traditional systems often suffered from delays in responding to strong customer emotions due to the results of sentiment analysis, making it difficult to respond quickly to such situations. Furthermore, manual handling by staff was required to generate appropriate responses, resulting in inefficiency. Consequently, rapid responses to customer satisfaction and security concerns remained a challenge.

[0172] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0173] In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of the sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating appropriate draft text, means for providing the generated draft text to the person in charge, means for analyzing the text in real time, and means for displaying the generated draft text on the person in charge's terminal. This makes it possible to respond quickly and efficiently to strong customer emotions.

[0174] "Means for saving received text" refers to a function for storing messages and text received from customers in storage such as a database.

[0175] "Means for analyzing saved text and performing sentiment analysis" refers to a function that analyzes saved text data and identifies and evaluates the emotions (joy, anger, sadness, etc.) contained within it.

[0176] A "means for evaluating the intensity of an emotion based on the results of an emotion analysis" refers to a function that uses the results of an emotion analysis to evaluate the intensity of that emotion as a numerical value or score.

[0177] "A means of sending a notification to the person in charge when strong emotions are detected" refers to a function that informs the person in charge in real time when the intensity of emotions exceeds a set threshold.

[0178] "Means for generating appropriate wording" refers to a function that uses generative AI to automatically create optimal reply messages and responses in order to manage customer emotions.

[0179] "Means of providing the generated text to the person in charge" refers to a function for displaying or sending the generated text and corresponding messages to the person in charge's terminal.

[0180] "A means of analyzing text in real time" refers to a processing function that immediately applies sentiment analysis to newly received text and quickly obtains the results.

[0181] "Means for displaying generated drafts on the employee's terminal" refers to a function for displaying generated reply drafts on the screen of the employee's smartphone, computer, or other terminal.

[0182] This invention is a system that performs sentiment analysis based on received customer messages and emails, generates appropriate response drafts, and provides rapid and efficient customer service. This system is implemented using the following hardware and software.

[0183] Hardware and software configuration

[0184] 1. Server:

[0185] Hardware: Standard server equipment (e.g., high-performance CPU, memory, storage)

[0186] Software: Python, MySQL, Firebase

[0187] 2. Terminal:

[0188] Hardware: Smartphones, PCs, tablets, etc.

[0189] Software: Specific notification applications or web browsers

[0190] System Processing Overview

[0191] 1. Data reception and storage:

[0192] When a user receives a message from a customer, the message is sent to the server and stored in a database (MySQL). This ensures the secure storage of email and chat data.

[0193] 2. Sentiment analysis:

[0194] The server retrieves the stored messages and inputs them into a generative AI model for sentiment analysis (e.g., OpenAI® GPT-4®). The generative AI model performs sentiment analysis using the following prompts:

[0195] "Perform an emotional analysis of the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[0196] The analyzed results are sent back to the server as an emotion score.

[0197] 3. Sentiment evaluation and notification:

[0198] The server receives the sentiment score and evaluates its strength. If it exceeds a threshold, it sends a notification to the agent's device using Firebase Cloud Messaging, allowing the agent to immediately understand the situation.

[0199] 4. Draft generation:

[0200] The server instructs the AI ​​model to generate an appropriate response. The following prompts are used:

[0201] "Please create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm extremely frustrated!'"

[0202] The generated draft is sent to the person in charge's terminal.

[0203] Specific example

[0204] For example, if a customer sends a message saying, "The delivery is delayed and it's very frustrating!", the server will execute the following process.

[0205] Receive messages and save them to the database.

[0206] OpenAI GPT-4 is used for sentiment analysis, and a strong emotion of "very irritated" is detected.

[0207] Since the emotion score exceeded the threshold, a notification will be sent to the person in charge.

[0208] Using OpenAI GPT-4 again, we generate a suitable reply message like the following:

[0209] "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible."

[0210] The generated draft is displayed on the employee's terminal, and the employee uses it to quickly reply to the customer.

[0211] In this way, this system can respond quickly and effectively to customers' strong emotions, and can contribute to improving customer satisfaction.

[0212] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0213] Step 1:

[0214] A user receives a message from a customer. The text data received by the user via email or messaging app is sent to the server. The server stores this received data in a database. The input is the customer's message text, and the output is the message data stored in the database. Specifically, the server inserts the received text into a MySQL database using SQL commands.

[0215] Step 2:

[0216] The server analyzes stored messages and performs sentiment analysis. It retrieves the stored message data and inputs it into a generative AI model (e.g., OpenAI GPT-4). The input is the stored message text, which is a request for sentiment analysis using a prompt. The output is a sentiment score. Specifically, the server sends the following prompt to the generative AI model: "Perform sentiment analysis on the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[0217] Step 3:

[0218] The server receives sentiment scores from the generated AI model and evaluates their strength. The input is the sentiment score, and the output is the result of determining whether it exceeds a threshold. Specifically, the server compares the sentiment score to a threshold, and if it exceeds the threshold, it proceeds to the next step.

[0219] Step 4:

[0220] If a sentiment score exceeding a threshold is detected, the server uses Firebase Cloud Messaging to send a notification to the employee's device. The input is the sentiment score exceeding the threshold and the original message text, and the output is the notification to the employee's device. Specifically, the server sends a notification message via the Firebase API indicating that strong sentiment has been detected.

[0221] Step 5:

[0222] The server instructs the generative AI model to generate an appropriate response. The input is the customer message and the prompt "Create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm very frustrated!'" and the output is the generated response. Specifically, the server sends another request to the generative AI model to retrieve an appropriate response.

[0223] Step 6:

[0224] The server provides the generated text to the user's terminal. The input is the text generated by the AI ​​model, and the output is the text displayed on the user's terminal screen. Specifically, the server sends the generated text to the user's terminal and displays it through a web application or notification application.

[0225] Step 7:

[0226] The person in charge replies to the customer based on the provided draft. The input is the generated draft sent from the server, and the output is the reply message sent to the customer. Specifically, the person in charge reviews the generated draft, makes adjustments as needed, and sends the final reply message to the customer.

[0227] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0228] This invention provides a system that enables more appropriate customer service by inputting customer emails and messages into a generative AI, performing sentiment analysis, and combining it with an emotion engine that recognizes user emotions. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has a function to quickly notify the relevant personnel of this information and automatically generate and provide appropriate wording to the personnel to help them manage their emotions.

[0229] Specifically, the following processes are performed.

[0230] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[0231] Subsequently, the server reads the emails stored in the database and prepares them for input into the generative AI. Specifically, it extracts the text data from the emails and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[0232] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[0233] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[0234] At the same time, the server instructs the generative AI to generate a text that suppresses strong emotions. The generative AI generates a text that incorporates appropriate language and a solution, and sends it back to the server. For example, it might generate a text that says, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0235] Furthermore, this invention incorporates an emotion engine to analyze the user's emotions in real time and adjust the tone of the generated text accordingly. The emotion engine analyzes the user's facial expressions, voice tone, and keyboard input speed to determine their current emotional state. This allows for adjustments such as generating text with a softer tone if, for example, the person in charge is feeling stressed.

[0236] The server sends the generated draft to the employee's terminal, and the employee uses that draft as a reference to reply to the customer. This allows for a quick and appropriate response to strong customer emotions, and provides a draft that is tailored to the user's emotional state.

[0237] This allows staff to handle customer inquiries more efficiently, which is expected to lead to improved customer satisfaction. For example, by using the generated template to quickly respond to customer complaint emails, customer anger can be mitigated, and staff can handle the situation with less mental strain.

[0238] The following describes the processing flow.

[0239] Step 1:

[0240] Users receive emails and messages from customers.

[0241] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[0242] Step 2:

[0243] The server saves received emails to the database.

[0244] Specifically, the server records information such as the body of the received email, sender information, and the date and time of receipt in a database.

[0245] Step 3:

[0246] The server reads emails stored in the database and prepares them for input into the generative AI.

[0247] Specifically, the text data from the email is extracted and converted into a format that can be processed by generative AI.

[0248] Step 4:

[0249] The server converts the email content and sends it to a generation AI.

[0250] Specifically, the server sends email data to the generative AI via an API.

[0251] Step 5:

[0252] Generative AI analyzes the content of emails and performs sentiment analysis.

[0253] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[0254] Step 6:

[0255] The server receives the emotion score and evaluates the intensity of the emotion.

[0256] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[0257] Step 7:

[0258] The server sends a notification to the person in charge when strong emotions are detected.

[0259] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[0260] Step 8:

[0261] The device displays a notification to the person in charge.

[0262] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[0263] Step 9:

[0264] The server requests the generative AI to generate appropriate text.

[0265] Specifically, the server specifies the requirements for the emotionally restrained text and sends them again to the generative AI via the API.

[0266] Step 10:

[0267] The generative AI generates text based on the specified conditions.

[0268] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0269] Step 11:

[0270] The server prepares to provide the draft text generated by the generative AI to the person in charge.

[0271] Specifically, this involves preparing the generated draft document for transmission to the responsible person's terminal.

[0272] Step 12:

[0273] The terminal displays the draft document to the person in charge.

[0274] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[0275] Step 13:

[0276] The person in charge will use the generated draft as a reference and reply to the customer.

[0277] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[0278] Furthermore, the processing when combining the emotion engine is as follows.

[0279] Step 14:

[0280] The emotion engine recognizes the user's emotion in real time.

[0281] Specifically, it analyzes the user's expression, voice, or keyboard input speed to determine the current emotional state.

[0282] Step 15:

[0283] Based on the user's emotional state, the server instructs the generative AI to adjust the tone of the text.

[0284] Specifically, if it is determined that the user is feeling stressed, it specifies to generate text with a softer tone.

[0285] This enables quick and appropriate responses to the strong emotions of customers, not only improving customer satisfaction but also reducing the mental burden on the staff.

[0286] (Example 2)

[0287] Next, Example 2 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart device 14 is referred to as the "terminal".

[0288] In the conventional system, it was difficult to respond respond respondappropriately and quickly respond to emails and messages containing strong emotions from customers. Also, it was difficult for the staff to handle customer inquiries while evaluating their own emotional state, resulting in problems such as a decrease in customer satisfaction and an increase in stress for the staff. Furthermore, the quality and appropriateness of the generated text were not guaranteed, and the burden on the staff was significant.

[0289] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving emails and messages from customers, means for storing the received text in a database, means for analyzing the stored text and performing sentiment analysis using a generative AI model, means for sending a notification to the person in charge when the sentiment score exceeds a threshold, means for generating appropriate text based on the generative AI model, means for analyzing the user's current emotional state in real time, and means for providing the generated text to the person in charge. This makes it possible to respond quickly and appropriately to strong customer emotions, and provides text that matches the emotional state of the person in charge, which can be expected to improve customer satisfaction and reduce stress for the person in charge.

[0290] "Customer" refers to individuals or companies that purchase or use goods or services.

[0291] "Emails and messages" refer to electronic means of communication that customers send to companies or their representatives.

[0292] A "database" refers to a system for efficiently storing, searching, and managing information.

[0293] A "generative AI model" refers to an algorithm that uses artificial intelligence to perform specific tasks, such as language generation or sentiment analysis.

[0294] "Sentiment analysis" refers to the process of identifying emotions from text data and quantifying them.

[0295] An "emotion score" is a numerical value obtained from the results of an emotion analysis that indicates the intensity and type of emotion.

[0296] A "threshold" is a value that indicates a specific standard; exceeding this value triggers a particular action.

[0297] "Notifications" refer to messages or alerts used to convey important information or warnings to users.

[0298] "Document draft" refers to a preliminary version or draft of a document used to communicate with a customer.

[0299] "User" refers to a person who operates this system, or a person who uses the system to handle customer inquiries.

[0300] "Real-time" refers to processing or responding to an event as close to the moment it occurs as possible.

[0301] This invention relates to a system that uses artificial intelligence technology to analyze emails and messages from customers and evaluate their emotions in order to provide more appropriate customer service. In this system, the server, terminal, and user each play important roles.

[0302] First, users receive emails and messages from customers. This reception occurs through the company's email server or messaging platform.

[0303] Next, the server saves the received emails to a database. Specifically, it uses a relational database management system (RDBMS) such as MySQL.

[0304] The server converts the stored emails into a format that can be input into a generative AI model (e.g., the BERT model in Hugging Face) for analysis. This conversion involves changing the text data to a format such as JSON.

[0305] The server sends the transformed data to the generating AI model. This data is sent via an HTTP request to the generating AI model's API endpoint.

[0306] The generative AI model analyzes the received data and performs sentiment analysis. Based on the analysis, a sentiment score is calculated. For example, a strong feeling of anger is detected from the text "very angry."

[0307] The server evaluates the sentiment score returned by the generative AI model. If the score exceeds a pre-set threshold, the server sends a notification to the terminal. This notification is in the form of a pop-up display or an alert email.

[0308] Based on the sentiment score, the server instructs the generative AI model to generate appropriate text. This text is for taking appropriate actions towards customers, and for example, sentences like "We apologize." are generated.

[0309] Furthermore, the emotion engine analyzes the user's current emotional state in real time. Specifically, it analyzes the user's facial expressions and voice using a webcam or microphone, and also evaluates factors such as the keyboard input speed.

[0310] The generated text is sent by the server to the terminal, and the responsible person replies to the customer referring to this text. This enables a quick and appropriate response to the customer's strong emotions, and also helps reduce the stress of the responsible person.

[0311] As a specific example, consider the case where a customer sends an email saying "I'm extremely angry about the product defect." The server that receives this email saves the content of the email in the database and performs sentiment analysis using the generative AI model. Since the sentiment score is very high at 75, the server sends an emergency notification to the responsible person and at the same time instructs the generative AI model to generate appropriate text. The generated text is "We sincerely apologize for the inconvenience caused by the product defect. We are currently checking the situation and will promptly inform you of the solution." Furthermore, if the emotion engine detects that the responsible person is feeling stressed, a text with a softer tone is generated.

[0312] As described above, by combining the generative AI model and the emotion engine, this system can quickly and appropriately respond to emails containing strong emotions from customers, and also contribute to reducing the stress of the responsible person.

[0313] Example of a prompt:

[0314] Content of email received from customer: "I am very upset about the defect in the product."

[0315] Generative AI model to use: A general-purpose generative AI model (e.g., BERT model)

[0316] Instructions: "Perform sentiment analysis on the customer's email text and return the sentiment score."

[0317] Analysis result: Emotion score = 75 (Strong anger)

[0318] Notification: "We have received an email expressing strong emotions. Customer support is required."

[0319] Generated text: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible."

[0320] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0321] Step 1:

[0322] Users receive emails and messages from customers.

[0323] Specific operation: Users receive emails and messages through the company's mail server or messaging platform. For example, a customer sends an email stating, "I am very upset about a product defect."

[0324] Input: Emails and messages from customers

[0325] Output: Data of received emails and messages

[0326] Step 2:

[0327] The server saves the email to the database.

[0328] Specific operation: The server saves the content of received emails to a relational database (e.g., MySQL). At the same time, it also saves the email's sending date and sender information.

[0329] Input: Data of received emails and messages

[0330] Output: Email data stored in the database

[0331] Step 3:

[0332] The server converts the email into a format that can be input into the AI ​​model that generates emails.

[0333] Specific operation: The server converts the email text data into JSON format. It also formats it so that it can be sent to the API endpoint of the generated AI model.

[0334] Input: Email data stored in the database

[0335] Output: JSON format data for input into the generated AI model.

[0336] Step 4:

[0337] The server sends the converted data to the AI ​​model.

[0338] Specific operation: The server uses an HTTP request to send data to the API endpoint of the generated AI model. The request includes a generated prompt (e.g., "Perform sentiment analysis on customer email text and return the sentiment score.").

[0339] Input: JSON formatted data and prompt text

[0340] Output: Data sent to the generating AI model

[0341] Step 5:

[0342] The generative AI model performs sentiment analysis on the received data.

[0343] Specific operation: The generative AI model (e.g., a general generative AI model) analyzes the received text data and performs a process of quantifying emotions. It calculates an anger score from "very annoyed" as an emotion score and returns it as a numerical value (e.g., 75).

[0344] Input: Converted JSON data, prompt text

[0345] Output: Calculation result of the emotion score

[0346] Step 6:

[0347] The server evaluates the sentiment analysis results and sends notifications to the relevant personnel.

[0348] Specific operation: The server evaluates the sentiment score received from the generated AI model, and if the score exceeds a threshold (e.g., 50), it sends a pop-up notification or alert email to the person in charge.

[0349] Input: Sentiment score from a generated AI model

[0350] Output: Notification to the person in charge

[0351] Step 7:

[0352] The server instructs the AI ​​model to generate appropriate text.

[0353] Specific operation: The server sends a prompt message to the AI ​​model again, instructing it to generate a text, and then generates an appropriate response. For example, a sentence such as "We sincerely apologize for the inconvenience caused by the product defect" is generated.

[0354] Input: Prompt based on sentiment score

[0355] Output: Generated text

[0356] Step 8:

[0357] The emotion engine analyzes the user's current emotional state in real time.

[0358] Specific operation: The emotion engine uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed. Based on this, it assesses the stress level of the person in charge and adjusts the tone of the drafted text accordingly.

[0359] Input: Real-time data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[0360] Output: Evaluation results of the person in charge's emotional state

[0361] Step 9:

[0362] The server sends the optimal draft to the person in charge's terminal.

[0363] Specific operation: The server combines the generated text with adjustments based on the user's emotional state and sends the optimal text to the terminal. The person in charge uses this text as a reference to reply to the customer.

[0364] Input: Generated text, user's emotional state

[0365] Output: Optimal text sent to the terminal

[0366] This enables quick and appropriate responses to customers' strong emotions, and by providing wording tailored to the emotional state of the person handling the request, it leads to improved customer satisfaction and reduced stress for the person handling the request.

[0367] (Application Example 2)

[0368] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0369] Logistics centers are required to respond quickly and appropriately to customer complaints and inquiries. However, manually analyzing each message and providing an appropriate response requires a tremendous amount of effort and time, potentially leading to decreased customer satisfaction. Furthermore, the quality of responses is inconsistent because they are influenced by the emotional state of the person handling them.

[0370] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating an appropriate draft, means for providing the generated draft to the person in charge, means for receiving and processing customer messages on a smart device, and means for analyzing the emotional state of the person in charge and adjusting the tone of the draft. This makes it possible to respond quickly and appropriately to customer complaints and inquiries, and is expected to improve customer satisfaction and reduce the burden on the person in charge.

[0371] "Means for saving received text" refers to a device or program for saving messages and emails sent by customers to a database.

[0372] "Methods for analyzing saved text and performing sentiment analysis" refers to programs that use natural language processing techniques to analyze saved text data and identify emotions and their intensity.

[0373] "Means for evaluating the intensity of emotions based on the results of emotion analysis" refers to a device or program that quantifies the data obtained from emotion analysis and evaluates its intensity.

[0374] "Means for sending notifications to responsible personnel when strong emotions are detected" refers to a device or program that sends alerts or messages to the relevant personnel when strong emotions are detected as a result of the analysis.

[0375] "Means for generating appropriate text" refers to a device or program that uses natural language generation technology to generate appropriate response text according to the customer's emotions.

[0376] "Means of providing the generated draft to the person in charge" refers to a device or program for sending the generated reply draft to the person in charge's device.

[0377] "Means for receiving and processing customer messages on a smart device" refers to a device or program for receiving, analyzing, and processing messages from customers on a mobile device such as a smartphone or tablet.

[0378] "Means for analyzing the emotional state of the person in charge and adjusting the tone of the draft" refers to a device or program for analyzing the emotions of the person in charge and adjusting the tone of the draft generated based on the results.

[0379] This invention relates to a system for quickly and appropriately responding to customer complaints and inquiries in a logistics center. This system consists of a server, smart devices, a generative AI model, and an emotion engine.

[0380] The server receives messages and emails sent by customers and stores them in a database. The stored text data is sent to a generative AI for analysis. The generative AI uses natural language processing techniques to analyze the text and perform sentiment analysis. For example, from a message such as "I am very angry about the product defect," it identifies the emotion of anger and quantifies its intensity. This sentiment score is then sent back to the server.

[0381] The server evaluates the returned sentiment score and compares it to a pre-set threshold. If a strong sentiment exceeding the threshold is detected, an alert is immediately sent to the employee's smart device. It also instructs a generative AI to generate an appropriate message. The generated message is then provided to the employee's smart device. For example, a message such as, "We sincerely apologize for the inconvenience caused by the delay in receiving your order. We are currently investigating the situation and will address it as soon as possible," might be generated.

[0382] Furthermore, the emotion engine analyzes the emotional state of the person handling the task and adjusts the tone of the generated text accordingly. The emotion engine determines emotions in real time based on factors such as the person's facial expressions, voice tone, and keyboard typing speed. For example, if the person is feeling stressed, a softer tone of text will be generated.

[0383] The hardware used to implement this system includes high-performance servers, smartphones, or tablets. The software used includes TENSORFLOW®, natural language processing libraries (SpaCy, NLTK), a reactive framework (React Native), Docker, and a RESTful API (Flask / Django).

[0384] As a concrete example, input the following prompt into the generative AI.

[0385] "The delivery of my order is delayed. I'm extremely frustrated."

[0386] "Perform a sentiment analysis on this message and generate appropriate wording."

[0387] The sentiment analysis results indicate anger (score: 9.2). The generated text based on these results is as follows:

[0388] "We sincerely apologize for the inconvenience regarding the delivery of your order. We will address this issue as quickly as possible, so please bear with us for a little while longer."

[0389] This will enable quick and accurate responses to customer complaints and inquiries, leading to improved customer satisfaction and reduced workload for staff.

[0390] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0391] Step 1:

[0392] The server receives messages and emails from customers. The received text data is stored in a database.

[0393] Input: Customer messages or emails

[0394] Output: Text data stored in the database

[0395] Step 2:

[0396] The server retrieves the stored text data for analysis and converts it into a format that the generating AI model can process. Specifically, it cleans the text and performs any necessary preprocessing.

[0397] Input: Text data extracted from a database

[0398] Output: Text data converted into a format that can be processed by the generative AI model.

[0399] Step 3:

[0400] The generative AI model receives the transformed text data and performs sentiment analysis. For example, it identifies the emotion of anger from the phrase "very angry" and quantifies its intensity.

[0401] Input: Text data sent to a generative AI.

[0402] Output: Emotion score (Example: Anger intensity score 9.2)

[0403] Step 4:

[0404] The server evaluates the sentiment score returned from the generated AI model and compares it to a pre-set threshold. If the threshold is exceeded, a notification is sent to the person in charge's terminal.

[0405] Input: Sentiment score

[0406] Output: Notification when threshold is exceeded

[0407] Step 5:

[0408] The server instructs the AI ​​model to generate appropriate text. The text includes expressions that alleviate anger.

[0409] Input: Sentiment score and message content

[0410] Output: Generated text (Example: "We apologize for the inconvenience this has caused. We will address this issue as soon as possible.")

[0411] Step 6:

[0412] The generated draft is sent to the employee's smart device. The employee uses this draft as a reference to quickly reply to the customer.

[0413] Input: Generated text

[0414] Output: Draft sent to the person in charge's terminal

[0415] Step 7:

[0416] The emotion engine analyzes the emotional state of the person in charge and adjusts the tone of the generated text accordingly. If the person in charge is stressed, the text will be set to a softer tone.

[0417] Input: Emotional data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[0418] Output: Adjusted tone of the text

[0419] Step 8:

[0420] The server provides the staff with a pre-formatted response to use in replying to the customer. This allows for a response that soothes the customer's emotions and reduces the burden on the staff.

[0421] Input: Revised text

[0422] Output: Final reply message sent to the customer

[0423] By following these steps, the entire system will be able to respond quickly and appropriately to customer complaints and inquiries, which is expected to improve customer satisfaction and reduce the burden on staff.

[0424] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0425] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0426] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0427] [Second Embodiment]

[0428] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0429] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0430] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0431] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0432] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0433] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0434] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0435] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0436] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0437] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0438] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0439] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0440] This invention relates to a system that inputs customer emails and messages into a generative AI and performs sentiment analysis. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has the function of quickly notifying the relevant personnel of this information and automatically generating and providing appropriate drafts of messages to help them manage their emotions.

[0441] Specifically, the following processes are performed.

[0442] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[0443] Next, the server prepares the stored email content for input into the generative AI. Specifically, it extracts the text data from the email and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[0444] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[0445] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[0446] At the same time, the server instructs the generative AI to generate a message that suppresses strong emotions. The generative AI generates a message that includes appropriate language and solutions, and sends it back to the server. For example, it might say, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0447] The server sends the generated draft to the employee's terminal, who then uses that draft to respond to the customer. This allows for a quick and appropriate response to strong customer emotions.

[0448] By using this system, staff can handle customer inquiries efficiently, which is expected to lead to improved customer satisfaction. A concrete example is the ability to respond quickly to complaint emails. If a customer sends an email saying, "The delivery is delayed and I'm very frustrated!", the system can quickly respond using a generated message such as, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," thereby calming the customer's anger.

[0449] The following describes the processing flow.

[0450] Step 1:

[0451] Users receive emails and messages from customers.

[0452] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[0453] Step 2:

[0454] The server saves received emails to the database.

[0455] Specifically, the server records information such as the email body, sender information, and the date and time of receipt in a database.

[0456] Step 3:

[0457] The server reads emails stored in the database and prepares them for input into the generative AI.

[0458] Specifically, the server extracts the text data from the email and converts it into a format for analysis by generative AI.

[0459] Step 4:

[0460] The server converts the email content and sends it to a generation AI.

[0461] Specifically, the server sends email data to the generative AI via an API.

[0462] Step 5:

[0463] Generative AI analyzes the content of emails and performs sentiment analysis.

[0464] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[0465] Step 6:

[0466] The server receives the emotion score and evaluates the intensity of the emotion.

[0467] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[0468] Step 7:

[0469] The server sends a notification to the person in charge when strong emotions are detected.

[0470] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[0471] Step 8:

[0472] The device displays a notification to the person in charge.

[0473] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[0474] Step 9:

[0475] The server requests the generative AI to generate appropriate text.

[0476] Specifically, the server specifies the requirements for the emotionally restrained text and sends them again to the generative AI via the API.

[0477] Step 10:

[0478] The generative AI generates text based on the specified conditions.

[0479] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0480] Step 11:

[0481] The server provides the generated draft to the person in charge.

[0482] Specifically, the server generates a draft document, which is then sent to the person in charge, allowing them to view it on their terminal.

[0483] Step 12:

[0484] The terminal displays the draft document to the person in charge.

[0485] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[0486] Step 13:

[0487] The person in charge will use the generated draft as a reference and reply to the customer.

[0488] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[0489] This makes it possible to respond quickly and effectively to customers' strong emotions.

[0490] (Example 1)

[0491] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0492] In customer communication, a challenge is responding quickly and appropriately to messages that evoke strong emotions. Conventional systems have been unable to respond quickly to such messages, potentially resulting in decreased customer satisfaction. In contrast, this invention aims to improve the efficiency and quality of customer service by automatically analyzing emotions and providing appropriate responses.

[0493] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0494] In this invention, the server includes means for storing received communications, means for analyzing the stored communications and performing sentiment analysis, means for evaluating the intensity of emotions based on the results of the sentiment analysis, means for sending notifications when the intensity of emotions exceeds a pre-set threshold, means for generating selected example sentences, and means for providing the generated example sentences to the relevant personnel. This makes it possible to quickly detect messages from customers with strong emotions and automatically generate and provide appropriate countermeasures.

[0495] "Received communications" refers to messages and data received by the target system or device.

[0496] "Means of preservation" refers to devices or programs for recording received communications in a database or storage medium.

[0497] "Means of analysis and sentiment analysis" refer to algorithms and software that process stored communications and identify emotions from their content.

[0498] A "means for evaluating the intensity of emotions" is a mechanism that determines, either numerically or qualitatively, how strong an emotion is, based on the results of an emotion analysis.

[0499] "Means of sending notifications" refers to communication methods and protocols used to transmit warnings and information to relevant personnel in real time, based on evaluation results and other factors.

[0500] "Means for generating selected example sentences" refers to artificial intelligence or programs that automatically create appropriate sentences to correspond to specific emotions.

[0501] "Means of providing" refers to means of sending or displaying generated example sentences, etc., for the purpose of showing them to the person in charge.

[0502] This invention is a system for analyzing customer communications and providing appropriate responses. This system stores received communications, performs sentiment analysis, evaluates the intensity of emotions, and generates appropriate example sentences. The following hardware and software are used to implement these functions.

[0503] System Configuration

[0504] This system consists of a user's terminal, a server, and a generative AI model.

[0505] User's device:

[0506] Receive communications from customers using email client software (e.g., Outlook, Gmail).

[0507] server:

[0508] The system uses a mail server (e.g., Postfix, Sendmail) to receive communications and saves them to a database server (e.g., MySQL, PostgreSQL).

[0509] Use a data transformation library (e.g., Pandas) to convert the data into a format suitable for sentiment analysis.

[0510] A notification system (e.g., Slack API, email notification system) is used to notify the responsible person when a sentiment score exceeding a threshold is detected.

[0511] Generative AI models:

[0512] We analyze communication content using natural language processing libraries (e.g., NLTK, SpaCy) and perform sentiment analysis.

[0513] The analysis results are returned to the server as an emotion score.

[0514] Based on the prompt, it generates example sentences that include appropriate wording and solutions.

[0515] Processing Overview

[0516] 1. Check for communications received by the user (e.g., emails such as "I am very upset about the product defect").

[0517] 2. The server stores the emails on the database server.

[0518] 3. The server uses the Python Pandas library to convert the email text data into a format that can be processed by the generative AI.

[0519] 4. The generative AI uses a natural language processing library to detect strong emotions from the context of the email. This detection result is returned to the server as a numerical emotion score.

[0520] 5. The server compares the sentiment score to a threshold, and if it exceeds the threshold, it sends a real-time notification to the person in charge using the notification system.

[0521] 6. The server sends prompt text to the generative AI, instructing it to generate example sentences to suppress emotions.

[0522] Example prompt: For the sentence "Customer email: I am very upset about the product defect," please generate an example sentence that includes an apology and a solution.

[0523] 7. The generative AI generates appropriate example sentences and sends them back to the server.

[0524] 8. The server sends the generated example sentences to the employee's terminal and displays the draft sentences on the notification system.

[0525] 9. The person in charge uses email client software to reply to the customer based on example sentences provided by the generative AI.

[0526] Specific example

[0527] This is an example of how to handle a customer who sends an email saying, "The delivery is delayed and I'm extremely frustrated!" The user who receives this email checks it using their email client software, and the server saves its contents to a database server. The server then converts the email text into a format that can be processed by a generative AI and sends it to the generative AI. The generative AI performs sentiment analysis, detects a strong emotion such as "extremely frustrated," and sends a sentiment score back to the server.

[0528] The server evaluates the sentiment score and sends a notification to the employee's terminal if it exceeds the threshold. At the same time, it sends a prompt to the generative AI, instructing it to generate example sentences. The generative AI generates an example sentence, for example, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," and sends it back to the server. The server sends the generated example sentence to the employee's terminal, and the employee uses email client software to reply to the customer based on that example sentence.

[0529] This will enable us to respond quickly and appropriately to customers' strong emotions, which is expected to improve customer satisfaction.

[0530] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0531] Step 1:

[0532] The user receives communications from the customer.

[0533] (Specific actions) The user's email client software (e.g., Outlook, Gmail) checks for emails from customers (e.g., "I am very upset about the product defect").

[0534] (Input) Email from customer

[0535] (Output) Received email content

[0536] Step 2:

[0537] The server saves the received communications to the database.

[0538] (Specific operation) The mail server saves received emails to a database server (e.g., MySQL, PostgreSQL) and records the email content.

[0539] (Input) Received email content

[0540] (Output) Email data stored in the database

[0541] Step 3:

[0542] The server performs a format conversion so that the email content can be input into the generation AI.

[0543] (Specific operation) The server uses the Python Pandas library to convert the saved email text data into a format that can be processed by the generative AI (e.g., JSON format).

[0544] (Input) Email data stored in the database

[0545] (Output) Data converted into a format that can be processed by a generative AI.

[0546] Step 4:

[0547] A generative AI analyzes the formatted email data and performs sentiment analysis.

[0548] (Specific operation) A generative AI uses a natural language processing library (e.g., SpaCy, NLTK) to detect a strong emotion such as "very angry" from the context of the email and quantifies it as an emotion score (e.g., anger 0.8).

[0549] (Input) Data converted into a format that can be processed by a generative AI.

[0550] (Output) Emotion score (Example: Anger 0.8)

[0551] Step 5:

[0552] The server evaluates the sentiment score and notifies the person in charge if it exceeds the threshold.

[0553] (Specific operation) The server compares the emotion score to a threshold (e.g., anger 0.7), and if the threshold is exceeded, it sends a real-time notification to the person in charge's terminal using a notification system (e.g., Slack API).

[0554] (Input) Emotion score

[0555] (Output) Notification sent to the responsible person's terminal

[0556] Step 6:

[0557] The server instructs the generative AI to generate text that suppresses emotions.

[0558] (Specific operation) The server sends a prompt message to the generation AI: "Customer email: Please generate a draft email that includes an apology and a solution regarding a customer's email stating that they are very upset about a product defect."

[0559] (Input) Sentiment score, prompt text

[0560] (Output) Instructions for text generation sent to the generation system AI.

[0561] Step 7:

[0562] A generative AI generates the text.

[0563] (Specific operation) The generative AI creates an appropriate message (e.g., "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible.") and sends it back to the server.

[0564] (Input) Instructions for generating text draft

[0565] (Output) Generated text

[0566] Step 8:

[0567] The server sends the generated draft to the user's terminal.

[0568] (Specific operation) The server sends the draft document to the person in charge's PC or tablet, and the document is displayed in the notification system (e.g., Microsoft Teams, Slack).

[0569] (Input) Generated text

[0570] (Output) Draft sent to the terminal of the person in charge

[0571] Step 9:

[0572] The person in charge will use the generated draft as a reference to reply to the customer.

[0573] (Specific Action) The person in charge uses email client software to reply, based on a template provided by the generation AI, with the message: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0574] (Input) Generated text

[0575] (Output) Reply email sent to the customer

[0576] (Application Example 1)

[0577] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0578] Traditional systems often suffered from delays in responding to strong customer emotions due to the results of sentiment analysis, making it difficult to respond quickly to such situations. Furthermore, manual handling by staff was required to generate appropriate responses, resulting in inefficiency. Consequently, rapid responses to customer satisfaction and security concerns remained a challenge.

[0579] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0580] In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of the sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating appropriate draft text, means for providing the generated draft text to the person in charge, means for analyzing the text in real time, and means for displaying the generated draft text on the person in charge's terminal. This makes it possible to respond quickly and efficiently to strong customer emotions.

[0581] "Means for saving received text" refers to a function for storing messages and text received from customers in storage such as a database.

[0582] "Means for analyzing saved text and performing sentiment analysis" refers to a function that analyzes saved text data and identifies and evaluates the emotions (joy, anger, sadness, etc.) contained within it.

[0583] A "means for evaluating the intensity of emotions based on the results of emotion analysis" refers to a function that uses the results of emotion analysis to evaluate the intensity of those emotions as a numerical value or score.

[0584] "A means of sending a notification to the person in charge when strong emotions are detected" refers to a function that informs the person in charge in real time when the intensity of emotions exceeds a set threshold.

[0585] "Means for generating appropriate wording" refers to a function that uses generative AI to automatically create optimal reply messages and responses in order to manage customer emotions.

[0586] "Means of providing the generated text to the person in charge" refers to a function for displaying or sending the generated text and corresponding messages to the person in charge's terminal.

[0587] "A means of analyzing text in real time" refers to a processing function that immediately applies sentiment analysis to newly received text and quickly obtains the results.

[0588] "Means for displaying the generated draft on the employee's terminal" refers to a function for displaying the generated reply draft on the screen of the employee's smartphone, computer, or other terminal.

[0589] This invention is a system that performs sentiment analysis based on received customer messages and emails, generates appropriate response drafts, and provides rapid and efficient customer service. This system is implemented using the following hardware and software.

[0590] Hardware and software configuration

[0591] 1. Server:

[0592] Hardware: Standard server equipment (e.g., high-performance CPU, memory, storage)

[0593] Software: Python, MySQL, Firebase

[0594] 2. Terminal:

[0595] Hardware: Smartphones, PCs, tablets, etc.

[0596] Software: Specific notification applications or web browsers

[0597] System Processing Overview

[0598] 1. Data reception and storage:

[0599] When a user receives a message from a customer, the message is sent to the server and stored in a database (MySQL). This ensures the secure storage of email and chat data.

[0600] 2. Sentiment analysis:

[0601] The server retrieves the stored messages and inputs them into a generative AI model for sentiment analysis (e.g., OpenAI GPT-4). The generative AI model performs sentiment analysis using the following prompts:

[0602] "Perform an emotional analysis of the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[0603] The analyzed results are sent back to the server as an emotion score.

[0604] 3. Sentiment evaluation and notification:

[0605] The server receives the sentiment score and evaluates its strength. If it exceeds a threshold, it sends a notification to the agent's device using Firebase Cloud Messaging, allowing the agent to immediately understand the situation.

[0606] 4. Draft generation:

[0607] The server instructs the AI ​​model to generate an appropriate response. The following prompts are used:

[0608] "Please create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm extremely frustrated!'"

[0609] The generated draft is sent to the person in charge's terminal.

[0610] Specific example

[0611] For example, if a customer sends a message saying, "The delivery is delayed and it's very frustrating!", the server will execute the following process.

[0612] Receive messages and save them to the database.

[0613] OpenAI GPT-4 is used for sentiment analysis, and a strong emotion of "very irritated" is detected.

[0614] Since the emotion score exceeded the threshold, a notification will be sent to the person in charge.

[0615] Using OpenAI GPT-4 again, we generate a suitable reply message like the following:

[0616] "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible."

[0617] The generated draft is displayed on the employee's terminal, and the employee uses it to quickly reply to the customer.

[0618] In this way, this system can respond quickly and effectively to customers' strong emotions, and can contribute to improving customer satisfaction.

[0619] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0620] Step 1:

[0621] A user receives a message from a customer. The text data received by the user via email or messaging app is sent to the server. The server stores this received data in a database. The input is the customer's message text, and the output is the message data stored in the database. Specifically, the server inserts the received text into a MySQL database using SQL commands.

[0622] Step 2:

[0623] The server analyzes stored messages and performs sentiment analysis. It retrieves the stored message data and inputs it into a generative AI model (e.g., OpenAI GPT-4). The input is the stored message text, which is a request for sentiment analysis using a prompt. The output is a sentiment score. Specifically, the server sends the following prompt to the generative AI model: "Perform sentiment analysis on the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[0624] Step 3:

[0625] The server receives sentiment scores from the generated AI model and evaluates their strength. The input is the sentiment score, and the output is the result of determining whether it exceeds a threshold. Specifically, the server compares the sentiment score to a threshold, and if it exceeds the threshold, it proceeds to the next step.

[0626] Step 4:

[0627] If a sentiment score exceeding a threshold is detected, the server uses Firebase Cloud Messaging to send a notification to the employee's device. The input is the sentiment score exceeding the threshold and the original message text, and the output is the notification to the employee's device. Specifically, the server sends a notification message via the Firebase API indicating that strong sentiment has been detected.

[0628] Step 5:

[0629] The server instructs the generative AI model to generate an appropriate response. The input is the customer message and the prompt "Create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm very frustrated!'" and the output is the generated response. Specifically, the server sends another request to the generative AI model to retrieve an appropriate response.

[0630] Step 6:

[0631] The server provides the generated text to the user's terminal. The input is the text generated by the AI ​​model, and the output is the text displayed on the user's terminal screen. Specifically, the server sends the generated text to the user's terminal and displays it through a web application or notification application.

[0632] Step 7:

[0633] The person in charge replies to the customer based on the provided draft. The input is the generated draft sent from the server, and the output is the reply message sent to the customer. Specifically, the person in charge reviews the generated draft, makes adjustments as needed, and sends the final reply message to the customer.

[0634] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0635] This invention provides a system that enables more appropriate customer service by inputting customer emails and messages into a generative AI, performing sentiment analysis, and combining it with an emotion engine that recognizes user emotions. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has a function to quickly notify the relevant personnel of this information and automatically generate and provide appropriate wording to the personnel to help them manage their emotions.

[0636] Specifically, the following processes are performed.

[0637] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[0638] Subsequently, the server reads the emails stored in the database and prepares them for input into the generative AI. Specifically, it extracts the text data from the emails and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[0639] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[0640] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[0641] At the same time, the server instructs the generative AI to generate a text that suppresses strong emotions. The generative AI generates a text that incorporates appropriate language and a solution, and sends it back to the server. For example, it might generate a text that says, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0642] Furthermore, this invention incorporates an emotion engine to analyze the user's emotions in real time and adjust the tone of the generated text accordingly. The emotion engine analyzes the user's facial expressions, voice tone, and keyboard input speed to determine their current emotional state. This allows for adjustments such as generating text with a softer tone if, for example, the person in charge is feeling stressed.

[0643] The server sends the generated draft to the employee's terminal, and the employee uses that draft as a reference to reply to the customer. This allows for a quick and appropriate response to strong customer emotions, and provides a draft that is tailored to the user's emotional state.

[0644] This allows staff to handle customer inquiries more efficiently, which is expected to lead to improved customer satisfaction. For example, by using the generated template to quickly respond to customer complaint emails, customer anger can be mitigated, and staff can handle the situation with less mental strain.

[0645] The following describes the processing flow.

[0646] Step 1:

[0647] Users receive emails and messages from customers.

[0648] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[0649] Step 2:

[0650] The server saves received emails to the database.

[0651] Specifically, the server records information such as the body of the received email, sender information, and the date and time of receipt in a database.

[0652] Step 3:

[0653] The server reads emails stored in the database and prepares them for input into the generative AI.

[0654] Specifically, the text data from the email is extracted and converted into a format that can be processed by generative AI.

[0655] Step 4:

[0656] The server converts the email content and sends it to a generation AI.

[0657] Specifically, the server sends email data to the generative AI via an API.

[0658] Step 5:

[0659] Generative AI analyzes the content of emails and performs sentiment analysis.

[0660] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[0661] Step 6:

[0662] The server receives the emotion score and evaluates the intensity of the emotion.

[0663] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[0664] Step 7:

[0665] The server sends a notification to the person in charge when strong emotions are detected.

[0666] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[0667] Step 8:

[0668] The device displays a notification to the person in charge.

[0669] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[0670] Step 9:

[0671] The server requests the generative AI to generate appropriate text.

[0672] Specifically, the server specifies the requirements for the emotionally restrained text and sends them again to the generative AI via the API.

[0673] Step 10:

[0674] The generative AI generates text based on the specified conditions.

[0675] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0676] Step 11:

[0677] The server prepares to provide the draft text generated by the generative AI to the person in charge.

[0678] Specifically, this involves preparing the generated draft document for transmission to the responsible person's terminal.

[0679] Step 12:

[0680] The terminal displays the draft document to the person in charge.

[0681] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[0682] Step 13:

[0683] The person in charge will use the generated draft as a reference and reply to the customer.

[0684] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[0685] Furthermore, the processing when combining it with the emotion engine is as follows:

[0686] Step 14:

[0687] The emotion engine recognizes the user's emotions in real time.

[0688] Specifically, it analyzes the user's facial expressions, voice tone, or keyboard input speed to determine their current emotional state.

[0689] Step 15:

[0690] The server instructs the generative AI to adjust the tone of the text based on the user's emotional state.

[0691] Specifically, if the system determines that the user is experiencing stress, it will be instructed to generate a softer-toned version of the message.

[0692] This makes it possible to respond quickly and appropriately to customers' strong emotions, which not only improves customer satisfaction but also reduces the mental burden on staff.

[0693] (Example 2)

[0694] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0695] Traditional systems struggled to respond appropriately and promptly to emails and messages from customers that contained strong emotions. Furthermore, it was difficult for staff to manage their own emotional state while interacting with customers, resulting in decreased customer satisfaction and increased stress for staff. Additionally, the quality and appropriateness of the generated messages were not guaranteed, placing a heavy burden on staff.

[0696] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving emails and messages from customers, means for storing the received text in a database, means for analyzing the stored text and performing sentiment analysis using a generative AI model, means for sending a notification to the person in charge when the sentiment score exceeds a threshold, means for generating appropriate text based on the generative AI model, means for analyzing the user's current emotional state in real time, and means for providing the generated text to the person in charge. This makes it possible to respond quickly and appropriately to strong customer emotions, and provides text that matches the emotional state of the person in charge, which can be expected to improve customer satisfaction and reduce stress for the person in charge.

[0697] "Customer" refers to individuals or companies that purchase or use goods or services.

[0698] "Emails and messages" refer to electronic means of communication that customers send to companies or their representatives.

[0699] A "database" refers to a system for efficiently storing, searching, and managing information.

[0700] A "generative AI model" refers to an algorithm that uses artificial intelligence to perform specific tasks, such as language generation or sentiment analysis.

[0701] "Sentiment analysis" refers to the process of identifying emotions from text data and quantifying them.

[0702] An "emotion score" is a numerical value obtained from the results of an emotion analysis that indicates the intensity and type of emotion.

[0703] A "threshold" is a value that indicates a specific standard; exceeding this value triggers a particular action.

[0704] "Notifications" refer to messages or alerts used to convey important information or warnings to users.

[0705] "Document draft" refers to a preliminary version or draft of a document used to communicate with a customer.

[0706] "User" refers to a person who operates this system, or a person who uses the system to handle customer inquiries.

[0707] "Real-time" refers to processing or responding to an event as close to the moment it occurs as possible.

[0708] This invention relates to a system that uses artificial intelligence technology to analyze emails and messages from customers and evaluate their emotions in order to provide more appropriate customer service. In this system, the server, terminal, and user each play important roles.

[0709] First, users receive emails and messages from customers. This reception occurs through the company's email server or messaging platform.

[0710] Next, the server saves the received emails to a database. Specifically, it uses a relational database management system (RDBMS) such as MySQL.

[0711] The server converts the stored emails into a format that can be input into a generative AI model (e.g., the BERT model in Hugging Face) for analysis. This conversion involves changing the text data to a format such as JSON.

[0712] The server sends the transformed data to the generating AI model. This data is sent via an HTTP request to the generating AI model's API endpoint.

[0713] The generative AI model analyzes the received data and performs sentiment analysis. Based on the analysis, a sentiment score is calculated. For example, a strong feeling of anger is detected from the text "very angry."

[0714] The server evaluates the sentiment score returned from the generated AI model, and if the score exceeds a pre-set threshold, it sends a notification to the device. This notification is delivered via a pop-up or alert email.

[0715] Based on the sentiment score, the server instructs the generative AI model to generate appropriate text. This text is designed to respond appropriately to the customer and may include phrases such as "We are sorry."

[0716] Furthermore, the emotion engine analyzes the user's current emotional state in real time. Specifically, it uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed.

[0717] The generated text is sent from the server to the terminal, and the person in charge uses it as a reference to reply to the customer. This allows for a quick and appropriate response to customers' strong emotions and also reduces stress for the person in charge.

[0718] As a concrete example, consider a scenario where a customer sends an email stating, "I am extremely upset about the product defect." Upon receiving this email, the server saves its contents to a database and performs sentiment analysis using a generative AI model. Because the sentiment score is very high at 75, the server sends an urgent notification to the relevant person in charge and simultaneously instructs the generative AI model to generate an appropriate message. The generated message is: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible." Furthermore, if the sentiment engine determines that the person in charge is feeling stressed, a softer tone of message will be generated.

[0719] As described above, by combining a generative AI model and an emotion engine, this system can respond quickly and appropriately to emails from customers that contain strong emotions, and also contribute to reducing stress for the staff in charge.

[0720] Example of a prompt:

[0721] Content of email received from customer: "I am very upset about the defect in the product."

[0722] Generative AI model to use: A general-purpose generative AI model (e.g., BERT model)

[0723] Instructions: "Perform sentiment analysis on the customer's email text and return the sentiment score."

[0724] Analysis result: Emotion score = 75 (Strong anger)

[0725] Notification: "We have received an email expressing strong emotions. Customer support is required."

[0726] Generated text: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible."

[0727] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0728] Step 1:

[0729] Users receive emails and messages from customers.

[0730] Specific operation: Users receive emails and messages through the company's mail server or messaging platform. For example, a customer sends an email stating, "I am very upset about a defect in the product."

[0731] Input: Emails and messages from customers

[0732] Output: Data of received emails and messages

[0733] Step 2:

[0734] The server saves the email to the database.

[0735] Specific operation: The server saves the content of received emails to a relational database (e.g., MySQL). At the same time, it also saves the email's sending date and sender information.

[0736] Input: Data of received emails and messages

[0737] Output: Email data stored in the database

[0738] Step 3:

[0739] The server converts the email into a format that can be input into the AI ​​model that generates emails.

[0740] Specific operation: The server converts the email text data into JSON format. It also formats it so that it can be sent to the API endpoint of the generated AI model.

[0741] Input: Email data stored in the database

[0742] Output: JSON format data for input into the generated AI model.

[0743] Step 4:

[0744] The server sends the converted data to the AI ​​model.

[0745] Specific operation: The server uses an HTTP request to send data to the API endpoint of the generated AI model. The request includes a generated prompt (e.g., "Perform sentiment analysis on customer email text and return the sentiment score.").

[0746] Input: JSON formatted data and prompt text

[0747] Output: Data sent to the generating AI model

[0748] Step 5:

[0749] The generative AI model performs sentiment analysis on the received data.

[0750] Specific operation: The generative AI model (e.g., a general generative AI model) analyzes the received text data and performs a process of quantifying emotions. It calculates an anger score from "very annoyed" as an emotion score and returns it as a numerical value (e.g., 75).

[0751] Input: Converted JSON data, prompt text

[0752] Output: Calculation result of the emotion score

[0753] Step 6:

[0754] The server evaluates the sentiment analysis results and sends notifications to the relevant personnel.

[0755] Specific operation: The server evaluates the sentiment score received from the generated AI model, and if the score exceeds a threshold (e.g., 50), it sends a pop-up notification or alert email to the person in charge.

[0756] Input: Sentiment score from a generated AI model

[0757] Output: Notification to the person in charge

[0758] Step 7:

[0759] The server instructs the AI ​​model to generate appropriate text.

[0760] Specific operation: The server sends a prompt message to the AI ​​model again, instructing it to generate a text, and then generates an appropriate response. For example, a sentence such as "We sincerely apologize for the inconvenience caused by the product defect" is generated.

[0761] Input: Prompt based on sentiment score

[0762] Output: Generated text

[0763] Step 8:

[0764] The emotion engine analyzes the user's current emotional state in real time.

[0765] Specific operation: The emotion engine uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed. Based on this, it assesses the stress level of the person in charge and adjusts the tone of the drafted text accordingly.

[0766] Input: Real-time data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[0767] Output: Evaluation results of the person in charge's emotional state

[0768] Step 9:

[0769] The server sends the optimal draft to the person in charge's terminal.

[0770] Specific operation: The server combines the generated text with adjustments based on the user's emotional state and sends the optimal text to the terminal. The person in charge uses this text as a reference to reply to the customer.

[0771] Input: Generated text, user's emotional state

[0772] Output: Optimal text sent to the terminal

[0773] This enables quick and appropriate responses to customers' strong emotions, and by providing wording tailored to the emotional state of the person handling the request, it leads to improved customer satisfaction and reduced stress for the person handling the request.

[0774] (Application Example 2)

[0775] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0776] Logistics centers are required to respond quickly and appropriately to customer complaints and inquiries. However, manually analyzing each message and providing an appropriate response requires a tremendous amount of effort and time, potentially leading to decreased customer satisfaction. Furthermore, the quality of responses is inconsistent because they are influenced by the emotional state of the person handling them.

[0777] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating an appropriate draft, means for providing the generated draft to the person in charge, means for receiving and processing customer messages on a smart device, and means for analyzing the emotional state of the person in charge and adjusting the tone of the draft. This makes it possible to respond quickly and appropriately to customer complaints and inquiries, and is expected to improve customer satisfaction and reduce the burden on the person in charge.

[0778] "Means for saving received text" refers to a device or program for saving messages and emails sent by customers to a database.

[0779] "Methods for analyzing saved text and performing sentiment analysis" refers to programs that use natural language processing techniques to analyze saved text data and identify emotions and their intensity.

[0780] "Means for evaluating the intensity of emotions based on the results of emotion analysis" refers to a device or program that quantifies the data obtained from emotion analysis and evaluates its intensity.

[0781] "Means for sending notifications to responsible personnel when strong emotions are detected" refers to a device or program that sends alerts or messages to the relevant personnel when strong emotions are detected as a result of the analysis.

[0782] "Means for generating appropriate text" refers to a device or program that uses natural language generation technology to generate appropriate response text according to the customer's emotions.

[0783] "Means of providing the generated draft to the person in charge" refers to a device or program for sending the generated reply draft to the person in charge's device.

[0784] "Means for receiving and processing customer messages on a smart device" refers to a device or program for receiving, analyzing, and processing messages from customers on a mobile device such as a smartphone or tablet.

[0785] "Means for analyzing the emotional state of the person in charge and adjusting the tone of the draft" refers to a device or program for analyzing the emotions of the person in charge and adjusting the tone of the draft generated based on the results.

[0786] This invention relates to a system for quickly and appropriately responding to customer complaints and inquiries in a logistics center. This system consists of a server, smart devices, a generative AI model, and an emotion engine.

[0787] The server receives messages and emails sent by customers and stores them in a database. The stored text data is sent to a generative AI for analysis. The generative AI uses natural language processing techniques to analyze the text and perform sentiment analysis. For example, from a message such as "I am very angry about the product defect," it identifies the emotion of anger and quantifies its intensity. This sentiment score is then sent back to the server.

[0788] The server evaluates the returned sentiment score and compares it to a pre-set threshold. If a strong sentiment exceeding the threshold is detected, an alert is immediately sent to the employee's smart device. It also instructs a generative AI to generate an appropriate message. The generated message is then provided to the employee's smart device. For example, a message such as, "We sincerely apologize for the inconvenience caused by the delay in receiving your order. We are currently investigating the situation and will address it as soon as possible," might be generated.

[0789] Furthermore, the emotion engine analyzes the emotional state of the person handling the task and adjusts the tone of the generated text accordingly. The emotion engine determines emotions in real time based on factors such as the person's facial expressions, voice tone, and keyboard typing speed. For example, if the person is feeling stressed, a softer tone of text will be generated.

[0790] The hardware used to implement this system includes high-performance servers, smartphones, or tablets. The software used includes TensorFlow, natural language processing libraries (SpaCy, NLTK), a reactive framework (React Native), Docker, and a RESTful API (Flask / Django).

[0791] As a concrete example, input the following prompt into the generative AI.

[0792] "The delivery of my order is delayed. I'm extremely frustrated."

[0793] "Perform a sentiment analysis on this message and generate appropriate wording."

[0794] The sentiment analysis results indicate anger (score: 9.2). The generated text based on these results is as follows:

[0795] "We sincerely apologize for the inconvenience regarding the delivery of your order. We will address this issue as quickly as possible, so please bear with us for a little while longer."

[0796] This will enable quick and accurate responses to customer complaints and inquiries, leading to improved customer satisfaction and reduced workload for staff.

[0797] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0798] Step 1:

[0799] The server receives messages and emails from customers. The received text data is stored in a database.

[0800] Input: Customer messages or emails

[0801] Output: Text data stored in the database

[0802] Step 2:

[0803] The server retrieves the stored text data for analysis and converts it into a format that the generating AI model can process. Specifically, it cleans the text and performs any necessary preprocessing.

[0804] Input: Text data extracted from a database

[0805] Output: Text data converted into a format that can be processed by the generative AI model.

[0806] Step 3:

[0807] The generative AI model receives the transformed text data and performs sentiment analysis. For example, it identifies the emotion of anger from the phrase "very angry" and quantifies its intensity.

[0808] Input: Text data sent to a generative AI.

[0809] Output: Emotion score (Example: Anger intensity score 9.2)

[0810] Step 4:

[0811] The server evaluates the sentiment score returned from the generated AI model and compares it to a pre-set threshold. If the threshold is exceeded, a notification is sent to the person in charge's terminal.

[0812] Input: Sentiment score

[0813] Output: Notification when threshold is exceeded

[0814] Step 5:

[0815] The server instructs the AI ​​model to generate appropriate text. The text includes expressions that alleviate anger.

[0816] Input: Sentiment score and message content

[0817] Output: Generated text (Example: "We apologize for the inconvenience this has caused. We will address this issue as soon as possible.")

[0818] Step 6:

[0819] The generated draft is sent to the employee's smart device. The employee uses this draft as a reference to quickly reply to the customer.

[0820] Input: Generated text

[0821] Output: Draft sent to the person in charge's terminal

[0822] Step 7:

[0823] The emotion engine analyzes the emotional state of the person in charge and adjusts the tone of the generated text accordingly. If the person in charge is stressed, the text will be set to a softer tone.

[0824] Input: Emotional data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[0825] Output: Adjusted tone of the text

[0826] Step 8:

[0827] The server provides the staff with a pre-formatted response to use in replying to the customer. This allows for a response that soothes the customer's emotions and reduces the burden on the staff.

[0828] Input: Revised text

[0829] Output: Final reply message sent to the customer

[0830] By following these steps, the entire system will be able to respond quickly and appropriately to customer complaints and inquiries, which is expected to improve customer satisfaction and reduce the burden on staff.

[0831] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0832] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0833] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0834] [Third Embodiment]

[0835] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0836] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0837] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0838] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0839] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0840] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0841] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0842] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0843] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0844] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0845] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0846] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0847] This invention relates to a system that inputs emails and messages from customers into a generative AI and performs sentiment analysis. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has the function of quickly notifying the relevant personnel of this information and automatically generating and providing appropriate drafts of messages to help them manage their emotions.

[0848] Specifically, the following processes are performed.

[0849] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[0850] Next, the server prepares the stored email content for input into the generative AI. Specifically, it extracts the text data from the email and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[0851] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[0852] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[0853] At the same time, the server instructs the generative AI to generate a message that suppresses strong emotions. The generative AI generates a message that includes appropriate language and solutions, and sends it back to the server. For example, it might say, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0854] The server sends the generated draft to the employee's terminal, who then uses that draft to respond to the customer. This allows for a quick and appropriate response to strong customer emotions.

[0855] By using this system, staff can handle customer inquiries efficiently, which is expected to lead to improved customer satisfaction. A concrete example is the ability to respond quickly to complaint emails. If a customer sends an email saying, "The delivery is delayed and I'm very frustrated!", the system can quickly respond using a generated message such as, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," thereby calming the customer's anger.

[0856] The following describes the processing flow.

[0857] Step 1:

[0858] Users receive emails and messages from customers.

[0859] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[0860] Step 2:

[0861] The server saves received emails to the database.

[0862] Specifically, the server records information such as the email body, sender information, and the date and time of receipt in a database.

[0863] Step 3:

[0864] The server reads emails stored in the database and prepares them for input into the generative AI.

[0865] Specifically, the server extracts the text data from the email and converts it into a format for analysis by generative AI.

[0866] Step 4:

[0867] The server converts the email content and sends it to a generation AI.

[0868] Specifically, the server sends email data to the generative AI via an API.

[0869] Step 5:

[0870] Generative AI analyzes the content of emails and performs sentiment analysis.

[0871] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[0872] Step 6:

[0873] The server receives the emotion score and evaluates the intensity of the emotion.

[0874] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[0875] Step 7:

[0876] The server sends a notification to the person in charge when strong emotions are detected.

[0877] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[0878] Step 8:

[0879] The device displays a notification to the person in charge.

[0880] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[0881] Step 9:

[0882] The server requests the generative AI to generate appropriate text.

[0883] Specifically, the server specifies the requirements for the text to suppress emotions and sends them again to the generative AI via the API.

[0884] Step 10:

[0885] The generative AI generates text based on the specified conditions.

[0886] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0887] Step 11:

[0888] The server provides the generated draft to the person in charge.

[0889] Specifically, the server generates a draft document, which is then sent to the person in charge, allowing them to view it on their terminal.

[0890] Step 12:

[0891] The terminal displays the draft document to the person in charge.

[0892] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[0893] Step 13:

[0894] The person in charge will use the generated draft as a reference and reply to the customer.

[0895] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[0896] This makes it possible to respond quickly and effectively to customers' strong emotions.

[0897] (Example 1)

[0898] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0899] In customer communication, a challenge is responding quickly and appropriately to messages that evoke strong emotions. Conventional systems have been unable to respond quickly to such messages, potentially resulting in decreased customer satisfaction. In contrast, this invention aims to improve the efficiency and quality of customer service by automatically analyzing emotions and providing appropriate responses.

[0900] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0901] In this invention, the server includes means for storing received communications, means for analyzing the stored communications and performing sentiment analysis, means for evaluating the intensity of emotions based on the results of the sentiment analysis, means for sending notifications when the intensity of emotions exceeds a pre-set threshold, means for generating selected example sentences, and means for providing the generated example sentences to the relevant personnel. This makes it possible to quickly detect messages from customers with strong emotions and automatically generate and provide appropriate countermeasures.

[0902] "Received communications" refers to messages and data received by the target system or device.

[0903] "Means of preservation" refers to devices or programs for recording received communications in a database or storage medium.

[0904] "Means of analysis and sentiment analysis" refer to algorithms and software that process stored communications and identify emotions from their content.

[0905] A "means for evaluating the intensity of emotions" is a mechanism that determines, either numerically or qualitatively, how strong an emotion is, based on the results of an emotion analysis.

[0906] "Means of sending notifications" refers to communication methods and protocols used to transmit warnings and information to relevant personnel in real time, based on evaluation results and other factors.

[0907] "Means for generating selected example sentences" refers to artificial intelligence or programs that automatically create appropriate sentences to correspond to specific emotions.

[0908] "Means of providing" refers to means of sending or displaying generated example sentences, etc., for the purpose of showing them to the person in charge.

[0909] This invention is a system for analyzing customer communications and providing appropriate responses. This system stores received communications, performs sentiment analysis, evaluates the intensity of emotions, and generates appropriate example sentences. The following hardware and software are used to implement these functions.

[0910] System Configuration

[0911] This system consists of a user's terminal, a server, and a generative AI model.

[0912] User's device:

[0913] Receive communications from customers using email client software (e.g., Outlook, Gmail).

[0914] server:

[0915] The system uses a mail server (e.g., Postfix, Sendmail) to receive communications and saves them to a database server (e.g., MySQL, PostgreSQL).

[0916] Use a data transformation library (e.g., Pandas) to convert the data into a format suitable for sentiment analysis.

[0917] A notification system (e.g., Slack API, email notification system) is used to notify the responsible person when a sentiment score exceeding a threshold is detected.

[0918] Generative AI models:

[0919] We analyze communication content using natural language processing libraries (e.g., NLTK, SpaCy) and perform sentiment analysis.

[0920] The analysis results are returned to the server as an emotion score.

[0921] Based on the prompt, it generates example sentences that include appropriate wording and solutions.

[0922] Processing Overview

[0923] 1. Check for communications received by the user (e.g., emails such as "I am very upset about the product defect").

[0924] 2. The server stores the emails on the database server.

[0925] 3. The server uses the Python Pandas library to convert the email text data into a format that can be processed by the generative AI.

[0926] 4. The generative AI uses a natural language processing library to detect strong emotions from the context of the email. This detection result is returned to the server as a numerical emotion score.

[0927] 5. The server compares the sentiment score to a threshold, and if it exceeds the threshold, it sends a real-time notification to the person in charge using the notification system.

[0928] 6. The server sends prompt text to the generative AI, instructing it to generate example sentences to suppress emotions.

[0929] Example prompt: For the sentence "Customer email: I am very upset about the product defect," please generate an example sentence that includes an apology and a solution.

[0930] 7. The generative AI generates appropriate example sentences and sends them back to the server.

[0931] 8. The server sends the generated example sentences to the employee's terminal and displays the draft sentences on the notification system.

[0932] 9. The person in charge uses email client software to reply to the customer based on example sentences provided by the generative AI.

[0933] Specific example

[0934] This is an example of how to handle a customer who sends an email saying, "The delivery is delayed and I'm extremely frustrated!" The user who receives this email checks it using their email client software, and the server saves its contents to a database server. The server then converts the email text into a format that can be processed by a generative AI and sends it to the generative AI. The generative AI performs sentiment analysis, detects a strong emotion such as "extremely frustrated," and sends a sentiment score back to the server.

[0935] The server evaluates the sentiment score and sends a notification to the employee's terminal if it exceeds the threshold. At the same time, it sends a prompt to the generative AI, instructing it to generate example sentences. The generative AI generates an example sentence, for example, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," and sends it back to the server. The server sends the generated example sentence to the employee's terminal, and the employee uses email client software to reply to the customer based on that example sentence.

[0936] This will enable us to respond quickly and appropriately to customers' strong emotions, which is expected to improve customer satisfaction.

[0937] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0938] Step 1:

[0939] The user receives communications from the customer.

[0940] (Specific actions) The user's email client software (e.g., Outlook, Gmail) checks for emails from customers (e.g., "I am very upset about the product defect").

[0941] (Input) Email from customer

[0942] (Output) Received email content

[0943] Step 2:

[0944] The server saves the received communications to the database.

[0945] (Specific operation) The mail server saves received emails to a database server (e.g., MySQL, PostgreSQL) and records the email content.

[0946] (Input) Received email content

[0947] (Output) Email data stored in the database

[0948] Step 3:

[0949] The server performs a format conversion so that the email content can be input into the generation AI.

[0950] (Specific operation) The server uses the Python Pandas library to convert the saved email text data into a format that can be processed by the generative AI (e.g., JSON format).

[0951] (Input) Email data stored in the database

[0952] (Output) Data converted into a format that can be processed by a generative AI.

[0953] Step 4:

[0954] A generative AI analyzes the formatted email data and performs sentiment analysis.

[0955] (Specific operation) A generative AI uses a natural language processing library (e.g., SpaCy, NLTK) to detect a strong emotion such as "very angry" from the context of the email and quantifies it as an emotion score (e.g., anger 0.8).

[0956] (Input) Data converted into a format that can be processed by a generative AI.

[0957] (Output) Emotion score (Example: Anger 0.8)

[0958] Step 5:

[0959] The server evaluates the sentiment score and notifies the person in charge if it exceeds the threshold.

[0960] (Specific operation) The server compares the emotion score to a threshold (e.g., anger 0.7), and if the threshold is exceeded, it sends a real-time notification to the person in charge's terminal using a notification system (e.g., Slack API).

[0961] (Input) Emotion score

[0962] (Output) Notification sent to the responsible person's terminal

[0963] Step 6:

[0964] The server instructs the generative AI to generate text that suppresses emotions.

[0965] (Specific operation) The server sends a prompt message to the generation AI: "Customer email: Please generate a draft email that includes an apology and a solution regarding a customer's email stating that they are very upset about a product defect."

[0966] (Input) Sentiment score, prompt text

[0967] (Output) Instructions for text generation sent to the generation system AI.

[0968] Step 7:

[0969] A generative AI generates the text.

[0970] (Specific operation) The generative AI creates an appropriate message (e.g., "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible.") and sends it back to the server.

[0971] (Input) Instructions for generating text draft

[0972] (Output) Generated text

[0973] Step 8:

[0974] The server sends the generated draft to the user's terminal.

[0975] (Specific operation) The server sends the draft document to the person in charge's PC or tablet, and the document is displayed in the notification system (e.g., Microsoft Teams, Slack).

[0976] (Input) Generated text

[0977] (Output) Draft sent to the terminal of the person in charge

[0978] Step 9:

[0979] The person in charge will use the generated draft as a reference to reply to the customer.

[0980] (Specific Action) The person in charge uses email client software to reply, based on a template provided by the generation AI, with the message: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[0981] (Input) Generated text

[0982] (Output) Reply email sent to the customer

[0983] (Application Example 1)

[0984] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0985] Traditional systems often suffered from delays in responding to strong customer emotions due to the results of sentiment analysis, making it difficult to respond quickly to such situations. Furthermore, manual handling by staff was required to generate appropriate responses, resulting in inefficiency. Consequently, rapid responses to customer satisfaction and security concerns remained a challenge.

[0986] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0987] In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of the sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating appropriate draft text, means for providing the generated draft text to the person in charge, means for analyzing the text in real time, and means for displaying the generated draft text on the person in charge's terminal. This makes it possible to respond quickly and efficiently to strong customer emotions.

[0988] "Means for saving received text" refers to a function for storing messages and text received from customers in storage such as a database.

[0989] "Means for analyzing saved text and performing sentiment analysis" refers to a function that analyzes saved text data and identifies and evaluates the emotions (joy, anger, sadness, etc.) contained within it.

[0990] A "means for evaluating the intensity of emotions based on the results of emotion analysis" refers to a function that uses the results of emotion analysis to evaluate the intensity of those emotions as a numerical value or score.

[0991] "A means of sending a notification to the person in charge when strong emotions are detected" refers to a function that informs the person in charge in real time when the intensity of emotions exceeds a set threshold.

[0992] "Means for generating appropriate wording" refers to a function that uses generative AI to automatically create optimal reply messages and responses in order to manage customer emotions.

[0993] "Means of providing the generated text to the person in charge" refers to a function for displaying or sending the generated text and corresponding messages to the person in charge's terminal.

[0994] "A means of analyzing text in real time" refers to a processing function that immediately applies sentiment analysis to newly received text and quickly obtains the results.

[0995] "Means for displaying the generated draft on the employee's terminal" refers to a function for displaying the generated reply draft on the screen of the employee's smartphone, computer, or other terminal.

[0996] This invention is a system that performs sentiment analysis based on received customer messages and emails, generates appropriate response drafts, and provides rapid and efficient customer service. This system is implemented using the following hardware and software.

[0997] Hardware and software configuration

[0998] 1. Server:

[0999] Hardware: Standard server equipment (e.g., high-performance CPU, memory, storage)

[1000] Software: Python, MySQL, Firebase

[1001] 2. Terminal:

[1002] Hardware: Smartphones, PCs, tablets, etc.

[1003] Software: Specific notification applications or web browsers

[1004] System Processing Overview

[1005] 1. Data reception and storage:

[1006] When a user receives a message from a customer, the message is sent to the server and stored in a database (MySQL). This ensures the secure storage of email and chat data.

[1007] 2. Sentiment analysis:

[1008] The server retrieves the stored messages and inputs them into a generative AI model for sentiment analysis (e.g., OpenAI GPT-4). The generative AI model performs sentiment analysis using the following prompts:

[1009] "Perform an emotional analysis of the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[1010] The analyzed results are sent back to the server as an emotion score.

[1011] 3. Sentiment evaluation and notification:

[1012] The server receives the sentiment score and evaluates its strength. If it exceeds a threshold, it sends a notification to the agent's device using Firebase Cloud Messaging, allowing the agent to immediately understand the situation.

[1013] 4. Draft generation:

[1014] The server instructs the AI ​​model to generate an appropriate response. The following prompts are used:

[1015] "Please create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm extremely frustrated!'"

[1016] The generated draft is sent to the person in charge's terminal.

[1017] Specific example

[1018] For example, if a customer sends a message saying, "The delivery is delayed and it's very frustrating!", the server will execute the following process.

[1019] Receive messages and save them to the database.

[1020] OpenAI GPT-4 is used for sentiment analysis, and a strong emotion of "very irritated" is detected.

[1021] Since the emotion score exceeded the threshold, a notification will be sent to the person in charge.

[1022] Using OpenAI GPT-4 again, we generate a suitable reply message like the following:

[1023] "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible."

[1024] The generated draft is displayed on the employee's terminal, and the employee uses it to quickly reply to the customer.

[1025] In this way, this system can respond quickly and effectively to customers' strong emotions, and can contribute to improving customer satisfaction.

[1026] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1027] Step 1:

[1028] A user receives a message from a customer. The text data received by the user via email or messaging app is sent to the server. The server stores this received data in a database. The input is the customer's message text, and the output is the message data stored in the database. Specifically, the server inserts the received text into a MySQL database using SQL commands.

[1029] Step 2:

[1030] The server analyzes stored messages and performs sentiment analysis. It retrieves the stored message data and inputs it into a generative AI model (e.g., OpenAI GPT-4). The input is the stored message text, which is a request for sentiment analysis using a prompt. The output is a sentiment score. Specifically, the server sends the following prompt to the generative AI model: "Perform sentiment analysis on the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[1031] Step 3:

[1032] The server receives sentiment scores from the generated AI model and evaluates their strength. The input is the sentiment score, and the output is the result of determining whether it exceeds a threshold. Specifically, the server compares the sentiment score to a threshold, and if it exceeds the threshold, it proceeds to the next step.

[1033] Step 4:

[1034] If a sentiment score exceeding a threshold is detected, the server uses Firebase Cloud Messaging to send a notification to the employee's device. The input is the sentiment score exceeding the threshold and the original message text, and the output is the notification to the employee's device. Specifically, the server sends a notification message via the Firebase API indicating that strong sentiment has been detected.

[1035] Step 5:

[1036] The server instructs the generative AI model to generate an appropriate response. The input is the customer message and the prompt "Create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm very frustrated!'" and the output is the generated response. Specifically, the server sends another request to the generative AI model to retrieve an appropriate response.

[1037] Step 6:

[1038] The server provides the generated text to the user's terminal. The input is the text generated by the AI ​​model, and the output is the text displayed on the user's terminal screen. Specifically, the server sends the generated text to the user's terminal and displays it through a web application or notification application.

[1039] Step 7:

[1040] The person in charge replies to the customer based on the provided draft. The input is the generated draft sent from the server, and the output is the reply message sent to the customer. Specifically, the person in charge reviews the generated draft, makes adjustments as needed, and sends the final reply message to the customer.

[1041] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1042] This invention provides a system that enables more appropriate customer service by inputting customer emails and messages into a generative AI, performing sentiment analysis, and combining it with an emotion engine that recognizes user emotions. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has a function to quickly notify the relevant personnel of this information and automatically generate and provide appropriate wording to the personnel to help them manage their emotions.

[1043] Specifically, the following processes are performed.

[1044] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[1045] Subsequently, the server reads the emails stored in the database and prepares them for input into the generative AI. Specifically, it extracts the text data from the emails and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[1046] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[1047] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[1048] At the same time, the server instructs the generative AI to generate a text that suppresses strong emotions. The generative AI generates a text that incorporates appropriate language and a solution, and sends it back to the server. For example, it might generate a text that says, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1049] Furthermore, this invention incorporates an emotion engine to analyze the user's emotions in real time and adjust the tone of the generated text accordingly. The emotion engine analyzes the user's facial expressions, voice tone, and keyboard input speed to determine their current emotional state. This allows for adjustments such as generating text with a softer tone if, for example, the person in charge is feeling stressed.

[1050] The server sends the generated draft to the employee's terminal, and the employee uses that draft as a reference to reply to the customer. This allows for a quick and appropriate response to strong customer emotions, and provides a draft that is tailored to the user's emotional state.

[1051] This allows staff to handle customer inquiries more efficiently, which is expected to lead to improved customer satisfaction. For example, by using the generated template to quickly respond to customer complaint emails, customer anger can be mitigated, and staff can handle the situation with less mental strain.

[1052] The following describes the processing flow.

[1053] Step 1:

[1054] Users receive emails and messages from customers.

[1055] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[1056] Step 2:

[1057] The server saves received emails to the database.

[1058] Specifically, the server records information such as the body of the received email, sender information, and the date and time of receipt in a database.

[1059] Step 3:

[1060] The server reads emails stored in the database and prepares them for input into the generative AI.

[1061] Specifically, the text data from the email is extracted and converted into a format that can be processed by a generative AI.

[1062] Step 4:

[1063] The server converts the email content and sends it to a generation AI.

[1064] Specifically, the server sends email data to the generative AI via an API.

[1065] Step 5:

[1066] Generative AI analyzes the content of emails and performs sentiment analysis.

[1067] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[1068] Step 6:

[1069] The server receives the emotion score and evaluates the intensity of the emotion.

[1070] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[1071] Step 7:

[1072] The server sends a notification to the person in charge when strong emotions are detected.

[1073] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[1074] Step 8:

[1075] The device displays a notification to the person in charge.

[1076] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[1077] Step 9:

[1078] The server requests the generative AI to generate appropriate text.

[1079] Specifically, the server specifies the requirements for the text to suppress emotions and sends them again to the generative AI via the API.

[1080] Step 10:

[1081] The generative AI generates text based on the specified conditions.

[1082] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1083] Step 11:

[1084] The server prepares to provide the draft text generated by the generative AI to the person in charge.

[1085] Specifically, this involves preparing the generated draft document for transmission to the responsible person's terminal.

[1086] Step 12:

[1087] The terminal displays the draft document to the person in charge.

[1088] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[1089] Step 13:

[1090] The person in charge will use the generated draft as a reference and reply to the customer.

[1091] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[1092] Furthermore, the processing when combining it with the emotion engine is as follows:

[1093] Step 14:

[1094] The emotion engine recognizes the user's emotions in real time.

[1095] Specifically, the system analyzes the user's facial expressions, voice, or keyboard input speed to determine their current emotional state.

[1096] Step 15:

[1097] The server instructs the generative AI to adjust the tone of the text based on the user's emotional state.

[1098] Specifically, if the system determines that the user is experiencing stress, it will be instructed to generate a softer-toned version of the message.

[1099] This makes it possible to respond quickly and appropriately to customers' strong emotions, which not only improves customer satisfaction but also reduces the mental burden on staff.

[1100] (Example 2)

[1101] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1102] Traditional systems struggled to respond appropriately and promptly to emails and messages from customers that contained strong emotions. Furthermore, it was difficult for staff to manage their own emotional state while interacting with customers, resulting in decreased customer satisfaction and increased stress for staff. Additionally, the quality and appropriateness of the generated messages were not guaranteed, placing a heavy burden on staff.

[1103] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving emails and messages from customers, means for storing the received text in a database, means for analyzing the stored text and performing sentiment analysis using a generative AI model, means for sending a notification to the person in charge when the sentiment score exceeds a threshold, means for generating appropriate text based on the generative AI model, means for analyzing the user's current emotional state in real time, and means for providing the generated text to the person in charge. This makes it possible to respond quickly and appropriately to strong customer emotions, and provides text that matches the emotional state of the person in charge, which can be expected to improve customer satisfaction and reduce stress for the person in charge.

[1104] "Customer" refers to individuals or companies that purchase or use goods or services.

[1105] "Emails and messages" refer to electronic means of communication that customers send to companies or their representatives.

[1106] A "database" refers to a system for efficiently storing, searching, and managing information.

[1107] A "generative AI model" refers to an algorithm that uses artificial intelligence to perform specific tasks, such as language generation or sentiment analysis.

[1108] "Sentiment analysis" refers to the process of identifying emotions from text data and quantifying them.

[1109] An "emotion score" is a numerical value obtained from the results of an emotion analysis that indicates the intensity and type of emotion.

[1110] A "threshold" is a value that indicates a specific standard; exceeding this value triggers a particular action.

[1111] "Notifications" refer to messages or alerts used to convey important information or warnings to users.

[1112] "Document draft" refers to a preliminary version or draft of a document used to communicate with a customer.

[1113] "User" refers to a person who operates this system, or a person who uses the system to handle customer inquiries.

[1114] "Real-time" refers to processing or responding to an event as close to the moment it occurs as possible.

[1115] This invention relates to a system that uses artificial intelligence technology to analyze emails and messages from customers and evaluate their emotions in order to provide more appropriate customer service. In this system, the server, terminal, and user each play important roles.

[1116] First, users receive emails and messages from customers. This reception occurs through the company's email server or messaging platform.

[1117] Next, the server saves the received emails to a database. Specifically, it uses a relational database management system (RDBMS) such as MySQL.

[1118] The server converts the stored emails into a format that can be input into a generative AI model (e.g., the BERT model in Hugging Face) for analysis. This conversion involves changing the text data to a format such as JSON.

[1119] The server sends the transformed data to the generating AI model. This data is sent via an HTTP request to the generating AI model's API endpoint.

[1120] The generative AI model analyzes the received data and performs sentiment analysis. Based on the analysis, a sentiment score is calculated. For example, a strong feeling of anger is detected from the text "very angry."

[1121] The server evaluates the sentiment score returned from the generated AI model, and if the score exceeds a pre-set threshold, it sends a notification to the device. This notification is delivered via a pop-up or alert email.

[1122] Based on the sentiment score, the server instructs the generative AI model to generate appropriate text. This text is designed to respond appropriately to the customer and may include phrases such as "We are sorry."

[1123] Furthermore, the emotion engine analyzes the user's current emotional state in real time. Specifically, it uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed.

[1124] The generated text is sent from the server to the terminal, and the person in charge uses it as a reference to reply to the customer. This allows for a quick and appropriate response to customers' strong emotions and also reduces stress for the person in charge.

[1125] As a concrete example, consider a scenario where a customer sends an email stating, "I am extremely upset about the product defect." Upon receiving this email, the server saves its contents to a database and performs sentiment analysis using a generative AI model. Because the sentiment score is very high at 75, the server sends an urgent notification to the relevant person in charge and simultaneously instructs the generative AI model to generate an appropriate message. The generated message is: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible." Furthermore, if the sentiment engine determines that the person in charge is feeling stressed, a softer tone of message will be generated.

[1126] As described above, by combining a generative AI model and an emotion engine, this system can respond quickly and appropriately to emails from customers that contain strong emotions, and also contribute to reducing stress for the staff in charge.

[1127] Example of a prompt:

[1128] Content of email received from customer: "I am very upset about the defect in the product."

[1129] Generative AI model to use: A general-purpose generative AI model (e.g., BERT model)

[1130] Instructions: "Perform sentiment analysis on the customer's email text and return the sentiment score."

[1131] Analysis result: Emotion score = 75 (Strong anger)

[1132] Notification: "We have received an email expressing strong emotions. Customer support is required."

[1133] Generated text: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible."

[1134] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1135] Step 1:

[1136] Users receive emails and messages from customers.

[1137] Specific operation: Users receive emails and messages through the company's mail server or messaging platform. For example, a customer sends an email stating, "I am very upset about a defect in the product."

[1138] Input: Emails and messages from customers

[1139] Output: Data of received emails and messages

[1140] Step 2:

[1141] The server saves the email to the database.

[1142] Specific operation: The server saves the content of received emails to a relational database (e.g., MySQL). At the same time, it also saves the email's sending date and sender information.

[1143] Input: Data of received emails and messages

[1144] Output: Email data stored in the database

[1145] Step 3:

[1146] The server converts the email into a format that can be input into the AI ​​model that generates emails.

[1147] Specific operation: The server converts the email text data into JSON format. It also formats it so that it can be sent to the API endpoint of the generated AI model.

[1148] Input: Email data stored in the database

[1149] Output: JSON format data for input into the generated AI model.

[1150] Step 4:

[1151] The server sends the converted data to the AI ​​model.

[1152] Specific operation: The server uses an HTTP request to send data to the API endpoint of the generated AI model. The request includes a generated prompt (e.g., "Perform sentiment analysis on customer email text and return the sentiment score.").

[1153] Input: JSON formatted data and prompt text

[1154] Output: Data sent to the generating AI model

[1155] Step 5:

[1156] The generative AI model performs sentiment analysis on the received data.

[1157] Specific operation: The generative AI model (e.g., a general generative AI model) analyzes the received text data and performs a process of quantifying emotions. It calculates an anger score from "very annoyed" as an emotion score and returns it as a numerical value (e.g., 75).

[1158] Input: Converted JSON data, prompt text

[1159] Output: Calculation result of the emotion score

[1160] Step 6:

[1161] The server evaluates the sentiment analysis results and sends notifications to the relevant personnel.

[1162] Specific operation: The server evaluates the sentiment score received from the generated AI model, and if the score exceeds a threshold (e.g., 50), it sends a pop-up notification or alert email to the person in charge.

[1163] Input: Sentiment score from a generated AI model

[1164] Output: Notification to the person in charge

[1165] Step 7:

[1166] The server instructs the AI ​​model to generate appropriate text.

[1167] Specific operation: The server sends a prompt message to the AI ​​model again, instructing it to generate a text, and then generates an appropriate response. For example, a sentence such as "We sincerely apologize for the inconvenience caused by the product defect" is generated.

[1168] Input: Prompt based on sentiment score

[1169] Output: Generated text

[1170] Step 8:

[1171] The emotion engine analyzes the user's current emotional state in real time.

[1172] Specific operation: The emotion engine uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed. Based on this, it assesses the stress level of the person in charge and adjusts the tone of the drafted text accordingly.

[1173] Input: Real-time data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[1174] Output: Evaluation results of the person in charge's emotional state

[1175] Step 9:

[1176] The server sends the optimal draft to the person in charge's terminal.

[1177] Specific operation: The server combines the generated text with adjustments based on the user's emotional state and sends the optimal text to the terminal. The person in charge uses this text as a reference to reply to the customer.

[1178] Input: Generated text, user's emotional state

[1179] Output: Optimal text sent to the terminal

[1180] This enables quick and appropriate responses to strong customer emotions, and by providing wording tailored to the emotional state of the person handling the request, it leads to improved customer satisfaction and reduced stress for the person handling the request.

[1181] (Application Example 2)

[1182] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1183] Logistics centers are required to respond quickly and appropriately to customer complaints and inquiries. However, manually analyzing each message and providing an appropriate response requires a tremendous amount of effort and time, potentially leading to decreased customer satisfaction. Furthermore, the quality of responses is inconsistent because they are influenced by the emotional state of the person handling them.

[1184] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating an appropriate draft, means for providing the generated draft to the person in charge, means for receiving and processing customer messages on a smart device, and means for analyzing the emotional state of the person in charge and adjusting the tone of the draft. This makes it possible to respond quickly and appropriately to customer complaints and inquiries, and is expected to improve customer satisfaction and reduce the burden on the person in charge.

[1185] "Means for saving received text" refers to a device or program for saving messages and emails sent by customers to a database.

[1186] "Methods for analyzing saved text and performing sentiment analysis" refers to programs that use natural language processing techniques to analyze saved text data and identify emotions and their intensity.

[1187] "Means for evaluating the intensity of emotions based on the results of emotion analysis" refers to a device or program that quantifies the data obtained from emotion analysis and evaluates its intensity.

[1188] "Means for sending notifications to the person in charge when strong emotions are detected" refers to a device or program that sends an alert or message to the relevant person in charge when strong emotions are detected as a result of the analysis.

[1189] "Means for generating appropriate text" refers to a device or program that uses natural language generation technology to generate appropriate response text according to the customer's emotions.

[1190] "Means of providing the generated draft to the person in charge" refers to a device or program for sending the generated reply draft to the person in charge's device.

[1191] "Means for receiving and processing customer messages on a smart device" refers to a device or program for receiving, analyzing, and processing messages from customers on a mobile device such as a smartphone or tablet.

[1192] "Means for analyzing the emotional state of the person in charge and adjusting the tone of the draft" refers to a device or program for analyzing the emotions of the person in charge and adjusting the tone of the draft generated based on the results.

[1193] This invention relates to a system for quickly and appropriately responding to customer complaints and inquiries in a logistics center. This system consists of a server, smart devices, a generative AI model, and an emotion engine.

[1194] The server receives messages and emails sent by customers and stores them in a database. The stored text data is sent to a generative AI for analysis. The generative AI uses natural language processing techniques to analyze the text and perform sentiment analysis. For example, from a message such as "I am very angry about the product defect," it identifies the emotion of anger and quantifies its intensity. This sentiment score is then sent back to the server.

[1195] The server evaluates the returned sentiment score and compares it to a pre-set threshold. If a strong sentiment exceeding the threshold is detected, an alert is immediately sent to the employee's smart device. It also instructs a generative AI to generate an appropriate message. The generated message is then provided to the employee's smart device. For example, a message such as, "We sincerely apologize for the inconvenience caused by the delay in receiving your order. We are currently investigating the situation and will address it as soon as possible," might be generated.

[1196] Furthermore, the emotion engine analyzes the emotional state of the person handling the task and adjusts the tone of the generated text accordingly. The emotion engine determines emotions in real time based on factors such as the person's facial expressions, voice tone, and keyboard typing speed. For example, if the person is feeling stressed, a softer tone of text will be generated.

[1197] The hardware used to implement this system includes high-performance servers, smartphones, or tablets. The software used includes TensorFlow, natural language processing libraries (SpaCy, NLTK), a reactive framework (React Native), Docker, and a RESTful API (Flask / Django).

[1198] As a concrete example, input the following prompt into the generative AI.

[1199] "The delivery of my order is delayed. I'm extremely frustrated."

[1200] "Perform a sentiment analysis on this message and generate appropriate wording."

[1201] The sentiment analysis results indicate anger (score: 9.2). The generated text based on these results is as follows:

[1202] "We sincerely apologize for the inconvenience regarding the delivery of your order. We will address this issue as quickly as possible, so please bear with us for a little while longer."

[1203] This will enable quick and accurate responses to customer complaints and inquiries, leading to improved customer satisfaction and reduced workload for staff.

[1204] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1205] Step 1:

[1206] The server receives messages and emails from customers. The received text data is stored in a database.

[1207] Input: Customer messages or emails

[1208] Output: Text data stored in the database

[1209] Step 2:

[1210] The server retrieves the stored text data for analysis and converts it into a format that the generating AI model can process. Specifically, it cleans the text and performs any necessary preprocessing.

[1211] Input: Text data extracted from a database

[1212] Output: Text data converted into a format that can be processed by the generative AI model.

[1213] Step 3:

[1214] The generative AI model receives the transformed text data and performs sentiment analysis. For example, it identifies the emotion of anger from the phrase "very angry" and quantifies its intensity.

[1215] Input: Text data sent to a generative AI.

[1216] Output: Emotion score (Example: Anger intensity score 9.2)

[1217] Step 4:

[1218] The server evaluates the sentiment score returned from the generated AI model and compares it to a pre-set threshold. If the threshold is exceeded, a notification is sent to the person in charge's terminal.

[1219] Input: Sentiment score

[1220] Output: Notification when threshold is exceeded

[1221] Step 5:

[1222] The server instructs the AI ​​model to generate appropriate text. The text includes expressions that alleviate anger.

[1223] Input: Sentiment score and message content

[1224] Output: Generated text (Example: "We apologize for the inconvenience this has caused. We will address this issue as soon as possible.")

[1225] Step 6:

[1226] The generated draft is sent to the employee's smart device. The employee uses this draft as a reference to quickly reply to the customer.

[1227] Input: Generated text

[1228] Output: Draft sent to the person in charge's terminal

[1229] Step 7:

[1230] The emotion engine analyzes the emotional state of the person in charge and adjusts the tone of the generated text accordingly. If the person in charge is stressed, the text will be set to a softer tone.

[1231] Input: Emotional data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[1232] Output: Adjusted tone of the text

[1233] Step 8:

[1234] The server provides the staff with a pre-formatted response to use in replying to the customer. This allows for a response that soothes the customer's emotions and reduces the burden on the staff.

[1235] Input: Revised text

[1236] Output: Final reply message sent to the customer

[1237] By following these steps, the entire system will be able to respond quickly and appropriately to customer complaints and inquiries, which is expected to improve customer satisfaction and reduce the burden on staff.

[1238] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1239] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1240] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1241] [Fourth Embodiment]

[1242] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1243] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1244] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1245] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1246] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1247] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1248] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1249] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1250] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1251] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1252] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1253] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1254] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1255] This invention relates to a system that inputs customer emails and messages into a generative AI and performs sentiment analysis. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has the function of quickly notifying the relevant personnel of this information and automatically generating and providing appropriate drafts of messages to help them manage their emotions.

[1256] Specifically, the following processes are performed.

[1257] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[1258] Next, the server prepares the stored email content for input into the generative AI. Specifically, it extracts the text data from the email and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[1259] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[1260] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[1261] At the same time, the server instructs the generative AI to generate a message that suppresses strong emotions. The generative AI generates a message that includes appropriate language and solutions, and sends it back to the server. For example, it might say, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1262] The server sends the generated draft to the employee's terminal, who then uses that draft to respond to the customer. This allows for a quick and appropriate response to strong customer emotions.

[1263] By using this system, staff can handle customer inquiries efficiently, which is expected to lead to improved customer satisfaction. A concrete example is the ability to respond quickly to complaint emails. If a customer sends an email saying, "The delivery is delayed and I'm very frustrated!", the system can quickly respond using a generated message such as, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," thereby calming the customer's anger.

[1264] The following describes the processing flow.

[1265] Step 1:

[1266] Users receive emails and messages from customers.

[1267] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[1268] Step 2:

[1269] The server saves received emails to the database.

[1270] Specifically, the server records information such as the email body, sender information, and the date and time of receipt in a database.

[1271] Step 3:

[1272] The server reads emails stored in the database and prepares them for input into the generative AI.

[1273] Specifically, the server extracts the text data from the email and converts it into a format for analysis by generative AI.

[1274] Step 4:

[1275] The server converts the email content and sends it to a generation AI.

[1276] Specifically, the server sends email data to the generative AI via an API.

[1277] Step 5:

[1278] Generative AI analyzes the content of emails and performs sentiment analysis.

[1279] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[1280] Step 6:

[1281] The server receives the emotion score and evaluates the intensity of the emotion.

[1282] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[1283] Step 7:

[1284] The server sends a notification to the person in charge when strong emotions are detected.

[1285] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[1286] Step 8:

[1287] The device displays a notification to the person in charge.

[1288] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[1289] Step 9:

[1290] The server requests the generative AI to generate appropriate text.

[1291] Specifically, the server specifies the requirements for the emotionally restrained text and sends them again to the generative AI via the API.

[1292] Step 10:

[1293] The generative AI generates text based on the specified conditions.

[1294] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1295] Step 11:

[1296] The server provides the generated draft to the person in charge.

[1297] Specifically, the server generates a draft document, which is then sent to the person in charge, allowing them to view it on their terminal.

[1298] Step 12:

[1299] The terminal displays the draft document to the person in charge.

[1300] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[1301] Step 13:

[1302] The person in charge will use the generated draft as a reference and reply to the customer.

[1303] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[1304] This makes it possible to respond quickly and effectively to customers' strong emotions.

[1305] (Example 1)

[1306] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1307] In customer communication, a challenge is responding quickly and appropriately to messages that evoke strong emotions. Conventional systems have been unable to respond quickly to such messages, potentially resulting in decreased customer satisfaction. In contrast, this invention aims to improve the efficiency and quality of customer service by automatically analyzing emotions and providing appropriate responses.

[1308] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1309] In this invention, the server includes means for storing received communications, means for analyzing the stored communications and performing sentiment analysis, means for evaluating the intensity of emotions based on the results of the sentiment analysis, means for sending notifications when the intensity of emotions exceeds a pre-set threshold, means for generating selected example sentences, and means for providing the generated example sentences to the relevant personnel. This makes it possible to quickly detect messages from customers with strong emotions and automatically generate and provide appropriate countermeasures.

[1310] "Received communications" refers to messages and data received by the target system or device.

[1311] "Means of preservation" refers to devices or programs for recording received communications in a database or storage medium.

[1312] "Means of analysis and sentiment analysis" refer to algorithms and software that process stored communications and identify emotions from their content.

[1313] A "means for evaluating the intensity of emotions" is a mechanism that determines, either numerically or qualitatively, how strong an emotion is, based on the results of an emotion analysis.

[1314] "Means of sending notifications" refers to communication methods and protocols used to transmit warnings and information to relevant personnel in real time, based on evaluation results and other factors.

[1315] "Means for generating selected example sentences" refers to artificial intelligence or programs that automatically create appropriate sentences to correspond to specific emotions.

[1316] "Means of providing" refers to means of sending or displaying generated example sentences, etc., for the purpose of showing them to the person in charge.

[1317] This invention is a system for analyzing customer communications and providing appropriate responses. This system stores received communications, performs sentiment analysis, evaluates the intensity of emotions, and generates appropriate example sentences. The following hardware and software are used to implement these functions.

[1318] System Configuration

[1319] This system consists of a user's terminal, a server, and a generative AI model.

[1320] User's device:

[1321] Receive communications from customers using email client software (e.g., Outlook, Gmail).

[1322] server:

[1323] The system uses a mail server (e.g., Postfix, Sendmail) to receive communications and saves them to a database server (e.g., MySQL, PostgreSQL).

[1324] Use a data transformation library (e.g., Pandas) to convert the data into a format suitable for sentiment analysis.

[1325] A notification system (e.g., Slack API, email notification system) is used to notify the responsible person when a sentiment score exceeding a threshold is detected.

[1326] Generative AI models:

[1327] We analyze communication content using natural language processing libraries (e.g., NLTK, SpaCy) and perform sentiment analysis.

[1328] The analysis results are returned to the server as an emotion score.

[1329] Based on the prompt, it generates example sentences that include appropriate wording and solutions.

[1330] Processing Overview

[1331] 1. Check for communications received by the user (e.g., emails such as "I am very upset about the product defect").

[1332] 2. The server stores the emails on the database server.

[1333] 3. The server uses the Python Pandas library to convert the email text data into a format that can be processed by the generative AI.

[1334] 4. The generative AI uses a natural language processing library to detect strong emotions from the context of the email. This detection result is returned to the server as a numerical emotion score.

[1335] 5. The server compares the sentiment score to a threshold, and if it exceeds the threshold, it sends a real-time notification to the person in charge using the notification system.

[1336] 6. The server sends prompt text to the generative AI, instructing it to generate example sentences to suppress emotions.

[1337] Example prompt: For the sentence "Customer email: I am very upset about the product defect," please generate an example sentence that includes an apology and a solution.

[1338] 7. The generative AI generates appropriate example sentences and sends them back to the server.

[1339] 8. The server sends the generated example sentences to the employee's terminal and displays the draft sentences on the notification system.

[1340] 9. The person in charge uses email client software to reply to the customer based on example sentences provided by the generative AI.

[1341] Specific example

[1342] This is an example of how to handle a customer who sends an email saying, "The delivery is delayed and I'm extremely frustrated!" The user who receives this email checks it using their email client software, and the server saves its contents to a database server. The server then converts the email text into a format that can be processed by a generative AI and sends it to the generative AI. The generative AI performs sentiment analysis, detects a strong emotion such as "extremely frustrated," and sends a sentiment score back to the server.

[1343] The server evaluates the sentiment score and sends a notification to the employee's terminal if it exceeds the threshold. At the same time, it sends a prompt to the generative AI, instructing it to generate example sentences. The generative AI generates an example sentence, for example, "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible," and sends it back to the server. The server sends the generated example sentence to the employee's terminal, and the employee uses email client software to reply to the customer based on that example sentence.

[1344] This will enable us to respond quickly and appropriately to customers' strong emotions, which is expected to improve customer satisfaction.

[1345] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1346] Step 1:

[1347] The user receives communications from the customer.

[1348] (Specific actions) The user's email client software (e.g., Outlook, Gmail) checks for emails from customers (e.g., "I am very upset about the product defect").

[1349] (Input) Email from customer

[1350] (Output) Received email content

[1351] Step 2:

[1352] The server saves the received communications to the database.

[1353] (Specific operation) The mail server saves received emails to a database server (e.g., MySQL, PostgreSQL) and records the email content.

[1354] (Input) Received email content

[1355] (Output) Email data stored in the database

[1356] Step 3:

[1357] The server performs a format conversion so that the email content can be input into the generation AI.

[1358] (Specific operation) The server uses the Python Pandas library to convert the saved email text data into a format that can be processed by the generative AI (e.g., JSON format).

[1359] (Input) Email data stored in the database

[1360] (Output) Data converted into a format that can be processed by a generative AI.

[1361] Step 4:

[1362] A generative AI analyzes the formatted email data and performs sentiment analysis.

[1363] (Specific operation) A generative AI uses a natural language processing library (e.g., SpaCy, NLTK) to detect a strong emotion such as "very angry" from the context of the email and quantifies the emotion score (e.g., anger 0.8).

[1364] (Input) Data converted into a format that can be processed by a generative AI.

[1365] (Output) Emotion score (Example: Anger 0.8)

[1366] Step 5:

[1367] The server evaluates the sentiment score and notifies the person in charge if it exceeds the threshold.

[1368] (Specific operation) The server compares the emotion score to a threshold (e.g., anger 0.7), and if the threshold is exceeded, it sends a real-time notification to the person in charge's terminal using a notification system (e.g., Slack API).

[1369] (Input) Emotion score

[1370] (Output) Notification sent to the responsible person's terminal

[1371] Step 6:

[1372] The server instructs the generative AI to generate text that suppresses emotions.

[1373] (Specific operation) The server sends a prompt message to the generation AI: "Customer email: Please generate a draft email that includes an apology and a solution regarding a customer's email stating that they are very upset about a product defect."

[1374] (Input) Sentiment score, prompt text

[1375] (Output) Instructions for text generation sent to the generation system AI.

[1376] Step 7:

[1377] A generative AI generates the text.

[1378] (Specific operation) The generative AI creates an appropriate message (e.g., "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible.") and sends it back to the server.

[1379] (Input) Instructions for generating text draft

[1380] (Output) Generated text

[1381] Step 8:

[1382] The server sends the generated draft to the user's terminal.

[1383] (Specific operation) The server sends the draft document to the person in charge's PC or tablet, and the document is displayed in the notification system (e.g., Microsoft Teams, Slack).

[1384] (Input) Generated text

[1385] (Output) Draft sent to the terminal of the person in charge

[1386] Step 9:

[1387] The person in charge will use the generated draft as a reference to reply to the customer.

[1388] (Specific Action) The person in charge uses email client software to reply, based on a template provided by the generation AI, with the message: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1389] (Input) Generated text

[1390] (Output) Reply email sent to the customer

[1391] (Application Example 1)

[1392] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1393] Traditional systems often suffered from delays in responding to strong customer emotions due to the results of sentiment analysis, making it difficult to respond quickly to such situations. Furthermore, manual handling by staff was required to generate appropriate responses, resulting in inefficiency. Consequently, rapid responses to customer satisfaction and security concerns remained a challenge.

[1394] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1395] In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of the sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating appropriate draft text, means for providing the generated draft text to the person in charge, means for analyzing the text in real time, and means for displaying the generated draft text on the person in charge's terminal. This makes it possible to respond quickly and efficiently to strong customer emotions.

[1396] "Means for saving received text" refers to a function for storing messages and text received from customers in storage such as a database.

[1397] "Means for analyzing saved text and performing sentiment analysis" refers to a function that analyzes saved text data and identifies and evaluates the emotions (joy, anger, sadness, etc.) contained within it.

[1398] A "means for evaluating the intensity of an emotion based on the results of an emotion analysis" refers to a function that uses the results of an emotion analysis to evaluate the intensity of that emotion as a numerical value or score.

[1399] "A means of sending a notification to the person in charge when strong emotions are detected" refers to a function that informs the person in charge in real time when the intensity of emotions exceeds a set threshold.

[1400] "Means for generating appropriate wording" refers to a function that uses generative AI to automatically create optimal reply messages and responses in order to manage customer emotions.

[1401] "Means of providing the generated text to the person in charge" refers to a function for displaying or sending the generated text and corresponding messages to the person in charge's terminal.

[1402] "A means of analyzing text in real time" refers to a processing function that immediately applies sentiment analysis to newly received text and quickly obtains the results.

[1403] "Means for displaying the generated draft on the employee's terminal" refers to a function for displaying the generated reply draft on the screen of the employee's smartphone, computer, or other terminal.

[1404] This invention is a system that performs sentiment analysis based on received customer messages and emails, generates appropriate response drafts, and provides rapid and efficient customer service. This system is implemented using the following hardware and software.

[1405] Hardware and software configuration

[1406] 1. Server:

[1407] Hardware: Standard server equipment (e.g., high-performance CPU, memory, storage)

[1408] Software: Python, MySQL, Firebase

[1409] 2. Terminal:

[1410] Hardware: Smartphones, PCs, tablets, etc.

[1411] Software: Specific notification applications or web browsers

[1412] System Processing Overview

[1413] 1. Data reception and storage:

[1414] When a user receives a message from a customer, the message is sent to the server and stored in a database (MySQL). This ensures the secure storage of email and chat data.

[1415] 2. Sentiment analysis:

[1416] The server retrieves the stored messages and inputs them into a generative AI model for sentiment analysis (e.g., OpenAI GPT-4). The generative AI model performs sentiment analysis using the following prompts:

[1417] "Perform an emotional analysis of the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[1418] The analyzed results are sent back to the server as an emotion score.

[1419] 3. Sentiment evaluation and notification:

[1420] The server receives the sentiment score and evaluates its strength. If it exceeds a threshold, it sends a notification to the agent's device using Firebase Cloud Messaging, allowing the agent to immediately understand the situation.

[1421] 4. Draft generation:

[1422] The server instructs the AI ​​model to generate an appropriate response. The following prompts are used:

[1423] "Please create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm extremely frustrated!'"

[1424] The generated draft is sent to the person in charge's terminal.

[1425] Specific example

[1426] For example, if a customer sends a message saying, "The delivery is delayed and it's very frustrating!", the server will execute the following process.

[1427] Receive messages and save them to the database.

[1428] OpenAI GPT-4 is used for sentiment analysis, and a strong emotion of "very irritated" is detected.

[1429] Since the emotion score exceeded the threshold, a notification will be sent to the person in charge.

[1430] Using OpenAI GPT-4 again, we generate a suitable reply message like the following:

[1431] "We sincerely apologize for the inconvenience caused by the delivery delay. We are currently investigating the situation and will inform you of a solution as soon as possible."

[1432] The generated draft is displayed on the employee's terminal, and the employee uses it to quickly reply to the customer.

[1433] In this way, this system can respond quickly and effectively to customers' strong emotions, and can contribute to improving customer satisfaction.

[1434] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1435] Step 1:

[1436] A user receives a message from a customer. The text data received by the user via email or messaging app is sent to the server. The server stores this received data in a database. The input is the customer's message text, and the output is the message data stored in the database. Specifically, the server inserts the received text into a MySQL database using SQL commands.

[1437] Step 2:

[1438] The server analyzes stored messages and performs sentiment analysis. It retrieves the stored message data and inputs it into a generative AI model (e.g., OpenAI GPT-4). The input is the stored message text, which is a request for sentiment analysis using a prompt. The output is a sentiment score. Specifically, the server sends the following prompt to the generative AI model: "Perform sentiment analysis on the following sentence and identify strong emotions such as anger or frustration: 'The delivery is delayed and I'm extremely frustrated!'"

[1439] Step 3:

[1440] The server receives sentiment scores from the generated AI model and evaluates their strength. The input is the sentiment score, and the output is the result of determining whether it exceeds a threshold. Specifically, the server compares the sentiment score to a threshold, and if it exceeds the threshold, it proceeds to the next step.

[1441] Step 4:

[1442] If a sentiment score exceeding a threshold is detected, the server uses Firebase Cloud Messaging to send a notification to the employee's device. The input is the sentiment score exceeding the threshold and the original message text, and the output is the notification to the employee's device. Specifically, the server sends a notification message via the Firebase API indicating that strong sentiment has been detected.

[1443] Step 5:

[1444] The server instructs the generative AI model to generate an appropriate response. The input is the customer message and the prompt "Create the best response message for a customer who is angry about the following situation: 'The delivery is delayed and I'm very frustrated!'" and the output is the generated response. Specifically, the server sends another request to the generative AI model to retrieve an appropriate response.

[1445] Step 6:

[1446] The server provides the generated text to the user's terminal. The input is the text generated by the AI ​​model, and the output is the text displayed on the user's terminal screen. Specifically, the server sends the generated text to the user's terminal and displays it through a web application or notification application.

[1447] Step 7:

[1448] The person in charge replies to the customer based on the provided draft. The input is the generated draft sent from the server, and the output is the reply message sent to the customer. Specifically, the person in charge reviews the generated draft, makes adjustments as needed, and sends the final reply message to the customer.

[1449] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1450] This invention provides a system that enables more appropriate customer service by inputting customer emails and messages into a generative AI, performing sentiment analysis, and combining it with an emotion engine that recognizes user emotions. If strong emotions such as anger or dissatisfaction are detected in the received text, the system has a function to quickly notify the relevant personnel of this information and automatically generate and provide appropriate wording to the personnel to help them manage their emotions.

[1451] Specifically, the following processes are performed.

[1452] First, the user receives emails or messages from customers. For example, suppose a customer sends an email saying, "I'm very upset about a defect in the product." Next, the server receives this email and saves its contents to the database.

[1453] Subsequently, the server reads the emails stored in the database and prepares them for input into the generative AI. Specifically, it extracts the text data from the emails and converts it into a format that the generative AI can process. This converted data is then sent to the generative AI.

[1454] The generative AI analyzes the received data and performs sentiment analysis. For example, this analysis might detect strong anger from phrases like "very angry." The results of the sentiment analysis are quantified as a sentiment score and sent back to the server.

[1455] The server receives the analysis results from the generative AI and evaluates the intensity of the emotion. If the emotion score is higher than a pre-set threshold, it notifies the relevant person. A notification is displayed on the terminal, allowing the person to immediately understand the situation.

[1456] At the same time, the server instructs the generative AI to generate a text that suppresses strong emotions. The generative AI generates a text that incorporates appropriate language and a solution, and sends it back to the server. For example, it might generate a text that says, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1457] Furthermore, this invention incorporates an emotion engine to analyze the user's emotions in real time and adjust the tone of the generated text accordingly. The emotion engine analyzes the user's facial expressions, voice tone, and keyboard input speed to determine their current emotional state. This allows for adjustments such as generating text with a softer tone if, for example, the person in charge is feeling stressed.

[1458] The server sends the generated draft to the employee's terminal, and the employee uses that draft as a reference to reply to the customer. This allows for a quick and appropriate response to strong customer emotions, and provides a draft that is tailored to the user's emotional state.

[1459] This allows staff to handle customer inquiries more efficiently, which is expected to lead to improved customer satisfaction. For example, by using the generated template to quickly respond to customer complaint emails, customer anger can be mitigated, and staff can handle the situation with less mental strain.

[1460] The following describes the processing flow.

[1461] Step 1:

[1462] Users receive emails and messages from customers.

[1463] Specifically, users receive messages from customers through email software or messaging platforms. For example, they might receive an email stating, "I am very upset about a defect in the product."

[1464] Step 2:

[1465] The server saves received emails to the database.

[1466] Specifically, the server records information such as the body of the received email, sender information, and the date and time of receipt in a database.

[1467] Step 3:

[1468] The server reads emails stored in the database and prepares them for input into the generative AI.

[1469] Specifically, the text data from the email is extracted and converted into a format that can be processed by generative AI.

[1470] Step 4:

[1471] The server converts the email content and sends it to a generation AI.

[1472] Specifically, the server sends email data to the generative AI via an API.

[1473] Step 5:

[1474] Generative AI analyzes the content of emails and performs sentiment analysis.

[1475] Specifically, a generative AI uses natural language processing techniques to determine the emotional state of an email and generate an emotional score. For example, it might detect strong anger from the phrase "very angry."

[1476] Step 6:

[1477] The server receives the emotion score and evaluates the intensity of the emotion.

[1478] Specifically, the server evaluates the emotion score received from the generative AI and compares it to a pre-set threshold. If the emotion score exceeds the threshold, the process proceeds to the next step.

[1479] Step 7:

[1480] The server sends a notification to the person in charge when strong emotions are detected.

[1481] Specifically, the server generates a notification message and sends it to the responsible person via email or the internal messaging system.

[1482] Step 8:

[1483] The device displays a notification to the person in charge.

[1484] Specifically, notification alerts and messages displayed on the device will allow the person in charge to immediately understand the situation.

[1485] Step 9:

[1486] The server requests the generative AI to generate appropriate text.

[1487] Specifically, the server specifies the requirements for the emotionally restrained text and sends them again to the generative AI via the API.

[1488] Step 10:

[1489] The generative AI generates text based on the specified conditions.

[1490] Specifically, the generative AI generates sentences that include appropriate wording and solutions. For example, it might generate a sentence like, "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible."

[1491] Step 11:

[1492] The server prepares to provide the draft text generated by the generative AI to the person in charge.

[1493] Specifically, this involves preparing the generated draft document for transmission to the responsible person's terminal.

[1494] Step 12:

[1495] The terminal displays the draft document to the person in charge.

[1496] Specifically, the terminal will display the generated draft text to the person in charge so that they can refer to it.

[1497] Step 13:

[1498] The person in charge will use the generated draft as a reference and reply to the customer.

[1499] Specifically, the person in charge will revise the draft as needed and send the email or message to the customer. The reply will include appropriate language and solutions.

[1500] Furthermore, the processing when combining it with the emotion engine is as follows:

[1501] Step 14:

[1502] The emotion engine recognizes the user's emotions in real time.

[1503] Specifically, it analyzes the user's facial expressions, voice tone, or keyboard input speed to determine their current emotional state.

[1504] Step 15:

[1505] The server instructs the generative AI to adjust the tone of the text based on the user's emotional state.

[1506] Specifically, if the system determines that the user is experiencing stress, it will be instructed to generate a softer-toned version of the message.

[1507] This makes it possible to respond quickly and appropriately to customers' strong emotions, which not only improves customer satisfaction but also reduces the mental burden on staff.

[1508] (Example 2)

[1509] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1510] Traditional systems struggled to respond appropriately and promptly to emails and messages from customers that contained strong emotions. Furthermore, it was difficult for staff to manage their own emotional state while interacting with customers, resulting in decreased customer satisfaction and increased stress for staff. Additionally, the quality and appropriateness of the generated messages were not guaranteed, placing a heavy burden on staff.

[1511] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving emails and messages from customers, means for storing the received text in a database, means for analyzing the stored text and performing sentiment analysis using a generative AI model, means for sending a notification to the person in charge when the sentiment score exceeds a threshold, means for generating appropriate text based on the generative AI model, means for analyzing the user's current emotional state in real time, and means for providing the generated text to the person in charge. This makes it possible to respond quickly and appropriately to strong customer emotions, and provides text that matches the emotional state of the person in charge, which can be expected to improve customer satisfaction and reduce stress for the person in charge.

[1512] "Customer" refers to individuals or companies that purchase or use goods or services.

[1513] "Emails and messages" refer to electronic means of communication that customers send to companies or their representatives.

[1514] A "database" refers to a system for efficiently storing, searching, and managing information.

[1515] A "generative AI model" refers to an algorithm that uses artificial intelligence to perform specific tasks, such as language generation or sentiment analysis.

[1516] "Sentiment analysis" refers to the process of identifying emotions from text data and quantifying them.

[1517] An "emotion score" is a numerical value obtained from the results of an emotion analysis that indicates the intensity and type of emotion.

[1518] A "threshold" is a value that indicates a specific standard; exceeding this value triggers a particular action.

[1519] "Notifications" refer to messages or alerts used to convey important information or warnings to users.

[1520] "Document draft" refers to a preliminary version or draft of a document used to communicate with a customer.

[1521] "User" refers to a person who operates this system, or a person who uses the system to handle customer inquiries.

[1522] "Real-time" refers to processing or responding to an event as close to the moment it occurs as possible.

[1523] This invention relates to a system that uses artificial intelligence technology to analyze emails and messages from customers and evaluate their emotions in order to provide more appropriate customer service. In this system, the server, terminal, and user each play important roles.

[1524] First, users receive emails and messages from customers. This reception occurs through the company's email server or messaging platform.

[1525] Next, the server saves the received emails to a database. Specifically, it uses a relational database management system (RDBMS) such as MySQL.

[1526] The server converts the stored emails into a format that can be input into a generative AI model (e.g., the BERT model in Hugging Face) for analysis. This conversion involves changing the text data to a format such as JSON.

[1527] The server sends the transformed data to the generating AI model. This data is sent via an HTTP request to the generating AI model's API endpoint.

[1528] The generative AI model analyzes the received data and performs sentiment analysis. Based on the analysis, a sentiment score is calculated. For example, a strong feeling of anger is detected from the text "very angry."

[1529] The server evaluates the sentiment score returned from the generated AI model, and if the score exceeds a pre-set threshold, it sends a notification to the device. This notification is delivered via a pop-up or alert email.

[1530] Based on the sentiment score, the server instructs the generative AI model to generate appropriate text. This text is designed to respond appropriately to the customer and may include phrases such as "We are sorry."

[1531] Furthermore, the emotion engine analyzes the user's current emotional state in real time. Specifically, it uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed.

[1532] The generated text is sent from the server to the terminal, and the person in charge uses it as a reference to reply to the customer. This allows for a quick and appropriate response to customers' strong emotions and also reduces stress for the person in charge.

[1533] As a concrete example, consider a scenario where a customer sends an email stating, "I am extremely upset about the product defect." Upon receiving this email, the server saves its contents to a database and performs sentiment analysis using a generative AI model. Because the sentiment score is very high at 75, the server sends an urgent notification to the relevant person in charge and simultaneously instructs the generative AI model to generate an appropriate message. The generated message is: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will provide you with a solution as soon as possible." Furthermore, if the sentiment engine determines that the person in charge is feeling stressed, a softer tone of message will be generated.

[1534] As described above, by combining a generative AI model and an emotion engine, this system can respond quickly and appropriately to emails from customers that contain strong emotions, and also contribute to reducing stress for the staff in charge.

[1535] Example of a prompt:

[1536] Content of email received from customer: "I am very upset about the defect in the product."

[1537] Generative AI model to use: A general-purpose generative AI model (e.g., BERT model)

[1538] Instructions: "Perform sentiment analysis on the customer's email text and return the sentiment score."

[1539] Analysis result: Emotion score = 75 (Strong anger)

[1540] Notification: "We have received an email expressing strong emotions. Customer support is required."

[1541] Generated text: "We sincerely apologize for the inconvenience caused by the product defect. We are currently investigating the situation and will inform you of a solution as soon as possible."

[1542] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1543] Step 1:

[1544] Users receive emails and messages from customers.

[1545] Specific operation: Users receive emails and messages through the company's mail server or messaging platform. For example, a customer sends an email stating, "I am very upset about a defect in the product."

[1546] Input: Emails and messages from customers

[1547] Output: Data of received emails and messages

[1548] Step 2:

[1549] The server saves the email to the database.

[1550] Specific operation: The server saves the content of received emails to a relational database (e.g., MySQL). At the same time, it also saves the email's sending date and sender information.

[1551] Input: Data of received emails and messages

[1552] Output: Email data stored in the database

[1553] Step 3:

[1554] The server converts the email into a format that can be input into the AI ​​model that generates emails.

[1555] Specific operation: The server converts the email text data into JSON format. It also formats it so that it can be sent to the API endpoint of the generated AI model.

[1556] Input: Email data stored in the database

[1557] Output: JSON format data for input into the generated AI model.

[1558] Step 4:

[1559] The server sends the converted data to the AI ​​model.

[1560] Specific operation: The server uses an HTTP request to send data to the API endpoint of the generated AI model. The request includes a generated prompt (e.g., "Perform sentiment analysis on customer email text and return the sentiment score.").

[1561] Input: JSON formatted data and prompt text

[1562] Output: Data sent to the generating AI model

[1563] Step 5:

[1564] The generative AI model performs sentiment analysis on the received data.

[1565] Specific operation: The generative AI model (e.g., a general generative AI model) analyzes the received text data and performs a process of quantifying emotions. It calculates an anger score from "very annoyed" as an emotion score and returns it as a numerical value (e.g., 75).

[1566] Input: Converted JSON data, prompt text

[1567] Output: Calculation result of the emotion score

[1568] Step 6:

[1569] The server evaluates the sentiment analysis results and sends notifications to the relevant personnel.

[1570] Specific operation: The server evaluates the sentiment score received from the generated AI model, and if the score exceeds a threshold (e.g., 50), it sends a pop-up notification or alert email to the person in charge.

[1571] Input: Sentiment score from a generated AI model

[1572] Output: Notification to the person in charge

[1573] Step 7:

[1574] The server instructs the AI ​​model to generate appropriate text.

[1575] Specific operation: The server sends a prompt message to the AI ​​model again, instructing it to generate a text, and then generates an appropriate response. For example, a sentence such as "We sincerely apologize for the inconvenience caused by the product defect" is generated.

[1576] Input: Prompt based on sentiment score

[1577] Output: Generated text

[1578] Step 8:

[1579] The emotion engine analyzes the user's current emotional state in real time.

[1580] Specific operation: The emotion engine uses a webcam and microphone to analyze the user's facial expressions and voice, and also evaluates keyboard input speed. Based on this, it assesses the stress level of the person in charge and adjusts the tone of the drafted text accordingly.

[1581] Input: Real-time data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[1582] Output: Evaluation results of the person in charge's emotional state

[1583] Step 9:

[1584] The server sends the optimal draft to the person in charge's terminal.

[1585] Specific operation: The server combines the generated text with adjustments based on the user's emotional state and sends the optimal text to the terminal. The person in charge uses this text as a reference to reply to the customer.

[1586] Input: Generated text, user's emotional state

[1587] Output: Optimal text sent to the terminal

[1588] This enables quick and appropriate responses to strong customer emotions, and by providing wording tailored to the emotional state of the person handling the request, it leads to improved customer satisfaction and reduced stress for the person handling the request.

[1589] (Application Example 2)

[1590] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1591] Logistics centers are required to respond quickly and appropriately to customer complaints and inquiries. However, manually analyzing each message and providing an appropriate response requires a tremendous amount of effort and time, potentially leading to decreased customer satisfaction. Furthermore, the quality of responses is inconsistent because they are influenced by the emotional state of the person handling them.

[1592] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for storing received text, means for analyzing the stored text and performing sentiment analysis, means for evaluating the intensity of emotion based on the results of sentiment analysis, means for sending a notification to the person in charge when strong emotion is detected, means for generating an appropriate draft, means for providing the generated draft to the person in charge, means for receiving and processing customer messages on a smart device, and means for analyzing the emotional state of the person in charge and adjusting the tone of the draft. This makes it possible to respond quickly and appropriately to customer complaints and inquiries, and is expected to improve customer satisfaction and reduce the burden on the person in charge.

[1593] "Means for saving received text" refers to a device or program for saving messages and emails sent by customers to a database.

[1594] "Methods for analyzing saved text and performing sentiment analysis" refers to programs that use natural language processing techniques to analyze saved text data and identify emotions and their intensity.

[1595] "Means for evaluating the intensity of emotions based on the results of emotion analysis" refers to a device or program that quantifies the data obtained from emotion analysis and evaluates its intensity.

[1596] "Means for sending notifications to responsible personnel when strong emotions are detected" refers to a device or program that sends alerts or messages to the relevant personnel when strong emotions are detected as a result of the analysis.

[1597] "Means for generating appropriate text" refers to a device or program that uses natural language generation technology to generate appropriate response text according to the customer's emotions.

[1598] "Means of providing the generated draft to the person in charge" refers to a device or program for sending the generated reply draft to the person in charge's device.

[1599] "Means for receiving and processing customer messages on a smart device" refers to a device or program for receiving, analyzing, and processing messages from customers on a mobile device such as a smartphone or tablet.

[1600] "Means for analyzing the emotional state of the person in charge and adjusting the tone of the draft" refers to a device or program for analyzing the emotions of the person in charge and adjusting the tone of the draft generated based on the results.

[1601] This invention relates to a system for quickly and appropriately responding to customer complaints and inquiries in a logistics center. This system consists of a server, smart devices, a generative AI model, and an emotion engine.

[1602] The server receives messages and emails sent by customers and stores them in a database. The stored text data is sent to a generative AI for analysis. The generative AI uses natural language processing techniques to analyze the text and perform sentiment analysis. For example, from a message such as "I am very angry about the product defect," it identifies the emotion of anger and quantifies its intensity. This sentiment score is then sent back to the server.

[1603] The server evaluates the returned sentiment score and compares it to a pre-set threshold. If a strong sentiment exceeding the threshold is detected, an alert is immediately sent to the employee's smart device. It also instructs a generative AI to generate an appropriate message. The generated message is then provided to the employee's smart device. For example, a message such as, "We sincerely apologize for the inconvenience caused by the delay in receiving your order. We are currently investigating the situation and will address it as soon as possible," might be generated.

[1604] Furthermore, the emotion engine analyzes the emotional state of the person handling the task and adjusts the tone of the generated text accordingly. The emotion engine determines emotions in real time based on factors such as the person's facial expressions, voice tone, and keyboard typing speed. For example, if the person is feeling stressed, a softer tone of text will be generated.

[1605] The hardware used to implement this system includes high-performance servers, smartphones, or tablets. The software used includes TensorFlow, natural language processing libraries (SpaCy, NLTK), a reactive framework (React Native), Docker, and a RESTful API (Flask / Django).

[1606] As a concrete example, input the following prompt into the generative AI.

[1607] "The delivery of my order is delayed. I'm extremely frustrated."

[1608] "Perform a sentiment analysis on this message and generate appropriate wording."

[1609] The sentiment analysis results indicate anger (score: 9.2). The generated text based on these results is as follows:

[1610] "We sincerely apologize for the inconvenience regarding the delivery of your order. We will address this issue as quickly as possible, so please bear with us for a little while longer."

[1611] This will enable quick and accurate responses to customer complaints and inquiries, leading to improved customer satisfaction and reduced workload for staff.

[1612] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1613] Step 1:

[1614] The server receives messages and emails from customers. The received text data is stored in a database.

[1615] Input: Customer messages or emails

[1616] Output: Text data stored in the database

[1617] Step 2:

[1618] The server retrieves the stored text data for analysis and converts it into a format that the generating AI model can process. Specifically, it cleans the text and performs any necessary preprocessing.

[1619] Input: Text data extracted from a database

[1620] Output: Text data converted into a format that can be processed by the generative AI model.

[1621] Step 3:

[1622] The generative AI model receives the transformed text data and performs sentiment analysis. For example, it identifies the emotion of anger from the phrase "very angry" and quantifies its intensity.

[1623] Input: Text data sent to a generative AI.

[1624] Output: Emotion score (Example: Anger intensity score 9.2)

[1625] Step 4:

[1626] The server evaluates the sentiment score returned from the generated AI model and compares it to a pre-set threshold. If the threshold is exceeded, a notification is sent to the person in charge's terminal.

[1627] Input: Sentiment score

[1628] Output: Notification when threshold is exceeded

[1629] Step 5:

[1630] The server instructs the AI ​​model to generate appropriate text. The text includes expressions that alleviate anger.

[1631] Input: Sentiment score and message content

[1632] Output: Generated text (Example: "We apologize for the inconvenience this has caused. We will address this issue as soon as possible.")

[1633] Step 6:

[1634] The generated draft is sent to the employee's smart device. The employee uses this draft as a reference to quickly reply to the customer.

[1635] Input: Generated text

[1636] Output: Draft sent to the person in charge's terminal

[1637] Step 7:

[1638] The emotion engine analyzes the emotional state of the person in charge and adjusts the tone of the generated text accordingly. If the person in charge is stressed, the text will be set to a softer tone.

[1639] Input: Emotional data of the person in charge (facial expressions, voice tone, keyboard input speed, etc.)

[1640] Output: Adjusted tone of the text

[1641] Step 8:

[1642] The server provides the staff with a pre-formatted response to use in replying to the customer. This allows for a response that soothes the customer's emotions and reduces the burden on the staff.

[1643] Input: Revised text

[1644] Output: Final reply message sent to the customer

[1645] By following these steps, the entire system will be able to respond quickly and appropriately to customer complaints and inquiries, which is expected to improve customer satisfaction and reduce the burden on staff.

[1646] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1647] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1648] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1649] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1650] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1651] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1652] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1653] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1654] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1655] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1656] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1657] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1658] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1659] 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.

[1660] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1661] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1662] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1663] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1664] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1665] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1666] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[1667] The following is further disclosed regarding the embodiments described above.

[1668] (Claim 1)

[1669] A means of saving received text,

[1670] A means of analyzing saved texts and performing sentiment analysis,

[1671] A means of evaluating the intensity of emotions based on the results of emotion analysis,

[1672] A means of sending a notification to the person in charge when strong emotions are detected,

[1673] A means of generating appropriate text,

[1674] A means of providing the generated draft to the person in charge,

[1675] A system that includes this.

[1676] (Claim 2)

[1677] The system according to claim 1, further comprising means for sending a notification to the person in charge when the result of the sentiment analysis exceeds a threshold.

[1678] (Claim 3)

[1679] The system according to claim 1, further comprising means for responding to a customer based on a generated draft.

[1680] "Example 1"

[1681] (Claim 1)

[1682] A means of saving received communications,

[1683] A means of analyzing stored communications and performing sentiment analysis,

[1684] A means of evaluating the intensity of emotions based on the results of emotion analysis,

[1685] A means of sending a notification when the intensity of emotion exceeds a pre-set threshold,

[1686] A means for generating selected example sentences,

[1687] A means of providing the generated example sentences to the relevant personnel,

[1688] A system that includes this.

[1689] (Claim 2)

[1690] The system according to claim 1, further comprising means for sending a notification to the employee's terminal when the intensity of the emotion exceeds a threshold.

[1691] (Claim 3)

[1692] The system according to claim 1, further comprising means for responding to the user based on the generated example sentences.

[1693] "Application Example 1"

[1694] (Claim 1)

[1695] A means of saving received text,

[1696] A means of analyzing saved texts and performing sentiment analysis,

[1697] A means of evaluating the intensity of emotions based on the results of emotion analysis,

[1698] A means of sending a notification to the person in charge when strong emotions are detected,

[1699] A means of generating appropriate text,

[1700] A means of providing the generated draft to the person in charge,

[1701] A means of analyzing text in real time,

[1702] A means of displaying the generated draft on the person in charge's terminal,

[1703] A system that includes this.

[1704] (Claim 2)

[1705] The system according to claim 1, further comprising means for sending a notification to the person in charge when the result of the sentiment analysis exceeds a threshold.

[1706] (Claim 3)

[1707] The system according to claim 1, further comprising means for responding to a customer based on a generated draft.

[1708] "Example 2 of combining an emotion engine"

[1709] (Claim 1)

[1710] Means of receiving emails and messages from customers,

[1711] A means of saving received text to a database,

[1712] A method for analyzing saved text and performing sentiment analysis using a generative AI model,

[1713] A means of evaluating the intensity of emotions based on the results of emotion analysis,

[1714] A means of sending a notification to the person in charge when the emotional score exceeds a threshold,

[1715] A means for generating appropriate text based on a generative AI model,

[1716] A means of analyzing the user's current emotional state in real time,

[1717] A means of providing the generated draft to the person in charge,

[1718] A system that includes this.

[1719] (Claim 2)

[1720] The system according to claim 1, further comprising means for sending a notification to a person in charge when the result of sentiment analysis exceeds a threshold.

[1721] (Claim 3)

[1722] The system according to claim 1, further comprising means for responding to a customer based on a generated draft.

[1723] "Application example 2 when combining with an emotional engine"

[1724] (Claim 1)

[1725] A means of saving received text,

[1726] A means of analyzing saved texts and performing sentiment analysis,

[1727] A means of evaluating the intensity of emotions based on the results of emotion analysis,

[1728] A means of sending a notification to the person in charge when strong emotions are detected,

[1729] A means of generating appropriate text,

[1730] A means of providing the generated draft to the person in charge,

[1731] A means for receiving and processing customer messages on a smart device,

[1732] A method for analyzing the emotional state of the person in charge and adjusting the tone of the draft,

[1733] A system that includes this.

[1734] (Claim 2)

[1735] The system according to claim 1, further comprising means for sending a notification to the person in charge when the result of the sentiment analysis exceeds a threshold.

[1736] (Claim 3)

[1737] The system according to claim 1, further comprising means for responding to a customer based on a generated draft. [Explanation of Symbols]

[1738] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of saving received text, A means of analyzing saved texts and performing sentiment analysis, A means of evaluating the intensity of emotions based on the results of emotion analysis, A means of sending a notification to the person in charge when strong emotions are detected, A means of generating appropriate text, A means of providing the generated draft to the person in charge, A system that includes this.

2. The system according to claim 1, further comprising means for sending a notification to the person in charge when the result of the sentiment analysis exceeds a threshold.

3. The system according to claim 1, further comprising means for responding to a customer based on a generated draft.

Images