System

Abstract

A system is provided.SOLUTION: A system comprising: means for storing minutes of a business negotiation in a database; means for analyzing the minutes of the business negotiation as text data by performing natural language processing using generative artificial intelligence; means for classifying a successful case and a failed case based on past business negotiation data; means for extracting a key phrase included in the successful case; means for predicting an agreement probability of a future business negotiation based on the extracted key phrase; and means for notifying a user of the predicted agreement probability.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] Traditional sales activities lack a method for accurately predicting the likelihood of closing a deal based on customer comments. This makes it difficult for sales representatives to develop appropriate strategies in response to ambiguous responses from customers. Furthermore, most sales management tools focus on quantitative data management and are unable to fully utilize qualitative information. This leads to issues such as reduced efficiency in sales activities and increased wasted sales activities. [Means for solving the problem]

[0005] The system includes a means for storing business meeting minutes in a database, a means for analyzing the business meeting minutes as text data using natural language processing with artificial intelligence, a means for classifying successful and unsuccessful cases based on past business meeting data, a means for extracting key phrases contained in successful cases, a means for predicting the likelihood of agreement in future business meetings based on the extracted key phrases, and a means for notifying the user of the predicted likelihood of agreement. This system scientifically analyzes business meeting minutes in sales activities, enabling the development of efficient and accurate strategies. Furthermore, by providing sales representatives with systematic know-how, it is possible to reduce waste in sales activities and improve the success rate of business meetings.

[0006] "Business negotiation minutes" are text and audio data that record the details of business negotiations with customers.

[0007] A "database" is a system for systematically storing and managing information.

[0008] "Generative AI" is a technology that uses natural language processing and machine learning to analyze and generate data.

[0009] "Natural language processing" is a technology that uses computers to understand and process human language.

[0010] "Text data" is character string data, and is a data format that includes sentences and phrases.

[0011] A "successful deal" is one in which negotiations have led to an agreement and an actual order or contract.

[0012] A "failed project" is a project in which negotiations did not result in an agreement and did not result in an order or contract.

[0013] A "key phrase" is a word or phrase that is considered to be particularly important in text data.

[0014] "Agreement probability" is a score or probability that indicates the likelihood of a deal being successful.

[0015] "User" refers to an individual or organization that uses the system, and in particular to sales personnel. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14]FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0017] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0018] First, the terms used in the following description will be explained.

[0019] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0020] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0021] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0022] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0023] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0027] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0028] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0029] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0030] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0031] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0034] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0035] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0037] This invention is a system that uses AI to analyze sales meeting minutes in sales activities and predict the likelihood of a deal being concluded. Specifically, it automates the process of collecting and analyzing sales meeting minutes and predicting the likelihood of an agreement.

[0038] Collection of business meeting minutes

[0039] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[0040] Data transmission and storage

[0041] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[0042] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[0043] Text analysis with generative artificial intelligence

[0044] Server: Natural language processing is performed on the business meeting minutes stored in the database using generation AI. Specific analysis methods include tokenization, part-of-speech tagging, and noun phrase extraction. For example, noun phrases such as "interest" and "proposal" are extracted.

[0045] Classification of successful and unsuccessful cases

[0046] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0047] Keyphrase Extraction

[0048] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[0049] Agreement probability prediction

[0050] User: The sales representative enters new minutes of the business meeting into the terminal.

[0051] Server: Checks whether newly entered minutes of business meetings contain pre-specified key phrases. If they do, predicts the likelihood of agreement based on the importance of those key phrases. For example, if the comment "high cost performance" is included, predicts the likelihood of agreement to be 80%.

[0052] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[0053] Specific examples

[0054] For example, consider the case where a sales representative holds an initial business meeting with a new client and writes in the minutes, "I am very interested in your proposal for a new solution." These minutes are sent from the device to the server, where they are analyzed. The generation AI extracts the key phrase "interested" and compares it with a database of past success stories. As a result, it confirms that this key phrase appears frequently in successful cases, and predicts an 80% probability of agreement.

[0055] The results are then sent to the terminal, allowing the sales representative to take appropriate action, such as preparing a concrete proposal as the next step.In this way, the system provides data-driven support for sales activities and an effective means of increasing the success rate.

[0056] The processing flow will be explained below.

[0057] Step 1: Data collection

[0058] Terminal: The sales representative inputs minutes of the sales meeting. Either text input or voice recognition input can be selected as the input method. For example, if a customer comments, "This solution is very interesting," the content of that comment is recorded.

[0059] Step 2: Send data

[0060] Terminal: The entered minutes of the business meeting are sent to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[0061] Step 3: Save data

[0062] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[0063] Step 4: Text Analysis

[0064] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0065] Step 5: Classification of successes and failures

[0066] Server: Based on past sales negotiation data, the generative AI classifies successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the word "interest" appears frequently in successful deals.

[0067] Step 6: Extracting Keyphrases

[0068] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores are used for extraction. For example, key phrases such as "cost performance" and "interesting" are extracted.

[0069] Step 7: Predicting agreement probability

[0070] User: A sales representative opens a new business deal and enters the details of the deal into the terminal as minutes.

[0071] Server: Checks key phrases against the input minutes, and the generation AI predicts the degree of agreement. For example, if the phrase "high cost performance" appears, it predicts an 80% degree of agreement.

[0072] Step 8: Notification of prediction results

[0073] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[0074] By using the above processing steps, the present invention provides a system for scientifically predicting the probability of success of a business negotiation and improving the efficiency of sales activities.

[0075] Example 1

[0076] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0077] In conventional sales activities, the collection, analysis, and closing probability prediction of sales meeting minutes are often done manually, which takes time and effort and reduces efficiency. Another issue is that data management is cumbersome, making it difficult to effectively utilize trends in successful deals. The present invention aims to solve these issues by automatically analyzing minutes and improving the accuracy of closing probability prediction.

[0078] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0079] In this invention, the server provides an input device for recording the details of business negotiations, and includes: means for a user to input minutes by text input or voice recognition; means for format-checking the input minutes and sending them; means for saving the minutes in a database; means for analyzing the minutes as text data using natural language processing with artificial intelligence; means for classifying successful and unsuccessful cases based on past business negotiation data; means for extracting important expressions contained in successful cases; means for predicting the likelihood of agreement in future business negotiations based on the extracted important expressions; and means for notifying the user of the predicted likelihood of agreement. This enables efficient collection and analysis of business negotiation minutes and highly accurate prediction of the likelihood of closing a deal.

[0080] "Business negotiation content" refers to information such as proposals, requests, and exchanges of opinions exchanged between customers and sales representatives during sales activities.

[0081] An "input device" is a device or software that a user uses to input information. Examples include a keyboard, microphone, and touch panel.

[0082] "Text input" refers to a method in which a user inputs characters or sentences using a keyboard or the like.

[0083] "Speech recognition" is a technology that inputs the words spoken by a user as voice and converts them into text information.

[0084] Minutes are documents that record the content exchanged, decisions made, and key points made during business negotiations.

[0085] "Format check" refers to the process of verifying whether the input data is in the correct format.

[0086] A "database" is a system for systematically storing large amounts of data and quickly searching and retrieving them.

[0087] "Generative AI" refers to an AI technology that analyzes data and performs pattern recognition and predictions.

[0088] "Natural language processing" refers to techniques that enable artificial intelligence to understand and analyze human language, including sentence tokenization, part-of-speech tagging, and semantic analysis.

[0089] "Past business negotiation data" refers to records and information related to past business negotiations, and is stored in a database.

[0090] A "successful case" is one in which negotiations proceed smoothly and an agreement or contract is reached with the customer.

[0091] A "failed project" is one in which negotiations did not progress and no agreement or contract was reached with the customer.

[0092] "Important expressions" are key phrases or noun phrases that are often found in successful deals and have a significant impact on the outcome of subsequent sales negotiations.

[0093] "Agreement probability" is an indicator that indicates the probability that a particular business deal will be successful.

[0094] "Notification" refers to informing a user of information or results, and is often done in real time via a terminal.

[0095] This invention is a system for analyzing sales negotiation minutes in sales activities using a generative AI model and predicting the probability of closing a deal. Specific embodiments of this system are described in detail below.

[0096] Entering business meeting minutes

[0097] Terminal: After a sales meeting, the sales representative uses the terminal to enter details of the sales meeting. Two methods of input are available: text input and voice recognition. For example, a sales representative can enter something like "I'm very interested in this proposal" as text or use voice recognition.

[0098] Data transmission and storage

[0099] Terminal: The entered minutes are sent to the server in real time, where a format check is performed to ensure the data is in the correct format.

[0100] Server: The received minutes of business meetings are stored in a database. Each minutes of business meetings is assigned a unique identifier and managed centrally in the database.

[0101] Text analytics

[0102] Server: Natural language processing is performed on the business meeting minutes stored in the database using a generative AI model. Specific analysis methods include the following processes:

[0103] 1. Tokenization: Splitting text into words and phrases.

[0104] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[0105] 3. Noun phrase extraction: Extract important noun phrases from the sentence. For example, noun phrases such as "interest" and "suggestion" are extracted.

[0106] Classification of successful and unsuccessful cases

[0107] Server: Classifies successful and unsuccessful deals based on past sales negotiation data. A generative AI model is used for this classification. The generative AI works as follows:

[0108] 1. Prepare a dataset: Use past sales data to prepare a dataset that includes successful and unsuccessful deals.

[0109] 2. Feature extraction: The generative AI model extracts commonalities between successful and unsuccessful deals from past sales negotiation data. For example, it confirms that the phrase "interested" appears frequently in successful deals.

[0110] Keyphrase Extraction

[0111] Server: Extracts frequently occurring key phrases from the meeting minutes of successful deals. This extraction uses TF-IDF scores. For example, phrases such as "cost performance" and "interested" are identified as important.

[0112] Agreement probability prediction

[0113] User: The sales representative enters new sales meeting minutes into the terminal.

[0114] server:

[0115] 1. Keyphrase check: Check whether newly entered meeting notes contain identified keyphrases.

[0116] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[0117] Terminal: Notifies the salesperson in real time of the predicted probability of closing sent from the server. For example, the notification may include a message such as "This opportunity has an 80% chance of closing."

[0118] Specific examples

[0119] For example, consider the case where a sales representative has an initial business meeting with a new customer and writes in the minutes, "They are very interested in proposing a new solution." This data is sent from the device to the server and analyzed by the generative AI model. The AI ​​extracts the key phrase "interested" and compares it with past success stories. As a result of this comparison, it predicts an 80% probability of closing the deal, and this result is notified to the device. The sales representative can then take action as the next step, such as preparing a concrete proposal.

[0120] Example prompts for generative AI models

[0121] "Analyze the following sales meeting transcript and predict the likelihood of closing the deal. Transcript: 'I'm very interested in your new solution proposal.'"

[0122] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0123] Step 1:

[0124] Terminal: After a sales meeting, the sales representative uses the terminal to enter minutes of the meeting. Specifically, the sales representative records the details of the meeting using text input or voice recognition. An example of input is, "I am very interested in this proposal." Once input is complete, the data is temporarily stored on the terminal.

[0125] Step 2:

[0126] Terminal: The entered minutes of the business meeting are sent to the server in real time. A format check is performed, and the data is sent in text format. For example, the format check verifies whether there are any spelling mistakes or inappropriate characters. If the format is determined to be correct, the data is sent to the server.

[0127] Step 3:

[0128] Server: The server stores the received minutes in a database. When stored, each minutes is assigned a unique identifier. The database also stores the minutes' contents along with a timestamp and the user information that entered them.

[0129] Step 4:

[0130] Server: The minutes of business meetings stored in the database are processed using natural language processing with a generative AI model. Specific analysis methods include the following processes:

[0131] 1. Tokenization: Split the text of the meeting notes into words.

[0132] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[0133] 3. Noun phrase extraction: Extract important noun phrases (e.g., “interest,” “suggestion”).

[0134] The input is the text data of the business meeting minutes, and the output is the parsed tokens, tags, and noun phrases.

[0135] Step 5:

[0136] Server: Classifies successful and unsuccessful deals based on past sales data. The generative AI model works as follows:

[0137] 1. Prepare the dataset: Divide past sales negotiation data into successful and unsuccessful cases.

[0138] 2. Feature extraction: Extract common features from successful and unsuccessful cases. Confirm that the phrase "interested" is frequently included in successful cases.

[0139] The input is past sales negotiation data, and the output is the features of successful and unsuccessful cases.

[0140] Step 6:

[0141] Server: Extracts frequent key phrases from successful business meeting minutes using a generative AI model. Specifically, it identifies important phrases such as "cost-effectiveness" and "interested" using TF-IDF scores.

[0142] The input is the sales negotiation data of successful cases, and the output is important key phrases.

[0143] Step 7:

[0144] User: The salesperson enters new meeting minutes into the device. The user again uses text input or voice recognition.

[0145] Step 8:

[0146] server:

[0147] 1. Keyphrase check: Check whether newly entered meeting minutes contain the identified keyphrases.

[0148] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[0149] The input is the newly entered minutes of the business meeting, and the output is the presence or absence of key phrases and the probability of closing the deal.

[0150] Step 9:

[0151] Terminal: The server notifies the salesperson in real time with the predicted close probability results, including a specific message such as "This opportunity has an 80% chance of closing."

[0152] (Application example 1)

[0153] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0154] In conventional sales activities, analysis of sales meeting minutes and prediction of the probability of closing a deal are often performed manually, resulting in inefficiency and inaccuracy. Furthermore, when audio recordings are performed, the process of converting the audio data into text is complicated, making real-time analysis and prediction difficult. The present invention aims to solve these problems and improve the efficiency and accuracy of customer service and sales activities in brick-and-mortar stores.

[0155] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0156] In this invention, the server includes a means for storing the minutes of business negotiations in a database, a means for analyzing the minutes as text data by natural language processing using artificial intelligence, and a means for converting speech to text, which enables real-time text analysis and the identification of key phrases with a high probability of closing a deal.

[0157]

[0158] "Business minutes" refers to information recorded by sales representatives or store clerks about business negotiations and customer service with customers.

[0159] A "database" is a system for storing data in an organized manner and for efficiently searching, extracting, and managing data.

[0160] "Generative AI" is a type of artificial intelligence that performs natural language processing, and is a technology responsible for analyzing and generating text data.

[0161] "Natural language processing" is a technology that allows computers to understand and analyze language spoken by humans.

[0162] "Text data" refers to data that stores character strings or sentences in digital format.

[0163] "Past business negotiation data" is information about business negotiations that have been conducted in the past, including records of successful and unsuccessful cases.

[0164] A "successful deal" refers to a case where a deal has been concluded.

[0165] A "failed deal" refers to a deal that did not result in a successful transaction.

[0166] A "keyphrase" is a word or phrase that has significant meaning in a particular document or conversation.

[0167] "Agreement probability" refers to the probability that a deal will be concluded.

[0168] "Notifying the user" means sending the analysis results and prediction results to the user's terminal and informing them.

[0169] "Speech to text" is the process of converting recorded audio data into a string or sentence format.

[0170] A "user interface that controls voice input" refers to software that controls the operation screen and voice recognition that allows users to input voice.

[0171]

[0172] The present invention is a system that records the details of business negotiations and customer service, analyzes the details, and predicts the likelihood of a deal. In particular, as an example of application in a physical store, it provides a means for recording conversations between store clerks and customers and predicting the likelihood of a deal in real time based on the data.

[0173] System Configuration

[0174] Hardware Configuration

[0175] The system of the present invention uses the following hardware:

[0176] 1. Terminal: A device capable of voice input and output, such as smart glasses or a head-mounted display.

[0177] 2. Server: A high-performance computer to run the database and generative AI.

[0178] Software Configuration

[0179] The system operates using the following software:

[0180] 1. Speech recognition software: Use the speech_recognition library to convert voice data into text data.

[0181] 2. Generative AI software: Hugging Face's Transformer model is used to analyze the text data.

[0182] 3. Text analysis software: Identify key phrases using TF-IDF and predict conversion probability using a logistic regression model.

[0183] Data processing and calculation

[0184] 1. Voice input:

[0185] The terminal records the conversation between the store clerk and the customer and transmits the audio data to the server.

[0186] 2. Audio conversion:

[0187] Speech recognition software running on the server converts the recorded voice data into text data.

[0188] 3. Text Analysis:

[0189] A generative AI model is used to perform natural language processing on the converted text data and extract key phrases, using the Hugging Face Transformer model.

[0190] 4. Data storage:

[0191] The text data and extracted key phrases are stored in a database.

[0192] 5. Win probability prediction:

[0193] Using a logistic regression model built on data from past successful and unsuccessful cases, the probability of closing a deal is predicted from the extracted key phrases.

[0194] 6. Notification of Results:

[0195] The prediction results are sent to the terminal in real time and presented to the store clerk.

[0196] Specific examples

[0197] For example, a salesperson can record the phrase "I'm very interested in this product" while serving a customer, and the recording is converted into text. This text is sent to the system, where generative AI and TF-IDF analysis are used to extract the key phrase "interested." This key phrase is pattern-matched with past successful cases, and the system predicts a high probability of success. For example, a result such as "85% probability of success" is displayed in real time on the smart glasses.

[0198] Prompt Sentence Examples

[0199] Customer conversation recording text: "I'm very interested in this product." Analyze the following text, extract key phrases, and predict the probability of closing a sale.

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

[0201]

[0202] Step 1:

[0203] A user records a conversation with a customer using smart glasses or a head-mounted display. The input is voice data, and the output is the recorded voice data. This voice data is stored on the device.

[0204] Step 2:

[0205] The device sends voice data to the server. The server receives the voice data and converts it into text data using speech recognition software (speech_recognition library). The input is voice data and the output is text data.

[0206] Step 3:

[0207] The server then performs natural language processing on the converted text data using a generative AI model (Hugging Face's Transformer model). This process involves tokenizing the input text data, tagging it with parts of speech, and extracting noun phrases. The input is text data, and the output is analyzed text data.

[0208] Step 4:

[0209] The server extracts important key phrases from the analyzed text data and identifies frequently occurring key phrases using TF-IDF scores. The input of this process is the analyzed text data, and the output is the extracted key phrases.

[0210] Step 5:

[0211] The server uses a logistic regression model based on data on past successful and unsuccessful cases to predict the probability of success from newly extracted key phrases. The input to this process is the extracted key phrases, and the output is the predicted probability of success.

[0212] Step 6:

[0213] The server sends the predicted probability of closing to the terminal. The terminal notifies the user (store clerk or salesperson) of the received predicted result of the agreement probability in real time. The input of this process is the predicted value of the probability of closing, and the output is a notification to the user.

[0214] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0215] This invention is a system for predicting the likelihood of a deal by analyzing sales meeting minutes using a generation AI and an emotion engine. Specifically, it automates the process of collecting, analyzing, and recognizing emotions in sales meeting minutes, as well as predicting the likelihood of an agreement.

[0216] Collection of business meeting minutes

[0217] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[0218] Data transmission and storage

[0219] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[0220] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[0221] Text analysis with generative artificial intelligence

[0222] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0223] Classification of successful and unsuccessful cases

[0224] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0225] Keyphrase Extraction

[0226] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[0227] Emotion recognition by emotion engine

[0228] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[0229] Server: The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[0230] Agreement probability prediction

[0231] User: The sales representative enters new minutes of the business meeting into the terminal.

[0232] Server: The generative AI predicts the likelihood of agreement for the input minutes based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[0233] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[0234] Specific examples

[0235] For example, consider the case where a salesperson conducts an initial business meeting with a new customer and writes in the minutes that they are "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts a 90% probability of agreement.

[0236] The results are then sent to the device, and the salesperson can use this information to take appropriate action, such as providing further proposals or specific support.In this way, this system uses a data-driven approach combined with emotion recognition to more accurately predict the likelihood of success of a sales negotiation, providing an effective means of improving the efficiency of sales activities.

[0237] The processing flow will be explained below.

[0238] Step 1: Data collection

[0239] User: After a sales meeting, a sales representative uses the device to enter minutes. Two input methods are available: text input and voice recognition, allowing the sales representative to efficiently record the details of the meeting. For example, a comment such as "I'm very interested in this proposal" can be entered.

[0240] Step 2: Send data

[0241] Terminal: The entered minutes are sent from the terminal to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[0242] Step 3: Save data

[0243] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[0244] Step 4: Text Analysis

[0245] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0246] Step 5: Classification of successes and failures

[0247] Server: Based on past sales negotiation data, generative AI is used to classify successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the phrase "interested" appears frequently in successful deals.

[0248] Step 6: Extracting Keyphrases

[0249] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores and correlation analysis are used for extraction. For example, key phrases such as "cost performance" and "interested" are extracted.

[0250] Step 7: Emotion Recognition

[0251] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes the tone of voice, facial expressions, and text content to identify the customer's emotions. For example, if the customer is expecting something, a positive emotion such as "expecting" will be detected.

[0252] Step 8: Storing Emotion Data

[0253] Server: Records the detected emotion data in a database and uses it for subsequent sales negotiation analysis.

[0254] Step 9: Predicting agreement probability

[0255] User: A sales representative enters new sales meeting minutes into a terminal.

[0256] Server: For the minutes entered, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and the emotional data detected by the emotion engine. For example, if the phrase "cost performance" and the emotional data "expectations" are included, the likelihood of agreement is predicted to be 90%.

[0257] Step 10: Notification of prediction results

[0258] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[0259] Through the above processing steps, the present invention scientifically predicts the likelihood of success in business negotiations and achieves improved efficiency in sales activities by combining emotion recognition.

[0260] Example 2

[0261] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0262] Conventional sales negotiation support systems have difficulty taking into account customer emotional information when analyzing sales negotiation minutes and predicting the probability of success. As a result, there was a problem of reduced accuracy in predicting the probability of success. In addition, inputting sales negotiation minutes and sending and saving the data was time-consuming, making efficient operation difficult.

[0263] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0264] In this invention, the server includes means for storing the minutes of business negotiations in a database, means for analyzing the minutes of business negotiations as text data using natural language processing with artificial intelligence, means for classifying successful cases and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for analyzing emotional data during business negotiations using an emotion engine and predicting the likelihood of agreement with higher accuracy based on this, and means for notifying the user of the predicted likelihood of agreement. This enables more accurate prediction of business negotiation results including customer emotional data, and also improves the efficiency of inputting business negotiation minutes and transmitting and saving data.

[0265] A "business meeting minutes" is a document that records what was discussed and what decisions were made during a business meeting.

[0266] A "database" is a system for efficiently managing, storing, searching, and updating large amounts of data.

[0267] "Generative AI" is a technology that uses machine learning algorithms and natural language processing techniques to generate and analyze text data.

[0268] "Natural language processing" is a technology that enables computers to understand, interpret, and generate human language.

[0269] A "key phrase" is a word or phrase that has a particularly important meaning in text data.

[0270] "Agreement probability" is a probability that indicates the likelihood that a proposal will be accepted in a business negotiation.

[0271] "User" refers to an individual or organization that uses this system to input minutes of business negotiations and receive the analysis results.

[0272] An "emotion engine" is a technology that analyzes voice tone, facial expressions, and text content to identify human emotions.

[0273] "Format check" is the process of checking whether the entered data conforms to the specified format.

[0274] "Real-time" is a concept that indicates that data is generated and processed immediately without delay.

[0275] This invention is a system for analyzing sales meeting minutes in sales activities using a generative AI model and an emotion engine to predict the likelihood of a deal. The system automates the processes of collecting, analyzing, and recognizing emotions in sales meeting minutes, and predicting the likelihood of agreement. Specific embodiments for implementing this system are described below.

[0276] Collection of business meeting minutes

[0277] Terminal: The user (salesperson) uses a user interface to input minutes after a business meeting. There are two input methods: text input and voice recognition, which allows for efficient recording. The software used is a voice recognition engine (for example, Google Speech-to-Text). Specifically, the user inputs a comment such as "I'm very interested in this proposal."

[0278] Data transmission and storage

[0279] Terminal: Once the minutes are entered, the data is sent to the server after format checks are performed. The format checks described above are a means to ensure accurate and consistent data entry.

[0280] Server: The server analyzes the received data and stores it in a database. This database is a system that allows for the management and efficient searching of large amounts of data. For example, a common database management system such as MySQL or PostgreSQL is used.

[0281] Text analysis with generative artificial intelligence

[0282] Server: Generates minutes stored in a database and performs natural language processing (NLP) using an AI model. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. Software used includes NLP libraries such as NLTK and SpaCy. For example, the predicate "interesting" is extracted.

[0283] Classification of successful and unsuccessful cases

[0284] Server: Using past sales negotiation data, a generative AI model extracts commonalities between successful and unsuccessful deals. Software used includes machine learning libraries such as scikit-learn and TensorFlow. It is recognized that successful deals frequently include the phrase "interested."

[0285] Keyphrase Extraction

[0286] Server: Extracts frequently occurring key phrases from successful business meeting minutes. Specifically, it calculates TF-IDF scores and identifies weighted key phrases. For example, key phrases such as "cost performance" and "interested" are extracted.

[0287] Emotion recognition by emotion engine

[0288] Terminal: The conversation between the user and the customer during the sales negotiation is analyzed using an emotion engine. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. Software used may include a voice recognition engine (Google Speech-to-Text) or a facial recognition tool (OpenCV). For example, if the customer is excited, it will be identified as "expecting."

[0289] Server: Emotion data is recorded in a database and used for analysis. The collected emotional data is an important indicator for predicting the success rate of business negotiations.

[0290] Agreement probability prediction

[0291] User: After the business meeting, the user inputs new minutes into the terminal.

[0292] Server: Based on the input minutes, the generative AI model predicts the likelihood of agreement based on identified key phrases and emotional data. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the model predicts the likelihood of agreement to be 90%.

[0293] Terminal: Prediction results are notified to salespeople in real time, allowing them to plan their next actions based on this information.

[0294] Examples of concrete examples and prompts

[0295] For example, if a salesperson conducts an initial sales meeting with a new customer and notes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting, this data is sent to the server and analyzed there. The generative AI model predicts an agreement probability of 90% based on the key phrase "interested" and the emotion data "expectation." This result is notified to the device, and the salesperson can use it to consider further proposals and specific support.

[0296] Prompt Sentence Examples

[0297] You wrote in the minutes, "I'm very interested in your new solution proposal." During the negotiation, the emotion engine detected the customer's emotion, such as "I'm looking forward to it." Use this data to predict the likelihood of an agreement.

[0298] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0299] Step 1: Enter meeting notes

[0300] Terminal: After a sales meeting, the user (salesperson) uses the terminal's user interface to input minutes. There are two input formats: text input and voice recognition, and the user can choose either. Input is mainly done using a keyboard or microphone.

[0301] Input: Text or audio data about the business.

[0302] Output: Minutes data that has passed format check.

[0303] What it does: The user enters text into an input form or records a voice, which is then converted into text using a speech recognition engine (e.g., Google Speech-to-Text).

[0304] Step 2: Send and store data

[0305] Terminal: After the minutes are entered, the terminal checks the format and sends the data to the server. Only data that passes the format check is sent.

[0306] Input: Format-checked minutes data.

[0307] Output: The minutes data sent to the server.

[0308] Specific behavior: The device sends data to the server using an HTTP request, including appropriate format conversion if necessary.

[0309] Server: The server analyzes the minutes data received from the terminal and stores it in a database.

[0310] Input: Received minutes data.

[0311] Output: Meeting minutes data stored in a database.

[0312] Specific operation: The server executes an insert operation in the database management system (e.g., MySQL) to permanently store the data.

[0313] Step 3: Text analysis

[0314] Server: The server uses a generative AI model to perform natural language processing (NLP) on the stored minutes data, specifically splitting tokens, tagging parts of speech, and extracting noun phrases.

[0315] Input: Meeting minutes data stored in the database.

[0316] Output: Text data as the analysis result.

[0317] What it does: The server uses an NLP library (e.g., NLTK, SpaCy) to segment the text, tag parts of speech, and extract noun phrases, such as the phrase "interesting."

[0318] Step 4: Classification of successes and failures

[0319] Server: Extracts commonalities between successful and unsuccessful deals based on past sales negotiation data. Classifies meeting minutes data using a generative AI model.

[0320] Inputs: Past and current deal data.

[0321] Output: Classification results of successful and unsuccessful cases.

[0322] What it does: The server uses machine learning algorithms (e.g., scikit-learn, TensorFlow) to learn patterns from past datasets and classify new data, identifying common key phrases and success factors and generating classification results.

[0323] Step 5: Keyphrase Extraction

[0324] Server: Calculate TF-IDF scores to identify frequently occurring key phrases from successful meeting notes, and extract important key phrases.

[0325] Input: Successful deal opportunity data.

[0326] Output: Extracted key phrases.

[0327] Specific operation: The server uses the TF-IDF algorithm to calculate important phrases in the sales negotiation data and extracts the top-ranked key phrases. For example, the phrase "cost performance" is identified as important.

[0328] Step 6: Emotion Recognition

[0329] Terminal: During sales negotiations, conversations between users and customers are collected and analyzed using an emotion engine. Voice tone, facial expressions, and text content are used to identify customer emotions.

[0330] Input: Voice, facial expression, and text data during business negotiations.

[0331] Output: Parsed emotion data.

[0332] What it does: The device uses an emotion engine (e.g., a voice recognition engine or facial recognition tool) to analyze the data and identify the customer's emotions. Positive emotions such as "expecting" are detected.

[0333] Server: The analyzed emotion data is recorded in a database and used for subsequent analysis.

[0334] Input: Parsed emotion data.

[0335] Output: Emotion data stored in a database.

[0336] Specific operation: The server inserts the emotion data into a database, making it available for future analysis.

[0337] Step 7: Predicting agreement probability

[0338] Server: Based on key phrases and sentiment data, a generative AI model is used to predict the likelihood of agreement. It uses a model learned from past successful cases.

[0339] Input: New sales meeting minutes data, identified key phrases, analyzed sentiment data.

[0340] Output: Consensus prediction results.

[0341] Specific operation: The server uses the generative AI model to analyze the input data and calculate the agreement probability. For example, if the comment "good cost performance" and the emotional data "expected" are included, the agreement probability is predicted to be 90%.

[0342] Step 8: Notification

[0343] Terminal: The terminal notifies the user in real time of the predicted consensus probability sent from the server. The user can plan their next action based on this information.

[0344] Input: Prediction results sent from the server.

[0345] Output: The prediction results displayed on the device screen.

[0346] Specific operation: The device uses a notification system to notify the user of the prediction results in real time, allowing the user to plan their next steps.

[0347] (Application example 2)

[0348] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0349] During sales negotiations and customer service in brick-and-mortar stores, it is difficult to accurately predict the customer's level of interest in a product or the likelihood of a sale. Conventional sales negotiation minutes and customer service records rely on subjective judgment, making it difficult to formulate effective proposals and sales strategies. Furthermore, it is difficult to accurately grasp the customer's emotions and make appropriate product proposals based on them.

[0350] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for saving business negotiation minutes in a database, means for analyzing the business negotiation minutes as text data using natural language processing with generative artificial intelligence, means for classifying successful and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in the successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for notifying the user of the predicted likelihood of agreement, means for analyzing customer emotion data using an emotion engine, means for predicting customer satisfaction based on the analyzed emotion data, and means for proposing appropriate products to customers based on the key phrases extracted from customer emotions and the successful cases. This enables product proposals that increase the likelihood of closing business negotiations and customer service and increase customer satisfaction based on the customer's emotions and interests.

[0351] A "business negotiation minutes" is a document that records the details of a business negotiation that a sales representative conducted with a customer.

[0352] A "database" is a system designed to store information in an organized manner and make it easy to access and search.

[0353] "Generative AI" is a system that uses artificial intelligence technology to generate and analyze text and data.

[0354] "Natural language processing" is a technology that allows computers to understand and process human language.

[0355] "Past business negotiation data" refers to data that includes records and results of previous business negotiations.

[0356] A "successful case" refers to a case where a business deal has been concluded.

[0357] A "failed deal" refers to a business deal that did not result in a successful transaction.

[0358] "Key phrases" are particularly important words or phrases in business negotiations or customer comments.

[0359] "Agreement probability" is an index that indicates the degree of possibility that a business negotiation will lead to a contract.

[0360] An "emotion engine" is a system that analyzes and identifies human emotions from voice and text.

[0361] "Emotion data" is information about the customer's emotional state analyzed by the emotion engine.

[0362] "Customer satisfaction" is an indicator of how satisfied customers are with a product or service.

[0363] "Product proposal" is the act of recommending appropriate products or services to customers.

[0364] This invention relates to a system for efficiently handling customer inquiries and conducting business negotiations in brick-and-mortar stores. This system collects business negotiation minutes, analyzes them using generative AI and an emotion engine, and predicts the probability of closing a deal and customer satisfaction. The components and procedures required to implement this invention are described below.

[0365] Collection of business meeting minutes

[0366] Device:

[0367] After the sales meeting, the sales representative will use a smartphone or tablet to input minutes of the meeting. Two input methods are available: text input and voice recognition, allowing sales representatives to efficiently record the details of the meeting.

[0368] Data transmission and storage

[0369] Device:

[0370] The entered minutes of the business meeting are sent from the terminal to the server, where they are checked for format and saved in the appropriate format.

[0371] server:

[0372] The server stores the received minutes of the business negotiations in a database, which allows for centralized management of data for later analysis.

[0373] Text analysis with generative artificial intelligence

[0374] server:

[0375] The minutes stored in the database are then processed using generative AI (e.g., Hugging Face Transformers) for natural language processing. Specifically, tokenization, part-of-speech tagging, and noun phrase extraction are performed. For example, the predicate "interesting" is extracted.

[0376] Classification of successful and unsuccessful cases

[0377] server:

[0378] Based on past sales negotiation data, successful and unsuccessful deals are classified. Generative AI is used to extract commonalities between successful and unsuccessful deals. For example, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0379] Keyphrase Extraction

[0380] server:

[0381] Extract frequently occurring key phrases from successful business meeting minutes. For example, key phrases such as "cost performance" and "interested" are identified as important using TF-IDF scores.

[0382] Emotion recognition by emotion engine

[0383] Device:

[0384] During a sales meeting, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[0385] server:

[0386] The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[0387] Agreement probability prediction

[0388] User:

[0389] The sales representative enters new minutes of the business meeting into the terminal.

[0390] server:

[0391] For input minutes, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[0392] Device:

[0393] The server sends the predicted results of the agreement probability to the sales representative in real time. Based on this information, the sales representative can plan the next action. It also enables the sales representative to propose appropriate products based on the customer's emotions and interests.

[0394] Examples of concrete examples and prompts

[0395] For example, consider the case where a salesperson conducts an initial sales meeting with a new customer and writes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts an agreement probability of 90%. This result is notified to the device, and the salesperson can use this information to take appropriate action, such as making further proposals or specific support.

[0396] Example prompt sentence:

[0397] Convert the conversation between the customer and the store clerk into text and use generative AI to analyze the following:

[0398] 1. Extracting important key phrases

[0399] 2. Recognizing customer sentiment

[0400] 3. Closing probability prediction

[0401] In this way, the system uses a data-driven approach that combines customer sentiment to more accurately predict the likelihood of a deal's success and provide an effective means of improving the efficiency of sales activities.

[0402] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0403] Step 1: Collect meeting notes

[0404] Device:

[0405] Salespeople, who are users, use smartphones or tablets to input minutes of sales negotiations after a sales meeting. There are two input methods: text input and voice recognition. With voice recognition, the user's voice is analyzed and the contents of the sales negotiation are converted into text. These input minutes of sales negotiations become the input data for the next processing step.

[0406] Step 2: Send and store data

[0407] Device:

[0408] Once the input of the business meeting minutes is complete, the terminal sends the minutes to the server. The sent data undergoes a format check and is saved in the appropriate format on the server. The format check verifies the data format and required fields. If successful, the data is saved in the server's database (output).

[0409] Step 3: Text analysis with generative AI

[0410] server:

[0411] Using saved business meeting minutes as input, natural language processing is performed using generative AI (e.g., Hugging Face Transformers). Tokenization, part-of-speech tagging, and noun phrase extraction are performed, and the analysis results are output as text data. For example, the predicate "interesting" is extracted.

[0412] Step 4: Classification of successes and failures

[0413] server:

[0414] By referencing past sales negotiation data, the system uses generative AI to extract commonalities between successful and unsuccessful deals. This allows the minutes of sales negotiations to be classified as either successful or unsuccessful. The classification results are output as labeled data indicating whether the deal was successful or unsuccessful.

[0415] Step 5: Extracting Keyphrases

[0416] server:

[0417] The input is a successful business meeting transcript, and frequent keyphrases are extracted. This process uses TF-IDF scores to identify important keyphrases. Keyphrases such as "cost-effectiveness" and "interested" are output.

[0418] Step 6: Emotion Recognition with the Emotion Engine

[0419] Device:

[0420] During a sales negotiation, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The inputs are voice tone, facial expressions, and text content. The analyzed emotion data is output as an emotion label, such as "expecting."

[0421] server:

[0422] The detected emotion data is recorded in a database, which helps predict the probability of success.

[0423] Step 7: Predicting agreement probability

[0424] server:

[0425] Based on newly input minutes of business meetings, the generation AI predicts the likelihood of agreement. Inputs include the text data of the minutes, pre-specified key phrases, and emotional data. The generation AI analyzes each input data and outputs the likelihood of agreement as a numerical value (percentage). For example, if the input data includes the comment "high cost performance" and the emotional data "we are looking forward to it," the generation AI predicts the likelihood of agreement to be 90%.

[0426] Step 8: Notification of results

[0427] Device:

[0428] The server then sends the predicted agreement probability to the device. The salesperson receives this information in real time and plans the next steps. For example, they can create a specific sales strategy, such as considering additional proposals.

[0429] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0430] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0431] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0432] [Second embodiment]

[0433] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0434] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0435] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0436] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0437] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0438] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0439] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0440] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0441] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0442] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0443] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0444] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0445] This invention is a system that uses AI to analyze sales meeting minutes in sales activities and predict the likelihood of a deal being concluded. Specifically, it automates the process of collecting and analyzing sales meeting minutes and predicting the likelihood of an agreement.

[0446] Collection of business meeting minutes

[0447] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[0448] Data transmission and storage

[0449] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[0450] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[0451] Text analysis with generative artificial intelligence

[0452] Server: Natural language processing is performed on the business meeting minutes stored in the database using generation AI. Specific analysis methods include tokenization, part-of-speech tagging, and noun phrase extraction. For example, noun phrases such as "interest" and "proposal" are extracted.

[0453] Classification of successful and unsuccessful cases

[0454] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0455] Keyphrase Extraction

[0456] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[0457] Agreement probability prediction

[0458] User: The sales representative enters new minutes of the business meeting into the terminal.

[0459] Server: Checks whether newly entered minutes of business meetings contain pre-specified key phrases. If they do, predicts the likelihood of agreement based on the importance of those key phrases. For example, if the comment "high cost performance" is included, predicts the likelihood of agreement to be 80%.

[0460] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[0461] Specific examples

[0462] For example, consider the case where a sales representative holds an initial business meeting with a new client and writes in the minutes, "I am very interested in your proposal for a new solution." These minutes are sent from the device to the server, where they are analyzed. The generation AI extracts the key phrase "interested" and compares it with a database of past success stories. As a result, it confirms that this key phrase appears frequently in successful cases, and predicts an 80% probability of agreement.

[0463] The results are then sent to the terminal, allowing the sales representative to take appropriate action, such as preparing a concrete proposal as the next step.In this way, the system provides data-driven support for sales activities and an effective means of increasing the success rate.

[0464] The processing flow will be explained below.

[0465] Step 1: Data collection

[0466] Terminal: The sales representative inputs minutes of the sales meeting. Either text input or voice recognition input can be selected as the input method. For example, if a customer comments, "This solution is very interesting," the content of that comment is recorded.

[0467] Step 2: Send data

[0468] Terminal: The entered minutes of the business meeting are sent to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[0469] Step 3: Save data

[0470] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[0471] Step 4: Text Analysis

[0472] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0473] Step 5: Classification of successes and failures

[0474] Server: Based on past sales negotiation data, the generative AI classifies successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the word "interest" appears frequently in successful deals.

[0475] Step 6: Extracting Keyphrases

[0476] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores are used for extraction. For example, key phrases such as "cost performance" and "interesting" are extracted.

[0477] Step 7: Predicting agreement probability

[0478] User: A sales representative opens a new business deal and enters the details of the deal into the terminal as minutes.

[0479] Server: Checks key phrases against the input minutes, and the generation AI predicts the degree of agreement. For example, if the phrase "high cost performance" appears, it predicts an 80% degree of agreement.

[0480] Step 8: Notification of prediction results

[0481] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[0482] By using the above processing steps, the present invention provides a system for scientifically predicting the probability of success of a business negotiation and improving the efficiency of sales activities.

[0483] Example 1

[0484] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0485] In conventional sales activities, the collection, analysis, and closing probability prediction of sales meeting minutes are often done manually, which takes time and effort and reduces efficiency. Another issue is that data management is cumbersome, making it difficult to effectively utilize trends in successful deals. The present invention aims to solve these issues by automatically analyzing minutes and improving the accuracy of closing probability prediction.

[0486] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0487] In this invention, the server provides an input device for recording the details of business negotiations, and includes: means for a user to input minutes by text input or voice recognition; means for format-checking the input minutes and sending them; means for saving the minutes in a database; means for analyzing the minutes as text data using natural language processing with artificial intelligence; means for classifying successful and unsuccessful cases based on past business negotiation data; means for extracting important expressions contained in successful cases; means for predicting the likelihood of agreement in future business negotiations based on the extracted important expressions; and means for notifying the user of the predicted likelihood of agreement. This enables efficient collection and analysis of business negotiation minutes and highly accurate prediction of the likelihood of closing a deal.

[0488] "Business negotiation content" refers to information such as proposals, requests, and exchanges of opinions exchanged between customers and sales representatives during sales activities.

[0489] An "input device" is a device or software that a user uses to input information. Examples include a keyboard, microphone, and touch panel.

[0490] "Text input" refers to a method in which a user inputs characters or sentences using a keyboard or the like.

[0491] "Speech recognition" is a technology that inputs the words spoken by a user as voice and converts them into text information.

[0492] Minutes are documents that record the content exchanged, decisions made, and key points made during business negotiations.

[0493] "Format check" refers to the process of verifying whether the input data is in the correct format.

[0494] A "database" is a system for systematically storing large amounts of data and quickly searching and retrieving them.

[0495] "Generative AI" refers to an AI technology that analyzes data and performs pattern recognition and predictions.

[0496] "Natural language processing" refers to techniques that enable artificial intelligence to understand and analyze human language, including sentence tokenization, part-of-speech tagging, and semantic analysis.

[0497] "Past business negotiation data" refers to records and information related to past business negotiations, and is stored in a database.

[0498] A "successful case" is one in which negotiations proceed smoothly and an agreement or contract is reached with the customer.

[0499] A "failed project" is one in which negotiations did not progress and no agreement or contract was reached with the customer.

[0500] "Important expressions" are key phrases or noun phrases that are often found in successful deals and have a significant impact on the outcome of subsequent sales negotiations.

[0501] "Agreement probability" is an indicator that indicates the probability that a particular business deal will be successful.

[0502] "Notification" refers to informing a user of information or results, and is often done in real time via a terminal.

[0503] This invention is a system for analyzing sales negotiation minutes in sales activities using a generative AI model and predicting the probability of closing a deal. Specific embodiments of this system are described in detail below.

[0504] Entering business meeting minutes

[0505] Terminal: After a sales meeting, the sales representative uses the terminal to enter details of the sales meeting. Two methods of input are available: text input and voice recognition. For example, a sales representative can enter something like "I'm very interested in this proposal" as text or use voice recognition.

[0506] Data transmission and storage

[0507] Terminal: The entered minutes are sent to the server in real time, where a format check is performed to ensure the data is in the correct format.

[0508] Server: The received minutes of business meetings are stored in a database. Each minutes of business meetings is assigned a unique identifier and managed centrally in the database.

[0509] Text analytics

[0510] Server: Natural language processing is performed on the business meeting minutes stored in the database using a generative AI model. Specific analysis methods include the following processes:

[0511] 1. Tokenization: Splitting text into words and phrases.

[0512] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[0513] 3. Noun phrase extraction: Extract important noun phrases from the sentence. For example, noun phrases such as "interest" and "suggestion" are extracted.

[0514] Classification of successful and unsuccessful cases

[0515] Server: Classifies successful and unsuccessful deals based on past sales negotiation data. A generative AI model is used for this classification. The generative AI works as follows:

[0516] 1. Prepare a dataset: Use past sales data to prepare a dataset that includes successful and unsuccessful deals.

[0517] 2. Feature extraction: The generative AI model extracts commonalities between successful and unsuccessful deals from past sales negotiation data. For example, it confirms that the phrase "interested" appears frequently in successful deals.

[0518] Keyphrase Extraction

[0519] Server: Extracts frequently occurring key phrases from the meeting minutes of successful deals. This extraction uses TF-IDF scores. For example, phrases such as "cost performance" and "interested" are identified as important.

[0520] Agreement probability prediction

[0521] User: The sales representative enters new sales meeting minutes into the terminal.

[0522] server:

[0523] 1. Keyphrase check: Check whether newly entered meeting notes contain identified keyphrases.

[0524] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[0525] Terminal: Notifies the salesperson in real time of the predicted probability of closing sent from the server. For example, the notification may include a message such as "This opportunity has an 80% chance of closing."

[0526] Specific examples

[0527] For example, consider the case where a sales representative has an initial business meeting with a new customer and writes in the minutes, "They are very interested in proposing a new solution." This data is sent from the device to the server and analyzed by the generative AI model. The AI ​​extracts the key phrase "interested" and compares it with past success stories. As a result of this comparison, it predicts an 80% probability of closing the deal, and this result is notified to the device. The sales representative can then take action as the next step, such as preparing a concrete proposal.

[0528] Example prompts for generative AI models

[0529] "Analyze the following sales meeting transcript and predict the likelihood of closing the deal. Transcript: 'I'm very interested in your new solution proposal.'"

[0530] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0531] Step 1:

[0532] Terminal: After a sales meeting, the sales representative uses the terminal to enter minutes of the meeting. Specifically, the sales representative records the details of the meeting using text input or voice recognition. An example of input is, "I am very interested in this proposal." Once input is complete, the data is temporarily stored on the terminal.

[0533] Step 2:

[0534] Terminal: The entered minutes of the business meeting are sent to the server in real time. A format check is performed, and the data is sent in text format. For example, the format check verifies whether there are any spelling mistakes or inappropriate characters. If the format is determined to be correct, the data is sent to the server.

[0535] Step 3:

[0536] Server: The server stores the received minutes in a database. When stored, each minutes is assigned a unique identifier. The database also stores the minutes' contents along with a timestamp and the user information that entered them.

[0537] Step 4:

[0538] Server: The minutes of business meetings stored in the database are processed using natural language processing with a generative AI model. Specific analysis methods include the following processes:

[0539] 1. Tokenization: Split the text of the meeting notes into words.

[0540] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[0541] 3. Noun phrase extraction: Extract important noun phrases (e.g., “interest,” “suggestion”).

[0542] The input is the text data of the business meeting minutes, and the output is the parsed tokens, tags, and noun phrases.

[0543] Step 5:

[0544] Server: Classifies successful and unsuccessful deals based on past sales data. The generative AI model works as follows:

[0545] 1. Prepare the dataset: Divide past sales negotiation data into successful and unsuccessful cases.

[0546] 2. Feature extraction: Extract common features from successful and unsuccessful cases. Confirm that the phrase "interested" is frequently included in successful cases.

[0547] The input is past sales negotiation data, and the output is the features of successful and unsuccessful cases.

[0548] Step 6:

[0549] Server: Extracts frequent key phrases from successful business meeting minutes using a generative AI model. Specifically, it identifies important phrases such as "cost-effectiveness" and "interested" using TF-IDF scores.

[0550] The input is the sales negotiation data of successful cases, and the output is important key phrases.

[0551] Step 7:

[0552] User: The salesperson enters new meeting minutes into the device. The user again uses text input or voice recognition.

[0553] Step 8:

[0554] server:

[0555] 1. Keyphrase check: Check whether newly entered meeting minutes contain the identified keyphrases.

[0556] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[0557] The input is the newly entered minutes of the business meeting, and the output is the presence or absence of key phrases and the probability of closing the deal.

[0558] Step 9:

[0559] Terminal: The server notifies the salesperson in real time with the predicted close probability results, including a specific message such as "This opportunity has an 80% chance of closing."

[0560] (Application example 1)

[0561] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0562] In conventional sales activities, analysis of sales meeting minutes and prediction of the probability of closing a deal are often performed manually, resulting in inefficiency and inaccuracy. Furthermore, when audio recordings are performed, the process of converting the audio data into text is complicated, making real-time analysis and prediction difficult. The present invention aims to solve these problems and improve the efficiency and accuracy of customer service and sales activities in brick-and-mortar stores.

[0563] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0564] In this invention, the server includes a means for storing the minutes of business negotiations in a database, a means for analyzing the minutes as text data by natural language processing using artificial intelligence, and a means for converting speech to text, which enables real-time text analysis and the identification of key phrases with a high probability of closing a deal.

[0565]

[0566] "Business minutes" refers to information recorded by sales representatives or store clerks about business negotiations and customer service with customers.

[0567] A "database" is a system for storing data in an organized manner and for efficiently searching, extracting, and managing data.

[0568] "Generative AI" is a type of artificial intelligence that performs natural language processing, and is a technology responsible for analyzing and generating text data.

[0569] "Natural language processing" is a technology that allows computers to understand and analyze language spoken by humans.

[0570] "Text data" refers to data that stores character strings or sentences in digital format.

[0571] "Past business negotiation data" is information about business negotiations that have been conducted in the past, including records of successful and unsuccessful cases.

[0572] A "successful deal" refers to a case where a deal has been concluded.

[0573] A "failed deal" refers to a deal that did not result in a successful transaction.

[0574] A "keyphrase" is a word or phrase that has significant meaning in a particular document or conversation.

[0575] "Agreement probability" refers to the probability that a deal will be concluded.

[0576] "Notifying the user" means sending the analysis results and prediction results to the user's terminal and informing them.

[0577] "Speech to text" is the process of converting recorded audio data into a string or sentence format.

[0578] A "user interface that controls voice input" refers to software that controls the operation screen and voice recognition that allows users to input voice.

[0579]

[0580] The present invention is a system that records the details of business negotiations and customer service, analyzes the details, and predicts the likelihood of a deal. In particular, as an example of application in a physical store, it provides a means for recording conversations between store clerks and customers and predicting the likelihood of a deal in real time based on the data.

[0581] System Configuration

[0582] Hardware Configuration

[0583] The system of the present invention uses the following hardware:

[0584] 1. Terminal: A device capable of voice input and output, such as smart glasses or a head-mounted display.

[0585] 2. Server: A high-performance computer to run the database and generative AI.

[0586] Software Configuration

[0587] The system operates using the following software:

[0588] 1. Speech recognition software: Use the speech_recognition library to convert voice data into text data.

[0589] 2. Generative AI software: Hugging Face's Transformer model is used to analyze the text data.

[0590] 3. Text analysis software: Identify key phrases using TF-IDF and predict conversion probability using a logistic regression model.

[0591] Data processing and calculation

[0592] 1. Voice input:

[0593] The terminal records the conversation between the store clerk and the customer and transmits the audio data to the server.

[0594] 2. Audio conversion:

[0595] Speech recognition software running on the server converts the recorded voice data into text data.

[0596] 3. Text Analysis:

[0597] A generative AI model is used to perform natural language processing on the converted text data and extract key phrases, using the Hugging Face Transformer model.

[0598] 4. Data storage:

[0599] The text data and extracted key phrases are stored in a database.

[0600] 5. Win probability prediction:

[0601] Using a logistic regression model built on data from past successful and unsuccessful cases, the probability of closing a deal is predicted from the extracted key phrases.

[0602] 6. Notification of Results:

[0603] The prediction results are sent to the terminal in real time and presented to the store clerk.

[0604] Specific examples

[0605] For example, a salesperson can record the phrase "I'm very interested in this product" while serving a customer, and the recording is converted into text. This text is sent to the system, where generative AI and TF-IDF analysis are used to extract the key phrase "interested." This key phrase is pattern-matched with past successful cases, and the system predicts a high probability of success. For example, a result such as "85% probability of success" is displayed in real time on the smart glasses.

[0606] Prompt Sentence Examples

[0607] Customer conversation recording text: "I'm very interested in this product." Analyze the following text, extract key phrases, and predict the probability of closing a sale.

[0608] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0609]

[0610] Step 1:

[0611] A user records a conversation with a customer using smart glasses or a head-mounted display. The input is voice data, and the output is the recorded voice data. This voice data is stored on the device.

[0612] Step 2:

[0613] The device sends voice data to the server. The server receives the voice data and converts it into text data using speech recognition software (speech_recognition library). The input is voice data and the output is text data.

[0614] Step 3:

[0615] The server then performs natural language processing on the converted text data using a generative AI model (Hugging Face's Transformer model). This process involves tokenizing the input text data, tagging it with parts of speech, and extracting noun phrases. The input is text data, and the output is analyzed text data.

[0616] Step 4:

[0617] The server extracts important key phrases from the analyzed text data and identifies frequently occurring key phrases using TF-IDF scores. The input of this process is the analyzed text data, and the output is the extracted key phrases.

[0618] Step 5:

[0619] The server uses a logistic regression model based on data on past successful and unsuccessful cases to predict the probability of success from newly extracted key phrases. The input to this process is the extracted key phrases, and the output is the predicted probability of success.

[0620] Step 6:

[0621] The server sends the predicted probability of closing to the terminal. The terminal notifies the user (store clerk or salesperson) of the received predicted result of the agreement probability in real time. The input of this process is the predicted value of the probability of closing, and the output is a notification to the user.

[0622] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0623] This invention is a system for predicting the likelihood of a deal by analyzing sales meeting minutes using a generation AI and an emotion engine. Specifically, it automates the process of collecting, analyzing, and recognizing emotions in sales meeting minutes, as well as predicting the likelihood of an agreement.

[0624] Collection of business meeting minutes

[0625] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[0626] Data transmission and storage

[0627] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[0628] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[0629] Text analysis with generative artificial intelligence

[0630] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0631] Classification of successful and unsuccessful cases

[0632] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0633] Keyphrase Extraction

[0634] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[0635] Emotion recognition by emotion engine

[0636] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[0637] Server: The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[0638] Agreement probability prediction

[0639] User: The sales representative enters new minutes of the business meeting into the terminal.

[0640] Server: The generative AI predicts the likelihood of agreement for the input minutes based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[0641] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[0642] Specific examples

[0643] For example, consider the case where a salesperson conducts an initial business meeting with a new customer and writes in the minutes that they are "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts a 90% probability of agreement.

[0644] The results are then sent to the device, and the salesperson can use this information to take appropriate action, such as providing further proposals or specific support.In this way, this system uses a data-driven approach combined with emotion recognition to more accurately predict the likelihood of success of a sales negotiation, providing an effective means of improving the efficiency of sales activities.

[0645] The processing flow will be explained below.

[0646] Step 1: Data collection

[0647] User: After a sales meeting, a sales representative uses the device to enter minutes. Two input methods are available: text input and voice recognition, allowing the sales representative to efficiently record the details of the meeting. For example, a comment such as "I'm very interested in this proposal" can be entered.

[0648] Step 2: Send data

[0649] Terminal: The entered minutes are sent from the terminal to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[0650] Step 3: Save data

[0651] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[0652] Step 4: Text Analysis

[0653] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0654] Step 5: Classification of successes and failures

[0655] Server: Based on past sales negotiation data, generative AI is used to classify successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the phrase "interested" appears frequently in successful deals.

[0656] Step 6: Extracting Keyphrases

[0657] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores and correlation analysis are used for extraction. For example, key phrases such as "cost performance" and "interested" are extracted.

[0658] Step 7: Emotion Recognition

[0659] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes the tone of voice, facial expressions, and text content to identify the customer's emotions. For example, if the customer is expecting something, a positive emotion such as "expecting" will be detected.

[0660] Step 8: Storing Emotion Data

[0661] Server: Records the detected emotion data in a database and uses it for subsequent sales negotiation analysis.

[0662] Step 9: Predicting agreement probability

[0663] User: A sales representative enters new sales meeting minutes into a terminal.

[0664] Server: For the minutes entered, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and the emotional data detected by the emotion engine. For example, if the phrase "cost performance" and the emotional data "expectations" are included, the likelihood of agreement is predicted to be 90%.

[0665] Step 10: Notification of prediction results

[0666] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[0667] Through the above processing steps, the present invention scientifically predicts the likelihood of success in business negotiations and achieves improved efficiency in sales activities by combining emotion recognition.

[0668] Example 2

[0669] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0670] Conventional sales negotiation support systems have difficulty taking into account customer emotional information when analyzing sales negotiation minutes and predicting the probability of success. As a result, there was a problem of reduced accuracy in predicting the probability of success. In addition, inputting sales negotiation minutes and sending and saving the data was time-consuming, making efficient operation difficult.

[0671] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0672] In this invention, the server includes means for storing the minutes of business negotiations in a database, means for analyzing the minutes of business negotiations as text data using natural language processing with artificial intelligence, means for classifying successful cases and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for analyzing emotional data during business negotiations using an emotion engine and predicting the likelihood of agreement with higher accuracy based on this, and means for notifying the user of the predicted likelihood of agreement. This enables more accurate prediction of business negotiation results including customer emotional data, and also improves the efficiency of inputting business negotiation minutes and transmitting and saving data.

[0673] A "business meeting minutes" is a document that records what was discussed and what decisions were made during a business meeting.

[0674] A "database" is a system for efficiently managing, storing, searching, and updating large amounts of data.

[0675] "Generative AI" is a technology that uses machine learning algorithms and natural language processing techniques to generate and analyze text data.

[0676] "Natural language processing" is a technology that enables computers to understand, interpret, and generate human language.

[0677] A "key phrase" is a word or phrase that has a particularly important meaning in text data.

[0678] "Agreement probability" is a probability that indicates the likelihood that a proposal will be accepted in a business negotiation.

[0679] "User" refers to an individual or organization that uses this system to input minutes of business negotiations and receive the analysis results.

[0680] An "emotion engine" is a technology that analyzes voice tone, facial expressions, and text content to identify human emotions.

[0681] "Format check" is the process of checking whether the entered data conforms to the specified format.

[0682] "Real-time" is a concept that indicates that data is generated and processed immediately without delay.

[0683] This invention is a system for analyzing sales meeting minutes in sales activities using a generative AI model and an emotion engine to predict the likelihood of a deal. The system automates the processes of collecting, analyzing, and recognizing emotions in sales meeting minutes, and predicting the likelihood of agreement. Specific embodiments for implementing this system are described below.

[0684] Collection of business meeting minutes

[0685] Terminal: The user (salesperson) uses a user interface to input minutes after a business meeting. There are two input methods: text input and voice recognition, which allows for efficient recording. The software used is a voice recognition engine (for example, Google Speech-to-Text). Specifically, the user inputs a comment such as "I'm very interested in this proposal."

[0686] Data transmission and storage

[0687] Terminal: Once the minutes are entered, the data is sent to the server after format checks are performed. The format checks described above are a means to ensure accurate and consistent data entry.

[0688] Server: The server analyzes the received data and stores it in a database. This database is a system that allows for the management and efficient searching of large amounts of data. For example, a common database management system such as MySQL or PostgreSQL is used.

[0689] Text analysis with generative artificial intelligence

[0690] Server: Generates minutes stored in a database and performs natural language processing (NLP) using an AI model. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. Software used includes NLP libraries such as NLTK and SpaCy. For example, the predicate "interesting" is extracted.

[0691] Classification of successful and unsuccessful cases

[0692] Server: Using past sales negotiation data, a generative AI model extracts commonalities between successful and unsuccessful deals. Software used includes machine learning libraries such as scikit-learn and TensorFlow. It is recognized that successful deals frequently include the phrase "interested."

[0693] Keyphrase Extraction

[0694] Server: Extracts frequently occurring key phrases from successful business meeting minutes. Specifically, it calculates TF-IDF scores and identifies weighted key phrases. For example, key phrases such as "cost performance" and "interested" are extracted.

[0695] Emotion recognition by emotion engine

[0696] Terminal: The conversation between the user and the customer during the sales negotiation is analyzed using an emotion engine. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. Software used may include a voice recognition engine (Google Speech-to-Text) or a facial recognition tool (OpenCV). For example, if the customer is excited, it will be identified as "expecting."

[0697] Server: Emotion data is recorded in a database and used for analysis. The collected emotional data is an important indicator for predicting the success rate of business negotiations.

[0698] Agreement probability prediction

[0699] User: After the business meeting, the user inputs new minutes into the terminal.

[0700] Server: Based on the input minutes, the generative AI model predicts the likelihood of agreement based on identified key phrases and emotional data. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the model predicts the likelihood of agreement to be 90%.

[0701] Terminal: Prediction results are notified to salespeople in real time, allowing them to plan their next actions based on this information.

[0702] Examples of concrete examples and prompts

[0703] For example, if a salesperson conducts an initial sales meeting with a new customer and notes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting, this data is sent to the server and analyzed there. The generative AI model predicts an agreement probability of 90% based on the key phrase "interested" and the emotion data "expectation." This result is notified to the device, and the salesperson can use it to consider further proposals and specific support.

[0704] Prompt Sentence Examples

[0705] You wrote in the minutes, "I'm very interested in your new solution proposal." During the negotiation, the emotion engine detected the customer's emotion, such as "I'm looking forward to it." Use this data to predict the likelihood of an agreement.

[0706] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0707] Step 1: Enter meeting notes

[0708] Terminal: After a sales meeting, the user (salesperson) uses the terminal's user interface to input minutes. There are two input formats: text input and voice recognition, and the user can choose either. Input is mainly done using a keyboard or microphone.

[0709] Input: Text or audio data about the business.

[0710] Output: Minutes data that has passed format check.

[0711] What it does: The user enters text into an input form or records a voice, which is then converted into text using a speech recognition engine (e.g., Google Speech-to-Text).

[0712] Step 2: Send and store data

[0713] Terminal: After the minutes are entered, the terminal checks the format and sends the data to the server. Only data that passes the format check is sent.

[0714] Input: Format-checked minutes data.

[0715] Output: The minutes data sent to the server.

[0716] Specific behavior: The device sends data to the server using an HTTP request, including appropriate format conversion if necessary.

[0717] Server: The server analyzes the minutes data received from the terminal and stores it in a database.

[0718] Input: Received minutes data.

[0719] Output: Meeting minutes data stored in a database.

[0720] Specific operation: The server executes an insert operation in the database management system (e.g., MySQL) to permanently store the data.

[0721] Step 3: Text analysis

[0722] Server: The server uses a generative AI model to perform natural language processing (NLP) on the stored minutes data, specifically splitting tokens, tagging parts of speech, and extracting noun phrases.

[0723] Input: Meeting minutes data stored in the database.

[0724] Output: Text data as the analysis result.

[0725] What it does: The server uses an NLP library (e.g., NLTK, SpaCy) to segment the text, tag parts of speech, and extract noun phrases, such as the phrase "interesting."

[0726] Step 4: Classification of successes and failures

[0727] Server: Extracts commonalities between successful and unsuccessful deals based on past sales negotiation data. Classifies meeting minutes data using a generative AI model.

[0728] Inputs: Past and current deal data.

[0729] Output: Classification results of successful and unsuccessful cases.

[0730] What it does: The server uses machine learning algorithms (e.g., scikit-learn, TensorFlow) to learn patterns from past datasets and classify new data, identifying common key phrases and success factors and generating classification results.

[0731] Step 5: Keyphrase Extraction

[0732] Server: Calculate TF-IDF scores to identify frequently occurring key phrases from successful meeting notes, and extract important key phrases.

[0733] Input: Successful deal opportunity data.

[0734] Output: Extracted key phrases.

[0735] Specific operation: The server uses the TF-IDF algorithm to calculate important phrases in the sales negotiation data and extracts the top-ranked key phrases. For example, the phrase "cost performance" is identified as important.

[0736] Step 6: Emotion Recognition

[0737] Terminal: During sales negotiations, conversations between users and customers are collected and analyzed using an emotion engine. Voice tone, facial expressions, and text content are used to identify customer emotions.

[0738] Input: Voice, facial expression, and text data during business negotiations.

[0739] Output: Parsed emotion data.

[0740] What it does: The device uses an emotion engine (e.g., a voice recognition engine or facial recognition tool) to analyze the data and identify the customer's emotions. Positive emotions such as "expecting" are detected.

[0741] Server: The analyzed emotion data is recorded in a database and used for subsequent analysis.

[0742] Input: Parsed emotion data.

[0743] Output: Emotion data stored in a database.

[0744] Specific operation: The server inserts the emotion data into a database, making it available for future analysis.

[0745] Step 7: Predicting agreement probability

[0746] Server: Based on key phrases and sentiment data, a generative AI model is used to predict the likelihood of agreement. It uses a model learned from past successful cases.

[0747] Input: New sales meeting minutes data, identified key phrases, analyzed sentiment data.

[0748] Output: Consensus prediction results.

[0749] Specific operation: The server uses the generative AI model to analyze the input data and calculate the agreement probability. For example, if the comment "good cost performance" and the emotional data "expected" are included, the agreement probability is predicted to be 90%.

[0750] Step 8: Notification

[0751] Terminal: The terminal notifies the user in real time of the predicted consensus probability sent from the server. The user can plan their next action based on this information.

[0752] Input: Prediction results sent from the server.

[0753] Output: The prediction results displayed on the device screen.

[0754] Specific operation: The device uses a notification system to notify the user of the prediction results in real time, allowing the user to plan their next steps.

[0755] (Application example 2)

[0756] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0757] During sales negotiations and customer service in brick-and-mortar stores, it is difficult to accurately predict the customer's level of interest in a product or the likelihood of a sale. Conventional sales negotiation minutes and customer service records rely on subjective judgment, making it difficult to formulate effective proposals and sales strategies. Furthermore, it is difficult to accurately grasp the customer's emotions and make appropriate product proposals based on them.

[0758] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for saving business negotiation minutes in a database, means for analyzing the business negotiation minutes as text data using natural language processing with generative artificial intelligence, means for classifying successful and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in the successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for notifying the user of the predicted likelihood of agreement, means for analyzing customer emotion data using an emotion engine, means for predicting customer satisfaction based on the analyzed emotion data, and means for proposing appropriate products to customers based on the key phrases extracted from customer emotions and the successful cases. This enables product proposals that increase the likelihood of closing business negotiations and customer service and increase customer satisfaction based on the customer's emotions and interests.

[0759] A "business negotiation minutes" is a document that records the details of a business negotiation that a sales representative conducted with a customer.

[0760] A "database" is a system designed to store information in an organized manner and make it easy to access and search.

[0761] "Generative AI" is a system that uses artificial intelligence technology to generate and analyze text and data.

[0762] "Natural language processing" is a technology that allows computers to understand and process human language.

[0763] "Past business negotiation data" refers to data that includes records and results of previous business negotiations.

[0764] A "successful case" refers to a case where a business deal has been concluded.

[0765] A "failed deal" refers to a business deal that did not result in a successful transaction.

[0766] "Key phrases" are particularly important words or phrases in business negotiations or customer comments.

[0767] "Agreement probability" is an index that indicates the degree of possibility that a business negotiation will lead to a contract.

[0768] An "emotion engine" is a system that analyzes and identifies human emotions from voice and text.

[0769] "Emotion data" is information about the customer's emotional state analyzed by the emotion engine.

[0770] "Customer satisfaction" is an indicator of how satisfied customers are with a product or service.

[0771] "Product proposal" is the act of recommending appropriate products or services to customers.

[0772] This invention relates to a system for efficiently handling customer inquiries and conducting business negotiations in brick-and-mortar stores. This system collects business negotiation minutes, analyzes them using generative AI and an emotion engine, and predicts the probability of closing a deal and customer satisfaction. The components and procedures required to implement this invention are described below.

[0773] Collection of business meeting minutes

[0774] Device:

[0775] After the sales meeting, the sales representative will use a smartphone or tablet to input minutes of the meeting. Two input methods are available: text input and voice recognition, allowing sales representatives to efficiently record the details of the meeting.

[0776] Data transmission and storage

[0777] Device:

[0778] The entered minutes of the business meeting are sent from the terminal to the server, where they are checked for format and saved in the appropriate format.

[0779] server:

[0780] The server stores the received minutes of the business negotiations in a database, which allows for centralized management of data for later analysis.

[0781] Text analysis with generative artificial intelligence

[0782] server:

[0783] The minutes stored in the database are then processed using generative AI (e.g., Hugging Face Transformers) for natural language processing. Specifically, tokenization, part-of-speech tagging, and noun phrase extraction are performed. For example, the predicate "interesting" is extracted.

[0784] Classification of successful and unsuccessful cases

[0785] server:

[0786] Based on past sales negotiation data, successful and unsuccessful deals are classified. Generative AI is used to extract commonalities between successful and unsuccessful deals. For example, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0787] Keyphrase Extraction

[0788] server:

[0789] Extract frequently occurring key phrases from successful business meeting minutes. For example, key phrases such as "cost performance" and "interested" are identified as important using TF-IDF scores.

[0790] Emotion recognition by emotion engine

[0791] Device:

[0792] During a sales meeting, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[0793] server:

[0794] The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[0795] Agreement probability prediction

[0796] User:

[0797] The sales representative enters new minutes of the business meeting into the terminal.

[0798] server:

[0799] For input minutes, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[0800] Device:

[0801] The server sends the predicted results of the agreement probability to the sales representative in real time. Based on this information, the sales representative can plan the next action. It also enables the sales representative to propose appropriate products based on the customer's emotions and interests.

[0802] Examples of concrete examples and prompts

[0803] For example, consider the case where a salesperson conducts an initial sales meeting with a new customer and writes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts an agreement probability of 90%. This result is notified to the device, and the salesperson can use this information to take appropriate action, such as making further proposals or specific support.

[0804] Example prompt sentence:

[0805] Convert the conversation between the customer and the store clerk into text and use generative AI to analyze the following:

[0806] 1. Extracting important key phrases

[0807] 2. Recognizing customer sentiment

[0808] 3. Closing probability prediction

[0809] In this way, the system uses a data-driven approach that combines customer sentiment to more accurately predict the likelihood of a deal's success and provide an effective means of improving the efficiency of sales activities.

[0810] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0811] Step 1: Collect meeting notes

[0812] Device:

[0813] Salespeople, who are users, use smartphones or tablets to input minutes of sales negotiations after a sales meeting. There are two input methods: text input and voice recognition. With voice recognition, the user's voice is analyzed and the contents of the sales negotiation are converted into text. These input minutes of sales negotiations become the input data for the next processing step.

[0814] Step 2: Send and store data

[0815] Device:

[0816] Once the input of the business meeting minutes is complete, the terminal sends the minutes to the server. The sent data undergoes a format check and is saved in the appropriate format on the server. The format check verifies the data format and required fields. If successful, the data is saved in the server's database (output).

[0817] Step 3: Text analysis with generative AI

[0818] server:

[0819] Using saved business meeting minutes as input, natural language processing is performed using generative AI (e.g., Hugging Face Transformers). Tokenization, part-of-speech tagging, and noun phrase extraction are performed, and the analysis results are output as text data. For example, the predicate "interesting" is extracted.

[0820] Step 4: Classification of successes and failures

[0821] server:

[0822] By referencing past sales negotiation data, the system uses generative AI to extract commonalities between successful and unsuccessful deals. This allows the minutes of sales negotiations to be classified as either successful or unsuccessful. The classification results are output as labeled data indicating whether the deal was successful or unsuccessful.

[0823] Step 5: Extracting Keyphrases

[0824] server:

[0825] The input is a successful business meeting transcript, and frequent keyphrases are extracted. This process uses TF-IDF scores to identify important keyphrases. Keyphrases such as "cost-effectiveness" and "interested" are output.

[0826] Step 6: Emotion Recognition with the Emotion Engine

[0827] Device:

[0828] During a sales negotiation, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The inputs are voice tone, facial expressions, and text content. The analyzed emotion data is output as an emotion label, such as "expecting."

[0829] server:

[0830] The detected emotion data is recorded in a database, which helps predict the probability of success.

[0831] Step 7: Predicting agreement probability

[0832] server:

[0833] Based on newly input minutes of business meetings, the generation AI predicts the likelihood of agreement. Inputs include the text data of the minutes, pre-specified key phrases, and emotional data. The generation AI analyzes each input data and outputs the likelihood of agreement as a numerical value (percentage). For example, if the input data includes the comment "high cost performance" and the emotional data "we are looking forward to it," the generation AI predicts the likelihood of agreement to be 90%.

[0834] Step 8: Notification of results

[0835] Device:

[0836] The server then sends the predicted agreement probability to the device. The salesperson receives this information in real time and plans the next steps. For example, they can create a specific sales strategy, such as considering additional proposals.

[0837] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0838] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0839] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0840] [Third embodiment]

[0841] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0842] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0843] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0844] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0845] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0846] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0847] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0848] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0849] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0850] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0851] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0852] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0853] This invention is a system that uses AI to analyze sales meeting minutes in sales activities and predict the likelihood of a deal being concluded. Specifically, it automates the process of collecting and analyzing sales meeting minutes and predicting the likelihood of an agreement.

[0854] Collection of business meeting minutes

[0855] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[0856] Data transmission and storage

[0857] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[0858] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[0859] Text analysis with generative artificial intelligence

[0860] Server: Natural language processing is performed on the business meeting minutes stored in the database using generation AI. Specific analysis methods include tokenization, part-of-speech tagging, and noun phrase extraction. For example, noun phrases such as "interest" and "proposal" are extracted.

[0861] Classification of successful and unsuccessful cases

[0862] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[0863] Keyphrase Extraction

[0864] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[0865] Agreement probability prediction

[0866] User: The sales representative enters new minutes of the business meeting into the terminal.

[0867] Server: Checks whether newly entered minutes of business meetings contain pre-specified key phrases. If they do, predicts the likelihood of agreement based on the importance of those key phrases. For example, if the comment "high cost performance" is included, predicts the likelihood of agreement to be 80%.

[0868] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[0869] Specific examples

[0870] For example, consider the case where a sales representative holds an initial business meeting with a new client and writes in the minutes, "I am very interested in your proposal for a new solution." These minutes are sent from the device to the server, where they are analyzed. The generation AI extracts the key phrase "interested" and compares it with a database of past success stories. As a result, it confirms that this key phrase appears frequently in successful cases, and predicts an 80% probability of agreement.

[0871] The results are then sent to the terminal, allowing the sales representative to take appropriate action, such as preparing a concrete proposal as the next step.In this way, the system provides data-driven support for sales activities and an effective means of increasing the success rate.

[0872] The processing flow will be explained below.

[0873] Step 1: Data collection

[0874] Terminal: The sales representative inputs minutes of the sales meeting. Either text input or voice recognition input can be selected as the input method. For example, if a customer comments, "This solution is very interesting," the content of that comment is recorded.

[0875] Step 2: Send data

[0876] Terminal: The entered minutes of the business meeting are sent to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[0877] Step 3: Save data

[0878] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[0879] Step 4: Text Analysis

[0880] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[0881] Step 5: Classification of successes and failures

[0882] Server: Based on past sales negotiation data, the generative AI classifies successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the word "interest" appears frequently in successful deals.

[0883] Step 6: Extracting Keyphrases

[0884] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores are used for extraction. For example, key phrases such as "cost performance" and "interesting" are extracted.

[0885] Step 7: Predicting agreement probability

[0886] User: A sales representative opens a new business deal and enters the details of the deal into the terminal as minutes.

[0887] Server: Checks key phrases against the input minutes, and the generation AI predicts the degree of agreement. For example, if the phrase "high cost performance" appears, it predicts an 80% degree of agreement.

[0888] Step 8: Notification of prediction results

[0889] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[0890] By using the above processing steps, the present invention provides a system for scientifically predicting the probability of success of a business negotiation and improving the efficiency of sales activities.

[0891] Example 1

[0892] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0893] In conventional sales activities, the collection, analysis, and closing probability prediction of sales meeting minutes are often done manually, which takes time and effort and reduces efficiency. Another issue is that data management is cumbersome, making it difficult to effectively utilize trends in successful deals. The present invention aims to solve these issues by automatically analyzing minutes and improving the accuracy of closing probability prediction.

[0894] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0895] In this invention, the server provides an input device for recording the details of business negotiations, and includes: means for a user to input minutes by text input or voice recognition; means for format-checking the input minutes and sending them; means for saving the minutes in a database; means for analyzing the minutes as text data using natural language processing with artificial intelligence; means for classifying successful and unsuccessful cases based on past business negotiation data; means for extracting important expressions contained in successful cases; means for predicting the likelihood of agreement in future business negotiations based on the extracted important expressions; and means for notifying the user of the predicted likelihood of agreement. This enables efficient collection and analysis of business negotiation minutes and highly accurate prediction of the likelihood of closing a deal.

[0896] "Business negotiation content" refers to information such as proposals, requests, and exchanges of opinions exchanged between customers and sales representatives during sales activities.

[0897] An "input device" is a device or software that a user uses to input information. Examples include a keyboard, microphone, and touch panel.

[0898] "Text input" refers to a method in which a user inputs characters or sentences using a keyboard or the like.

[0899] "Speech recognition" is a technology that inputs the words spoken by a user as voice and converts them into text information.

[0900] Minutes are documents that record the content exchanged, decisions made, and key points made during business negotiations.

[0901] "Format check" refers to the process of verifying whether the input data is in the correct format.

[0902] A "database" is a system for systematically storing large amounts of data and quickly searching and retrieving them.

[0903] "Generative AI" refers to an AI technology that analyzes data and performs pattern recognition and predictions.

[0904] "Natural language processing" refers to techniques that enable artificial intelligence to understand and analyze human language, including sentence tokenization, part-of-speech tagging, and semantic analysis.

[0905] "Past business negotiation data" refers to records and information related to past business negotiations, and is stored in a database.

[0906] A "successful case" is one in which negotiations proceed smoothly and an agreement or contract is reached with the customer.

[0907] A "failed project" is one in which negotiations did not progress and no agreement or contract was reached with the customer.

[0908] "Important expressions" are key phrases or noun phrases that are often found in successful deals and have a significant impact on the outcome of subsequent sales negotiations.

[0909] "Agreement probability" is an indicator that indicates the probability that a particular business deal will be successful.

[0910] "Notification" refers to informing a user of information or results, and is often done in real time via a terminal.

[0911] This invention is a system for analyzing sales negotiation minutes in sales activities using a generative AI model and predicting the probability of closing a deal. Specific embodiments of this system are described in detail below.

[0912] Entering business meeting minutes

[0913] Terminal: After a sales meeting, the sales representative uses the terminal to enter details of the sales meeting. Two methods of input are available: text input and voice recognition. For example, a sales representative can enter something like "I'm very interested in this proposal" as text or use voice recognition.

[0914] Data transmission and storage

[0915] Terminal: The entered minutes are sent to the server in real time, where a format check is performed to ensure the data is in the correct format.

[0916] Server: The received minutes of business meetings are stored in a database. Each minutes of business meetings is assigned a unique identifier and managed centrally in the database.

[0917] Text analytics

[0918] Server: Natural language processing is performed on the business meeting minutes stored in the database using a generative AI model. Specific analysis methods include the following processes:

[0919] 1. Tokenization: Splitting text into words and phrases.

[0920] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[0921] 3. Noun phrase extraction: Extract important noun phrases from the sentence. For example, noun phrases such as "interest" and "suggestion" are extracted.

[0922] Classification of successful and unsuccessful cases

[0923] Server: Classifies successful and unsuccessful deals based on past sales negotiation data. A generative AI model is used for this classification. The generative AI works as follows:

[0924] 1. Prepare a dataset: Use past sales data to prepare a dataset that includes successful and unsuccessful deals.

[0925] 2. Feature extraction: The generative AI model extracts commonalities between successful and unsuccessful deals from past sales negotiation data. For example, it confirms that the phrase "interested" appears frequently in successful deals.

[0926] Keyphrase Extraction

[0927] Server: Extracts frequently occurring key phrases from the meeting minutes of successful deals. This extraction uses TF-IDF scores. For example, phrases such as "cost performance" and "interested" are identified as important.

[0928] Agreement probability prediction

[0929] User: The sales representative enters new sales meeting minutes into the terminal.

[0930] server:

[0931] 1. Keyphrase check: Check whether newly entered meeting notes contain identified keyphrases.

[0932] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[0933] Terminal: Notifies the salesperson in real time of the predicted probability of closing sent from the server. For example, the notification may include a message such as "This opportunity has an 80% chance of closing."

[0934] Specific examples

[0935] For example, consider the case where a sales representative has an initial business meeting with a new customer and writes in the minutes, "They are very interested in proposing a new solution." This data is sent from the device to the server and analyzed by the generative AI model. The AI ​​extracts the key phrase "interested" and compares it with past success stories. As a result of this comparison, it predicts an 80% probability of closing the deal, and this result is notified to the device. The sales representative can then take action as the next step, such as preparing a concrete proposal.

[0936] Example prompts for generative AI models

[0937] "Analyze the following sales meeting transcript and predict the likelihood of closing the deal. Transcript: 'I'm very interested in your new solution proposal.'"

[0938] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0939] Step 1:

[0940] Terminal: After a sales meeting, the sales representative uses the terminal to enter minutes of the meeting. Specifically, the sales representative records the details of the meeting using text input or voice recognition. An example of input is, "I am very interested in this proposal." Once input is complete, the data is temporarily stored on the terminal.

[0941] Step 2:

[0942] Terminal: The entered minutes of the business meeting are sent to the server in real time. A format check is performed, and the data is sent in text format. For example, the format check verifies whether there are any spelling mistakes or inappropriate characters. If the format is determined to be correct, the data is sent to the server.

[0943] Step 3:

[0944] Server: The server stores the received minutes in a database. When stored, each minutes is assigned a unique identifier. The database also stores the minutes' contents along with a timestamp and the user information that entered them.

[0945] Step 4:

[0946] Server: The minutes of business meetings stored in the database are processed using natural language processing with a generative AI model. Specific analysis methods include the following processes:

[0947] 1. Tokenization: Split the text of the meeting notes into words.

[0948] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[0949] 3. Noun phrase extraction: Extract important noun phrases (e.g., “interest,” “suggestion”).

[0950] The input is the text data of the business meeting minutes, and the output is the parsed tokens, tags, and noun phrases.

[0951] Step 5:

[0952] Server: Classifies successful and unsuccessful deals based on past sales data. The generative AI model works as follows:

[0953] 1. Prepare the dataset: Divide past sales negotiation data into successful and unsuccessful cases.

[0954] 2. Feature extraction: Extract common features from successful and unsuccessful cases. Confirm that the phrase "interested" is frequently included in successful cases.

[0955] The input is past sales negotiation data, and the output is the features of successful and unsuccessful cases.

[0956] Step 6:

[0957] Server: Extracts frequent key phrases from successful business meeting minutes using a generative AI model. Specifically, it identifies important phrases such as "cost-effectiveness" and "interested" using TF-IDF scores.

[0958] The input is the sales negotiation data of successful cases, and the output is important key phrases.

[0959] Step 7:

[0960] User: The salesperson enters new meeting minutes into the device. The user again uses text input or voice recognition.

[0961] Step 8:

[0962] server:

[0963] 1. Keyphrase check: Check whether newly entered meeting minutes contain the identified keyphrases.

[0964] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[0965] The input is the newly entered minutes of the business meeting, and the output is the presence or absence of key phrases and the probability of closing the deal.

[0966] Step 9:

[0967] Terminal: The server notifies the salesperson in real time with the predicted close probability results, including a specific message such as "This opportunity has an 80% chance of closing."

[0968] (Application example 1)

[0969] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0970] In conventional sales activities, analysis of sales meeting minutes and prediction of the probability of closing a deal are often performed manually, resulting in inefficiency and inaccuracy. Furthermore, when audio recordings are performed, the process of converting the audio data into text is complicated, making real-time analysis and prediction difficult. The present invention aims to solve these problems and improve the efficiency and accuracy of customer service and sales activities in brick-and-mortar stores.

[0971] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0972] In this invention, the server includes a means for storing the minutes of business negotiations in a database, a means for analyzing the minutes as text data by natural language processing using artificial intelligence, and a means for converting speech to text, which enables real-time text analysis and the identification of key phrases with a high probability of closing a deal.

[0973]

[0974] "Business minutes" refers to information recorded by sales representatives or store clerks about business negotiations and customer service with customers.

[0975] A "database" is a system for storing data in an organized manner and for efficiently searching, extracting, and managing data.

[0976] "Generative AI" is a type of artificial intelligence that performs natural language processing, and is a technology responsible for analyzing and generating text data.

[0977] "Natural language processing" is a technology that allows computers to understand and analyze language spoken by humans.

[0978] "Text data" refers to data that stores character strings or sentences in digital format.

[0979] "Past business negotiation data" is information about business negotiations that have been conducted in the past, including records of successful and unsuccessful cases.

[0980] A "successful deal" refers to a case where a deal has been concluded.

[0981] A "failed deal" refers to a deal that did not result in a successful transaction.

[0982] A "keyphrase" is a word or phrase that has significant meaning in a particular document or conversation.

[0983] "Agreement probability" refers to the probability that a deal will be concluded.

[0984] "Notifying the user" means sending the analysis results and prediction results to the user's terminal and informing them.

[0985] "Speech to text" is the process of converting recorded audio data into a string or sentence format.

[0986] A "user interface that controls voice input" refers to software that controls the operation screen and voice recognition that allows users to input voice.

[0987]

[0988] The present invention is a system that records the details of business negotiations and customer service, analyzes the details, and predicts the likelihood of a deal. In particular, as an example of application in a physical store, it provides a means for recording conversations between store clerks and customers and predicting the likelihood of a deal in real time based on the data.

[0989] System Configuration

[0990] Hardware Configuration

[0991] The system of the present invention uses the following hardware:

[0992] 1. Terminal: A device capable of voice input and output, such as smart glasses or a head-mounted display.

[0993] 2. Server: A high-performance computer to run the database and generative AI.

[0994] Software Configuration

[0995] The system operates using the following software:

[0996] 1. Speech recognition software: Use the speech_recognition library to convert voice data into text data.

[0997] 2. Generative AI software: Hugging Face's Transformer model is used to analyze the text data.

[0998] 3. Text analysis software: Identify key phrases using TF-IDF and predict conversion probability using a logistic regression model.

[0999] Data processing and calculation

[1000] 1. Voice input:

[1001] The terminal records the conversation between the store clerk and the customer and transmits the audio data to the server.

[1002] 2. Audio conversion:

[1003] Speech recognition software running on the server converts the recorded voice data into text data.

[1004] 3. Text Analysis:

[1005] A generative AI model is used to perform natural language processing on the converted text data and extract key phrases, using the Hugging Face Transformer model.

[1006] 4. Data storage:

[1007] The text data and extracted key phrases are stored in a database.

[1008] 5. Win probability prediction:

[1009] Using a logistic regression model built on data from past successful and unsuccessful cases, the probability of closing a deal is predicted from the extracted key phrases.

[1010] 6. Notification of Results:

[1011] The prediction results are sent to the terminal in real time and presented to the store clerk.

[1012] Specific examples

[1013] For example, a salesperson can record the phrase "I'm very interested in this product" while serving a customer, and the recording is converted into text. This text is sent to the system, where generative AI and TF-IDF analysis are used to extract the key phrase "interested." This key phrase is pattern-matched with past successful cases, and the system predicts a high probability of success. For example, a result such as "85% probability of success" is displayed in real time on the smart glasses.

[1014] Prompt Sentence Examples

[1015] Customer conversation recording text: "I'm very interested in this product." Analyze the following text, extract key phrases, and predict the probability of closing a sale.

[1016] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1017]

[1018] Step 1:

[1019] A user records a conversation with a customer using smart glasses or a head-mounted display. The input is voice data, and the output is the recorded voice data. This voice data is stored on the device.

[1020] Step 2:

[1021] The device sends voice data to the server. The server receives the voice data and converts it into text data using speech recognition software (speech_recognition library). The input is voice data and the output is text data.

[1022] Step 3:

[1023] The server then performs natural language processing on the converted text data using a generative AI model (Hugging Face's Transformer model). This process involves tokenizing the input text data, tagging it with parts of speech, and extracting noun phrases. The input is text data, and the output is analyzed text data.

[1024] Step 4:

[1025] The server extracts important key phrases from the analyzed text data and identifies frequently occurring key phrases using TF-IDF scores. The input of this process is the analyzed text data, and the output is the extracted key phrases.

[1026] Step 5:

[1027] The server uses a logistic regression model based on data on past successful and unsuccessful cases to predict the probability of success from newly extracted key phrases. The input to this process is the extracted key phrases, and the output is the predicted probability of success.

[1028] Step 6:

[1029] The server sends the predicted probability of closing to the terminal. The terminal notifies the user (store clerk or salesperson) of the received predicted result of the agreement probability in real time. The input of this process is the predicted value of the probability of closing, and the output is a notification to the user.

[1030] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1031] This invention is a system for predicting the likelihood of a deal by analyzing sales meeting minutes using a generation AI and an emotion engine. Specifically, it automates the process of collecting, analyzing, and recognizing emotions in sales meeting minutes, as well as predicting the likelihood of an agreement.

[1032] Collection of business meeting minutes

[1033] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[1034] Data transmission and storage

[1035] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[1036] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[1037] Text analysis with generative artificial intelligence

[1038] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[1039] Classification of successful and unsuccessful cases

[1040] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[1041] Keyphrase Extraction

[1042] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[1043] Emotion recognition by emotion engine

[1044] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[1045] Server: The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[1046] Agreement probability prediction

[1047] User: The sales representative enters new minutes of the business meeting into the terminal.

[1048] Server: The generative AI predicts the likelihood of agreement for the input minutes based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[1049] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[1050] Specific examples

[1051] For example, consider the case where a salesperson conducts an initial business meeting with a new customer and writes in the minutes that they are "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts a 90% probability of agreement.

[1052] The results are then sent to the device, and the salesperson can use this information to take appropriate action, such as providing further proposals or specific support.In this way, this system uses a data-driven approach combined with emotion recognition to more accurately predict the likelihood of success of a sales negotiation, providing an effective means of improving the efficiency of sales activities.

[1053] The processing flow will be explained below.

[1054] Step 1: Data collection

[1055] User: After a sales meeting, a sales representative uses the device to enter minutes. Two input methods are available: text input and voice recognition, allowing the sales representative to efficiently record the details of the meeting. For example, a comment such as "I'm very interested in this proposal" can be entered.

[1056] Step 2: Send data

[1057] Terminal: The entered minutes are sent from the terminal to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[1058] Step 3: Save data

[1059] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[1060] Step 4: Text Analysis

[1061] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[1062] Step 5: Classification of successes and failures

[1063] Server: Based on past sales negotiation data, generative AI is used to classify successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the phrase "interested" appears frequently in successful deals.

[1064] Step 6: Extracting Keyphrases

[1065] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores and correlation analysis are used for extraction. For example, key phrases such as "cost performance" and "interested" are extracted.

[1066] Step 7: Emotion Recognition

[1067] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes the tone of voice, facial expressions, and text content to identify the customer's emotions. For example, if the customer is expecting something, a positive emotion such as "expecting" will be detected.

[1068] Step 8: Storing Emotion Data

[1069] Server: Records the detected emotion data in a database and uses it for subsequent sales negotiation analysis.

[1070] Step 9: Predicting agreement probability

[1071] User: A sales representative enters new sales meeting minutes into a terminal.

[1072] Server: For the minutes entered, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and the emotional data detected by the emotion engine. For example, if the phrase "cost performance" and the emotional data "expectations" are included, the likelihood of agreement is predicted to be 90%.

[1073] Step 10: Notification of prediction results

[1074] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[1075] Through the above processing steps, the present invention scientifically predicts the likelihood of success in business negotiations and achieves improved efficiency in sales activities by combining emotion recognition.

[1076] Example 2

[1077] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1078] Conventional sales negotiation support systems have difficulty taking into account customer emotional information when analyzing sales negotiation minutes and predicting the probability of success. As a result, there was a problem of reduced accuracy in predicting the probability of success. In addition, inputting sales negotiation minutes and sending and saving the data was time-consuming, making efficient operation difficult.

[1079] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1080] In this invention, the server includes means for storing the minutes of business negotiations in a database, means for analyzing the minutes of business negotiations as text data using natural language processing with artificial intelligence, means for classifying successful cases and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for analyzing emotional data during business negotiations using an emotion engine and predicting the likelihood of agreement with higher accuracy based on this, and means for notifying the user of the predicted likelihood of agreement. This enables more accurate prediction of business negotiation results including customer emotional data, and also improves the efficiency of inputting business negotiation minutes and transmitting and saving data.

[1081] A "business meeting minutes" is a document that records what was discussed and what decisions were made during a business meeting.

[1082] A "database" is a system for efficiently managing, storing, searching, and updating large amounts of data.

[1083] "Generative AI" is a technology that uses machine learning algorithms and natural language processing techniques to generate and analyze text data.

[1084] "Natural language processing" is a technology that enables computers to understand, interpret, and generate human language.

[1085] A "key phrase" is a word or phrase that has a particularly important meaning in text data.

[1086] "Agreement probability" is a probability that indicates the likelihood that a proposal will be accepted in a business negotiation.

[1087] "User" refers to an individual or organization that uses this system to input minutes of business negotiations and receive the analysis results.

[1088] An "emotion engine" is a technology that analyzes voice tone, facial expressions, and text content to identify human emotions.

[1089] "Format check" is the process of checking whether the entered data conforms to the specified format.

[1090] "Real-time" is a concept that indicates that data is generated and processed immediately without delay.

[1091] This invention is a system for analyzing sales meeting minutes in sales activities using a generative AI model and an emotion engine to predict the likelihood of a deal. The system automates the processes of collecting, analyzing, and recognizing emotions in sales meeting minutes, and predicting the likelihood of agreement. Specific embodiments for implementing this system are described below.

[1092] Collection of business meeting minutes

[1093] Terminal: The user (salesperson) uses a user interface to input minutes after a business meeting. There are two input methods: text input and voice recognition, which allows for efficient recording. The software used is a voice recognition engine (for example, Google Speech-to-Text). Specifically, the user inputs a comment such as "I'm very interested in this proposal."

[1094] Data transmission and storage

[1095] Terminal: Once the minutes are entered, the data is sent to the server after format checks are performed. The format checks described above are a means to ensure accurate and consistent data entry.

[1096] Server: The server analyzes the received data and stores it in a database. This database is a system that allows for the management and efficient searching of large amounts of data. For example, a common database management system such as MySQL or PostgreSQL is used.

[1097] Text analysis with generative artificial intelligence

[1098] Server: Generates minutes stored in a database and performs natural language processing (NLP) using an AI model. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. Software used includes NLP libraries such as NLTK and SpaCy. For example, the predicate "interesting" is extracted.

[1099] Classification of successful and unsuccessful cases

[1100] Server: Using past sales negotiation data, a generative AI model extracts commonalities between successful and unsuccessful deals. Software used includes machine learning libraries such as scikit-learn and TensorFlow. It is recognized that successful deals frequently include the phrase "interested."

[1101] Keyphrase Extraction

[1102] Server: Extracts frequently occurring key phrases from successful business meeting minutes. Specifically, it calculates TF-IDF scores and identifies weighted key phrases. For example, key phrases such as "cost performance" and "interested" are extracted.

[1103] Emotion recognition by emotion engine

[1104] Terminal: The conversation between the user and the customer during the sales negotiation is analyzed using an emotion engine. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. Software used may include a voice recognition engine (Google Speech-to-Text) or a facial recognition tool (OpenCV). For example, if the customer is excited, it will be identified as "expecting."

[1105] Server: Emotion data is recorded in a database and used for analysis. The collected emotional data is an important indicator for predicting the success rate of business negotiations.

[1106] Agreement probability prediction

[1107] User: After the business meeting, the user inputs new minutes into the terminal.

[1108] Server: Based on the input minutes, the generative AI model predicts the likelihood of agreement based on identified key phrases and emotional data. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the model predicts the likelihood of agreement to be 90%.

[1109] Terminal: Prediction results are notified to salespeople in real time, allowing them to plan their next actions based on this information.

[1110] Examples of concrete examples and prompts

[1111] For example, if a salesperson conducts an initial sales meeting with a new customer and notes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting, this data is sent to the server and analyzed there. The generative AI model predicts an agreement probability of 90% based on the key phrase "interested" and the emotion data "expectation." This result is notified to the device, and the salesperson can use it to consider further proposals and specific support.

[1112] Prompt Sentence Examples

[1113] You wrote in the minutes, "I'm very interested in your new solution proposal." During the negotiation, the emotion engine detected the customer's emotion, such as "I'm looking forward to it." Use this data to predict the likelihood of an agreement.

[1114] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1115] Step 1: Enter meeting notes

[1116] Terminal: After a sales meeting, the user (salesperson) uses the terminal's user interface to input minutes. There are two input formats: text input and voice recognition, and the user can choose either. Input is mainly done using a keyboard or microphone.

[1117] Input: Text or audio data about the business.

[1118] Output: Minutes data that has passed format check.

[1119] What it does: The user enters text into an input form or records a voice, which is then converted into text using a speech recognition engine (e.g., Google Speech-to-Text).

[1120] Step 2: Send and store data

[1121] Terminal: After the minutes are entered, the terminal checks the format and sends the data to the server. Only data that passes the format check is sent.

[1122] Input: Format-checked minutes data.

[1123] Output: The minutes data sent to the server.

[1124] Specific behavior: The device sends data to the server using an HTTP request, including appropriate format conversion if necessary.

[1125] Server: The server analyzes the minutes data received from the terminal and stores it in a database.

[1126] Input: Received minutes data.

[1127] Output: Meeting minutes data stored in a database.

[1128] Specific operation: The server executes an insert operation in the database management system (e.g., MySQL) to permanently store the data.

[1129] Step 3: Text analysis

[1130] Server: The server uses a generative AI model to perform natural language processing (NLP) on the stored minutes data, specifically splitting tokens, tagging parts of speech, and extracting noun phrases.

[1131] Input: Meeting minutes data stored in the database.

[1132] Output: Text data as the analysis result.

[1133] What it does: The server uses an NLP library (e.g., NLTK, SpaCy) to segment the text, tag parts of speech, and extract noun phrases, such as the phrase "interesting."

[1134] Step 4: Classification of successes and failures

[1135] Server: Extracts commonalities between successful and unsuccessful deals based on past sales negotiation data. Classifies meeting minutes data using a generative AI model.

[1136] Inputs: Past and current deal data.

[1137] Output: Classification results of successful and unsuccessful cases.

[1138] What it does: The server uses machine learning algorithms (e.g., scikit-learn, TensorFlow) to learn patterns from past datasets and classify new data, identifying common key phrases and success factors and generating classification results.

[1139] Step 5: Keyphrase Extraction

[1140] Server: Calculate TF-IDF scores to identify frequently occurring key phrases from successful meeting notes, and extract important key phrases.

[1141] Input: Successful deal opportunity data.

[1142] Output: Extracted key phrases.

[1143] Specific operation: The server uses the TF-IDF algorithm to calculate important phrases in the sales negotiation data and extracts the top-ranked key phrases. For example, the phrase "cost performance" is identified as important.

[1144] Step 6: Emotion Recognition

[1145] Terminal: During sales negotiations, conversations between users and customers are collected and analyzed using an emotion engine. Voice tone, facial expressions, and text content are used to identify customer emotions.

[1146] Input: Voice, facial expression, and text data during business negotiations.

[1147] Output: Parsed emotion data.

[1148] What it does: The device uses an emotion engine (e.g., a voice recognition engine or facial recognition tool) to analyze the data and identify the customer's emotions. Positive emotions such as "expecting" are detected.

[1149] Server: The analyzed emotion data is recorded in a database and used for subsequent analysis.

[1150] Input: Parsed emotion data.

[1151] Output: Emotion data stored in a database.

[1152] Specific operation: The server inserts the emotion data into a database, making it available for future analysis.

[1153] Step 7: Predicting agreement probability

[1154] Server: Based on key phrases and sentiment data, a generative AI model is used to predict the likelihood of agreement. It uses a model learned from past successful cases.

[1155] Input: New sales meeting minutes data, identified key phrases, analyzed sentiment data.

[1156] Output: Consensus prediction results.

[1157] Specific operation: The server uses the generative AI model to analyze the input data and calculate the agreement probability. For example, if the comment "good cost performance" and the emotional data "expected" are included, the agreement probability is predicted to be 90%.

[1158] Step 8: Notification

[1159] Terminal: The terminal notifies the user in real time of the predicted consensus probability sent from the server. The user can plan their next action based on this information.

[1160] Input: Prediction results sent from the server.

[1161] Output: The prediction results displayed on the device screen.

[1162] Specific operation: The device uses a notification system to notify the user of the prediction results in real time, allowing the user to plan their next steps.

[1163] (Application example 2)

[1164] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1165] During sales negotiations and customer service in brick-and-mortar stores, it is difficult to accurately predict the customer's level of interest in a product or the likelihood of a sale. Conventional sales negotiation minutes and customer service records rely on subjective judgment, making it difficult to formulate effective proposals and sales strategies. Furthermore, it is difficult to accurately grasp the customer's emotions and make appropriate product proposals based on them.

[1166] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for saving business negotiation minutes in a database, means for analyzing the business negotiation minutes as text data using natural language processing with generative artificial intelligence, means for classifying successful and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in the successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for notifying the user of the predicted likelihood of agreement, means for analyzing customer emotion data using an emotion engine, means for predicting customer satisfaction based on the analyzed emotion data, and means for proposing appropriate products to customers based on the key phrases extracted from customer emotions and the successful cases. This enables product proposals that increase the likelihood of closing business negotiations and customer service and increase customer satisfaction based on the customer's emotions and interests.

[1167] A "business negotiation minutes" is a document that records the details of a business negotiation that a sales representative conducted with a customer.

[1168] A "database" is a system designed to store information in an organized manner and make it easy to access and search.

[1169] "Generative AI" is a system that uses artificial intelligence technology to generate and analyze text and data.

[1170] "Natural language processing" is a technology that allows computers to understand and process human language.

[1171] "Past business negotiation data" refers to data that includes records and results of previous business negotiations.

[1172] A "successful case" refers to a case where a business deal has been concluded.

[1173] A "failed deal" refers to a business deal that did not result in a successful transaction.

[1174] "Key phrases" are particularly important words or phrases in business negotiations or customer comments.

[1175] "Agreement probability" is an index that indicates the degree of possibility that a business negotiation will lead to a contract.

[1176] An "emotion engine" is a system that analyzes and identifies human emotions from voice and text.

[1177] "Emotion data" is information about the customer's emotional state analyzed by the emotion engine.

[1178] "Customer satisfaction" is an indicator of how satisfied customers are with a product or service.

[1179] "Product proposal" is the act of recommending appropriate products or services to customers.

[1180] This invention relates to a system for efficiently handling customer inquiries and conducting business negotiations in brick-and-mortar stores. This system collects business negotiation minutes, analyzes them using generative AI and an emotion engine, and predicts the probability of closing a deal and customer satisfaction. The components and procedures required to implement this invention are described below.

[1181] Collection of business meeting minutes

[1182] Device:

[1183] After the sales meeting, the sales representative will use a smartphone or tablet to input minutes of the meeting. Two input methods are available: text input and voice recognition, allowing sales representatives to efficiently record the details of the meeting.

[1184] Data transmission and storage

[1185] Device:

[1186] The entered minutes of the business meeting are sent from the terminal to the server, where they are checked for format and saved in the appropriate format.

[1187] server:

[1188] The server stores the received minutes of the business negotiations in a database, which allows for centralized management of data for later analysis.

[1189] Text analysis with generative artificial intelligence

[1190] server:

[1191] The minutes stored in the database are then processed using generative AI (e.g., Hugging Face Transformers) for natural language processing. Specifically, tokenization, part-of-speech tagging, and noun phrase extraction are performed. For example, the predicate "interesting" is extracted.

[1192] Classification of successful and unsuccessful cases

[1193] server:

[1194] Based on past sales negotiation data, successful and unsuccessful deals are classified. Generative AI is used to extract commonalities between successful and unsuccessful deals. For example, it is confirmed that the phrase "interested" is frequently included in successful deals.

[1195] Keyphrase Extraction

[1196] server:

[1197] Extract frequently occurring key phrases from successful business meeting minutes. For example, key phrases such as "cost performance" and "interested" are identified as important using TF-IDF scores.

[1198] Emotion recognition by emotion engine

[1199] Device:

[1200] During a sales meeting, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[1201] server:

[1202] The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[1203] Agreement probability prediction

[1204] User:

[1205] The sales representative enters new minutes of the business meeting into the terminal.

[1206] server:

[1207] For input minutes, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[1208] Device:

[1209] The server sends the predicted results of the agreement probability to the sales representative in real time. Based on this information, the sales representative can plan the next action. It also enables the sales representative to propose appropriate products based on the customer's emotions and interests.

[1210] Examples of concrete examples and prompts

[1211] For example, consider the case where a salesperson conducts an initial sales meeting with a new customer and writes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts an agreement probability of 90%. This result is notified to the device, and the salesperson can use this information to take appropriate action, such as making further proposals or specific support.

[1212] Example prompt sentence:

[1213] Convert the conversation between the customer and the store clerk into text and use generative AI to analyze the following:

[1214] 1. Extracting important key phrases

[1215] 2. Recognizing customer sentiment

[1216] 3. Closing probability prediction

[1217] In this way, the system uses a data-driven approach that combines customer sentiment to more accurately predict the likelihood of a deal's success and provide an effective means of improving the efficiency of sales activities.

[1218] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1219] Step 1: Collect meeting notes

[1220] Device:

[1221] Salespeople, who are users, use smartphones or tablets to input minutes of sales negotiations after a sales meeting. There are two input methods: text input and voice recognition. With voice recognition, the user's voice is analyzed and the contents of the sales negotiation are converted into text. These input minutes of sales negotiations become the input data for the next processing step.

[1222] Step 2: Send and store data

[1223] Device:

[1224] Once the input of the business meeting minutes is complete, the terminal sends the minutes to the server. The sent data undergoes a format check and is saved in the appropriate format on the server. The format check verifies the data format and required fields. If successful, the data is saved in the server's database (output).

[1225] Step 3: Text analysis with generative AI

[1226] server:

[1227] Using saved business meeting minutes as input, natural language processing is performed using generative AI (e.g., Hugging Face Transformers). Tokenization, part-of-speech tagging, and noun phrase extraction are performed, and the analysis results are output as text data. For example, the predicate "interesting" is extracted.

[1228] Step 4: Classification of successes and failures

[1229] server:

[1230] By referencing past sales negotiation data, the system uses generative AI to extract commonalities between successful and unsuccessful deals. This allows the minutes of sales negotiations to be classified as either successful or unsuccessful. The classification results are output as labeled data indicating whether the deal was successful or unsuccessful.

[1231] Step 5: Extracting Keyphrases

[1232] server:

[1233] The input is a successful business meeting transcript, and frequent keyphrases are extracted. This process uses TF-IDF scores to identify important keyphrases. Keyphrases such as "cost-effectiveness" and "interested" are output.

[1234] Step 6: Emotion Recognition with the Emotion Engine

[1235] Device:

[1236] During a sales negotiation, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The inputs are voice tone, facial expressions, and text content. The analyzed emotion data is output as an emotion label, such as "expecting."

[1237] server:

[1238] The detected emotion data is recorded in a database, which helps predict the probability of success.

[1239] Step 7: Predicting agreement probability

[1240] server:

[1241] Based on newly input minutes of business meetings, the generation AI predicts the likelihood of agreement. Inputs include the text data of the minutes, pre-specified key phrases, and emotional data. The generation AI analyzes each input data and outputs the likelihood of agreement as a numerical value (percentage). For example, if the input data includes the comment "high cost performance" and the emotional data "we are looking forward to it," the generation AI predicts the likelihood of agreement to be 90%.

[1242] Step 8: Notification of results

[1243] Device:

[1244] The server then sends the predicted agreement probability to the device. The salesperson receives this information in real time and plans the next steps. For example, they can create a specific sales strategy, such as considering additional proposals.

[1245] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1246] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1247] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1248] [Fourth embodiment]

[1249] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1250] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1251] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1252] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1253] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1254] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1255] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1256] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1257] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1258] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1259] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1260] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1261] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1262] This invention is a system that uses AI to analyze sales meeting minutes in sales activities and predict the likelihood of a deal being concluded. Specifically, it automates the process of collecting and analyzing sales meeting minutes and predicting the likelihood of an agreement.

[1263] Collection of business meeting minutes

[1264] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[1265] Data transmission and storage

[1266] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[1267] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[1268] Text analysis with generative artificial intelligence

[1269] Server: Natural language processing is performed on the business meeting minutes stored in the database using generation AI. Specific analysis methods include tokenization, part-of-speech tagging, and noun phrase extraction. For example, noun phrases such as "interest" and "proposal" are extracted.

[1270] Classification of successful and unsuccessful cases

[1271] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[1272] Keyphrase Extraction

[1273] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[1274] Agreement probability prediction

[1275] User: The sales representative enters new minutes of the business meeting into the terminal.

[1276] Server: Checks whether newly entered minutes of business meetings contain pre-specified key phrases. If they do, predicts the likelihood of agreement based on the importance of those key phrases. For example, if the comment "high cost performance" is included, predicts the likelihood of agreement to be 80%.

[1277] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[1278] Specific examples

[1279] For example, consider the case where a sales representative holds an initial business meeting with a new client and writes in the minutes, "I am very interested in your proposal for a new solution." These minutes are sent from the device to the server, where they are analyzed. The generation AI extracts the key phrase "interested" and compares it with a database of past success stories. As a result, it confirms that this key phrase appears frequently in successful cases, and predicts an 80% probability of agreement.

[1280] The results are then sent to the terminal, allowing the sales representative to take appropriate action, such as preparing a concrete proposal as the next step.In this way, the system provides data-driven support for sales activities and an effective means of increasing the success rate.

[1281] The processing flow will be explained below.

[1282] Step 1: Data collection

[1283] Terminal: The sales representative inputs minutes of the sales meeting. Either text input or voice recognition input can be selected as the input method. For example, if a customer comments, "This solution is very interesting," the content of that comment is recorded.

[1284] Step 2: Send data

[1285] Terminal: The entered minutes of the business meeting are sent to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[1286] Step 3: Save data

[1287] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[1288] Step 4: Text Analysis

[1289] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[1290] Step 5: Classification of successes and failures

[1291] Server: Based on past sales negotiation data, the generative AI classifies successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the word "interest" appears frequently in successful deals.

[1292] Step 6: Extracting Keyphrases

[1293] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores are used for extraction. For example, key phrases such as "cost performance" and "interesting" are extracted.

[1294] Step 7: Predicting agreement probability

[1295] User: A sales representative opens a new business deal and enters the details of the deal into the terminal as minutes.

[1296] Server: Checks key phrases against the input minutes, and the generation AI predicts the degree of agreement. For example, if the phrase "high cost performance" appears, it predicts an 80% degree of agreement.

[1297] Step 8: Notification of prediction results

[1298] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[1299] By using the above processing steps, the present invention provides a system for scientifically predicting the probability of success of a business negotiation and improving the efficiency of sales activities.

[1300] Example 1

[1301] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1302] In conventional sales activities, the collection, analysis, and closing probability prediction of sales meeting minutes are often done manually, which takes time and effort and reduces efficiency. Another issue is that data management is cumbersome, making it difficult to effectively utilize trends in successful deals. The present invention aims to solve these issues by automatically analyzing minutes and improving the accuracy of closing probability prediction.

[1303] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1304] In this invention, the server provides an input device for recording the details of business negotiations, and includes: means for a user to input minutes by text input or voice recognition; means for format-checking the input minutes and sending them; means for saving the minutes in a database; means for analyzing the minutes as text data using natural language processing with artificial intelligence; means for classifying successful and unsuccessful cases based on past business negotiation data; means for extracting important expressions contained in successful cases; means for predicting the likelihood of agreement in future business negotiations based on the extracted important expressions; and means for notifying the user of the predicted likelihood of agreement. This enables efficient collection and analysis of business negotiation minutes and highly accurate prediction of the likelihood of closing a deal.

[1305] "Business negotiation content" refers to information such as proposals, requests, and exchanges of opinions exchanged between customers and sales representatives during sales activities.

[1306] An "input device" is a device or software that a user uses to input information. Examples include a keyboard, microphone, and touch panel.

[1307] "Text input" refers to a method in which a user inputs characters or sentences using a keyboard or the like.

[1308] "Speech recognition" is a technology that inputs the words spoken by a user as voice and converts them into text information.

[1309] Minutes are documents that record the content exchanged, decisions made, and key points made during business negotiations.

[1310] "Format check" refers to the process of verifying whether the input data is in the correct format.

[1311] A "database" is a system for systematically storing large amounts of data and quickly searching and retrieving them.

[1312] "Generative AI" refers to an AI technology that analyzes data and performs pattern recognition and predictions.

[1313] "Natural language processing" refers to techniques that enable artificial intelligence to understand and analyze human language, including sentence tokenization, part-of-speech tagging, and semantic analysis.

[1314] "Past business negotiation data" refers to records and information related to past business negotiations, and is stored in a database.

[1315] A "successful case" is one in which negotiations proceed smoothly and an agreement or contract is reached with the customer.

[1316] A "failed project" is one in which negotiations did not progress and no agreement or contract was reached with the customer.

[1317] "Important expressions" are key phrases or noun phrases that are often found in successful deals and have a significant impact on the outcome of subsequent sales negotiations.

[1318] "Agreement probability" is an indicator that indicates the probability that a particular business deal will be successful.

[1319] "Notification" refers to informing a user of information or results, and is often done in real time via a terminal.

[1320] This invention is a system for analyzing sales negotiation minutes in sales activities using a generative AI model and predicting the probability of closing a deal. Specific embodiments of this system are described in detail below.

[1321] Entering business meeting minutes

[1322] Terminal: After a sales meeting, the sales representative uses the terminal to enter details of the sales meeting. Two methods of input are available: text input and voice recognition. For example, a sales representative can enter something like "I'm very interested in this proposal" as text or use voice recognition.

[1323] Data transmission and storage

[1324] Terminal: The entered minutes are sent to the server in real time, where a format check is performed to ensure the data is in the correct format.

[1325] Server: The received minutes of business meetings are stored in a database. Each minutes of business meetings is assigned a unique identifier and managed centrally in the database.

[1326] Text analytics

[1327] Server: Natural language processing is performed on the business meeting minutes stored in the database using a generative AI model. Specific analysis methods include the following processes:

[1328] 1. Tokenization: Splitting text into words and phrases.

[1329] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[1330] 3. Noun phrase extraction: Extract important noun phrases from the sentence. For example, noun phrases such as "interest" and "suggestion" are extracted.

[1331] Classification of successful and unsuccessful cases

[1332] Server: Classifies successful and unsuccessful deals based on past sales negotiation data. A generative AI model is used for this classification. The generative AI works as follows:

[1333] 1. Prepare a dataset: Use past sales data to prepare a dataset that includes successful and unsuccessful deals.

[1334] 2. Feature extraction: The generative AI model extracts commonalities between successful and unsuccessful deals from past sales negotiation data. For example, it confirms that the phrase "interested" appears frequently in successful deals.

[1335] Keyphrase Extraction

[1336] Server: Extracts frequently occurring key phrases from the meeting minutes of successful deals. This extraction uses TF-IDF scores. For example, phrases such as "cost performance" and "interested" are identified as important.

[1337] Agreement probability prediction

[1338] User: The sales representative enters new sales meeting minutes into the terminal.

[1339] server:

[1340] 1. Keyphrase check: Check whether newly entered meeting notes contain identified keyphrases.

[1341] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[1342] Terminal: Notifies the salesperson in real time of the predicted probability of closing sent from the server. For example, the notification may include a message such as "This opportunity has an 80% chance of closing."

[1343] Specific examples

[1344] For example, consider the case where a sales representative has an initial business meeting with a new customer and writes in the minutes, "They are very interested in proposing a new solution." This data is sent from the device to the server and analyzed by the generative AI model. The AI ​​extracts the key phrase "interested" and compares it with past success stories. As a result of this comparison, it predicts an 80% probability of closing the deal, and this result is notified to the device. The sales representative can then take action as the next step, such as preparing a concrete proposal.

[1345] Example prompts for generative AI models

[1346] "Analyze the following sales meeting transcript and predict the likelihood of closing the deal. Transcript: 'I'm very interested in your new solution proposal.'"

[1347] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1348] Step 1:

[1349] Terminal: After a sales meeting, the sales representative uses the terminal to enter minutes of the meeting. Specifically, the sales representative records the details of the meeting using text input or voice recognition. An example of input is, "I am very interested in this proposal." Once input is complete, the data is temporarily stored on the terminal.

[1350] Step 2:

[1351] Terminal: The entered minutes of the business meeting are sent to the server in real time. A format check is performed, and the data is sent in text format. For example, the format check verifies whether there are any spelling mistakes or inappropriate characters. If the format is determined to be correct, the data is sent to the server.

[1352] Step 3:

[1353] Server: The server stores the received minutes in a database. When stored, each minutes is assigned a unique identifier. The database also stores the minutes' contents along with a timestamp and the user information that entered them.

[1354] Step 4:

[1355] Server: The minutes of business meetings stored in the database are processed using natural language processing with a generative AI model. Specific analysis methods include the following processes:

[1356] 1. Tokenization: Split the text of the meeting notes into words.

[1357] 2. Part-of-speech tagging: tag each word with its part of speech (noun, verb, adjective, etc.).

[1358] 3. Noun phrase extraction: Extract important noun phrases (e.g., “interest,” “suggestion”).

[1359] The input is the text data of the business meeting minutes, and the output is the parsed tokens, tags, and noun phrases.

[1360] Step 5:

[1361] Server: Classifies successful and unsuccessful deals based on past sales data. The generative AI model works as follows:

[1362] 1. Prepare the dataset: Divide past sales negotiation data into successful and unsuccessful cases.

[1363] 2. Feature extraction: Extract common features from successful and unsuccessful cases. Confirm that the phrase "interested" is frequently included in successful cases.

[1364] The input is past sales negotiation data, and the output is the features of successful and unsuccessful cases.

[1365] Step 6:

[1366] Server: Extracts frequent key phrases from successful business meeting minutes using a generative AI model. Specifically, it identifies important phrases such as "cost-effectiveness" and "interested" using TF-IDF scores.

[1367] The input is the sales negotiation data of successful cases, and the output is important key phrases.

[1368] Step 7:

[1369] User: The salesperson enters new meeting minutes into the device. The user again uses text input or voice recognition.

[1370] Step 8:

[1371] server:

[1372] 1. Keyphrase check: Check whether newly entered meeting minutes contain the identified keyphrases.

[1373] 2. Agreement probability prediction: Predict the probability of closing based on the importance of the key phrase. For example, if the comment "high cost performance" is included, the probability of closing is predicted to be 80%.

[1374] The input is the newly entered minutes of the business meeting, and the output is the presence or absence of key phrases and the probability of closing the deal.

[1375] Step 9:

[1376] Terminal: The server notifies the salesperson in real time with the predicted close probability results, including a specific message such as "This opportunity has an 80% chance of closing."

[1377] (Application example 1)

[1378] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1379] In conventional sales activities, analysis of sales meeting minutes and prediction of the probability of closing a deal are often performed manually, resulting in inefficiency and inaccuracy. Furthermore, when audio recordings are performed, the process of converting the audio data into text is complicated, making real-time analysis and prediction difficult. The present invention aims to solve these problems and improve the efficiency and accuracy of customer service and sales activities in brick-and-mortar stores.

[1380] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1381] In this invention, the server includes a means for storing the minutes of business negotiations in a database, a means for analyzing the minutes as text data by natural language processing using artificial intelligence, and a means for converting speech to text, which enables real-time text analysis and the identification of key phrases with a high probability of closing a deal.

[1382]

[1383] "Business minutes" refers to information recorded by sales representatives or store clerks about business negotiations and customer service with customers.

[1384] A "database" is a system for storing data in an organized manner and for efficiently searching, extracting, and managing data.

[1385] "Generative AI" is a type of artificial intelligence that performs natural language processing, and is a technology responsible for analyzing and generating text data.

[1386] "Natural language processing" is a technology that allows computers to understand and analyze language spoken by humans.

[1387] "Text data" refers to data that stores character strings or sentences in digital format.

[1388] "Past business negotiation data" is information about business negotiations that have been conducted in the past, including records of successful and unsuccessful cases.

[1389] A "successful deal" refers to a case where a deal has been concluded.

[1390] A "failed deal" refers to a deal that did not result in a successful transaction.

[1391] A "keyphrase" is a word or phrase that has significant meaning in a particular document or conversation.

[1392] "Agreement probability" refers to the probability that a deal will be concluded.

[1393] "Notifying the user" means sending the analysis results and prediction results to the user's terminal and informing them.

[1394] "Speech to text" is the process of converting recorded audio data into a string or sentence format.

[1395] A "user interface that controls voice input" refers to software that controls the operation screen and voice recognition that allows users to input voice.

[1396]

[1397] The present invention is a system that records the details of business negotiations and customer service, analyzes the details, and predicts the likelihood of a deal. In particular, as an example of application in a physical store, it provides a means for recording conversations between store clerks and customers and predicting the likelihood of a deal in real time based on the data.

[1398] System Configuration

[1399] Hardware Configuration

[1400] The system of the present invention uses the following hardware:

[1401] 1. Terminal: A device capable of voice input and output, such as smart glasses or a head-mounted display.

[1402] 2. Server: A high-performance computer to run the database and generative AI.

[1403] Software Configuration

[1404] The system operates using the following software:

[1405] 1. Speech recognition software: Use the speech_recognition library to convert voice data into text data.

[1406] 2. Generative AI software: Hugging Face's Transformer model is used to analyze the text data.

[1407] 3. Text analysis software: Identify key phrases using TF-IDF and predict conversion probability using a logistic regression model.

[1408] Data processing and calculation

[1409] 1. Voice input:

[1410] The terminal records the conversation between the store clerk and the customer and transmits the audio data to the server.

[1411] 2. Audio conversion:

[1412] Speech recognition software running on the server converts the recorded voice data into text data.

[1413] 3. Text Analysis:

[1414] A generative AI model is used to perform natural language processing on the converted text data and extract key phrases, using the Hugging Face Transformer model.

[1415] 4. Data storage:

[1416] The text data and extracted key phrases are stored in a database.

[1417] 5. Win probability prediction:

[1418] Using a logistic regression model built on data from past successful and unsuccessful cases, the probability of closing a deal is predicted from the extracted key phrases.

[1419] 6. Notification of Results:

[1420] The prediction results are sent to the terminal in real time and presented to the store clerk.

[1421] Specific examples

[1422] For example, a salesperson can record the phrase "I'm very interested in this product" while serving a customer, and the recording is converted into text. This text is sent to the system, where generative AI and TF-IDF analysis are used to extract the key phrase "interested." This key phrase is pattern-matched with past successful cases, and the system predicts a high probability of success. For example, a result such as "85% probability of success" is displayed in real time on the smart glasses.

[1423] Prompt Sentence Examples

[1424] Customer conversation recording text: "I'm very interested in this product." Analyze the following text, extract key phrases, and predict the probability of closing a sale.

[1425] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1426]

[1427] Step 1:

[1428] A user records a conversation with a customer using smart glasses or a head-mounted display. The input is voice data, and the output is the recorded voice data. This voice data is stored on the device.

[1429] Step 2:

[1430] The device sends voice data to the server. The server receives the voice data and converts it into text data using speech recognition software (speech_recognition library). The input is voice data and the output is text data.

[1431] Step 3:

[1432] The server then performs natural language processing on the converted text data using a generative AI model (Hugging Face's Transformer model). This process involves tokenizing the input text data, tagging it with parts of speech, and extracting noun phrases. The input is text data, and the output is analyzed text data.

[1433] Step 4:

[1434] The server extracts important key phrases from the analyzed text data and identifies frequently occurring key phrases using TF-IDF scores. The input of this process is the analyzed text data, and the output is the extracted key phrases.

[1435] Step 5:

[1436] The server uses a logistic regression model based on data on past successful and unsuccessful cases to predict the probability of success from newly extracted key phrases. The input to this process is the extracted key phrases, and the output is the predicted probability of success.

[1437] Step 6:

[1438] The server sends the predicted probability of closing to the terminal. The terminal notifies the user (store clerk or salesperson) of the received predicted result of the agreement probability in real time. The input of this process is the predicted value of the probability of closing, and the output is a notification to the user.

[1439] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1440] This invention is a system for predicting the likelihood of a deal by analyzing sales meeting minutes using a generation AI and an emotion engine. Specifically, it automates the process of collecting, analyzing, and recognizing emotions in sales meeting minutes, as well as predicting the likelihood of an agreement.

[1441] Collection of business meeting minutes

[1442] Terminal: After a sales meeting, salespeople use the terminal to input minutes. Two input methods are available: text input and voice recognition, allowing salespeople to efficiently record the details of the meeting. For example, comments such as "I'm very interested in this proposal" can be entered.

[1443] Data transmission and storage

[1444] Terminal: The entered minutes of the business meeting are sent from the terminal to the server. The sent data undergoes a format check and is saved in the server in the appropriate format.

[1445] Server: The server stores the received minutes in a database, allowing data to be centrally managed for use in subsequent processing.

[1446] Text analysis with generative artificial intelligence

[1447] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[1448] Classification of successful and unsuccessful cases

[1449] Server: Classify successful and unsuccessful deals based on past sales negotiation data. Uses generative AI to extract commonalities between successful and unsuccessful deals. From past data, it is confirmed that the phrase "interested" is frequently included in successful deals.

[1450] Keyphrase Extraction

[1451] Server: Extract frequently occurring key phrases from successful business meeting minutes. For example, using TF-IDF scores, key phrases such as "cost performance" and "interested" are identified as important.

[1452] Emotion recognition by emotion engine

[1453] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[1454] Server: The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[1455] Agreement probability prediction

[1456] User: The sales representative enters new minutes of the business meeting into the terminal.

[1457] Server: The generative AI predicts the likelihood of agreement for the input minutes based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[1458] Terminal: The salesperson is notified in real time of the predicted agreement probability sent from the server, allowing the salesperson to plan their next action based on this information.

[1459] Specific examples

[1460] For example, consider the case where a salesperson conducts an initial business meeting with a new customer and writes in the minutes that they are "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts a 90% probability of agreement.

[1461] The results are then sent to the device, and the salesperson can use this information to take appropriate action, such as providing further proposals or specific support.In this way, this system uses a data-driven approach combined with emotion recognition to more accurately predict the likelihood of success of a sales negotiation, providing an effective means of improving the efficiency of sales activities.

[1462] The processing flow will be explained below.

[1463] Step 1: Data collection

[1464] User: After a sales meeting, a sales representative uses the device to enter minutes. Two input methods are available: text input and voice recognition, allowing the sales representative to efficiently record the details of the meeting. For example, a comment such as "I'm very interested in this proposal" can be entered.

[1465] Step 2: Send data

[1466] Terminal: The entered minutes are sent from the terminal to the server. Before being sent, a format check is performed to confirm the consistency of the data.

[1467] Step 3: Save data

[1468] Server: The received minutes of the business meeting are stored in a database. The stored data is used for future analysis and prediction.

[1469] Step 4: Text Analysis

[1470] Server: Generates minutes of business meetings stored in a database and performs natural language processing using AI. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. For example, it extracts the predicate "interesting."

[1471] Step 5: Classification of successes and failures

[1472] Server: Based on past sales negotiation data, generative AI is used to classify successful and unsuccessful deals. Cluster analysis and classifiers are used for classification, and commonalities are extracted. For example, it is confirmed that the phrase "interested" appears frequently in successful deals.

[1473] Step 6: Extracting Keyphrases

[1474] Server: Extracts frequently occurring key phrases from the minutes of successful business negotiations. Text mining techniques such as TF-IDF scores and correlation analysis are used for extraction. For example, key phrases such as "cost performance" and "interested" are extracted.

[1475] Step 7: Emotion Recognition

[1476] Terminal: The emotion engine analyzes the conversation between the salesperson and the customer during the sales negotiation. The emotion engine analyzes the tone of voice, facial expressions, and text content to identify the customer's emotions. For example, if the customer is expecting something, a positive emotion such as "expecting" will be detected.

[1477] Step 8: Storing Emotion Data

[1478] Server: Records the detected emotion data in a database and uses it for subsequent sales negotiation analysis.

[1479] Step 9: Predicting agreement probability

[1480] User: A sales representative enters new sales meeting minutes into a terminal.

[1481] Server: For the minutes entered, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and the emotional data detected by the emotion engine. For example, if the phrase "cost performance" and the emotional data "expectations" are included, the likelihood of agreement is predicted to be 90%.

[1482] Step 10: Notification of prediction results

[1483] Terminal: Salespeople are notified in real time of the agreement accuracy of the forecast results. Based on this information, salespeople can plan their next actions. For example, they can improve their proposals or prepare additional materials based on the forecast results.

[1484] Through the above processing steps, the present invention scientifically predicts the likelihood of success in business negotiations and achieves improved efficiency in sales activities by combining emotion recognition.

[1485] Example 2

[1486] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1487] Conventional sales negotiation support systems have difficulty taking into account customer emotional information when analyzing sales negotiation minutes and predicting the probability of success. As a result, there was a problem of reduced accuracy in predicting the probability of success. In addition, inputting sales negotiation minutes and sending and saving the data was time-consuming, making efficient operation difficult.

[1488] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1489] In this invention, the server includes means for storing the minutes of business negotiations in a database, means for analyzing the minutes of business negotiations as text data using natural language processing with artificial intelligence, means for classifying successful cases and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for analyzing emotional data during business negotiations using an emotion engine and predicting the likelihood of agreement with higher accuracy based on this, and means for notifying the user of the predicted likelihood of agreement. This enables more accurate prediction of business negotiation results including customer emotional data, and also improves the efficiency of inputting business negotiation minutes and transmitting and saving data.

[1490] A "business meeting minutes" is a document that records what was discussed and what decisions were made during a business meeting.

[1491] A "database" is a system for efficiently managing, storing, searching, and updating large amounts of data.

[1492] "Generative AI" is a technology that uses machine learning algorithms and natural language processing techniques to generate and analyze text data.

[1493] "Natural language processing" is a technology that enables computers to understand, interpret, and generate human language.

[1494] A "key phrase" is a word or phrase that has a particularly important meaning in text data.

[1495] "Agreement probability" is a probability that indicates the likelihood that a proposal will be accepted in a business negotiation.

[1496] "User" refers to an individual or organization that uses this system to input minutes of business negotiations and receive the analysis results.

[1497] An "emotion engine" is a technology that analyzes voice tone, facial expressions, and text content to identify human emotions.

[1498] "Format check" is the process of checking whether the entered data conforms to the specified format.

[1499] "Real-time" is a concept that indicates that data is generated and processed immediately without delay.

[1500] This invention is a system for analyzing sales meeting minutes in sales activities using a generative AI model and an emotion engine to predict the likelihood of a deal. The system automates the processes of collecting, analyzing, and recognizing emotions in sales meeting minutes, and predicting the likelihood of agreement. Specific embodiments for implementing this system are described below.

[1501] Collection of business meeting minutes

[1502] Terminal: The user (salesperson) uses a user interface to input minutes after a business meeting. There are two input methods: text input and voice recognition, which allows for efficient recording. The software used is a voice recognition engine (for example, Google Speech-to-Text). Specifically, the user inputs a comment such as "I'm very interested in this proposal."

[1503] Data transmission and storage

[1504] Terminal: Once the minutes are entered, the data is sent to the server after format checks are performed. The format checks described above are a means to ensure accurate and consistent data entry.

[1505] Server: The server analyzes the received data and stores it in a database. This database is a system that allows for the management and efficient searching of large amounts of data. For example, a common database management system such as MySQL or PostgreSQL is used.

[1506] Text analysis with generative artificial intelligence

[1507] Server: Generates minutes stored in a database and performs natural language processing (NLP) using an AI model. Specifically, it splits tokens, tags parts of speech, and extracts noun phrases. Software used includes NLP libraries such as NLTK and SpaCy. For example, the predicate "interesting" is extracted.

[1508] Classification of successful and unsuccessful cases

[1509] Server: Using past sales negotiation data, a generative AI model extracts commonalities between successful and unsuccessful deals. Software used includes machine learning libraries such as scikit-learn and TensorFlow. It is recognized that successful deals frequently include the phrase "interested."

[1510] Keyphrase Extraction

[1511] Server: Extracts frequently occurring key phrases from successful business meeting minutes. Specifically, it calculates TF-IDF scores and identifies weighted key phrases. For example, key phrases such as "cost performance" and "interested" are extracted.

[1512] Emotion recognition by emotion engine

[1513] Terminal: The conversation between the user and the customer during the sales negotiation is analyzed using an emotion engine. The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. Software used may include a voice recognition engine (Google Speech-to-Text) or a facial recognition tool (OpenCV). For example, if the customer is excited, it will be identified as "expecting."

[1514] Server: Emotion data is recorded in a database and used for analysis. The collected emotional data is an important indicator for predicting the success rate of business negotiations.

[1515] Agreement probability prediction

[1516] User: After the business meeting, the user inputs new minutes into the terminal.

[1517] Server: Based on the input minutes, the generative AI model predicts the likelihood of agreement based on identified key phrases and emotional data. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the model predicts the likelihood of agreement to be 90%.

[1518] Terminal: Prediction results are notified to salespeople in real time, allowing them to plan their next actions based on this information.

[1519] Examples of concrete examples and prompts

[1520] For example, if a salesperson conducts an initial sales meeting with a new customer and notes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting, this data is sent to the server and analyzed there. The generative AI model predicts an agreement probability of 90% based on the key phrase "interested" and the emotion data "expectation." This result is notified to the device, and the salesperson can use it to consider further proposals and specific support.

[1521] Prompt Sentence Examples

[1522] You wrote in the minutes, "I'm very interested in your new solution proposal." During the negotiation, the emotion engine detected the customer's emotion, such as "I'm looking forward to it." Use this data to predict the likelihood of an agreement.

[1523] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1524] Step 1: Enter meeting notes

[1525] Terminal: After a sales meeting, the user (salesperson) uses the terminal's user interface to input minutes. There are two input formats: text input and voice recognition, and the user can choose either. Input is mainly done using a keyboard or microphone.

[1526] Input: Text or audio data about the business.

[1527] Output: Minutes data that has passed format check.

[1528] What it does: The user enters text into an input form or records a voice, which is then converted into text using a speech recognition engine (e.g., Google Speech-to-Text).

[1529] Step 2: Send and store data

[1530] Terminal: After the minutes are entered, the terminal checks the format and sends the data to the server. Only data that passes the format check is sent.

[1531] Input: Format-checked minutes data.

[1532] Output: The minutes data sent to the server.

[1533] Specific behavior: The device sends data to the server using an HTTP request, including appropriate format conversion if necessary.

[1534] Server: The server analyzes the minutes data received from the terminal and stores it in a database.

[1535] Input: Received minutes data.

[1536] Output: Meeting minutes data stored in a database.

[1537] Specific operation: The server executes an insert operation in the database management system (e.g., MySQL) to permanently store the data.

[1538] Step 3: Text analysis

[1539] Server: The server uses a generative AI model to perform natural language processing (NLP) on the stored minutes data, specifically splitting tokens, tagging parts of speech, and extracting noun phrases.

[1540] Input: Meeting minutes data stored in the database.

[1541] Output: Text data as the analysis result.

[1542] What it does: The server uses an NLP library (e.g., NLTK, SpaCy) to segment the text, tag parts of speech, and extract noun phrases, such as the phrase "interesting."

[1543] Step 4: Classification of successes and failures

[1544] Server: Extracts commonalities between successful and unsuccessful deals based on past sales negotiation data. Classifies meeting minutes data using a generative AI model.

[1545] Inputs: Past and current deal data.

[1546] Output: Classification results of successful and unsuccessful cases.

[1547] What it does: The server uses machine learning algorithms (e.g., scikit-learn, TensorFlow) to learn patterns from past datasets and classify new data, identifying common key phrases and success factors and generating classification results.

[1548] Step 5: Keyphrase Extraction

[1549] Server: Calculate TF-IDF scores to identify frequently occurring key phrases from successful meeting notes, and extract important key phrases.

[1550] Input: Successful deal opportunity data.

[1551] Output: Extracted key phrases.

[1552] Specific operation: The server uses the TF-IDF algorithm to calculate important phrases in the sales negotiation data and extracts the top-ranked key phrases. For example, the phrase "cost performance" is identified as important.

[1553] Step 6: Emotion Recognition

[1554] Terminal: During sales negotiations, conversations between users and customers are collected and analyzed using an emotion engine. Voice tone, facial expressions, and text content are used to identify customer emotions.

[1555] Input: Voice, facial expression, and text data during business negotiations.

[1556] Output: Parsed emotion data.

[1557] What it does: The device uses an emotion engine (e.g., a voice recognition engine or facial recognition tool) to analyze the data and identify the customer's emotions. Positive emotions such as "expecting" are detected.

[1558] Server: The analyzed emotion data is recorded in a database and used for subsequent analysis.

[1559] Input: Parsed emotion data.

[1560] Output: Emotion data stored in a database.

[1561] Specific operation: The server inserts the emotion data into a database, making it available for future analysis.

[1562] Step 7: Predicting agreement probability

[1563] Server: Based on key phrases and sentiment data, a generative AI model is used to predict the likelihood of agreement. It uses a model learned from past successful cases.

[1564] Input: New sales meeting minutes data, identified key phrases, analyzed sentiment data.

[1565] Output: Consensus prediction results.

[1566] Specific operation: The server uses the generative AI model to analyze the input data and calculate the agreement probability. For example, if the comment "good cost performance" and the emotional data "expected" are included, the agreement probability is predicted to be 90%.

[1567] Step 8: Notification

[1568] Terminal: The terminal notifies the user in real time of the predicted consensus probability sent from the server. The user can plan their next action based on this information.

[1569] Input: Prediction results sent from the server.

[1570] Output: The prediction results displayed on the device screen.

[1571] Specific operation: The device uses a notification system to notify the user of the prediction results in real time, allowing the user to plan their next steps.

[1572] (Application example 2)

[1573] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1574] During sales negotiations and customer service in brick-and-mortar stores, it is difficult to accurately predict the customer's level of interest in a product or the likelihood of a sale. Conventional sales negotiation minutes and customer service records rely on subjective judgment, making it difficult to formulate effective proposals and sales strategies. Furthermore, it is difficult to accurately grasp the customer's emotions and make appropriate product proposals based on them.

[1575] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for saving business negotiation minutes in a database, means for analyzing the business negotiation minutes as text data using natural language processing with generative artificial intelligence, means for classifying successful and unsuccessful cases based on past business negotiation data, means for extracting key phrases contained in the successful cases, means for predicting the likelihood of agreement in future business negotiations based on the extracted key phrases, means for notifying the user of the predicted likelihood of agreement, means for analyzing customer emotion data using an emotion engine, means for predicting customer satisfaction based on the analyzed emotion data, and means for proposing appropriate products to customers based on the key phrases extracted from customer emotions and the successful cases. This enables product proposals that increase the likelihood of closing business negotiations and customer service and increase customer satisfaction based on the customer's emotions and interests.

[1576] A "business negotiation minutes" is a document that records the details of a business negotiation that a sales representative conducted with a customer.

[1577] A "database" is a system designed to store information in an organized manner and make it easy to access and search.

[1578] "Generative AI" is a system that uses artificial intelligence technology to generate and analyze text and data.

[1579] "Natural language processing" is a technology that allows computers to understand and process human language.

[1580] "Past business negotiation data" refers to data that includes records and results of previous business negotiations.

[1581] A "successful case" refers to a case where a business deal has been concluded.

[1582] A "failed deal" refers to a business deal that did not result in a successful transaction.

[1583] "Key phrases" are particularly important words or phrases in business negotiations or customer comments.

[1584] "Agreement probability" is an index that indicates the degree of possibility that a business negotiation will lead to a contract.

[1585] An "emotion engine" is a system that analyzes and identifies human emotions from voice and text.

[1586] "Emotion data" is information about the customer's emotional state analyzed by the emotion engine.

[1587] "Customer satisfaction" is an indicator of how satisfied customers are with a product or service.

[1588] "Product proposal" is the act of recommending appropriate products or services to customers.

[1589] This invention relates to a system for efficiently handling customer inquiries and conducting business negotiations in brick-and-mortar stores. This system collects business negotiation minutes, analyzes them using generative AI and an emotion engine, and predicts the probability of closing a deal and customer satisfaction. The components and procedures required to implement this invention are described below.

[1590] Collection of business meeting minutes

[1591] Device:

[1592] After the sales meeting, the sales representative will use a smartphone or tablet to input minutes of the meeting. Two input methods are available: text input and voice recognition, allowing sales representatives to efficiently record the details of the meeting.

[1593] Data transmission and storage

[1594] Device:

[1595] The entered minutes of the business meeting are sent from the terminal to the server, where they are checked for format and saved in the appropriate format.

[1596] server:

[1597] The server stores the received minutes of the business negotiations in a database, which allows for centralized management of data for later analysis.

[1598] Text analysis with generative artificial intelligence

[1599] server:

[1600] The minutes stored in the database are then processed using generative AI (e.g., Hugging Face Transformers) for natural language processing. Specifically, tokenization, part-of-speech tagging, and noun phrase extraction are performed. For example, the predicate "interesting" is extracted.

[1601] Classification of successful and unsuccessful cases

[1602] server:

[1603] Based on past sales negotiation data, successful and unsuccessful deals are classified. Generative AI is used to extract commonalities between successful and unsuccessful deals. For example, it is confirmed that the phrase "interested" is frequently included in successful deals.

[1604] Keyphrase Extraction

[1605] server:

[1606] Extract frequently occurring key phrases from successful business meeting minutes. For example, key phrases such as "cost performance" and "interested" are identified as important using TF-IDF scores.

[1607] Emotion recognition by emotion engine

[1608] Device:

[1609] During a sales meeting, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The emotion engine analyzes voice tone, facial expressions, and text content to identify the customer's emotions. For example, if the customer is excited, a positive emotion such as "expecting" is detected.

[1610] server:

[1611] The detected emotion data is recorded in a database and used as data to predict the likelihood of success of a business negotiation with even greater accuracy.

[1612] Agreement probability prediction

[1613] User:

[1614] The sales representative enters new minutes of the business meeting into the terminal.

[1615] server:

[1616] For input minutes, the generative AI predicts the likelihood of agreement based on pre-specified key phrases and emotional data detected by the emotion engine. For example, if the minutes contain the comment "high cost performance" and the emotional data "expectations," the AI ​​predicts the likelihood of agreement to be 90%.

[1617] Device:

[1618] The server sends the predicted results of the agreement probability to the sales representative in real time. Based on this information, the sales representative can plan the next action. It also enables the sales representative to propose appropriate products based on the customer's emotions and interests.

[1619] Examples of concrete examples and prompts

[1620] For example, consider the case where a salesperson conducts an initial sales meeting with a new customer and writes in the minutes that the customer is "very interested in the new solution proposal," and the emotion engine detects the customer's "expectation" emotion during the meeting. The minutes and emotion data are sent from the device to the server, where they are analyzed. Based on the key phrase "interested" and the emotion data "expectation," the generative AI predicts an agreement probability of 90%. This result is notified to the device, and the salesperson can use this information to take appropriate action, such as making further proposals or specific support.

[1621] Example prompt sentence:

[1622] Convert the conversation between the customer and the store clerk into text and use generative AI to analyze the following:

[1623] 1. Extracting important key phrases

[1624] 2. Recognizing customer sentiment

[1625] 3. Closing probability prediction

[1626] In this way, the system uses a data-driven approach that combines customer sentiment to more accurately predict the likelihood of a deal's success and provide an effective means of improving the efficiency of sales activities.

[1627] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1628] Step 1: Collect meeting notes

[1629] Device:

[1630] Salespeople, who are users, use smartphones or tablets to input minutes of sales negotiations after a sales meeting. There are two input methods: text input and voice recognition. With voice recognition, the user's voice is analyzed and the contents of the sales negotiation are converted into text. These input minutes of sales negotiations become the input data for the next processing step.

[1631] Step 2: Send and store data

[1632] Device:

[1633] Once the input of the business meeting minutes is complete, the terminal sends the minutes to the server. The sent data undergoes a format check and is saved in the appropriate format on the server. The format check verifies the data format and required fields. If successful, the data is saved in the server's database (output).

[1634] Step 3: Text analysis with generative AI

[1635] server:

[1636] Using saved business meeting minutes as input, natural language processing is performed using generative AI (e.g., Hugging Face Transformers). Tokenization, part-of-speech tagging, and noun phrase extraction are performed, and the analysis results are output as text data. For example, the predicate "interesting" is extracted.

[1637] Step 4: Classification of successes and failures

[1638] server:

[1639] By referencing past sales negotiation data, the system uses generative AI to extract commonalities between successful and unsuccessful deals. This allows the minutes of sales negotiations to be classified as either successful or unsuccessful. The classification results are output as labeled data indicating whether the deal was successful or unsuccessful.

[1640] Step 5: Extracting Keyphrases

[1641] server:

[1642] The input is a successful business meeting transcript, and frequent keyphrases are extracted. This process uses TF-IDF scores to identify important keyphrases. Keyphrases such as "cost-effectiveness" and "interested" are output.

[1643] Step 6: Emotion Recognition with the Emotion Engine

[1644] Device:

[1645] During a sales negotiation, the conversation between the salesperson and the customer is analyzed using an emotion engine (e.g., j-hartmann / emotion-english-distilroberta-base model). The inputs are voice tone, facial expressions, and text content. The analyzed emotion data is output as an emotion label, such as "expecting."

[1646] server:

[1647] The detected emotion data is recorded in a database, which helps predict the probability of success.

[1648] Step 7: Predicting agreement probability

[1649] server:

[1650] Based on newly input minutes of business meetings, the generation AI predicts the likelihood of agreement. Inputs include the text data of the minutes, pre-specified key phrases, and emotional data. The generation AI analyzes each input data and outputs the likelihood of agreement as a numerical value (percentage). For example, if the input data includes the comment "high cost performance" and the emotional data "we are looking forward to it," the generation AI predicts the likelihood of agreement to be 90%.

[1651] Step 8: Notification of results

[1652] Device:

[1653] The server then sends the predicted agreement probability to the device. The salesperson receives this information in real time and plans the next steps. For example, they can create a specific sales strategy, such as considering additional proposals.

[1654] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1655] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1656] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1657] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1658] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1659] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1660] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1661] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1662] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1663] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1664] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1665] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1666] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1667] 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.

[1668] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1669] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1670] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1671] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1672] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1673] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1674] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1675] The following is further disclosed regarding the above embodiment.

[1676] (Claim 1)

[1677] A means for storing the minutes of business meetings in a database;

[1678] A method for generating minutes of business meetings using artificial intelligence to process natural language and analyze it as text data,

[1679] A means to classify successful and unsuccessful cases based on past business negotiation data,

[1680] A means of extracting key phrases contained in successful cases;

[1681] A means of predicting the likelihood of agreement in future business negotiations based on the extracted key phrases;

[1682] a means for notifying a user of the predicted agreement probability;

[1683] A system including:

[1684] (Claim 2)

[1685] 2. The system according to claim 1, further comprising means for providing a user interface for inputting minutes of business negotiations, and for format-checking the input minutes and transmitting them to the server.

[1686] (Claim 3)

[1687] 10. The system of claim 1, further comprising means for analyzing the minutes of the business meeting in real time and predicting the likelihood of an agreement.

[1688] "Example 1"

[1689] (Claim 1)

[1690] An input device is provided for recording the contents of business negotiations, and means is provided for a user to input minutes of the negotiations by text input or voice recognition;

[1691] A means to check the format of entered minutes of business meetings and send them;

[1692] A means for storing the minutes of business meetings in a database;

[1693] A method for generating minutes of business meetings using artificial intelligence to process natural language and analyze it as text data,

[1694] A means to classify successful and unsuccessful cases based on past business negotiation data,

[1695] A means of extracting important expressions contained in successful cases;

[1696] A means for predicting the likelihood of agreement in future business negotiations based on the extracted important expressions;

[1697] a means for notifying a user of the predicted agreement probability;

[1698] A system including:

[1699] (Claim 2)

[1700] 10. The system of claim 1, further comprising means for analyzing the minutes of the business meeting in real time and predicting the likelihood of an agreement.

[1701] (Claim 3)

[1702] 2. The system according to claim 1, further comprising means for providing a user interface for inputting minutes of business negotiations, and for format-checking the input minutes and transmitting them to the server.

[1703] "Application Example 1"

[1704] (Claim 1)

[1705] A means for storing the minutes of business meetings in a database;

[1706] A method for generating minutes of business meetings using artificial intelligence to process natural language and analyze it as text data,

[1707] A means to classify successful and unsuccessful cases based on past business negotiation data,

[1708] A means of extracting key phrases contained in successful cases;

[1709] A means of predicting the likelihood of agreement in future business negotiations based on the extracted key phrases;

[1710] a means for notifying a user of the predicted agreement probability;

[1711] a means for converting speech to text;

[1712] A means for providing a user interface for controlling voice input;

[1713] A system including:

[1714] (Claim 2)

[1715] 2. The system according to claim 1, further comprising means for providing a user interface for inputting minutes of business negotiations, and for format-checking the input minutes and transmitting them to the server.

[1716] (Claim 3)

[1717] 10. The system of claim 1, further comprising means for analyzing the minutes of the business meeting in real time and predicting the likelihood of an agreement.

[1718]

[1719] "Example 2: Combining Emotion Engines"

[1720] (Claim 1)

[1721] A means for storing the minutes of business meetings in a database;

[1722] A method for generating minutes of business meetings using artificial intelligence to process natural language and analyze it as text data,

[1723] A means to classify successful and unsuccessful cases based on past business negotiation data,

[1724] A means of extracting key phrases contained in successful cases;

[1725] A means of predicting the likelihood of agreement in future business negotiations based on the extracted key phrases;

[1726] a means for notifying a user of the predicted agreement probability;

[1727] A method for analyzing emotional data during negotiations using an emotion engine and predicting the likelihood of agreement with even greater accuracy based on this data;

[1728] A system including:

[1729] (Claim 2)

[1730] 2. The system according to claim 1, further comprising means for providing a user interface for inputting minutes of business negotiations, and for format-checking the input minutes and transmitting them to the server.

[1731] (Claim 3)

[1732] 10. The system of claim 1, further comprising means for analyzing the minutes of the business meeting in real time and predicting the likelihood of an agreement.

[1733] "Application example 2 when combining emotion engines"

[1734] (Claim 1)

[1735] A means for storing the minutes of business meetings in a database;

[1736] A method for generating minutes of business meetings using artificial intelligence to process natural language and analyze it as text data,

[1737] A means to classify successful and unsuccessful cases based on past business negotiation data,

[1738] A means of extracting key phrases contained in successful cases;

[1739] A means of predicting the likelihood of agreement in future business negotiations based on the extracted key phrases;

[1740] a means for notifying a user of the predicted agreement probability;

[1741] means for analyzing customer emotion data using an emotion engine;

[1742] A means for predicting customer satisfaction based on the analyzed emotion data;

[1743] A method to propose appropriate products to customers based on key phrases and success stories extracted from customer sentiment,

[1744] A system including:

[1745] (Claim 2)

[1746] 2. The system according to claim 1, further comprising means for providing a user interface for inputting minutes of business negotiations, and for format-checking the input minutes and transmitting them to the server.

[1747] (Claim 3)

[1748] 10. The system of claim 1, further comprising means for analyzing the minutes of the business meeting in real time and predicting the likelihood of an agreement. [Explanation of symbols]

[1749] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for storing the minutes of business meetings in a database; A method for generating minutes of business meetings using artificial intelligence to process natural language and analyze it as text data, A means to classify successful and unsuccessful cases based on past business negotiation data, A means of extracting key phrases contained in successful cases; A means of predicting the likelihood of agreement in future business negotiations based on the extracted key phrases; a means for notifying a user of the predicted agreement probability; A system including:

2. 2. The system according to claim 1, further comprising means for providing a user interface for inputting minutes of business negotiations, and for checking the format of the input minutes and transmitting them to the server.

3. The system of claim 1 , further comprising means for analyzing the minutes of the business meeting in real time and predicting the likelihood of agreement.

Images