system

Abstract

We provide the system. [Solution] A means for collecting sales data, Means for pre-processing the aforementioned sales data, A means for training a machine learning model based on the aforementioned pre-processed sales data, A means for inputting new sales opportunity data into the aforementioned machine learning model and inferring the next action to take, Means for notifying the sales representative of the inferred action A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] In business activities, it is said to be difficult to appropriately determine what actions should be taken for which customers. As a result, ineffective business activities may be carried out, and order-receiving opportunities may be lost. In addition, since there is no technology for predicting actions to increase the order-receiving rate using past business data, salespersons have to spend a lot of time and effort on these determinations. Existing systems cover only part of business activities, so it is difficult to achieve an overall improvement in business efficiency. To solve these problems, a new system is required.

Means for Solving the Problems

[0005] The present invention solves the aforementioned problems by providing a system that includes means for collecting sales data, means for preprocessing the sales data, means for training a machine learning model based on the preprocessed sales data, means for inputting new sales opportunity data into the machine learning model and inferring the next action to take, and means for notifying the sales representative of the inferred action. Specifically, the system learns the characteristics of successful and unsuccessful sales based on data collected from a customer management system and a schedule management system, and recommends the optimal sales action. This system allows sales representatives to quickly grasp the next effective action, significantly improving sales efficiency.

[0006] "Sales data" refers to information related to sales activities, and specifically includes case information, activity records, visit history, and interaction history obtained from customer management systems and schedule management systems.

[0007] "Means of collection" refers to the methods and technologies used to collect sales data, and specifically includes data acquisition using APIs and data extraction from databases.

[0008] "Preprocessing methods" refer to methods and techniques for converting collected sales data into a format suitable for analysis and model training, and specifically include data cleaning, imputation of missing values, normalization, and standardization.

[0009] A "machine learning model" refers to algorithms and technologies that learn from sales data and make predictions and recommendations for future sales activities. Specifically, this includes methods such as logistic regression, random forests, and neural networks.

[0010] "Means of inference" refers to methods and techniques that use machine learning models to predict appropriate actions for new sales opportunities, and specifically includes the process of inputting data into the model and analyzing the resulting predictions.

[0011] "Means of notification" refers to methods and technologies for informing sales representatives of anticipated actions, and specifically includes smartphone push notifications, PC desktop notifications, and email notifications.

[0012] A "customer management system" refers to a system for managing customer information, and specifically includes databases and software for managing customer contact information, project information, sales activity history, etc.

[0013] A "schedule management system" refers to a system for managing the schedules of sales representatives, and specifically includes calendar systems and software for managing visit dates, meeting schedules, task management, etc. [Brief explanation of the drawing]

[0014] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10]Shows an emotion map to which multiple emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Embodiment 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Embodiment 2 when the emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when the emotion engine is combined.

Mode for Carrying Out the Invention

[0015] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described according to the accompanying drawings.

[0016] First, the terms used in the following description will be explained.

[0017] In the following embodiments, a labeled processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0018] In the following embodiments, a labeled RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0019] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0020] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0021] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0022] [First Embodiment]

[0023] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0024] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0025] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0026] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0027] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0028] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0029] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0030] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0032] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0033] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0034] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0035] The present invention provides a system for suggesting the next course of action to sales representatives based on sales data. This system is primarily implemented through the following steps.

[0036] Data collection

[0037] 1. The server collects sales data from customer relationship management (CRM) and scheduling management systems. Specifically, the server uses APIs to retrieve deal information (e.g., deal name, size, progress, person in charge) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[0038] Data preprocessing

[0039] 2. The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[0040] Model Learning

[0041] 3. The server receives the pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between various features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[0042] Inference and Action Suggestions

[0043] 4. Every time the server acquires new sales opportunity data, it inputs the data into a machine learning model to infer the next action to take. For example, the model might suggest a specific action such as "a visit would be beneficial for this customer."

[0044] notification

[0045] 5. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive insights from the system in real time.

[0046] Specific example

[0047] 1. Data collection:

[0048] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[0049] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[0050] 2. Data preprocessing:

[0051] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[0052] 3. Model learning:

[0053] The server uses a logistic regression algorithm to learn features from past data, specifically identifying that a high number of visits correlates with a high conversion rate.

[0054] 4. Inferences and Action Suggestions:

[0055] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in charge: Sato)" into the model and infers the action "This project requires 3 visits."

[0056] 5. Notice:

[0057] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[0058] As described above, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and suggesting the optimal action to sales representatives in real time.

[0059] The following describes the processing flow.

[0060] Step 1:

[0061] The server collects sales data from the customer management system and the scheduling management system. Specifically, the server calls the API of the customer management system to obtain case information (e.g., case ID, case name, size, person in charge ID, progress status). It also uses the API of the scheduling management system to obtain sales activity schedules (e.g., visit date and time, meeting time, participants).

[0062] Step 2:

[0063] The server cleans the collected data. First, it removes duplicate entries from the data. Next, it appropriately fills in any missing items. For example, if there is a missing visit date and time, it automatically fills in the nearest visit date.

[0064] Step 3:

[0065] The server standardizes and normalizes the preprocessed data. Specifically, it scales numerical data (e.g., project size) to a certain range. This transforms the data into a format suitable for machine learning models.

[0066] Step 4:

[0067] The server trains a machine learning model based on pre-processed data. Specifically, it uses data from past successful and unsuccessful deals to build and train a model using algorithms such as logistic regression and random forest. This model learns the relationship between features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[0068] Step 5:

[0069] The server inputs new case data into the model and infers the next action to take. For example, for a new case, specific actions such as the number of visits or phone calls may be suggested. The inference results will output actions such as "increase the number of visits" or "follow up with a phone call."

[0070] Step 6:

[0071] Based on the inference results, the server generates data that notifies sales representatives of specific actions to take. This data includes detailed instructions on who to contact and what actions to take.

[0072] Step 7:

[0073] The terminal receives notifications from the server and informs the sales representative. Specifically, it reports specific actions to the sales representative, such as "You should visit Mr. Yamada, the manager at Company B, next," via smartphone push notifications, PC desktop notifications, or email notifications.

[0074] Through these steps, the system significantly improves the efficiency and effectiveness of sales activities by providing sales representatives with specific instructions on what to do next.

[0075] (Example 1)

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

[0077] In traditional sales processes, sales representatives often make decisions based on intuition and past experience, which can lead to inefficiencies and a lack of effectiveness. Furthermore, insufficient collection and analysis of sales data can result in lower closing rates. Additionally, the absence of a system that quickly suggests appropriate actions for new sales opportunities prevents the optimization of sales activities.

[0078] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0079] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for training a machine learning model based on the preprocessed sales data, means for inputting new sales data into the machine learning model and inferring the next business action to be taken, and means for notifying the business person in charge of the inferred business action. This enables efficient collection and analysis of sales data, and optimizes sales activities by quickly suggesting appropriate actions for new sales opportunities.

[0080] "Sales data" refers to all information related to sales activities conducted by a company with its customers, including customer information, deal information, and activity history.

[0081] "Preprocessing" refers to processes such as organizing, removing duplicates, imputing missing values, and standardizing formats that are performed to improve the quality of collected data.

[0082] A "machine learning model" refers to a collection of algorithms that learn patterns and trends based on past data and use them to make predictions and classifications on new data.

[0083] "Inference" refers to the process of inputting new data into a trained machine learning model and predicting appropriate actions or outcomes based on that data.

[0084] A "sales representative" refers to a person who is engaged in sales activities with customers within a company.

[0085] "Notification" refers to the act of informing sales representatives in real time about inferred results or suggested actions.

[0086] A "customer management system" refers to a software system used by companies to manage customer information, and includes customer databases and project management functions.

[0087] A "schedule management system" refers to a system used by companies to manage various schedules and appointments, and includes features such as calendar functions and reminder functions.

[0088] A "closed deal" refers to a deal for which a contract has been concluded as a result of sales activities.

[0089] "Non-contracted deals" refer to cases where a contract was not concluded as a result of sales activities.

[0090] This invention is a system that suggests the next action to take to sales representatives based on sales data. This system is primarily implemented using the following hardware and software.

[0091] System Overview

[0092] Hardware and software:

[0093] Server: A cloud server equipped with high-performance computing power and large-capacity storage.

[0094] Customer Relationship Management (CRM): A CRM system that provides APIs, such as Salesforce.

[0095] Schedule Management System: A schedule management system that provides APIs such as Google Calendar (registered trademark).

[0096] Device: A smartphone or PC used by the sales representative.

[0097] Data collection

[0098] The server collects sales data from customer management systems and scheduling management systems via APIs. This includes obtaining case information (case name, size, progress, and assigned person) and sales activity schedules (visit date and time, meeting time, and participants). This data is temporarily stored in the server's internal database.

[0099] Data preprocessing

[0100] The server performs preprocessing on the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. For example, it standardizes the date format to "YYYY-MM-DD" and standardizes the units for contract amounts. Standardization and normalization processes are also carried out here, converting the data into a format suitable for machine learning models.

[0101] Model Learning

[0102] The server trains a machine learning model based on preprocessed data. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between the number of visits or contacts and whether a deal was successful or unsuccessful. Model performance evaluation and hyperparameter tuning are also performed during this process.

[0103] Inference and Action Suggestions

[0104] The server collects new sales data and inputs it into a machine learning model to infer the next action to take. For example, it generates specific action instructions such as, "A visit would be beneficial for this customer." This result is then communicated to the sales representative in real time.

[0105] notification

[0106] The server generates data based on the inference results to notify sales representatives of specific actions and sends it to their terminals. The terminals display the received notifications to the user via pop-ups or alerts, and the user plans and executes their next sales activities based on this information.

[0107] Specific example

[0108] Specific examples of data collection:

[0109] The server uses an API to retrieve data for "Project A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the "Customer Management System".

[0110] The server retrieves data from the "schedule management system" regarding the assigned person's "visit scheduled for 10:00 on September 1st."

[0111] Specific examples of data preprocessing:

[0112] The server deletes duplicate data for "Case A" and fills in the missing "Visit Date and Time" from previous visit records.

[0113] The server will standardize all date formats to "YYYY-MM-DD".

[0114] Specific examples of model learning:

[0115] The server uses past sales data to train a model that predicts the conversion rate using a logistic regression algorithm, learning the feature that a higher number of visits corresponds to a higher conversion rate.

[0116] Specific examples of inference and action suggestions:

[0117] The server inputs data for a new project, "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)," into the model and infers the action "This project requires 3 visits."

[0118] Specific examples of notifications:

[0119] The server generates instruction data, such as "The next action should be to visit a representative at a certain company," and sends it to the terminal.

[0120] The device notifies the sales representative that they should visit a company representative at 10:00 tomorrow.

[0121] Examples of prompts for generative AI models

[0122] Examples of prompt statements to input into a generative AI model include the following:

[0123] "Based on this month's sales performance, please suggest the next steps we should take, taking into account recent visits, emails, etc., and recommend appropriate sales activities."

[0124] In this way, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and notifying sales representatives in real time of the optimal next action to take.

[0125] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0126] Step 1:

[0127] The server collects sales data from customer relationship management (CRM) and scheduling management systems.

[0128] Specific operation: The server uses an API to send requests to the customer management system. This retrieves case information (case name, scale, progress, person in charge). It also retrieves sales activity data such as visit date and time, meeting time, and participants from the schedule management system.

[0129] Input: API endpoints and access keys for the customer management system and scheduling management system.

[0130] Output: Acquired sales data (deal information, schedule data).

[0131] Step 2:

[0132] The server performs preprocessing on the collected sales data.

[0133] Specific actions: The server removes duplicate data. If missing values ​​exist, it uses the nearest visit date and time as a reference to complete the data. To unify all data formats, it converts date formats to "YYYY-MM-DD" and standardizes currency units. Furthermore, it standardizes and normalizes the data.

[0134] Input: Collected sales data.

[0135] Output: Preprocessed data (deduplication, imputation of missing values, formatting standardization, and normalization).

[0136] Step 3:

[0137] The server trains a machine learning model based on the pre-processed data.

[0138] Specific operation: The server splits the preprocessed data into training and test data. Then, it trains a model using algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful projects. The trained model is evaluated, and hyperparameters are adjusted as needed.

[0139] Input: Preprocessed data.

[0140] Output: A trained machine learning model.

[0141] Step 4:

[0142] The server infers the next action to take based on the new sales data.

[0143] Specific operation: The server collects the latest sales opportunity data and performs preprocessing. Then, it inputs the preprocessed data into a machine learning model and performs inference. As a result of the inference, it generates specific actions to take next (e.g., "A visit would be beneficial for this customer").

[0144] Input: New sales data, trained machine learning model.

[0145] Output: Inference result (specific action to take next).

[0146] Step 5:

[0147] The server notifies the sales representative of specific actions to take based on the inference results.

[0148] Specific operation: The server generates notification data based on the inference results and sends it to the sales representative's terminal. The terminal displays the received notification to the sales representative via a pop-up or alert.

[0149] Input: Inference results, information about the sales representative's device.

[0150] Output: Notification sent to the sales representative's terminal.

[0151] Examples of specific actions

[0152] Specific examples of data collection:

[0153] The server uses an API to retrieve data for "Case A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the customer management system and data for "Scheduled visit on September 1st at 10:00" from the schedule management system.

[0154] Specific examples of data preprocessing:

[0155] The server will delete duplicate data for "Case A" and fill in the missing "Visit Date and Time" from previous visit records. All date formats will be standardized to "YYYY-MM-DD".

[0156] Specific examples of model learning:

[0157] The server uses historical sales data to train a model that predicts the conversion rate using a logistic regression algorithm, and learns the feature that a higher number of visits is associated with a higher conversion rate.

[0158] Specific examples of inference and action suggestions:

[0159] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)" into the model and infers the action "This project requires 3 visits."

[0160] Specific examples of notifications:

[0161] The server sends the generated instruction data, "The next action should be to visit a representative at a certain company," to the terminal, which then notifies the sales representative that "You should visit a representative at a certain company tomorrow at 10:00."

[0162] Thus, the system of the present invention can optimize sales activities through efficient data collection and preprocessing, inference using machine learning, and notification of appropriate actions.

[0163] (Application Example 1)

[0164] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0165] While conventional sales systems offer suggestions for optimal sales actions, systems that similarly automatically suggest maintenance for factory machinery are limited. Furthermore, many factory machines are large and complex, making proper maintenance crucial before breakdowns occur. Neglecting this can lead to unexpected downtime and costly repairs. Therefore, a system that automatically suggests the next maintenance action to take is needed to improve factory production efficiency and prevent unexpected failures.

[0166] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0167] In this invention, the server includes means for collecting sales data, means for pre-processing the sales data, means for training a machine learning model based on the pre-processed sales data, means for inputting new sales opportunity data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, means for collecting and pre-processing machine operation data in the factory and inferring maintenance actions, and means for notifying the inferred maintenance action. This makes it possible to optimize not only sales activities but also machine maintenance in the factory.

[0168] "Sales data" refers to data collected by companies during their sales activities, such as customer information, project information, and sales progress.

[0169] A "machine learning model" is an algorithm that learns patterns and rules from data and uses that knowledge to perform inferences and predictions on new data.

[0170] A "customer management system" is software that centrally manages detailed customer information, transaction history, and contact history to efficiently support sales activities.

[0171] A "schedule management system" is software that centralizes the management of individual and team schedules and tasks, supporting efficient schedule management.

[0172] "Preprocessing" refers to the process of transforming raw data, which cannot be used for analysis in its original form, into an analyzable format. This includes removing duplicate data, imputing missing values, and standardizing data formats.

[0173] "Inference" refers to making predictions and decisions based on new data using machine learning models.

[0174] "Notification" refers to the act of a system informing a user of its inference results or other important information.

[0175] "Factory operation data" refers to data including the operating status of machinery and equipment within the factory, failure history, and usage status of consumables.

[0176] "Maintenance action" refers to specific maintenance and repair actions taken to maintain the proper operation of machinery and equipment.

[0177] The system for implementing this invention is a system that suggests the appropriate next action to take based on sales data and factory operation data. The system is implemented in the following steps.

[0178] Data collection

[0179] The server collects sales data from customer management and scheduling management systems. For example, the server uses APIs to retrieve deal information and sales activity schedules. Similarly, the server collects factory operation data from sensor devices, including data such as machine uptime, failure history, and consumable usage.

[0180] Data preprocessing

[0181] The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also performs processes such as data standardization and normalization to transform the data into a format suitable for machine learning models. This improves the accuracy of the analysis.

[0182] Model Learning

[0183] The server receives pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of sales deals that have been won and those that have been lost. Similarly, it applies a similar algorithm to learn predictive maintenance actions from factory operation data. In this way, the model learns the relationship between various features and outcomes.

[0184] Inference and Action Suggestions

[0185] The server inputs new sales opportunity data and factory operation data into a machine learning model to infer the next action to take. For example, in sales, it might suggest a specific action such as "a visit would be beneficial for this customer." In the factory, it might suggest a maintenance action such as "this machine needs a part replaced."

[0186] notification

[0187] Based on the inference results, the server generates data that notifies sales representatives and maintenance teams of specific actions, and sends notifications to their terminals. This allows users to receive real-time insights from the system. For example, a sales representative might receive a notification on their smartphone or PC saying, "Next action: Visit," and the maintenance team might receive an email saying, "Next action: Replace parts."

[0188] Hardware / software used and specific examples

[0189] The hardware required includes a customer management system, a scheduling system, sensor devices, and smartphones and PCs for sales representatives. The software used will include Python, the Requests library, Pandas, Scikit-learn, and smtplib. An API will be used for data collection, and Pandas will be used for data preprocessing. A machine learning model will be built using Scikit-learn for model training, and notifications will be sent using the SMTP library.

[0190] As a concrete example, information on "Deal A (ID: A123, Size: 10 million yen, Contact Person: Tanaka)" is obtained from sales data, and the model suggests that "three visits would be effective for this customer." Similarly, information on "Machine X (ID: X456, Operating Hours: 10,000 hours, Consumable Usage: 70%)" is obtained from factory operation data, and the model suggests that "the next maintenance action should be parts replacement."

[0191] Example of a prompt

[0192] Example of a prompt message for sales: "Based on customer information and sales activity data, infer and notify us of the optimal next sales action to take."

[0193] Example of a prompt message for a factory: "Based on machine operating data (operating hours, failure history, consumable usage, etc.), infer and notify the next maintenance action to be taken."

[0194] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0195] Step 1:

[0196] The server collects sales data from customer management and scheduling management systems. It uses APIs to retrieve deal information and sales activity schedules (e.g., deal name, size, progress, visit date and time, meeting time). Input data is obtained in JSON format and converted to a Pandas DataFrame. Output is sales data in Pandas DataFrame format.

[0197] Step 2:

[0198] The server collects factory operation data from sensor devices. Using the sensor device's API, it retrieves information such as machine operating time, failure history, and consumable usage. The input data is real-time data from the sensors, and the output is operation data in Pandas DataFrame format.

[0199] Step 3:

[0200] The server preprocesses the collected sales and operational data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. It also performs standardization and normalization to transform the data into a format suitable for machine learning models. The input is each acquired data frame, and the output is the preprocessed data frame.

[0201] Step 4:

[0202] The server trains a machine learning model based on pre-processed sales data. It uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. The input is pre-processed sales data, and the output is the trained machine learning model.

[0203] Step 5:

[0204] The server trains a machine learning model based on pre-processed operational data. A similar algorithm is applied to learn predictive maintenance actions from factory operational data. The input is pre-processed operational data, and the output is the trained machine learning model.

[0205] Step 6:

[0206] The server acquires new sales opportunity data and new operational data, inputs it into a machine learning model, and infers the next action to take. The input is the newly acquired data, and the output is the inferred sales action and maintenance action.

[0207] Step 7:

[0208] The server generates notification messages based on the inferred actions and sends them to the terminals of sales representatives and maintenance teams. Sales representatives are informed of their next sales action, and maintenance teams are informed of their next maintenance action. The input is the inference result, and the output is the notification message.

[0209] Step 8:

[0210] Users (sales representatives and maintenance teams) receive notification messages sent to their devices and act according to the next required action. The input is the notification message, and the output is the actual action.

[0211] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0212] The present invention is a system that collects and analyzes user sentiment data in addition to sales data, and suggests the next action to be taken to sales representatives. It is mainly carried out through the following steps: sales data collection, data preprocessing, sentiment data collection and analysis, machine learning model training, action inference, and notification.

[0213] Data collection

[0214] 1. The server collects sales data from the customer management system and the scheduling management system. Specifically, the server uses APIs to retrieve deal information (e.g., deal ID, deal name, size, contact person ID, progress) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[0215] Data preprocessing

[0216] 2. The server cleans the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[0217] Collection of emotional data

[0218] 3. The server uses an emotion engine to recognize the user's emotions. Specifically, the emotion engine analyzes the user's emotions from sensor devices and applications, such as voice tone, facial expressions, and text, and generates emotion data.

[0219] Model Learning

[0220] 4. The server trains a machine learning model based on pre-processed sales data and sentiment data. Algorithms used for training include logistic regression, random forest, and neural networks. This model learns the relationship between features such as the number of visits, the number of phone calls, and the number of emails, and the relationships between successful and unsuccessful deals, and also takes sentiment data into consideration.

[0221] Inference and Action Suggestions

[0222] 5. The server inputs new sales opportunity data and the latest sentiment data into the model and infers the next action to take. For example, based on sentiment data, if a customer shows strong interest, the server recommends actions such as a visit or phone call to that customer. The inference result may output "immediate follow-up inspection" or "provide additional materials."

[0223] notification

[0224] 6. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive real-time instructions from the system and take appropriate action.

[0225] Specific example

[0226] 1. Data collection:

[0227] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[0228] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[0229] 2. Data preprocessing:

[0230] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[0231] 3. Collection of emotional data:

[0232] The server analyzes the user's emotions using an emotion engine and obtains emotion data such as "Customer B is very satisfied."

[0233] 4. Model Learning:

[0234] The server trains a machine learning model based on historical sales and sentiment data. For example, it learns the relationship between features that result in high order conversion rates and specific emotions (such as satisfaction or excitement).

[0235] 5. Inferences and Action Suggestions:

[0236] The server inputs new data for "Project B (ID: B456, scale: 20 million yen, person in charge: Sato)" and the latest sentiment data, "Customer B is very excited," and infers the action "This customer needs immediate follow-up."

[0237] 6. Notice:

[0238] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[0239] By following the above steps, this system not only provides specific instructions to sales representatives on what to do next, but also takes user emotions into consideration, enabling more effective sales activities.

[0240] The following describes the processing flow.

[0241] Step 1:

[0242] The server collects sales data from the customer management system and the scheduling management system. The server calls the API of the customer management system to retrieve deal information (deal ID, deal name, size, person in charge ID, progress status, etc.). It also uses the API of the scheduling management system to retrieve sales activity schedules (visit date and time, meeting time, participants, etc.).

[0243] Step 2:

[0244] The server preprocesses the collected data. First, it removes duplicate data. Next, it fills in any missing data. For example, if the visit date and time are missing, it automatically fills in the most recent visit date and time. Furthermore, it standardizes the data format. For example, it converts all dates and times to ISO 8601 format.

[0245] Step 3:

[0246] The server standardizes and normalizes the data after preprocessing. Specifically, it scales numerical data (e.g., project size) to a certain range and encodes categorical data (e.g., person in charge ID).

[0247] Step 4:

[0248] The server uses an emotion engine to collect user emotion data. Specifically, the emotion engine recognizes the user's emotions using technologies such as voice tone analysis, facial expression recognition, and text analysis, and acquires that data.

[0249] Step 5:

[0250] The server trains machine learning models using pre-processed sales data and collected sentiment data. For example, logistic regression and random forests are used to learn the relationship between the number of visits, phone calls, emails, and sentiment data and whether a deal was won or lost.

[0251] Step 6:

[0252] The server inputs new sales opportunity data and the latest sentiment data into a machine learning model to infer the next action. For example, if the sentiment data indicates that the customer is satisfied, it recommends a visit or phone call to that customer.

[0253] Step 7:

[0254] Based on the inference results, the server generates instruction data for sales representatives to take specific actions. For example, it might generate data such as, "Next, you should visit Mr. Tanaka at Company A."

[0255] Step 8:

[0256] The terminal receives instruction data from the server and sends notifications to the sales representative. For example, it displays a message such as, "You should visit Mr. Tanaka, the manager of Company A, at 10:00 tomorrow," via smartphone push notifications or PC desktop notifications.

[0257] This allows sales representatives to clearly understand their next steps and conduct sales activities more efficiently. Furthermore, inference based on emotional data enables more effective approaches.

[0258] (Example 2)

[0259] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0260] The present invention aims to improve the efficiency of sales activities and enhance customer satisfaction. Specifically, it aims to enable sales representatives to respond appropriately and in a timely manner by inferring actions that take into account not only sales data but also customer sentiment data. Furthermore, it aims to optimize sales activities based on real-time sentiment analysis, which was not possible with conventional systems.

[0261] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0262] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for collecting the sales data and additional sentiment data, means for training a machine learning model based on the preprocessed sales data and sentiment data, means for inputting new sales opportunity data and the latest sentiment data into the machine learning model and inferring the next action to be taken, and means for notifying the sales representative of the inferred action. This enables effective and timely inference and notification of actions by combining sales data and sentiment data.

[0263] "Sales data" refers to information related to sales activities, such as project information, customer information, and schedule information.

[0264] "Emotional data" refers to data that indicates a customer's emotional state, including information such as satisfaction, excitement, and anger, obtained through voice tone, facial expressions, and text analysis.

[0265] A "machine learning model" refers to a set of algorithms that learn patterns based on past data and use them to make predictions and classifications about future data.

[0266] "Preprocessing" refers to the process of preparing collected data to be suitable for analysis and machine learning, and includes removing duplicate data, filling in missing data, and standardizing data formats.

[0267] "Inference" refers to the act of predicting the next action to take or the resulting outcome based on the input data.

[0268] "Notification" refers to communicating the action taken as a result of the inference to the sales representative.

[0269] "Means of collection" refers to the methods and techniques a system uses to obtain data from a specific source.

[0270] "Training methods" refer to the process of using collected data to train machine learning models and improve the accuracy of predictions and classifications.

[0271] The system of this invention collects and analyzes sales data and user sentiment data to suggest the next course of action to sales representatives. This system primarily utilizes servers and terminals and is implemented in the following steps.

[0272] Data collection

[0273] The server collects sales data from customer relationship management (CRM) and scheduling management (calendar app) systems using APIs. API requests retrieve sales opportunity information (opportunity ID, opportunity name, size, contact person ID, progress status) and sales activity schedule information (visit date and time, meeting time, participants).

[0274] Specific example:

[0275] The server retrieves information about "Project A (ID: A123, scale: 10 million yen)" from the CRM system.

[0276] The server obtains event information scheduled as "Visit Company A at 10:00 on September 1st" from the calendar app.

[0277] Data preprocessing

[0278] The server cleans the collected data and converts it into a format suitable for input to the machine learning model. It performs operations such as deleting duplicate data, complementing missing data, and unifying data formats. Additionally, it conducts processes like standardization and normalization to improve the quality of the data.

[0279] Specific example:

[0280] The server deletes duplicate data related to "Project A" and stores it in a unified format.

[0281] Collection of sentiment data

[0282] The server collects the user's sentiment using a sentiment engine. This sentiment engine analyzes the user's sentiment through sensors and applications such as voice recognition sensors, face recognition cameras, and chatbots, and generates sentiment data such as satisfaction, excitement, and anger.

[0283] Specific example:

[0284] The server analyzes the data collected from the voice recognition sensor and obtains sentiment data indicating that "Customer B is very satisfied."

[0285] Model learning

[0286] The server trains a machine learning model based on the preprocessed sales data and sentiment data. The algorithms used include logistic regression, random forest, neural network, etc., and these are used to build a prediction model for winning and losing orders.

[0287] Specific example:

[0288] The server uses historical sales data to learn features that have a high conversion rate and their association with specific sentiment data.

[0289] Action reasoning

[0290] The server inputs new sales opportunity data and the latest sentiment data to infer the next action to take. For example, if a customer is very excited, it might generate specific actions such as recommending an urgent follow-up call to that customer.

[0291] Specific example:

[0292] Based on the data for "Project B (ID: B456, Scale: 20 million yen)" and the sentiment data that "Customer B is very excited," the server infers that "this customer requires immediate follow-up."

[0293] notification

[0294] The server generates data to notify sales representatives of the inferred actions and sends it to their devices (smartphones, PCs, etc.). Upon receiving the notification, the sales representatives can respond quickly.

[0295] Specific example:

[0296] The server generates instruction data stating, "The next action is to visit customer C," and sends a notification to the terminal saying, "Visit customer C tomorrow at 10:00."

[0297] These steps allow the system to accurately suggest the next action sales representatives should take, while also considering user emotions, thereby supporting more effective sales activities.

[0298] Example of a prompt

[0299] For example, by inputting a prompt sentence such as "What actions should be taken if Customer B is very satisfied?" to the generative AI model, it is possible to infer appropriate actions.

[0300] As described above, the system of the present invention can significantly improve the efficiency and effectiveness of sales activities through advanced analysis and action suggestions that combine sales data and sentiment data.

[0301] The flow of the specific process in Example 2 will be described using FIG. 13.

[0302] Step 1:

[0303] The server collects sales data from the customer management system and the schedule management system.

[0304] As a specific operation, the server sends an API request and obtains project information (project ID, project name, scale, person in charge ID, progress status) and schedule information (visit date and time, meeting time, participants).

[0305] Input: Access information to the customer management system and the schedule management system.

[0306] Output: The obtained sales data and schedule data.

[0307] Specific example: The server collects data of Project A (ID: A123, scale: 10 million yen) from the customer management system. Then, it obtains the information of "scheduled to visit Company A at 10:00 on September 1st" of Tanaka from the schedule management system.

[0308] Step 2:

[0309] The server preprocesses the collected sales data.

[0310] Specifically, the server performs tasks such as deleting duplicate data, filling in missing data, and unifying data formats. It also carries out processes such as standardization and normalization.

[0311] Input: Sales data and schedule data obtained in Step 1.

[0312] Output: Preprocessed data.

[0313] Specific example: The server deletes duplicate data related to project A and rewrites the scale of project A in a unified format. Additionally, if visit dates and times are missing, it automatically fills in the nearest available date and time.

[0314] Step 3:

[0315] The server collects user sentiment data.

[0316] Specifically, it uses an emotion engine to analyze the user's emotions through voice recognition sensors, facial recognition cameras, text analysis, and other means.

[0317] Input: User's voice tone, facial expressions, and text data.

[0318] Output: Analyzed sentiment data.

[0319] Specific example: The server analyzes the user's voice tone to determine that customer B is very satisfied and saves this as sentiment data.

[0320] Step 4:

[0321] The server trains a machine learning model based on pre-processed sales data and sentiment data.

[0322] In terms of specific operations, the server builds and trains models using algorithms such as logistic regression, random forest, and neural networks.

[0323] Input: Pre-processed sales data and sentiment data.

[0324] Output: Trained machine learning model.

[0325] Specific example: The server learns the correlation between features that lead to high order conversion rates and sentiment data based on past sales data, and builds a highly accurate predictive model.

[0326] Step 5:

[0327] The server uses new sales opportunity data and the latest sentiment data to infer the next action to take.

[0328] Specifically, the process involves inputting new data into a machine learning model and generating an action plan based on the results.

[0329] Input: New sales opportunity data and latest sentiment data.

[0330] Output: Inferred action plan.

[0331] Specific example: The server takes data for "Project B (ID: B456, scale: 20 million yen)" and sentiment data for "Customer B is very excited" and infers the action "immediate follow-up is necessary".

[0332] Step 6:

[0333] The server notifies the sales representative of specific actions to take based on the inference results.

[0334] Specifically, the system generates notification data for the action plan and sends it to the device.

[0335] Input: Inferred action plan.

[0336] Output: Notification data for sales representatives.

[0337] Specific example: The server generates an instruction that says, "The next action is to visit customer C," and sends a notification to the sales representative's smartphone saying, "Visit customer C tomorrow at 10:00."

[0338] (Application Example 2)

[0339] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0340] Traditional sales support systems focused on collecting and analyzing sales data, but lacked action inference that considered customer sentiment data, making efficient sales activities difficult. Similarly, in factory robot management, the inability to integrate robot operation data with staff sentiment data resulted in inadequate maintenance and management. These situations highlight the need for sales activities that consider user sentiment and efficient management of factory robots.

[0341] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0342] In this invention, the server includes means for collecting sales data, means for collecting sentiment data, means for pre-processing the sales data and sentiment data, means for training a machine learning model based on the pre-processed sales data and sentiment data, means for inputting new sales opportunity data and the latest user sentiment data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, and means for collecting factory robot operating status data and staff sentiment data and suggesting appropriate actions to the robot manager based on this. As a result, sales activities can be taken efficiently while taking customer sentiment into consideration, and in the management of factory robots, operating status data and sentiment data can be analyzed in an integrated manner to enable appropriate maintenance and management.

[0343] "Sales data" refers to information such as project details and sales activity schedules obtained from customer management systems and schedule management systems.

[0344] "Emotional data" refers to user emotional information analyzed using an emotion engine based on the user's voice tone, facial expressions, text, etc.

[0345] A "machine learning model" is an algorithm trained on collected sales and sentiment data to infer the next action to take.

[0346] A "factory robot" is a machine used to automate manufacturing processes and provides operational status data.

[0347] "Staff" refers to personnel who manage and maintain robots within the factory.

[0348] "Operating status data" refers to information such as the operating time, error rate, and performance of factory robots.

[0349] "Preprocessing" refers to the process of cleaning, standardizing, and normalizing raw data to transform it into a format suitable for machine learning models.

[0350] "Inference" is the process of using a machine learning model to input new data and derive the next action to take.

[0351] "Notification" refers to sending a message to inform sales representatives or factory robot managers of the inferred action.

[0352] The present invention provides a system for collecting and analyzing sales data and sentiment data, inferring the next course of action, and notifying sales representatives and factory robot managers. This system makes sales activities and factory robot management more efficient and effective.

[0353] Hardware and software to be used

[0354] The hardware to be used is as follows:

[0355] 1. Sensor device: A device used to collect data on the operating status of factory robots.

[0356] 2. Camera and microphone: Devices for collecting voice tones and facial expressions of factory staff and users.

[0357] The software to be used is as follows:

[0358] 1. RESTful API: An interface for collecting data from multiple systems.

[0359] 2. Python Libraries: Libraries for data preprocessing and machine learning model training (e.g., scikit-learn).

[0360] 3. HTTP Request Library: A library for sending notifications (e.g., requests).

[0361] Data collection

[0362] The server collects sales data and factory robot operating status data from related systems. Specifically, the server obtains sales opportunity information, visit schedules, robot operating hours, error rates, etc., via APIs.

[0363] Data preprocessing

[0364] The server cleans the collected data and performs standardization and normalization as needed. This makes the data suitable for machine learning models.

[0365] Collection of emotional data

[0366] The server uses an emotion engine to collect user emotion data. For example, it obtains emotional information such as "satisfied" or "excited" from the voice tone and facial expressions of factory staff.

[0367] Model Learning

[0368] The server trains machine learning models based on pre-processed sales data, sentiment data, and factory robot operating status data. These models are trained using methods such as logistic regression, random forests, and neural networks.

[0369] Inference and Action Suggestions

[0370] The server inputs new sales data, the latest sentiment data, and factory robot operating status data into the model to infer the next action to take. For example, it might suggest actions such as "immediate follow-up" for sales or "maintenance required" for factory robots.

[0371] notification

[0372] The server sends notifications to sales representatives and factory robot managers based on the inference results. For example, specific instructions such as "You need to visit the customer as the next action" or "You should perform maintenance on the machine" are sent.

[0373] Specific example

[0374] Data collection: Obtain factory robot operating status data from the API endpoint "http: / / robot_data.api".

[0375] Collection of emotional data: Emotional data obtained from the voice tone and facial expressions of factory staff.

[0376] Inference and Action Suggestions: Infer the next action to take based on new sales opportunity data.

[0377] Notification: A notification is sent to the sales representative stating, "The next action required is to visit the customer."

[0378] Example of a prompt

[0379] "Implement a system that collects robot operating status data from an API endpoint (http: / / robot_data.api), takes into account emotional data collected from voice tone and facial expressions, and suggests the next action to take. Notify the robot administrator of the appropriate action and recommend maintenance if any abnormalities are detected."

[0380] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0381] Step 1:

[0382] The server collects sales data and factory robot operating status data via APIs. Specifically, it obtains sales opportunity information, visit schedules, robot operating hours, error rates, and more. This allows the server to obtain sales data and robot operating status data as input data.

[0383] Step 2:

[0384] The server preprocesses the collected data. Specifically, it performs data cleaning (removing duplicate data, imputing missing values), standardization, and normalization. Using the collected raw data as input, it obtains preprocessed data in a format suitable for machine learning models as output.

[0385] Step 3:

[0386] The server uses an emotion engine to collect user emotion data. Specifically, it analyzes the user's emotions from their voice tone and facial expressions and generates emotion data. This requires sensor devices and audio / video data as input, and the analyzed emotion data is obtained as output.

[0387] Step 4:

[0388] The server trains a machine learning model based on pre-processed sales and sentiment data. Specifically, it uses algorithms such as logistic regression, random forest, and neural networks to learn the relationship between features and successful / unsuccessful deals. It takes pre-processed data as input and produces a trained machine learning model as output.

[0389] Step 5:

[0390] The server inputs new sales opportunity data and the latest sentiment data into a trained machine learning model to infer the next action to take. Specifically, it feeds sales data and sentiment data into the model and obtains actions such as "follow-up required" or "maintenance required" as inference results. New data is used as input, and the inference result is obtained as output.

[0391] Step 6:

[0392] The server sends notifications to sales representatives and factory robot managers based on the inference results. Specifically, it generates a message containing the appropriate action and sends it to the terminal. In this way, the inference results are used as input, and the notification message is sent to the sales representatives and managers as output.

[0393] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0394] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0395] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0396] [Second Embodiment]

[0397] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0398] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0399] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0400] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0401] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0402] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0403] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0404] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0405] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0406] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0407] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0408] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0409] The present invention provides a system for suggesting the next course of action to sales representatives based on sales data. This system is primarily implemented through the following steps.

[0410] Data collection

[0411] 1. The server collects sales data from customer relationship management (CRM) and scheduling management systems. Specifically, the server uses APIs to retrieve deal information (e.g., deal name, size, progress, person in charge) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[0412] Data preprocessing

[0413] 2. The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[0414] Model Learning

[0415] 3. The server receives the pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between various features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[0416] Inference and Action Suggestions

[0417] 4. Every time the server acquires new sales opportunity data, it inputs the data into a machine learning model to infer the next action to take. For example, the model might suggest a specific action such as "a visit would be beneficial for this customer."

[0418] notification

[0419] 5. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive insights from the system in real time.

[0420] Specific example

[0421] 1. Data collection:

[0422] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[0423] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[0424] 2. Data preprocessing:

[0425] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[0426] 3. Model learning:

[0427] The server uses a logistic regression algorithm to learn features from past data, specifically identifying that a high number of visits correlates with a high conversion rate.

[0428] 4. Inferences and Action Suggestions:

[0429] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in charge: Sato)" into the model and infers the action "This project requires 3 visits."

[0430] 5. Notice:

[0431] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[0432] As described above, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and suggesting the optimal action to sales representatives in real time.

[0433] The following describes the processing flow.

[0434] Step 1:

[0435] The server collects sales data from the customer management system and the scheduling management system. Specifically, the server calls the API of the customer management system to obtain case information (e.g., case ID, case name, size, person in charge ID, progress status). It also uses the API of the scheduling management system to obtain sales activity schedules (e.g., visit date and time, meeting time, participants).

[0436] Step 2:

[0437] The server cleans the collected data. First, it removes duplicate entries from the data. Next, it appropriately fills in any missing items. For example, if there is a missing visit date and time, it automatically fills in the nearest visit date.

[0438] Step 3:

[0439] The server standardizes and normalizes the preprocessed data. Specifically, it scales numerical data (e.g., project size) to a certain range. This transforms the data into a format suitable for machine learning models.

[0440] Step 4:

[0441] The server trains a machine learning model based on pre-processed data. Specifically, it uses data from past successful and unsuccessful deals to build and train a model using algorithms such as logistic regression and random forest. This model learns the relationship between features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[0442] Step 5:

[0443] The server inputs new case data into the model and infers the next action to take. For example, for a new case, specific actions such as the number of visits or phone calls may be suggested. The inference results will output actions such as "increase the number of visits" or "follow up with a phone call."

[0444] Step 6:

[0445] Based on the inference results, the server generates data that notifies sales representatives of specific actions to take. This data includes detailed instructions on who to contact and what actions to take.

[0446] Step 7:

[0447] The terminal receives notifications from the server and informs the sales representative. Specifically, it reports specific actions to the sales representative, such as "You should visit Mr. Yamada, the manager at Company B, next," via smartphone push notifications, PC desktop notifications, or email notifications.

[0448] Through these steps, the system significantly improves the efficiency and effectiveness of sales activities by providing sales representatives with specific instructions on what to do next.

[0449] (Example 1)

[0450] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0451] In traditional sales processes, sales representatives often make decisions based on intuition and past experience, which can lead to inefficiencies and a lack of effectiveness. Furthermore, insufficient collection and analysis of sales data can result in lower closing rates. Additionally, the absence of a system that quickly suggests appropriate actions for new sales opportunities prevents the optimization of sales activities.

[0452] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0453] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for training a machine learning model based on the preprocessed sales data, means for inputting new sales data into the machine learning model and inferring the next business action to be taken, and means for notifying the business person in charge of the inferred business action. This enables efficient collection and analysis of sales data, and optimizes sales activities by quickly suggesting appropriate actions for new sales opportunities.

[0454] "Sales data" refers to all information related to sales activities conducted by a company with its customers, including customer information, deal information, and activity history.

[0455] "Preprocessing" refers to processes such as organizing, removing duplicates, imputing missing values, and standardizing formats that are performed to improve the quality of collected data.

[0456] A "machine learning model" refers to a collection of algorithms that learn patterns and trends based on past data and use them to make predictions and classifications on new data.

[0457] "Inference" refers to the process of inputting new data into a trained machine learning model and predicting appropriate actions or outcomes based on that data.

[0458] A "sales representative" refers to a person who is engaged in sales activities with customers within a company.

[0459] "Notification" refers to the act of informing sales representatives in real time about inferred results or suggested actions.

[0460] A "customer management system" refers to a software system used by companies to manage customer information, and includes customer databases and project management functions.

[0461] A "schedule management system" refers to a system used by companies to manage various schedules and appointments, and includes features such as calendar functions and reminder functions.

[0462] A "closed deal" refers to a deal for which a contract has been concluded as a result of sales activities.

[0463] "Non-contracted deals" refer to cases where a contract was not concluded as a result of sales activities.

[0464] This invention is a system that suggests the next action to take to sales representatives based on sales data. This system is primarily implemented using the following hardware and software.

[0465] System Overview

[0466] Hardware and software:

[0467] Server: A cloud server equipped with high-performance computing power and large-capacity storage.

[0468] Customer Relationship Management (CRM): A CRM system that provides APIs, such as Salesforce.

[0469] Schedule management system: A schedule management system that provides APIs such as Google Calendar.

[0470] Device: A smartphone or PC used by the sales representative.

[0471] Data collection

[0472] The server collects sales data from customer management systems and scheduling management systems via APIs. This includes obtaining case information (case name, size, progress, and assigned person) and sales activity schedules (visit date and time, meeting time, and participants). This data is temporarily stored in the server's internal database.

[0473] Data preprocessing

[0474] The server performs preprocessing on the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. For example, it standardizes the date format to "YYYY-MM-DD" and standardizes the units for contract amounts. Standardization and normalization processes are also carried out here, converting the data into a format suitable for machine learning models.

[0475] Model Learning

[0476] The server trains a machine learning model based on preprocessed data. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between the number of visits or contacts and whether a deal was successful or unsuccessful. Model performance evaluation and hyperparameter tuning are also performed during this process.

[0477] Inference and Action Suggestions

[0478] The server collects new sales data and inputs it into a machine learning model to infer the next action to take. For example, it generates specific action instructions such as, "A visit would be beneficial for this customer." This result is then communicated to the sales representative in real time.

[0479] notification

[0480] The server generates data based on the inference results to notify sales representatives of specific actions and sends it to their terminals. The terminals display the received notifications to the user via pop-ups or alerts, and the user plans and executes their next sales activities based on this information.

[0481] Specific example

[0482] Specific examples of data collection:

[0483] The server uses an API to retrieve data for "Project A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the "Customer Management System".

[0484] The server retrieves data from the "schedule management system" regarding the assigned person's "visit scheduled for 10:00 on September 1st."

[0485] Specific examples of data preprocessing:

[0486] The server deletes duplicate data for "Case A" and fills in the missing "Visit Date and Time" from previous visit records.

[0487] The server will standardize all date formats to "YYYY-MM-DD".

[0488] Specific examples of model learning:

[0489] The server uses past sales data to train a model that predicts the conversion rate using a logistic regression algorithm, learning the feature that a higher number of visits corresponds to a higher conversion rate.

[0490] Specific examples of inference and action suggestions:

[0491] The server inputs data for a new project, "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)," into the model and infers the action "This project requires 3 visits."

[0492] Specific examples of notifications:

[0493] The server generates instruction data, such as "The next action should be to visit a representative at a certain company," and sends it to the terminal.

[0494] The device notifies the sales representative that they should visit a company representative at 10:00 tomorrow.

[0495] Examples of prompts for generative AI models

[0496] Examples of prompt statements to input into a generative AI model include the following:

[0497] "Based on this month's sales performance, please suggest the next steps we should take, taking into account recent visits, emails, etc., and recommend appropriate sales activities."

[0498] In this way, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and notifying sales representatives in real time of the optimal next action to take.

[0499] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0500] Step 1:

[0501] The server collects sales data from customer relationship management (CRM) and scheduling management systems.

[0502] Specific operation: The server uses an API to send requests to the customer management system. This retrieves case information (case name, scale, progress, person in charge). It also retrieves sales activity data such as visit date and time, meeting time, and participants from the schedule management system.

[0503] Input: API endpoints and access keys for the customer management system and scheduling management system.

[0504] Output: Acquired sales data (deal information, schedule data).

[0505] Step 2:

[0506] The server performs preprocessing on the collected sales data.

[0507] Specific actions: The server removes duplicate data. If missing values ​​exist, it uses the nearest visit date and time as a reference to complete the data. To unify all data formats, it converts date formats to "YYYY-MM-DD" and standardizes currency units. Furthermore, it standardizes and normalizes the data.

[0508] Input: Collected sales data.

[0509] Output: Preprocessed data (deduplication, imputation of missing values, formatting standardization, and normalization).

[0510] Step 3:

[0511] The server trains a machine learning model based on the pre-processed data.

[0512] Specific operation: The server splits the preprocessed data into training and test data. Then, it trains a model using algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful projects. The trained model is evaluated, and hyperparameters are adjusted as needed.

[0513] Input: Preprocessed data.

[0514] Output: A trained machine learning model.

[0515] Step 4:

[0516] The server infers the next action to take based on the new sales data.

[0517] Specific operation: The server collects the latest sales opportunity data and performs preprocessing. Then, it inputs the preprocessed data into a machine learning model and performs inference. As a result of the inference, it generates specific actions to take next (e.g., "A visit would be beneficial for this customer").

[0518] Input: New sales data, trained machine learning model.

[0519] Output: Inference result (specific action to take next).

[0520] Step 5:

[0521] The server notifies the sales representative of specific actions to take based on the inference results.

[0522] Specific operation: The server generates notification data based on the inference results and sends it to the sales representative's terminal. The terminal displays the received notification to the sales representative via a pop-up or alert.

[0523] Input: Inference results, information about the sales representative's device.

[0524] Output: Notification sent to the sales representative's terminal.

[0525] Examples of specific actions

[0526] Specific examples of data collection:

[0527] The server uses an API to retrieve data for "Case A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the customer management system and data for "Scheduled visit on September 1st at 10:00" from the schedule management system.

[0528] Specific examples of data preprocessing:

[0529] The server will delete duplicate data for "Case A" and fill in the missing "Visit Date and Time" from previous visit records. All date formats will be standardized to "YYYY-MM-DD".

[0530] Specific examples of model learning:

[0531] The server uses historical sales data to train a model that predicts the conversion rate using a logistic regression algorithm, and learns the feature that a higher number of visits is associated with a higher conversion rate.

[0532] Specific examples of inference and action suggestions:

[0533] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)" into the model and infers the action "This project requires 3 visits."

[0534] Specific examples of notifications:

[0535] The server sends the generated instruction data, "The next action should be to visit a representative at a certain company," to the terminal, which then notifies the sales representative that "You should visit a representative at a certain company tomorrow at 10:00."

[0536] Thus, the system of the present invention can optimize sales activities through efficient data collection and preprocessing, inference using machine learning, and notification of appropriate actions.

[0537] (Application Example 1)

[0538] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0539] While conventional sales systems offer suggestions for optimal sales actions, systems that similarly automatically suggest maintenance for factory machinery are limited. Furthermore, many factory machines are large and complex, making proper maintenance crucial before breakdowns occur. Neglecting this can lead to unexpected downtime and costly repairs. Therefore, a system that automatically suggests the next maintenance action to take is needed to improve factory production efficiency and prevent unexpected failures.

[0540] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0541] In this invention, the server includes means for collecting sales data, means for pre-processing the sales data, means for training a machine learning model based on the pre-processed sales data, means for inputting new sales opportunity data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, means for collecting and pre-processing machine operation data in the factory and inferring maintenance actions, and means for notifying the inferred maintenance action. This makes it possible to optimize not only sales activities but also machine maintenance in the factory.

[0542] "Sales data" refers to data collected by companies during their sales activities, such as customer information, project information, and sales progress.

[0543] A "machine learning model" is an algorithm that learns patterns and rules from data and uses that knowledge to perform inferences and predictions on new data.

[0544] A "customer management system" is software that centrally manages detailed customer information, transaction history, and contact history to efficiently support sales activities.

[0545] A "schedule management system" is software that centralizes the management of individual and team schedules and tasks, supporting efficient schedule management.

[0546] "Preprocessing" refers to the process of transforming raw data, which cannot be used for analysis in its original form, into an analyzable format. This includes removing duplicate data, imputing missing values, and standardizing data formats.

[0547] "Inference" refers to making predictions and decisions based on new data using machine learning models.

[0548] "Notification" refers to the act of a system informing a user of its inference results or other important information.

[0549] "Factory operation data" refers to data including the operating status of machinery and equipment within the factory, failure history, and usage status of consumables.

[0550] "Maintenance action" refers to specific maintenance and repair actions taken to maintain the proper operation of machinery and equipment.

[0551] The system for implementing this invention is a system that suggests the appropriate next action to take based on sales data and factory operation data. The system is implemented in the following steps.

[0552] Data collection

[0553] The server collects sales data from customer management and scheduling management systems. For example, the server uses APIs to retrieve deal information and sales activity schedules. Similarly, the server collects factory operation data from sensor devices, including data such as machine uptime, failure history, and consumable usage.

[0554] Data preprocessing

[0555] The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also performs processes such as data standardization and normalization to transform the data into a format suitable for machine learning models. This improves the accuracy of the analysis.

[0556] Model Learning

[0557] The server receives pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of sales deals that have been won and those that have been lost. Similarly, it applies a similar algorithm to learn predictive maintenance actions from factory operation data. In this way, the model learns the relationship between various features and outcomes.

[0558] Inference and Action Suggestions

[0559] The server inputs new sales opportunity data and factory operation data into a machine learning model to infer the next action to take. For example, in sales, it might suggest a specific action such as "a visit would be beneficial for this customer." In the factory, it might suggest a maintenance action such as "this machine needs a part replaced."

[0560] notification

[0561] Based on the inference results, the server generates data that notifies sales representatives and maintenance teams of specific actions, and sends notifications to their terminals. This allows users to receive real-time insights from the system. For example, a sales representative might receive a notification on their smartphone or PC saying, "Next action: Visit," and the maintenance team might receive an email saying, "Next action: Replace parts."

[0562] Hardware / software used and specific examples

[0563] The hardware required includes a customer management system, a scheduling system, sensor devices, and smartphones and PCs for sales representatives. The software used will include Python, the Requests library, Pandas, Scikit-learn, and smtplib. An API will be used for data collection, and Pandas will be used for data preprocessing. A machine learning model will be built using Scikit-learn for model training, and notifications will be sent using the SMTP library.

[0564] As a concrete example, information on "Deal A (ID: A123, Size: 10 million yen, Contact Person: Tanaka)" is obtained from sales data, and the model suggests that "three visits would be effective for this customer." Similarly, information on "Machine X (ID: X456, Operating Hours: 10,000 hours, Consumable Usage: 70%)" is obtained from factory operation data, and the model suggests that "the next maintenance action should be parts replacement."

[0565] Example of a prompt

[0566] Example of a prompt message for sales: "Based on customer information and sales activity data, infer and notify us of the optimal next sales action to take."

[0567] Example of a prompt message for a factory: "Based on machine operating data (operating hours, failure history, consumable usage, etc.), infer and notify the next maintenance action to be taken."

[0568] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0569] Step 1:

[0570] The server collects sales data from customer management and scheduling management systems. It uses APIs to retrieve deal information and sales activity schedules (e.g., deal name, size, progress, visit date and time, meeting time). Input data is obtained in JSON format and converted to a Pandas DataFrame. Output is sales data in Pandas DataFrame format.

[0571] Step 2:

[0572] The server collects factory operation data from sensor devices. Using the sensor device's API, it retrieves information such as machine operating time, failure history, and consumable usage. The input data is real-time data from the sensors, and the output is operation data in Pandas DataFrame format.

[0573] Step 3:

[0574] The server preprocesses the collected sales and operational data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. It also performs standardization and normalization to transform the data into a format suitable for machine learning models. The input is each acquired data frame, and the output is the preprocessed data frame.

[0575] Step 4:

[0576] The server trains a machine learning model based on pre-processed sales data. It uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. The input is pre-processed sales data, and the output is the trained machine learning model.

[0577] Step 5:

[0578] The server trains a machine learning model based on pre-processed operational data. A similar algorithm is applied to learn predictive maintenance actions from factory operational data. The input is pre-processed operational data, and the output is the trained machine learning model.

[0579] Step 6:

[0580] The server acquires new sales opportunity data and new operational data, inputs it into a machine learning model, and infers the next action to take. The input is the newly acquired data, and the output is the inferred sales action and maintenance action.

[0581] Step 7:

[0582] The server generates notification messages based on the inferred actions and sends them to the terminals of sales representatives and maintenance teams. Sales representatives are informed of their next sales action, and maintenance teams are informed of their next maintenance action. The input is the inference result, and the output is the notification message.

[0583] Step 8:

[0584] Users (sales representatives and maintenance teams) receive notification messages sent to their devices and act according to the next required action. The input is the notification message, and the output is the actual action.

[0585] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0586] The present invention is a system that collects and analyzes user sentiment data in addition to sales data, and suggests the next action to be taken to sales representatives. It is mainly carried out through the following steps: sales data collection, data preprocessing, sentiment data collection and analysis, machine learning model training, action inference, and notification.

[0587] Data collection

[0588] 1. The server collects sales data from the customer management system and the scheduling management system. Specifically, the server uses APIs to retrieve deal information (e.g., deal ID, deal name, size, contact person ID, progress) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[0589] Data preprocessing

[0590] 2. The server cleans the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[0591] Collection of emotional data

[0592] 3. The server uses an emotion engine to recognize the user's emotions. Specifically, the emotion engine analyzes the user's emotions from sensor devices and applications, such as voice tone, facial expressions, and text, and generates emotion data.

[0593] Model Learning

[0594] 4. The server trains a machine learning model based on pre-processed sales data and sentiment data. Algorithms used for training include logistic regression, random forest, and neural networks. This model learns the relationship between features such as the number of visits, the number of phone calls, and the number of emails, and the relationships between successful and unsuccessful deals, and also takes sentiment data into consideration.

[0595] Inference and Action Suggestions

[0596] 5. The server inputs new sales opportunity data and the latest sentiment data into the model and infers the next action to take. For example, based on sentiment data, if a customer shows strong interest, the server recommends actions such as a visit or phone call to that customer. The inference result may output "immediate follow-up inspection" or "provide additional materials."

[0597] notification

[0598] 6. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive real-time instructions from the system and take appropriate action.

[0599] Specific example

[0600] 1. Data collection:

[0601] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[0602] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[0603] 2. Data preprocessing:

[0604] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[0605] 3. Collection of emotional data:

[0606] The server analyzes the user's emotions using an emotion engine and obtains emotion data such as "Customer B is very satisfied."

[0607] 4. Model Learning:

[0608] The server trains a machine learning model based on historical sales and sentiment data. For example, it learns the relationship between features that result in high order conversion rates and specific emotions (such as satisfaction or excitement).

[0609] 5. Inferences and Action Suggestions:

[0610] The server inputs new data for "Project B (ID: B456, scale: 20 million yen, person in charge: Sato)" and the latest sentiment data, "Customer B is very excited," and infers the action "This customer needs immediate follow-up."

[0611] 6. Notice:

[0612] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[0613] By following the above steps, this system not only provides specific instructions to sales representatives on what to do next, but also takes user emotions into consideration, enabling more effective sales activities.

[0614] The following describes the processing flow.

[0615] Step 1:

[0616] The server collects sales data from the customer management system and the scheduling management system. The server calls the API of the customer management system to retrieve deal information (deal ID, deal name, size, person in charge ID, progress status, etc.). It also uses the API of the scheduling management system to retrieve sales activity schedules (visit date and time, meeting time, participants, etc.).

[0617] Step 2:

[0618] The server preprocesses the collected data. First, it removes duplicate data. Next, it fills in any missing data. For example, if the visit date and time are missing, it automatically fills in the most recent visit date and time. Furthermore, it standardizes the data format. For example, it converts all dates and times to ISO 8601 format.

[0619] Step 3:

[0620] The server standardizes and normalizes the data after preprocessing. Specifically, it scales numerical data (e.g., project size) to a certain range and encodes categorical data (e.g., person in charge ID).

[0621] Step 4:

[0622] The server uses an emotion engine to collect user emotion data. Specifically, the emotion engine recognizes the user's emotions using technologies such as voice tone analysis, facial expression recognition, and text analysis, and acquires that data.

[0623] Step 5:

[0624] The server trains machine learning models using pre-processed sales data and collected sentiment data. For example, logistic regression and random forests are used to learn the relationship between the number of visits, phone calls, emails, and sentiment data and whether a deal was won or lost.

[0625] Step 6:

[0626] The server inputs new sales opportunity data and the latest sentiment data into a machine learning model to infer the next action. For example, if the sentiment data indicates that the customer is satisfied, it recommends a visit or phone call to that customer.

[0627] Step 7:

[0628] Based on the inference results, the server generates instruction data for sales representatives to take specific actions. For example, it might generate data such as, "Next, you should visit Mr. Tanaka at Company A."

[0629] Step 8:

[0630] The terminal receives instruction data from the server and sends notifications to the sales representative. For example, it displays a message such as, "You should visit Mr. Tanaka, the manager of Company A, at 10:00 tomorrow," via smartphone push notifications or PC desktop notifications.

[0631] This allows sales representatives to clearly understand their next steps and conduct sales activities more efficiently. Furthermore, inference based on emotional data enables more effective approaches.

[0632] (Example 2)

[0633] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0634] The present invention aims to improve the efficiency of sales activities and enhance customer satisfaction. Specifically, it aims to enable sales representatives to respond appropriately and in a timely manner by inferring actions that take into account not only sales data but also customer sentiment data. Furthermore, it aims to optimize sales activities based on real-time sentiment analysis, which was not possible with conventional systems.

[0635] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0636] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for collecting the sales data and additional sentiment data, means for training a machine learning model based on the preprocessed sales data and sentiment data, means for inputting new sales opportunity data and the latest sentiment data into the machine learning model and inferring the next action to be taken, and means for notifying the sales representative of the inferred action. This enables effective and timely inference and notification of actions by combining sales data and sentiment data.

[0637] "Sales data" refers to information related to sales activities, such as project information, customer information, and schedule information.

[0638] "Emotional data" refers to data that indicates a customer's emotional state, including information such as satisfaction, excitement, and anger, obtained through voice tone, facial expressions, and text analysis.

[0639] A "machine learning model" refers to a set of algorithms that learn patterns based on past data and use them to make predictions and classifications about future data.

[0640] "Preprocessing" refers to the process of preparing collected data to be suitable for analysis and machine learning, and includes removing duplicate data, filling in missing data, and standardizing data formats.

[0641] "Inference" refers to the act of predicting the next action to take or the resulting outcome based on the input data.

[0642] "Notification" refers to communicating the action taken as a result of the inference to the sales representative.

[0643] "Means of collection" refers to the methods and techniques a system uses to obtain data from a specific source.

[0644] "Training methods" refer to the process of using collected data to train machine learning models and improve the accuracy of predictions and classifications.

[0645] The system of this invention collects and analyzes sales data and user sentiment data to suggest the next course of action to sales representatives. This system primarily utilizes servers and terminals and is implemented in the following steps.

[0646] Data collection

[0647] The server collects sales data from customer relationship management (CRM) and scheduling management (calendar app) systems using APIs. API requests retrieve sales opportunity information (opportunity ID, opportunity name, size, contact person ID, progress status) and sales activity schedule information (visit date and time, meeting time, participants).

[0648] Specific example:

[0649] The server retrieves information about "Project A (ID: A123, scale: 10 million yen)" from the CRM system.

[0650] The server retrieves event information from the calendar app, indicating that a visit to Company A is scheduled for September 1st at 10:00 AM.

[0651] Data preprocessing

[0652] The server cleans the collected data and converts it into a format suitable for input to machine learning models. It removes duplicate data, imputes missing data, and standardizes data formats. Furthermore, it performs processes such as standardization and normalization to improve data quality.

[0653] Specific example:

[0654] The server removes duplicate data related to "Case A" and saves it in a unified format.

[0655] Collection of emotional data

[0656] The server uses an emotion engine to collect user emotions. This emotion engine analyzes user emotions through sensors and applications such as voice recognition sensors, facial recognition cameras, and chatbots, and generates emotion data such as satisfaction, excitement, and anger.

[0657] Specific example:

[0658] The server analyzes the data collected from the voice recognition sensor and obtains sentiment data indicating that "Customer B is very satisfied."

[0659] Model Learning

[0660] The server trains machine learning models based on pre-processed sales and sentiment data. The algorithms used include logistic regression, random forests, and neural networks, which are used to build predictive models for winning and losing deals.

[0661] Specific example:

[0662] The server uses historical sales data to learn features that have a high conversion rate and their association with specific sentiment data.

[0663] Action reasoning

[0664] The server inputs new sales opportunity data and the latest sentiment data to infer the next action to take. For example, if a customer is very excited, it might generate specific actions such as recommending an urgent follow-up call to that customer.

[0665] Specific example:

[0666] Based on the data for "Project B (ID: B456, Scale: 20 million yen)" and the sentiment data that "Customer B is very excited," the server infers that "this customer requires immediate follow-up."

[0667] notification

[0668] The server generates data to notify sales representatives of the inferred actions and sends it to their devices (smartphones, PCs, etc.). Upon receiving the notification, the sales representatives can respond quickly.

[0669] Specific example:

[0670] The server generates instruction data stating, "The next action is to visit customer C," and sends a notification to the terminal saying, "Visit customer C tomorrow at 10:00."

[0671] These steps allow the system to accurately suggest the next action sales representatives should take, while also considering user emotions, thereby supporting more effective sales activities.

[0672] Example of a prompt

[0673] For example, by inputting a prompt such as "If customer B is very satisfied, what action should be taken?" into the generative AI model, it is possible to infer the appropriate action.

[0674] As described above, the system of the present invention can significantly improve the efficiency and effectiveness of sales activities through advanced analysis and action suggestions that combine sales data and sentiment data.

[0675] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0676] Step 1:

[0677] The server collects sales data from the customer management system and the scheduling management system.

[0678] Specifically, the server sends an API request to retrieve case information (case ID, case name, scale, person in charge ID, progress status) and schedule information (visit date and time, meeting time, participants).

[0679] Input: Access information for the customer management system and scheduling management system.

[0680] Output: Acquired sales data and schedule data.

[0681] Specific example: The server collects data for project A (ID: A123, scale: 10 million yen) from the customer management system. Next, it retrieves information from the schedule management system that Tanaka is scheduled to visit Company A at 10:00 on September 1st.

[0682] Step 2:

[0683] The server preprocesses the collected sales data.

[0684] Specifically, the server performs tasks such as deleting duplicate data, filling in missing data, and unifying data formats. It also carries out processes such as standardization and normalization.

[0685] Input: Sales data and schedule data obtained in Step 1.

[0686] Output: Preprocessed data.

[0687] Specific example: The server deletes duplicate data related to project A and rewrites the scale of project A in a unified format. Additionally, if visit dates and times are missing, it automatically fills in the nearest available date and time.

[0688] Step 3:

[0689] The server collects user sentiment data.

[0690] Specifically, it uses an emotion engine to analyze the user's emotions through voice recognition sensors, facial recognition cameras, text analysis, and other means.

[0691] Input: User's voice tone, facial expressions, and text data.

[0692] Output: Analyzed sentiment data.

[0693] Specific example: The server analyzes the user's voice tone to determine that customer B is very satisfied and saves this as sentiment data.

[0694] Step 4:

[0695] The server trains a machine learning model based on pre-processed sales data and sentiment data.

[0696] In terms of specific operations, the server builds and trains models using algorithms such as logistic regression, random forest, and neural networks.

[0697] Input: Pre-processed sales data and sentiment data.

[0698] Output: Trained machine learning model.

[0699] Specific example: The server learns the correlation between features that lead to high order conversion rates and sentiment data based on past sales data, and builds a highly accurate predictive model.

[0700] Step 5:

[0701] The server uses new sales opportunity data and the latest sentiment data to infer the next action to take.

[0702] Specifically, the process involves inputting new data into a machine learning model and generating an action plan based on the results.

[0703] Input: New sales opportunity data and latest sentiment data.

[0704] Output: Inferred action plan.

[0705] Specific example: The server takes data for "Project B (ID: B456, scale: 20 million yen)" and sentiment data for "Customer B is very excited" and infers the action "immediate follow-up is necessary".

[0706] Step 6:

[0707] The server notifies the sales representative of specific actions to take based on the inference results.

[0708] Specifically, the system generates notification data for the action plan and sends it to the device.

[0709] Input: Inferred action plan.

[0710] Output: Notification data for sales representatives.

[0711] Specific example: The server generates an instruction that says, "The next action is to visit customer C," and sends a notification to the sales representative's smartphone saying, "Visit customer C tomorrow at 10:00."

[0712] (Application Example 2)

[0713] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0714] Traditional sales support systems focused on collecting and analyzing sales data, but lacked action inference that considered customer sentiment data, making efficient sales activities difficult. Similarly, in factory robot management, the inability to integrate robot operation data with staff sentiment data resulted in inadequate maintenance and management. These situations highlight the need for sales activities that consider user sentiment and efficient management of factory robots.

[0715] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0716] In this invention, the server includes means for collecting sales data, means for collecting sentiment data, means for pre-processing the sales data and sentiment data, means for training a machine learning model based on the pre-processed sales data and sentiment data, means for inputting new sales opportunity data and the latest user sentiment data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, and means for collecting factory robot operating status data and staff sentiment data and suggesting appropriate actions to the robot manager based on this. As a result, sales activities can be taken efficiently while taking customer sentiment into consideration, and in the management of factory robots, operating status data and sentiment data can be analyzed in an integrated manner to enable appropriate maintenance and management.

[0717] "Sales data" refers to information such as project details and sales activity schedules obtained from customer management systems and schedule management systems.

[0718] "Emotional data" refers to user emotional information analyzed using an emotion engine based on the user's voice tone, facial expressions, text, etc.

[0719] A "machine learning model" is an algorithm trained on collected sales and sentiment data to infer the next action to take.

[0720] A "factory robot" is a machine used to automate manufacturing processes and provides operational status data.

[0721] "Staff" refers to personnel who manage and maintain robots within the factory.

[0722] "Operating status data" refers to information such as the operating time, error rate, and performance of factory robots.

[0723] "Preprocessing" refers to the process of cleaning, standardizing, and normalizing raw data to transform it into a format suitable for machine learning models.

[0724] "Inference" is the process of using a machine learning model to input new data and derive the next action to take.

[0725] "Notification" refers to sending a message to inform sales representatives or factory robot managers of the inferred action.

[0726] The present invention provides a system for collecting and analyzing sales data and sentiment data, inferring the next course of action, and notifying sales representatives and factory robot managers. This system makes sales activities and factory robot management more efficient and effective.

[0727] Hardware and software to be used

[0728] The hardware to be used is as follows:

[0729] 1. Sensor device: A device used to collect data on the operating status of factory robots.

[0730] 2. Camera and microphone: Devices for collecting voice tones and facial expressions of factory staff and users.

[0731] The software to be used is as follows:

[0732] 1. RESTful API: An interface for collecting data from multiple systems.

[0733] 2. Python Libraries: Libraries for data preprocessing and machine learning model training (e.g., scikit-learn).

[0734] 3. HTTP Request Library: A library for sending notifications (e.g., requests).

[0735] Data collection

[0736] The server collects sales data and factory robot operating status data from related systems. Specifically, the server obtains sales opportunity information, visit schedules, robot operating hours, error rates, etc., via APIs.

[0737] Data preprocessing

[0738] The server cleans the collected data and performs standardization and normalization as needed. This makes the data suitable for machine learning models.

[0739] Collection of emotional data

[0740] The server uses an emotion engine to collect user emotion data. For example, it obtains emotional information such as "satisfied" or "excited" from the voice tone and facial expressions of factory staff.

[0741] Model Learning

[0742] The server trains machine learning models based on pre-processed sales data, sentiment data, and factory robot operating status data. These models are trained using methods such as logistic regression, random forests, and neural networks.

[0743] Inference and Action Suggestions

[0744] The server inputs new sales data, the latest sentiment data, and factory robot operating status data into the model to infer the next action to take. For example, it might suggest actions such as "immediate follow-up" for sales or "maintenance required" for factory robots.

[0745] notification

[0746] The server sends notifications to sales representatives and factory robot managers based on the inference results. For example, specific instructions such as "You need to visit the customer as the next action" or "You should perform maintenance on the machine" are sent.

[0747] Specific example

[0748] Data collection: Obtain factory robot operating status data from the API endpoint "http: / / robot_data.api".

[0749] Collection of emotional data: Emotional data obtained from the voice tone and facial expressions of factory staff.

[0750] Inference and Action Suggestions: Infer the next action to take based on new sales opportunity data.

[0751] Notification: A notification is sent to the sales representative stating, "The next action required is to visit the customer."

[0752] Example of a prompt

[0753] "Implement a system that collects robot operating status data from an API endpoint (http: / / robot_data.api), takes into account emotional data collected from voice tone and facial expressions, and suggests the next action to take. Notify the robot administrator of the appropriate action and recommend maintenance if any abnormalities are detected."

[0754] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0755] Step 1:

[0756] The server collects sales data and factory robot operating status data via APIs. Specifically, it obtains sales opportunity information, visit schedules, robot operating hours, error rates, and more. This allows the server to obtain sales data and robot operating status data as input data.

[0757] Step 2:

[0758] The server preprocesses the collected data. Specifically, it performs data cleaning (removing duplicate data, imputing missing values), standardization, and normalization. Using the collected raw data as input, it obtains preprocessed data in a format suitable for machine learning models as output.

[0759] Step 3:

[0760] The server uses an emotion engine to collect user emotion data. Specifically, it analyzes the user's emotions from their voice tone and facial expressions and generates emotion data. This requires sensor devices and audio / video data as input, and the analyzed emotion data is obtained as output.

[0761] Step 4:

[0762] The server trains a machine learning model based on pre-processed sales and sentiment data. Specifically, it uses algorithms such as logistic regression, random forest, and neural networks to learn the relationship between features and successful / unsuccessful deals. It takes pre-processed data as input and produces a trained machine learning model as output.

[0763] Step 5:

[0764] The server inputs new sales opportunity data and the latest sentiment data into a trained machine learning model to infer the next action to take. Specifically, it feeds sales data and sentiment data into the model and obtains actions such as "follow-up required" or "maintenance required" as inference results. New data is used as input, and the inference result is obtained as output.

[0765] Step 6:

[0766] The server sends notifications to sales representatives and factory robot managers based on the inference results. Specifically, it generates a message containing the appropriate action and sends it to the terminal. In this way, the inference results are used as input, and the notification message is sent to the sales representatives and managers as output.

[0767] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0768] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0769] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0770] [Third Embodiment]

[0771] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0772] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0773] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0774] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0775] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0776] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0777] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0778] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0779] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0780] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0781] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0782] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0783] The present invention provides a system for suggesting the next course of action to sales representatives based on sales data. This system is primarily implemented through the following steps.

[0784] Data collection

[0785] 1. The server collects sales data from customer relationship management (CRM) and scheduling management systems. Specifically, the server uses APIs to retrieve deal information (e.g., deal name, size, progress, person in charge) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[0786] Data preprocessing

[0787] 2. The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[0788] Model Learning

[0789] 3. The server receives the pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between various features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[0790] Inference and Action Suggestions

[0791] 4. Every time the server acquires new sales opportunity data, it inputs the data into a machine learning model to infer the next action to take. For example, the model might suggest a specific action such as "a visit would be beneficial for this customer."

[0792] notification

[0793] 5. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive insights from the system in real time.

[0794] Specific example

[0795] 1. Data collection:

[0796] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[0797] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[0798] 2. Data preprocessing:

[0799] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[0800] 3. Model learning:

[0801] The server uses a logistic regression algorithm to learn features from past data, specifically identifying that a high number of visits correlates with a high conversion rate.

[0802] 4. Inferences and Action Suggestions:

[0803] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in charge: Sato)" into the model and infers the action "This project requires 3 visits."

[0804] 5. Notice:

[0805] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[0806] As described above, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and suggesting the optimal action to sales representatives in real time.

[0807] The following describes the processing flow.

[0808] Step 1:

[0809] The server collects sales data from the customer management system and the scheduling management system. Specifically, the server calls the API of the customer management system to obtain case information (e.g., case ID, case name, size, person in charge ID, progress status). It also uses the API of the scheduling management system to obtain sales activity schedules (e.g., visit date and time, meeting time, participants).

[0810] Step 2:

[0811] The server cleans the collected data. First, it removes duplicate entries from the data. Next, it appropriately fills in any missing items. For example, if there is a missing visit date and time, it automatically fills in the nearest visit date.

[0812] Step 3:

[0813] The server standardizes and normalizes the preprocessed data. Specifically, it scales numerical data (e.g., project size) to a certain range. This transforms the data into a format suitable for machine learning models.

[0814] Step 4:

[0815] The server trains a machine learning model based on pre-processed data. Specifically, it uses data from past successful and unsuccessful deals to build and train a model using algorithms such as logistic regression and random forest. This model learns the relationship between features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[0816] Step 5:

[0817] The server inputs new case data into the model and infers the next action to take. For example, for a new case, specific actions such as the number of visits or phone calls may be suggested. The inference results will output actions such as "increase the number of visits" or "follow up with a phone call."

[0818] Step 6:

[0819] Based on the inference results, the server generates data that notifies sales representatives of specific actions to take. This data includes detailed instructions on who to contact and what actions to take.

[0820] Step 7:

[0821] The terminal receives notifications from the server and informs the sales representative. Specifically, it reports specific actions to the sales representative, such as "You should visit Mr. Yamada, the manager at Company B, next," via smartphone push notifications, PC desktop notifications, or email notifications.

[0822] Through these steps, the system significantly improves the efficiency and effectiveness of sales activities by providing sales representatives with specific instructions on what to do next.

[0823] (Example 1)

[0824] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0825] In traditional sales processes, sales representatives often make decisions based on intuition and past experience, which can lead to inefficiencies and a lack of effectiveness. Furthermore, insufficient collection and analysis of sales data can result in lower closing rates. Additionally, the absence of a system that quickly suggests appropriate actions for new sales opportunities prevents the optimization of sales activities.

[0826] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0827] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for training a machine learning model based on the preprocessed sales data, means for inputting new sales data into the machine learning model and inferring the next business action to be taken, and means for notifying the business person in charge of the inferred business action. This enables efficient collection and analysis of sales data, and optimizes sales activities by quickly suggesting appropriate actions for new sales opportunities.

[0828] "Sales data" refers to all information related to sales activities conducted by a company with its customers, including customer information, deal information, and activity history.

[0829] "Preprocessing" refers to processes such as organizing, removing duplicates, imputing missing values, and standardizing formats that are performed to improve the quality of collected data.

[0830] A "machine learning model" refers to a collection of algorithms that learn patterns and trends based on past data and use them to make predictions and classifications on new data.

[0831] "Inference" refers to the process of inputting new data into a trained machine learning model and predicting appropriate actions or outcomes based on that data.

[0832] A "sales representative" refers to a person who is engaged in sales activities with customers within a company.

[0833] "Notification" refers to the act of informing sales representatives in real time about inferred results or suggested actions.

[0834] A "customer management system" refers to a software system used by companies to manage customer information, and includes customer databases and project management functions.

[0835] A "schedule management system" refers to a system used by companies to manage various schedules and appointments, and includes features such as calendar functions and reminder functions.

[0836] A "closed deal" refers to a deal for which a contract has been concluded as a result of sales activities.

[0837] "Non-contracted deals" refer to cases where a contract was not concluded as a result of sales activities.

[0838] This invention is a system that suggests the next action to take to sales representatives based on sales data. This system is primarily implemented using the following hardware and software.

[0839] System Overview

[0840] Hardware and software:

[0841] Server: A cloud server equipped with high-performance computing power and large-capacity storage.

[0842] Customer Relationship Management (CRM): A CRM system that provides APIs, such as Salesforce.

[0843] Schedule management system: A schedule management system that provides APIs such as Google Calendar.

[0844] Device: A smartphone or PC used by the sales representative.

[0845] Data collection

[0846] The server collects sales data from customer management systems and scheduling management systems via APIs. This includes obtaining case information (case name, size, progress, and assigned person) and sales activity schedules (visit date and time, meeting time, and participants). This data is temporarily stored in the server's internal database.

[0847] Data preprocessing

[0848] The server performs preprocessing on the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. For example, it standardizes the date format to "YYYY-MM-DD" and standardizes the units for contract amounts. Standardization and normalization processes are also carried out here, converting the data into a format suitable for machine learning models.

[0849] Model Learning

[0850] The server trains a machine learning model based on preprocessed data. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between the number of visits or contacts and whether a deal was successful or unsuccessful. Model performance evaluation and hyperparameter tuning are also performed during this process.

[0851] Inference and Action Suggestions

[0852] The server collects new sales data and inputs it into a machine learning model to infer the next action to take. For example, it generates specific action instructions such as, "A visit would be beneficial for this customer." This result is then communicated to the sales representative in real time.

[0853] notification

[0854] The server generates data based on the inference results to notify sales representatives of specific actions and sends it to their terminals. The terminals display the received notifications to the user via pop-ups or alerts, and the user plans and executes their next sales activities based on this information.

[0855] Specific example

[0856] Specific examples of data collection:

[0857] The server uses an API to retrieve data for "Project A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the "Customer Management System".

[0858] The server retrieves data from the "schedule management system" regarding the assigned person's "visit scheduled for 10:00 on September 1st."

[0859] Specific examples of data preprocessing:

[0860] The server deletes duplicate data for "Case A" and fills in the missing "Visit Date and Time" from previous visit records.

[0861] The server will standardize all date formats to "YYYY-MM-DD".

[0862] Specific examples of model learning:

[0863] The server uses past sales data to train a model that predicts the conversion rate using a logistic regression algorithm, learning the feature that a higher number of visits corresponds to a higher conversion rate.

[0864] Specific examples of inference and action suggestions:

[0865] The server inputs data for a new project, "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)," into the model and infers the action "This project requires 3 visits."

[0866] Specific examples of notifications:

[0867] The server generates instruction data, such as "The next action should be to visit a representative at a certain company," and sends it to the terminal.

[0868] The device notifies the sales representative that they should visit a company representative at 10:00 tomorrow.

[0869] Examples of prompts for generative AI models

[0870] Examples of prompt statements to input into a generative AI model include the following:

[0871] "Based on this month's sales performance, please suggest the next steps we should take, taking into account recent visits, emails, etc., and recommend appropriate sales activities."

[0872] In this way, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and notifying sales representatives in real time of the optimal next action to take.

[0873] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0874] Step 1:

[0875] The server collects sales data from customer relationship management (CRM) and scheduling management systems.

[0876] Specific operation: The server uses an API to send requests to the customer management system. This retrieves case information (case name, scale, progress, person in charge). It also retrieves sales activity data such as visit date and time, meeting time, and participants from the schedule management system.

[0877] Input: API endpoints and access keys for the customer management system and scheduling management system.

[0878] Output: Acquired sales data (deal information, schedule data).

[0879] Step 2:

[0880] The server performs preprocessing on the collected sales data.

[0881] Specific actions: The server removes duplicate data. If missing values ​​exist, it uses the nearest visit date and time as a reference to complete the data. To unify all data formats, it converts date formats to "YYYY-MM-DD" and standardizes currency units. Furthermore, it standardizes and normalizes the data.

[0882] Input: Collected sales data.

[0883] Output: Preprocessed data (deduplication, imputation of missing values, formatting standardization, and normalization).

[0884] Step 3:

[0885] The server trains a machine learning model based on the pre-processed data.

[0886] Specific operation: The server splits the preprocessed data into training and test data. Then, it trains a model using algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful projects. The trained model is evaluated, and hyperparameters are adjusted as needed.

[0887] Input: Preprocessed data.

[0888] Output: A trained machine learning model.

[0889] Step 4:

[0890] The server infers the next action to take based on the new sales data.

[0891] Specific operation: The server collects the latest sales opportunity data and performs preprocessing. Then, it inputs the preprocessed data into a machine learning model and performs inference. As a result of the inference, it generates specific actions to take next (e.g., "A visit would be beneficial for this customer").

[0892] Input: New sales data, trained machine learning model.

[0893] Output: Inference result (specific action to take next).

[0894] Step 5:

[0895] The server notifies the sales representative of specific actions to take based on the inference results.

[0896] Specific operation: The server generates notification data based on the inference results and sends it to the sales representative's terminal. The terminal displays the received notification to the sales representative via a pop-up or alert.

[0897] Input: Inference results, information about the sales representative's device.

[0898] Output: Notification sent to the sales representative's terminal.

[0899] Examples of specific actions

[0900] Specific examples of data collection:

[0901] The server uses an API to retrieve data for "Case A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the customer management system and data for "Scheduled visit on September 1st at 10:00" from the schedule management system.

[0902] Specific examples of data preprocessing:

[0903] The server will delete duplicate data for "Case A" and fill in the missing "Visit Date and Time" from previous visit records. All date formats will be standardized to "YYYY-MM-DD".

[0904] Specific examples of model learning:

[0905] The server uses historical sales data to train a model that predicts the conversion rate using a logistic regression algorithm, and learns the feature that a higher number of visits is associated with a higher conversion rate.

[0906] Specific examples of inference and action suggestions:

[0907] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)" into the model and infers the action "This project requires 3 visits."

[0908] Specific examples of notifications:

[0909] The server sends the generated instruction data, "The next action should be to visit a representative at a certain company," to the terminal, which then notifies the sales representative that "You should visit a representative at a certain company tomorrow at 10:00."

[0910] Thus, the system of the present invention can optimize sales activities through efficient data collection and preprocessing, inference using machine learning, and notification of appropriate actions.

[0911] (Application Example 1)

[0912] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0913] While conventional sales systems offer suggestions for optimal sales actions, systems that similarly automatically suggest maintenance for factory machinery are limited. Furthermore, many factory machines are large and complex, making proper maintenance crucial before breakdowns occur. Neglecting this can lead to unexpected downtime and costly repairs. Therefore, a system that automatically suggests the next maintenance action to take is needed to improve factory production efficiency and prevent unexpected failures.

[0914] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0915] In this invention, the server includes means for collecting sales data, means for pre-processing the sales data, means for training a machine learning model based on the pre-processed sales data, means for inputting new sales opportunity data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, means for collecting and pre-processing machine operation data in the factory and inferring maintenance actions, and means for notifying the inferred maintenance action. This makes it possible to optimize not only sales activities but also machine maintenance in the factory.

[0916] "Sales data" refers to data collected by companies during their sales activities, such as customer information, project information, and sales progress.

[0917] A "machine learning model" is an algorithm that learns patterns and rules from data and uses that knowledge to perform inferences and predictions on new data.

[0918] A "customer management system" is software that centrally manages detailed customer information, transaction history, and contact history to efficiently support sales activities.

[0919] A "schedule management system" is software that centralizes the management of individual and team schedules and tasks, supporting efficient schedule management.

[0920] "Preprocessing" refers to the process of transforming raw data, which cannot be used for analysis in its original form, into an analyzable format. This includes removing duplicate data, imputing missing values, and standardizing data formats.

[0921] "Inference" refers to making predictions and decisions based on new data using machine learning models.

[0922] "Notification" refers to the act of a system informing a user of its inference results or other important information.

[0923] "Factory operation data" refers to data including the operating status of machinery and equipment within the factory, failure history, and usage status of consumables.

[0924] "Maintenance action" refers to specific maintenance and repair actions taken to maintain the proper operation of machinery and equipment.

[0925] The system for implementing this invention is a system that suggests the appropriate next action to take based on sales data and factory operation data. The system is implemented in the following steps.

[0926] Data collection

[0927] The server collects sales data from customer management and scheduling management systems. For example, the server uses APIs to retrieve deal information and sales activity schedules. Similarly, the server collects factory operation data from sensor devices, including data such as machine uptime, failure history, and consumable usage.

[0928] Data preprocessing

[0929] The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also performs processes such as data standardization and normalization to transform the data into a format suitable for machine learning models. This improves the accuracy of the analysis.

[0930] Model Learning

[0931] The server receives pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of sales deals that have been won and those that have been lost. Similarly, it applies a similar algorithm to learn predictive maintenance actions from factory operation data. In this way, the model learns the relationship between various features and outcomes.

[0932] Inference and Action Suggestions

[0933] The server inputs new sales opportunity data and factory operation data into a machine learning model to infer the next action to take. For example, in sales, it might suggest a specific action such as "a visit would be beneficial for this customer." In the factory, it might suggest a maintenance action such as "this machine needs a part replaced."

[0934] notification

[0935] Based on the inference results, the server generates data that notifies sales representatives and maintenance teams of specific actions, and sends notifications to their terminals. This allows users to receive real-time insights from the system. For example, a sales representative might receive a notification on their smartphone or PC saying, "Next action: Visit," and the maintenance team might receive an email saying, "Next action: Replace parts."

[0936] Hardware / software used and specific examples

[0937] The hardware required includes a customer management system, a scheduling system, sensor devices, and smartphones and PCs for sales representatives. The software used will include Python, the Requests library, Pandas, Scikit-learn, and smtplib. An API will be used for data collection, and Pandas will be used for data preprocessing. A machine learning model will be built using Scikit-learn for model training, and notifications will be sent using the SMTP library.

[0938] As a concrete example, information on "Deal A (ID: A123, Size: 10 million yen, Contact Person: Tanaka)" is obtained from sales data, and the model suggests that "three visits would be effective for this customer." Similarly, information on "Machine X (ID: X456, Operating Hours: 10,000 hours, Consumable Usage: 70%)" is obtained from factory operation data, and the model suggests that "the next maintenance action should be parts replacement."

[0939] Example of a prompt

[0940] Example of a prompt message for sales: "Based on customer information and sales activity data, infer and notify us of the optimal next sales action to take."

[0941] Example of a prompt message for a factory: "Based on machine operating data (operating hours, failure history, consumable usage, etc.), infer and notify the next maintenance action to be taken."

[0942] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0943] Step 1:

[0944] The server collects sales data from customer management and scheduling management systems. It uses APIs to retrieve deal information and sales activity schedules (e.g., deal name, size, progress, visit date and time, meeting time). Input data is obtained in JSON format and converted to a Pandas DataFrame. Output is sales data in Pandas DataFrame format.

[0945] Step 2:

[0946] The server collects factory operation data from sensor devices. Using the sensor device's API, it retrieves information such as machine operating time, failure history, and consumable usage. The input data is real-time data from the sensors, and the output is operation data in Pandas DataFrame format.

[0947] Step 3:

[0948] The server preprocesses the collected sales and operational data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. It also performs standardization and normalization to transform the data into a format suitable for machine learning models. The input is each acquired data frame, and the output is the preprocessed data frame.

[0949] Step 4:

[0950] The server trains a machine learning model based on pre-processed sales data. It uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. The input is pre-processed sales data, and the output is the trained machine learning model.

[0951] Step 5:

[0952] The server trains a machine learning model based on pre-processed operational data. A similar algorithm is applied to learn predictive maintenance actions from factory operational data. The input is pre-processed operational data, and the output is the trained machine learning model.

[0953] Step 6:

[0954] The server acquires new sales opportunity data and new operational data, inputs it into a machine learning model, and infers the next action to take. The input is the newly acquired data, and the output is the inferred sales action and maintenance action.

[0955] Step 7:

[0956] The server generates notification messages based on the inferred actions and sends them to the terminals of sales representatives and maintenance teams. Sales representatives are informed of their next sales action, and maintenance teams are informed of their next maintenance action. The input is the inference result, and the output is the notification message.

[0957] Step 8:

[0958] Users (sales representatives and maintenance teams) receive notification messages sent to their devices and act according to the next required action. The input is the notification message, and the output is the actual action.

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

[0960] The present invention is a system that collects and analyzes user sentiment data in addition to sales data, and suggests the next action to be taken to sales representatives. It is mainly carried out through the following steps: sales data collection, data preprocessing, sentiment data collection and analysis, machine learning model training, action inference, and notification.

[0961] Data collection

[0962] 1. The server collects sales data from the customer management system and the scheduling management system. Specifically, the server uses APIs to retrieve deal information (e.g., deal ID, deal name, size, contact person ID, progress) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[0963] Data preprocessing

[0964] 2. The server cleans the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[0965] Collection of emotional data

[0966] 3. The server uses an emotion engine to recognize the user's emotions. Specifically, the emotion engine analyzes the user's emotions from sensor devices and applications, such as voice tone, facial expressions, and text, and generates emotion data.

[0967] Model Learning

[0968] 4. The server trains a machine learning model based on pre-processed sales data and sentiment data. Algorithms used for training include logistic regression, random forest, and neural networks. This model learns the relationship between features such as the number of visits, the number of phone calls, and the number of emails, and the relationships between successful and unsuccessful deals, and also takes sentiment data into consideration.

[0969] Inference and Action Suggestions

[0970] 5. The server inputs new sales opportunity data and the latest sentiment data into the model and infers the next action to take. For example, based on sentiment data, if a customer shows strong interest, the server recommends actions such as a visit or phone call to that customer. The inference result may output "immediate follow-up inspection" or "provide additional materials."

[0971] notification

[0972] 6. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive real-time instructions from the system and take appropriate action.

[0973] Specific example

[0974] 1. Data collection:

[0975] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[0976] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[0977] 2. Data preprocessing:

[0978] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[0979] 3. Collection of emotional data:

[0980] The server analyzes the user's emotions using an emotion engine and obtains emotion data such as "Customer B is very satisfied."

[0981] 4. Model Learning:

[0982] The server trains a machine learning model based on historical sales and sentiment data. For example, it learns the relationship between features that result in high order conversion rates and specific emotions (such as satisfaction or excitement).

[0983] 5. Inferences and Action Suggestions:

[0984] The server inputs new data for "Project B (ID: B456, scale: 20 million yen, person in charge: Sato)" and the latest sentiment data, "Customer B is very excited," and infers the action "This customer needs immediate follow-up."

[0985] 6. Notice:

[0986] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[0987] By following the above steps, this system not only provides specific instructions to sales representatives on what to do next, but also takes user emotions into consideration, enabling more effective sales activities.

[0988] The following describes the processing flow.

[0989] Step 1:

[0990] The server collects sales data from the customer management system and the scheduling management system. The server calls the API of the customer management system to retrieve deal information (deal ID, deal name, size, person in charge ID, progress status, etc.). It also uses the API of the scheduling management system to retrieve sales activity schedules (visit date and time, meeting time, participants, etc.).

[0991] Step 2:

[0992] The server preprocesses the collected data. First, it removes duplicate data. Next, it fills in any missing data. For example, if the visit date and time are missing, it automatically fills in the most recent visit date and time. Furthermore, it standardizes the data format. For example, it converts all dates and times to ISO 8601 format.

[0993] Step 3:

[0994] The server standardizes and normalizes the data after preprocessing. Specifically, it scales numerical data (e.g., project size) to a certain range and encodes categorical data (e.g., person in charge ID).

[0995] Step 4:

[0996] The server uses an emotion engine to collect user emotion data. Specifically, the emotion engine recognizes the user's emotions using technologies such as voice tone analysis, facial expression recognition, and text analysis, and acquires that data.

[0997] Step 5:

[0998] The server trains machine learning models using pre-processed sales data and collected sentiment data. For example, logistic regression and random forests are used to learn the relationship between the number of visits, phone calls, emails, and sentiment data and whether a deal was won or lost.

[0999] Step 6:

[1000] The server inputs new sales opportunity data and the latest sentiment data into a machine learning model to infer the next action. For example, if the sentiment data indicates that the customer is satisfied, it recommends a visit or phone call to that customer.

[1001] Step 7:

[1002] Based on the inference results, the server generates instruction data for sales representatives to take specific actions. For example, it might generate data such as, "Next, you should visit Mr. Tanaka at Company A."

[1003] Step 8:

[1004] The terminal receives instruction data from the server and sends notifications to the sales representative. For example, it displays a message such as, "You should visit Mr. Tanaka, the manager of Company A, at 10:00 tomorrow," via smartphone push notifications or PC desktop notifications.

[1005] This allows sales representatives to clearly understand their next steps and conduct sales activities more efficiently. Furthermore, inference based on emotional data enables more effective approaches.

[1006] (Example 2)

[1007] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1008] The present invention aims to improve the efficiency of sales activities and enhance customer satisfaction. Specifically, it aims to enable sales representatives to respond appropriately and in a timely manner by inferring actions that take into account not only sales data but also customer sentiment data. Furthermore, it aims to optimize sales activities based on real-time sentiment analysis, which was not possible with conventional systems.

[1009] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1010] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for collecting the sales data and additional sentiment data, means for training a machine learning model based on the preprocessed sales data and sentiment data, means for inputting new sales opportunity data and the latest sentiment data into the machine learning model and inferring the next action to be taken, and means for notifying the sales representative of the inferred action. This enables effective and timely inference and notification of actions by combining sales data and sentiment data.

[1011] "Sales data" refers to information related to sales activities, such as project information, customer information, and schedule information.

[1012] "Emotional data" refers to data that indicates a customer's emotional state, including information such as satisfaction, excitement, and anger, obtained through voice tone, facial expressions, and text analysis.

[1013] A "machine learning model" refers to a set of algorithms that learn patterns based on past data and use them to make predictions and classifications about future data.

[1014] "Preprocessing" refers to the process of preparing collected data to be suitable for analysis and machine learning, and includes removing duplicate data, filling in missing data, and standardizing data formats.

[1015] "Inference" refers to the act of predicting the next action to take or the resulting outcome based on the input data.

[1016] "Notification" refers to communicating the action taken as a result of the inference to the sales representative.

[1017] "Means of collection" refers to the methods and techniques a system uses to obtain data from a specific source.

[1018] "Training methods" refer to the process of using collected data to train machine learning models and improve the accuracy of predictions and classifications.

[1019] The system of this invention collects and analyzes sales data and user sentiment data to suggest the next course of action to sales representatives. This system primarily utilizes servers and terminals and is implemented in the following steps.

[1020] Data collection

[1021] The server collects sales data from customer relationship management (CRM) and scheduling management (calendar app) systems using APIs. API requests retrieve sales opportunity information (opportunity ID, opportunity name, size, contact person ID, progress status) and sales activity schedule information (visit date and time, meeting time, participants).

[1022] Specific example:

[1023] The server retrieves information about "Project A (ID: A123, scale: 10 million yen)" from the CRM system.

[1024] The server retrieves event information from the calendar app, indicating that a visit to Company A is scheduled for September 1st at 10:00 AM.

[1025] Data preprocessing

[1026] The server cleans the collected data and converts it into a format suitable for input to machine learning models. It removes duplicate data, imputes missing data, and standardizes data formats. Furthermore, it performs processes such as standardization and normalization to improve data quality.

[1027] Specific example:

[1028] The server removes duplicate data related to "Case A" and saves it in a unified format.

[1029] Collection of emotional data

[1030] The server uses an emotion engine to collect user emotions. This emotion engine analyzes user emotions through sensors and applications such as voice recognition sensors, facial recognition cameras, and chatbots, and generates emotion data such as satisfaction, excitement, and anger.

[1031] Specific example:

[1032] The server analyzes the data collected from the voice recognition sensor and obtains sentiment data indicating that "Customer B is very satisfied."

[1033] Model Learning

[1034] The server trains machine learning models based on pre-processed sales and sentiment data. The algorithms used include logistic regression, random forests, and neural networks, which are used to build predictive models for winning and losing deals.

[1035] Specific example:

[1036] The server uses historical sales data to learn features that have a high conversion rate and their association with specific sentiment data.

[1037] Action reasoning

[1038] The server inputs new sales opportunity data and the latest sentiment data to infer the next action to take. For example, if a customer is very excited, it might generate specific actions such as recommending an urgent follow-up call to that customer.

[1039] Specific example:

[1040] Based on the data for "Project B (ID: B456, Scale: 20 million yen)" and the sentiment data that "Customer B is very excited," the server infers that "this customer requires immediate follow-up."

[1041] notification

[1042] The server generates data to notify sales representatives of the inferred actions and sends it to their devices (smartphones, PCs, etc.). Upon receiving the notification, the sales representatives can respond quickly.

[1043] Specific example:

[1044] The server generates instruction data stating, "The next action is to visit customer C," and sends a notification to the terminal saying, "Visit customer C tomorrow at 10:00."

[1045] These steps allow the system to accurately suggest the next action sales representatives should take, while also considering user emotions, thereby supporting more effective sales activities.

[1046] Example of a prompt

[1047] For example, by inputting a prompt such as "If customer B is very satisfied, what action should be taken?" into the generative AI model, it is possible to infer the appropriate action.

[1048] As described above, the system of the present invention can significantly improve the efficiency and effectiveness of sales activities through advanced analysis and action suggestions that combine sales data and sentiment data.

[1049] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1050] Step 1:

[1051] The server collects sales data from the customer management system and the scheduling management system.

[1052] Specifically, the server sends an API request to retrieve case information (case ID, case name, scale, person in charge ID, progress status) and schedule information (visit date and time, meeting time, participants).

[1053] Input: Access information for the customer management system and scheduling management system.

[1054] Output: Acquired sales data and schedule data.

[1055] Specific example: The server collects data for project A (ID: A123, scale: 10 million yen) from the customer management system. Next, it retrieves information from the schedule management system that Tanaka is scheduled to visit Company A at 10:00 on September 1st.

[1056] Step 2:

[1057] The server preprocesses the collected sales data.

[1058] Specifically, the server performs tasks such as deleting duplicate data, filling in missing data, and unifying data formats. It also carries out processes such as standardization and normalization.

[1059] Input: Sales data and schedule data obtained in Step 1.

[1060] Output: Preprocessed data.

[1061] Specific example: The server deletes duplicate data related to project A and rewrites the scale of project A in a unified format. Additionally, if visit dates and times are missing, it automatically fills in the nearest available date and time.

[1062] Step 3:

[1063] The server collects user sentiment data.

[1064] Specifically, it uses an emotion engine to analyze the user's emotions through voice recognition sensors, facial recognition cameras, text analysis, and other means.

[1065] Input: User's voice tone, facial expressions, and text data.

[1066] Output: Analyzed sentiment data.

[1067] Specific example: The server analyzes the user's voice tone to determine that customer B is very satisfied and saves this as sentiment data.

[1068] Step 4:

[1069] The server trains a machine learning model based on pre-processed sales data and sentiment data.

[1070] In terms of specific operations, the server builds and trains models using algorithms such as logistic regression, random forest, and neural networks.

[1071] Input: Pre-processed sales data and sentiment data.

[1072] Output: Trained machine learning model.

[1073] Specific example: The server learns the correlation between features that lead to high order conversion rates and sentiment data based on past sales data, and builds a highly accurate predictive model.

[1074] Step 5:

[1075] The server uses new sales opportunity data and the latest sentiment data to infer the next action to take.

[1076] Specifically, the process involves inputting new data into a machine learning model and generating an action plan based on the results.

[1077] Input: New sales opportunity data and latest sentiment data.

[1078] Output: Inferred action plan.

[1079] Specific example: The server takes data for "Project B (ID: B456, scale: 20 million yen)" and sentiment data for "Customer B is very excited" and infers the action "immediate follow-up is necessary".

[1080] Step 6:

[1081] The server notifies the sales representative of specific actions to take based on the inference results.

[1082] Specifically, the system generates notification data for the action plan and sends it to the device.

[1083] Input: Inferred action plan.

[1084] Output: Notification data for sales representatives.

[1085] Specific example: The server generates an instruction that says, "The next action is to visit customer C," and sends a notification to the sales representative's smartphone saying, "Visit customer C tomorrow at 10:00."

[1086] (Application Example 2)

[1087] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1088] Traditional sales support systems focused on collecting and analyzing sales data, but lacked action inference that considered customer sentiment data, making efficient sales activities difficult. Similarly, in factory robot management, the inability to integrate robot operation data with staff sentiment data resulted in inadequate maintenance and management. These situations highlight the need for sales activities that consider user sentiment and efficient management of factory robots.

[1089] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1090] In this invention, the server includes means for collecting sales data, means for collecting sentiment data, means for pre-processing the sales data and sentiment data, means for training a machine learning model based on the pre-processed sales data and sentiment data, means for inputting new sales opportunity data and the latest user sentiment data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, and means for collecting factory robot operating status data and staff sentiment data and suggesting appropriate actions to the robot manager based on this. As a result, sales activities can be taken efficiently while taking customer sentiment into consideration, and in the management of factory robots, operating status data and sentiment data can be analyzed in an integrated manner to enable appropriate maintenance and management.

[1091] "Sales data" refers to information such as project details and sales activity schedules obtained from customer management systems and schedule management systems.

[1092] "Emotional data" refers to user emotional information analyzed using an emotion engine based on the user's voice tone, facial expressions, text, etc.

[1093] A "machine learning model" is an algorithm trained on collected sales and sentiment data to infer the next action to take.

[1094] A "factory robot" is a machine used to automate manufacturing processes and provides operational status data.

[1095] "Staff" refers to personnel who manage and maintain robots within the factory.

[1096] "Operating status data" refers to information such as the operating time, error rate, and performance of factory robots.

[1097] "Preprocessing" refers to the process of cleaning, standardizing, and normalizing raw data to transform it into a format suitable for machine learning models.

[1098] "Inference" is the process of using a machine learning model to input new data and derive the next action to take.

[1099] "Notification" refers to sending a message to inform sales representatives or factory robot managers of the inferred action.

[1100] The present invention provides a system for collecting and analyzing sales data and sentiment data, inferring the next course of action, and notifying sales representatives and factory robot managers. This system makes sales activities and factory robot management more efficient and effective.

[1101] Hardware and software to be used

[1102] The hardware to be used is as follows:

[1103] 1. Sensor device: A device used to collect data on the operating status of factory robots.

[1104] 2. Camera and microphone: Devices for collecting voice tones and facial expressions of factory staff and users.

[1105] The software to be used is as follows:

[1106] 1. RESTful API: An interface for collecting data from multiple systems.

[1107] 2. Python Libraries: Libraries for data preprocessing and machine learning model training (e.g., scikit-learn).

[1108] 3. HTTP Request Library: A library for sending notifications (e.g., requests).

[1109] Data collection

[1110] The server collects sales data and factory robot operating status data from related systems. Specifically, the server obtains sales opportunity information, visit schedules, robot operating hours, error rates, etc., via APIs.

[1111] Data preprocessing

[1112] The server cleans the collected data and performs standardization and normalization as needed. This makes the data suitable for machine learning models.

[1113] Collection of emotional data

[1114] The server uses an emotion engine to collect user emotion data. For example, it obtains emotional information such as "satisfied" or "excited" from the voice tone and facial expressions of factory staff.

[1115] Model Learning

[1116] The server trains machine learning models based on pre-processed sales data, sentiment data, and factory robot operating status data. These models are trained using methods such as logistic regression, random forests, and neural networks.

[1117] Inference and Action Suggestions

[1118] The server inputs new sales data, the latest sentiment data, and factory robot operating status data into the model to infer the next action to take. For example, it might suggest actions such as "immediate follow-up" for sales or "maintenance required" for factory robots.

[1119] notification

[1120] The server sends notifications to sales representatives and factory robot managers based on the inference results. For example, specific instructions such as "You need to visit the customer as the next action" or "You should perform maintenance on the machine" are sent.

[1121] Specific example

[1122] Data collection: Obtain factory robot operating status data from the API endpoint "http: / / robot_data.api".

[1123] Collection of emotional data: Emotional data obtained from the voice tone and facial expressions of factory staff.

[1124] Inference and Action Suggestions: Infer the next action to take based on new sales opportunity data.

[1125] Notification: A notification is sent to the sales representative stating, "The next action required is to visit the customer."

[1126] Example of a prompt

[1127] "Implement a system that collects robot operating status data from an API endpoint (http: / / robot_data.api), takes into account emotional data collected from voice tone and facial expressions, and suggests the next action to take. Notify the robot administrator of the appropriate action and recommend maintenance if any abnormalities are detected."

[1128] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1129] Step 1:

[1130] The server collects sales data and factory robot operating status data via APIs. Specifically, it obtains sales opportunity information, visit schedules, robot operating hours, error rates, and more. This allows the server to obtain sales data and robot operating status data as input data.

[1131] Step 2:

[1132] The server preprocesses the collected data. Specifically, it performs data cleaning (removing duplicate data, imputing missing values), standardization, and normalization. Using the collected raw data as input, it obtains preprocessed data in a format suitable for machine learning models as output.

[1133] Step 3:

[1134] The server uses an emotion engine to collect user emotion data. Specifically, it analyzes the user's emotions from their voice tone and facial expressions and generates emotion data. This requires sensor devices and audio / video data as input, and the analyzed emotion data is obtained as output.

[1135] Step 4:

[1136] The server trains a machine learning model based on pre-processed sales and sentiment data. Specifically, it uses algorithms such as logistic regression, random forest, and neural networks to learn the relationship between features and successful / unsuccessful deals. It takes pre-processed data as input and produces a trained machine learning model as output.

[1137] Step 5:

[1138] The server inputs new sales opportunity data and the latest sentiment data into a trained machine learning model to infer the next action to take. Specifically, it feeds sales data and sentiment data into the model and obtains actions such as "follow-up required" or "maintenance required" as inference results. New data is used as input, and the inference result is obtained as output.

[1139] Step 6:

[1140] The server sends notifications to sales representatives and factory robot managers based on the inference results. Specifically, it generates a message containing the appropriate action and sends it to the terminal. In this way, the inference results are used as input, and the notification message is sent to the sales representatives and managers as output.

[1141] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1142] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1143] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1144] [Fourth Embodiment]

[1145] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1146] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1147] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1148] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1149] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1150] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1151] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1152] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1153] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1154] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1155] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1156] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1157] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1158] The present invention provides a system for suggesting the next course of action to sales representatives based on sales data. This system is primarily implemented through the following steps.

[1159] Data collection

[1160] 1. The server collects sales data from customer relationship management (CRM) and scheduling management systems. Specifically, the server uses APIs to retrieve deal information (e.g., deal name, size, progress, person in charge) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[1161] Data preprocessing

[1162] 2. The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[1163] Model Learning

[1164] 3. The server receives the pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between various features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[1165] Inference and Action Suggestions

[1166] 4. Every time the server acquires new sales opportunity data, it inputs the data into a machine learning model to infer the next action to take. For example, the model might suggest a specific action such as "a visit would be beneficial for this customer."

[1167] notification

[1168] 5. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive insights from the system in real time.

[1169] Specific example

[1170] 1. Data collection:

[1171] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[1172] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[1173] 2. Data preprocessing:

[1174] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[1175] 3. Model learning:

[1176] The server uses a logistic regression algorithm to learn features from past data, specifically identifying that a high number of visits correlates with a high conversion rate.

[1177] 4. Inferences and Action Suggestions:

[1178] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in charge: Sato)" into the model and infers the action "This project requires 3 visits."

[1179] 5. Notice:

[1180] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[1181] As described above, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and suggesting the optimal action to sales representatives in real time.

[1182] The following describes the processing flow.

[1183] Step 1:

[1184] The server collects sales data from the customer management system and the scheduling management system. Specifically, the server calls the API of the customer management system to obtain case information (e.g., case ID, case name, size, person in charge ID, progress status). It also uses the API of the scheduling management system to obtain sales activity schedules (e.g., visit date and time, meeting time, participants).

[1185] Step 2:

[1186] The server cleans the collected data. First, it removes duplicate entries from the data. Next, it appropriately fills in any missing items. For example, if there is a missing visit date and time, it automatically fills in the nearest visit date.

[1187] Step 3:

[1188] The server standardizes and normalizes the preprocessed data. Specifically, it scales numerical data (e.g., project size) to a certain range. This transforms the data into a format suitable for machine learning models.

[1189] Step 4:

[1190] The server trains a machine learning model based on pre-processed data. Specifically, it uses data from past successful and unsuccessful deals to build and train a model using algorithms such as logistic regression and random forest. This model learns the relationship between features (e.g., number of visits, number of phone calls, number of emails) and successful / unsuccessful deals.

[1191] Step 5:

[1192] The server inputs new case data into the model and infers the next action to take. For example, for a new case, specific actions such as the number of visits or phone calls may be suggested. The inference results will output actions such as "increase the number of visits" or "follow up with a phone call."

[1193] Step 6:

[1194] Based on the inference results, the server generates data that notifies sales representatives of specific actions to take. This data includes detailed instructions on who to contact and what actions to take.

[1195] Step 7:

[1196] The terminal receives notifications from the server and informs the sales representative. Specifically, it reports specific actions to the sales representative, such as "You should visit Mr. Yamada, the manager at Company B, next," via smartphone push notifications, PC desktop notifications, or email notifications.

[1197] Through these steps, the system significantly improves the efficiency and effectiveness of sales activities by providing sales representatives with specific instructions on what to do next.

[1198] (Example 1)

[1199] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1200] In traditional sales processes, sales representatives often make decisions based on intuition and past experience, which can lead to inefficiencies and a lack of effectiveness. Furthermore, insufficient collection and analysis of sales data can result in lower closing rates. Additionally, the absence of a system that quickly suggests appropriate actions for new sales opportunities prevents the optimization of sales activities.

[1201] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1202] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for training a machine learning model based on the preprocessed sales data, means for inputting new sales data into the machine learning model and inferring the next business action to be taken, and means for notifying the business person in charge of the inferred business action. This enables efficient collection and analysis of sales data, and optimizes sales activities by quickly suggesting appropriate actions for new sales opportunities.

[1203] "Sales data" refers to all information related to sales activities conducted by a company with its customers, including customer information, deal information, and activity history.

[1204] "Preprocessing" refers to processes such as organizing, removing duplicates, imputing missing values, and standardizing formats that are performed to improve the quality of collected data.

[1205] A "machine learning model" refers to a collection of algorithms that learn patterns and trends based on past data and use them to make predictions and classifications on new data.

[1206] "Inference" refers to the process of inputting new data into a trained machine learning model and predicting appropriate actions or outcomes based on that data.

[1207] A "sales representative" refers to a person who is engaged in sales activities with customers within a company.

[1208] "Notification" refers to the act of informing sales representatives in real time about inferred results or suggested actions.

[1209] A "customer management system" refers to a software system used by companies to manage customer information, and includes customer databases and project management functions.

[1210] A "schedule management system" refers to a system used by companies to manage various schedules and appointments, and includes features such as calendar functions and reminder functions.

[1211] A "closed deal" refers to a deal for which a contract has been concluded as a result of sales activities.

[1212] "Non-contracted deals" refer to cases where a contract was not concluded as a result of sales activities.

[1213] This invention is a system that suggests the next action to take to sales representatives based on sales data. This system is primarily implemented using the following hardware and software.

[1214] System Overview

[1215] Hardware and software:

[1216] Server: A cloud server equipped with high-performance computing power and large-capacity storage.

[1217] Customer Relationship Management (CRM): A CRM system that provides APIs, such as Salesforce.

[1218] Schedule management system: A schedule management system that provides APIs such as Google Calendar.

[1219] Device: A smartphone or PC used by the sales representative.

[1220] Data collection

[1221] The server collects sales data from customer management systems and scheduling management systems via APIs. This includes obtaining case information (case name, size, progress, and assigned person) and sales activity schedules (visit date and time, meeting time, and participants). This data is temporarily stored in the server's internal database.

[1222] Data preprocessing

[1223] The server performs preprocessing on the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. For example, it standardizes the date format to "YYYY-MM-DD" and standardizes the units for contract amounts. Standardization and normalization processes are also carried out here, converting the data into a format suitable for machine learning models.

[1224] Model Learning

[1225] The server trains a machine learning model based on preprocessed data. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. At this stage, the model learns the relationship between the number of visits or contacts and whether a deal was successful or unsuccessful. Model performance evaluation and hyperparameter tuning are also performed during this process.

[1226] Inference and Action Suggestions

[1227] The server collects new sales data and inputs it into a machine learning model to infer the next action to take. For example, it generates specific action instructions such as, "A visit would be beneficial for this customer." This result is then communicated to the sales representative in real time.

[1228] notification

[1229] The server generates data based on the inference results to notify sales representatives of specific actions and sends it to their terminals. The terminals display the received notifications to the user via pop-ups or alerts, and the user plans and executes their next sales activities based on this information.

[1230] Specific example

[1231] Specific examples of data collection:

[1232] The server uses an API to retrieve data for "Project A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the "Customer Management System".

[1233] The server retrieves data from the "schedule management system" regarding the assigned person's "visit scheduled for 10:00 on September 1st."

[1234] Specific examples of data preprocessing:

[1235] The server deletes duplicate data for "Case A" and fills in the missing "Visit Date and Time" from previous visit records.

[1236] The server will standardize all date formats to "YYYY-MM-DD".

[1237] Specific examples of model learning:

[1238] The server uses past sales data to train a model that predicts the conversion rate using a logistic regression algorithm, learning the feature that a higher number of visits corresponds to a higher conversion rate.

[1239] Specific examples of inference and action suggestions:

[1240] The server inputs data for a new project, "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)," into the model and infers the action "This project requires 3 visits."

[1241] Specific examples of notifications:

[1242] The server generates instruction data, such as "The next action should be to visit a representative at a certain company," and sends it to the terminal.

[1243] The device notifies the sales representative that they should visit a company representative at 10:00 tomorrow.

[1244] Examples of prompts for generative AI models

[1245] Examples of prompt statements to input into a generative AI model include the following:

[1246] "Based on this month's sales performance, please suggest the next steps we should take, taking into account recent visits, emails, etc., and recommend appropriate sales activities."

[1247] In this way, the system of the present invention can improve sales efficiency by training a machine learning model based on collected sales data and notifying sales representatives in real time of the optimal next action to take.

[1248] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1249] Step 1:

[1250] The server collects sales data from customer relationship management (CRM) and scheduling management systems.

[1251] Specific operation: The server uses an API to send requests to the customer management system. This retrieves case information (case name, scale, progress, person in charge). It also retrieves sales activity data such as visit date and time, meeting time, and participants from the schedule management system.

[1252] Input: API endpoints and access keys for the customer management system and scheduling management system.

[1253] Output: Acquired sales data (deal information, schedule data).

[1254] Step 2:

[1255] The server performs preprocessing on the collected sales data.

[1256] Specific actions: The server removes duplicate data. If missing values ​​exist, it uses the nearest visit date and time as a reference to complete the data. To unify all data formats, it converts date formats to "YYYY-MM-DD" and standardizes currency units. Furthermore, it standardizes and normalizes the data.

[1257] Input: Collected sales data.

[1258] Output: Preprocessed data (deduplication, imputation of missing values, formatting standardization, and normalization).

[1259] Step 3:

[1260] The server trains a machine learning model based on the pre-processed data.

[1261] Specific operation: The server splits the preprocessed data into training and test data. Then, it trains a model using algorithms such as logistic regression or random forest to learn the characteristics of successful and unsuccessful projects. The trained model is evaluated, and hyperparameters are adjusted as needed.

[1262] Input: Preprocessed data.

[1263] Output: A trained machine learning model.

[1264] Step 4:

[1265] The server infers the next action to take based on the new sales data.

[1266] Specific operation: The server collects the latest sales opportunity data and performs preprocessing. Then, it inputs the preprocessed data into a machine learning model and performs inference. As a result of the inference, it generates specific actions to take next (e.g., "A visit would be beneficial for this customer").

[1267] Input: New sales data, trained machine learning model.

[1268] Output: Inference result (specific action to take next).

[1269] Step 5:

[1270] The server notifies the sales representative of specific actions to take based on the inference results.

[1271] Specific operation: The server generates notification data based on the inference results and sends it to the sales representative's terminal. The terminal displays the received notification to the sales representative via a pop-up or alert.

[1272] Input: Inference results, information about the sales representative's device.

[1273] Output: Notification sent to the sales representative's terminal.

[1274] Examples of specific actions

[1275] Specific examples of data collection:

[1276] The server uses an API to retrieve data for "Case A (ID: A123, Scale: 10 million yen, Person in Charge: Person in Charge)" from the customer management system and data for "Scheduled visit on September 1st at 10:00" from the schedule management system.

[1277] Specific examples of data preprocessing:

[1278] The server will delete duplicate data for "Case A" and fill in the missing "Visit Date and Time" from previous visit records. All date formats will be standardized to "YYYY-MM-DD".

[1279] Specific examples of model learning:

[1280] The server uses historical sales data to train a model that predicts the conversion rate using a logistic regression algorithm, and learns the feature that a higher number of visits is associated with a higher conversion rate.

[1281] Specific examples of inference and action suggestions:

[1282] The server inputs the data for the new "Project B (ID: B456, Scale: 20 million yen, Person in Charge: Person in Charge)" into the model and infers the action "This project requires 3 visits."

[1283] Specific examples of notifications:

[1284] The server sends the generated instruction data, "The next action should be to visit a representative at a certain company," to the terminal, which then notifies the sales representative that "You should visit a representative at a certain company tomorrow at 10:00."

[1285] Thus, the system of the present invention can optimize sales activities through efficient data collection and preprocessing, inference using machine learning, and notification of appropriate actions.

[1286] (Application Example 1)

[1287] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1288] While conventional sales systems offer suggestions for optimal sales actions, systems that similarly automatically suggest maintenance for factory machinery are limited. Furthermore, many factory machines are large and complex, making proper maintenance crucial before breakdowns occur. Neglecting this can lead to unexpected downtime and costly repairs. Therefore, a system that automatically suggests the next maintenance action to take is needed to improve factory production efficiency and prevent unexpected failures.

[1289] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1290] In this invention, the server includes means for collecting sales data, means for pre-processing the sales data, means for training a machine learning model based on the pre-processed sales data, means for inputting new sales opportunity data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, means for collecting and pre-processing machine operation data in the factory and inferring maintenance actions, and means for notifying the inferred maintenance action. This makes it possible to optimize not only sales activities but also machine maintenance in the factory.

[1291] "Sales data" refers to data collected by companies during their sales activities, such as customer information, project information, and sales progress.

[1292] A "machine learning model" is an algorithm that learns patterns and rules from data and uses that knowledge to perform inferences and predictions on new data.

[1293] A "customer management system" is software that centrally manages detailed customer information, transaction history, and contact history to efficiently support sales activities.

[1294] A "schedule management system" is software that centralizes the management of individual and team schedules and tasks, supporting efficient schedule management.

[1295] "Preprocessing" refers to the process of transforming raw data, which cannot be used for analysis in its original form, into an analyzable format. This includes removing duplicate data, imputing missing values, and standardizing data formats.

[1296] "Inference" refers to making predictions and decisions based on new data using machine learning models.

[1297] "Notification" refers to the act of a system informing a user of its inference results or other important information.

[1298] "Factory operation data" refers to data including the operating status of machinery and equipment within the factory, failure history, and usage status of consumables.

[1299] "Maintenance action" refers to specific maintenance and repair actions taken to maintain the proper operation of machinery and equipment.

[1300] The system for implementing this invention is a system that suggests the appropriate next action to take based on sales data and factory operation data. The system is implemented in the following steps.

[1301] Data collection

[1302] The server collects sales data from customer management and scheduling management systems. For example, the server uses APIs to retrieve deal information and sales activity schedules. Similarly, the server collects factory operation data from sensor devices, including data such as machine uptime, failure history, and consumable usage.

[1303] Data preprocessing

[1304] The server preprocesses the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also performs processes such as data standardization and normalization to transform the data into a format suitable for machine learning models. This improves the accuracy of the analysis.

[1305] Model Learning

[1306] The server receives pre-processed data as input and trains a machine learning model. For example, it uses algorithms such as logistic regression and random forest to learn the characteristics of sales deals that have been won and those that have been lost. Similarly, it applies a similar algorithm to learn predictive maintenance actions from factory operation data. In this way, the model learns the relationship between various features and outcomes.

[1307] Inference and Action Suggestions

[1308] The server inputs new sales opportunity data and factory operation data into a machine learning model to infer the next action to take. For example, in sales, it might suggest a specific action such as "a visit would be beneficial for this customer." In the factory, it might suggest a maintenance action such as "this machine needs a part replaced."

[1309] notification

[1310] Based on the inference results, the server generates data that notifies sales representatives and maintenance teams of specific actions, and sends notifications to their terminals. This allows users to receive real-time insights from the system. For example, a sales representative might receive a notification on their smartphone or PC saying, "Next action: Visit," and the maintenance team might receive an email saying, "Next action: Replace parts."

[1311] Hardware / software used and specific examples

[1312] The hardware required includes a customer management system, a scheduling system, sensor devices, and smartphones and PCs for sales representatives. The software used will include Python, the Requests library, Pandas, Scikit-learn, and smtplib. An API will be used for data collection, and Pandas will be used for data preprocessing. A machine learning model will be built using Scikit-learn for model training, and notifications will be sent using the SMTP library.

[1313] As a concrete example, information on "Deal A (ID: A123, Size: 10 million yen, Contact Person: Tanaka)" is obtained from sales data, and the model suggests that "three visits would be effective for this customer." Similarly, information on "Machine X (ID: X456, Operating Hours: 10,000 hours, Consumable Usage: 70%)" is obtained from factory operation data, and the model suggests that "the next maintenance action should be parts replacement."

[1314] Example of a prompt

[1315] Example of a prompt message for sales: "Based on customer information and sales activity data, infer and notify us of the optimal next sales action to take."

[1316] Example of a prompt message for a factory: "Based on machine operating data (operating hours, failure history, consumable usage, etc.), infer and notify the next maintenance action to be taken."

[1317] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1318] Step 1:

[1319] The server collects sales data from customer management and scheduling management systems. It uses APIs to retrieve deal information and sales activity schedules (e.g., deal name, size, progress, visit date and time, meeting time). Input data is obtained in JSON format and converted to a Pandas DataFrame. Output is sales data in Pandas DataFrame format.

[1320] Step 2:

[1321] The server collects factory operation data from sensor devices. Using the sensor device's API, it retrieves information such as machine operating time, failure history, and consumable usage. The input data is real-time data from the sensors, and the output is operation data in Pandas DataFrame format.

[1322] Step 3:

[1323] The server preprocesses the collected sales and operational data. Specifically, it removes duplicate data, imputes missing values, and standardizes data formats. It also performs standardization and normalization to transform the data into a format suitable for machine learning models. The input is each acquired data frame, and the output is the preprocessed data frame.

[1324] Step 4:

[1325] The server trains a machine learning model based on pre-processed sales data. It uses algorithms such as logistic regression and random forest to learn the characteristics of successful and unsuccessful deals. The input is pre-processed sales data, and the output is the trained machine learning model.

[1326] Step 5:

[1327] The server trains a machine learning model based on pre-processed operational data. A similar algorithm is applied to learn predictive maintenance actions from factory operational data. The input is pre-processed operational data, and the output is the trained machine learning model.

[1328] Step 6:

[1329] The server acquires new sales opportunity data and new operational data, inputs it into a machine learning model, and infers the next action to take. The input is the newly acquired data, and the output is the inferred sales action and maintenance action.

[1330] Step 7:

[1331] The server generates notification messages based on the inferred actions and sends them to the terminals of sales representatives and maintenance teams. Sales representatives are informed of their next sales action, and maintenance teams are informed of their next maintenance action. The input is the inference result, and the output is the notification message.

[1332] Step 8:

[1333] Users (sales representatives and maintenance teams) receive notification messages sent to their devices and act according to the next required action. The input is the notification message, and the output is the actual action.

[1334] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1335] The present invention is a system that collects and analyzes user sentiment data in addition to sales data, and suggests the next action to be taken to sales representatives. It is mainly carried out through the following steps: sales data collection, data preprocessing, sentiment data collection and analysis, machine learning model training, action inference, and notification.

[1336] Data collection

[1337] 1. The server collects sales data from the customer management system and the scheduling management system. Specifically, the server uses APIs to retrieve deal information (e.g., deal ID, deal name, size, contact person ID, progress) and sales activity schedules (e.g., visit date and time, meeting time, participants).

[1338] Data preprocessing

[1339] 2. The server cleans the collected data. Specifically, it removes duplicate data, imputes missing values, and standardizes the data format. It also converts the data into a format suitable for machine learning models by performing processes such as standardization and normalization.

[1340] Collection of emotional data

[1341] 3. The server uses an emotion engine to recognize the user's emotions. Specifically, the emotion engine analyzes the user's emotions from sensor devices and applications, such as voice tone, facial expressions, and text, and generates emotion data.

[1342] Model Learning

[1343] 4. The server trains a machine learning model based on pre-processed sales data and sentiment data. Algorithms used for training include logistic regression, random forest, and neural networks. This model learns the relationship between features such as the number of visits, the number of phone calls, and the number of emails, and the relationships between successful and unsuccessful deals, and also takes sentiment data into consideration.

[1344] Inference and Action Suggestions

[1345] 5. The server inputs new sales opportunity data and the latest sentiment data into the model and infers the next action to take. For example, based on sentiment data, if a customer shows strong interest, the server recommends actions such as a visit or phone call to that customer. The inference result may output "immediate follow-up inspection" or "provide additional materials."

[1346] notification

[1347] 6. Based on the inference results, the server generates data that notifies the sales representative of specific actions to take and sends the notification to their device (e.g., the sales representative's smartphone or PC). This allows the sales representative to receive real-time instructions from the system and take appropriate action.

[1348] Specific example

[1349] 1. Data collection:

[1350] The server retrieves information about "Project A (ID: A123, Scale: 10 million yen, Person in charge: Tanaka)" from the customer management system.

[1351] The server retrieves information from the schedule management system indicating that Tanaka has a scheduled visit to Company A at 10:00 AM on September 1st.

[1352] 2. Data preprocessing:

[1353] The server removes duplicate data for "Project A" and converts information such as "Project A's scale is 10 million yen" into a standardized format. If a visit is missing, the server automatically fills in the nearest visit date and time.

[1354] 3. Collection of emotional data:

[1355] The server analyzes the user's emotions using an emotion engine and obtains emotion data such as "Customer B is very satisfied."

[1356] 4. Model Learning:

[1357] The server trains a machine learning model based on historical sales and sentiment data. For example, it learns the relationship between features that result in high order conversion rates and specific emotions (such as satisfaction or excitement).

[1358] 5. Inferences and Action Suggestions:

[1359] The server inputs new data for "Project B (ID: B456, scale: 20 million yen, person in charge: Sato)" and the latest sentiment data, "Customer B is very excited," and infers the action "This customer needs immediate follow-up."

[1360] 6. Notice:

[1361] The server generates instruction data stating, "The next action is to visit Mr. Yamada, the section chief at Company B," and sends a notification to the terminal saying, "You should visit Mr. Yamada, the section chief at Company B, at 10:00 tomorrow."

[1362] By following the above steps, this system not only provides specific instructions to sales representatives on what to do next, but also takes user emotions into consideration, enabling more effective sales activities.

[1363] The following describes the processing flow.

[1364] Step 1:

[1365] The server collects sales data from the customer management system and the scheduling management system. The server calls the API of the customer management system to retrieve deal information (deal ID, deal name, size, person in charge ID, progress status, etc.). It also uses the API of the scheduling management system to retrieve sales activity schedules (visit date and time, meeting time, participants, etc.).

[1366] Step 2:

[1367] The server preprocesses the collected data. First, it removes duplicate data. Next, it fills in any missing data. For example, if the visit date and time are missing, it automatically fills in the most recent visit date and time. Furthermore, it standardizes the data format. For example, it converts all dates and times to ISO 8601 format.

[1368] Step 3:

[1369] The server standardizes and normalizes the data after preprocessing. Specifically, it scales numerical data (e.g., project size) to a certain range and encodes categorical data (e.g., person in charge ID).

[1370] Step 4:

[1371] The server uses an emotion engine to collect user emotion data. Specifically, the emotion engine recognizes the user's emotions using technologies such as voice tone analysis, facial expression recognition, and text analysis, and acquires that data.

[1372] Step 5:

[1373] The server trains machine learning models using pre-processed sales data and collected sentiment data. For example, logistic regression and random forests are used to learn the relationship between the number of visits, phone calls, emails, and sentiment data and whether a deal was won or lost.

[1374] Step 6:

[1375] The server inputs new sales opportunity data and the latest sentiment data into a machine learning model to infer the next action. For example, if the sentiment data indicates that the customer is satisfied, it recommends a visit or phone call to that customer.

[1376] Step 7:

[1377] Based on the inference results, the server generates instruction data for sales representatives to take specific actions. For example, it might generate data such as, "Next, you should visit Mr. Tanaka at Company A."

[1378] Step 8:

[1379] The terminal receives instruction data from the server and sends notifications to the sales representative. For example, it displays a message such as, "You should visit Mr. Tanaka, the manager of Company A, at 10:00 tomorrow," via smartphone push notifications or PC desktop notifications.

[1380] This allows sales representatives to clearly understand their next steps and conduct sales activities more efficiently. Furthermore, inference based on emotional data enables more effective approaches.

[1381] (Example 2)

[1382] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1383] The present invention aims to improve the efficiency of sales activities and enhance customer satisfaction. Specifically, it aims to enable sales representatives to respond appropriately and in a timely manner by inferring actions that take into account not only sales data but also customer sentiment data. Furthermore, it aims to optimize sales activities based on real-time sentiment analysis, which was not possible with conventional systems.

[1384] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1385] In this invention, the server includes means for collecting sales data, means for preprocessing the sales data, means for collecting the sales data and additional sentiment data, means for training a machine learning model based on the preprocessed sales data and sentiment data, means for inputting new sales opportunity data and the latest sentiment data into the machine learning model and inferring the next action to be taken, and means for notifying the sales representative of the inferred action. This enables effective and timely inference and notification of actions by combining sales data and sentiment data.

[1386] "Sales data" refers to information related to sales activities, such as project information, customer information, and schedule information.

[1387] "Emotional data" refers to data that indicates a customer's emotional state, including information such as satisfaction, excitement, and anger, obtained through voice tone, facial expressions, and text analysis.

[1388] A "machine learning model" refers to a set of algorithms that learn patterns based on past data and use them to make predictions and classifications about future data.

[1389] "Preprocessing" refers to the process of preparing collected data to be suitable for analysis and machine learning, and includes removing duplicate data, filling in missing data, and standardizing data formats.

[1390] "Inference" refers to the act of predicting the next action to take or the resulting outcome based on the input data.

[1391] "Notification" refers to communicating the action taken as a result of the inference to the sales representative.

[1392] "Means of collection" refers to the methods and techniques a system uses to obtain data from a specific source.

[1393] "Training methods" refer to the process of using collected data to train machine learning models and improve the accuracy of predictions and classifications.

[1394] The system of this invention collects and analyzes sales data and user sentiment data to suggest the next course of action to sales representatives. This system primarily utilizes servers and terminals and is implemented in the following steps.

[1395] Data collection

[1396] The server collects sales data from customer relationship management (CRM) and scheduling management (calendar app) systems using APIs. API requests retrieve sales opportunity information (opportunity ID, opportunity name, size, contact person ID, progress status) and sales activity schedule information (visit date and time, meeting time, participants).

[1397] Specific example:

[1398] The server retrieves information about "Project A (ID: A123, scale: 10 million yen)" from the CRM system.

[1399] The server retrieves event information from the calendar app, indicating that a visit to Company A is scheduled for September 1st at 10:00 AM.

[1400] Data preprocessing

[1401] The server cleans the collected data and converts it into a format suitable for input to machine learning models. It removes duplicate data, imputes missing data, and standardizes data formats. Furthermore, it performs processes such as standardization and normalization to improve data quality.

[1402] Specific example:

[1403] The server removes duplicate data related to "Case A" and saves it in a unified format.

[1404] Collection of emotional data

[1405] The server uses an emotion engine to collect user emotions. This emotion engine analyzes user emotions through sensors and applications such as voice recognition sensors, facial recognition cameras, and chatbots, and generates emotion data such as satisfaction, excitement, and anger.

[1406] Specific example:

[1407] The server analyzes the data collected from the voice recognition sensor and obtains sentiment data indicating that "Customer B is very satisfied."

[1408] Model Learning

[1409] The server trains machine learning models based on pre-processed sales and sentiment data. The algorithms used include logistic regression, random forests, and neural networks, which are used to build predictive models for winning and losing deals.

[1410] Specific example:

[1411] The server uses historical sales data to learn features that have a high conversion rate and their association with specific sentiment data.

[1412] Action reasoning

[1413] The server inputs new sales opportunity data and the latest sentiment data to infer the next action to take. For example, if a customer is very excited, it might generate specific actions such as recommending an urgent follow-up call to that customer.

[1414] Specific example:

[1415] Based on the data for "Project B (ID: B456, Scale: 20 million yen)" and the sentiment data that "Customer B is very excited," the server infers that "this customer requires immediate follow-up."

[1416] notification

[1417] The server generates data to notify sales representatives of the inferred actions and sends it to their devices (smartphones, PCs, etc.). Upon receiving the notification, the sales representatives can respond quickly.

[1418] Specific example:

[1419] The server generates instruction data stating, "The next action is to visit customer C," and sends a notification to the terminal saying, "Visit customer C tomorrow at 10:00."

[1420] These steps allow the system to accurately suggest the next action sales representatives should take, while also considering user emotions, thereby supporting more effective sales activities.

[1421] Example of a prompt

[1422] For example, by inputting a prompt such as "If customer B is very satisfied, what action should be taken?" into the generative AI model, it is possible to infer the appropriate action.

[1423] As described above, the system of the present invention can significantly improve the efficiency and effectiveness of sales activities through advanced analysis and action suggestions that combine sales data and sentiment data.

[1424] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1425] Step 1:

[1426] The server collects sales data from the customer management system and the scheduling management system.

[1427] Specifically, the server sends an API request to retrieve case information (case ID, case name, scale, person in charge ID, progress status) and schedule information (visit date and time, meeting time, participants).

[1428] Input: Access information for the customer management system and scheduling management system.

[1429] Output: Acquired sales data and schedule data.

[1430] Specific example: The server collects data for project A (ID: A123, scale: 10 million yen) from the customer management system. Next, it retrieves information from the schedule management system that Tanaka is scheduled to visit Company A at 10:00 on September 1st.

[1431] Step 2:

[1432] The server preprocesses the collected sales data.

[1433] Specifically, the server performs tasks such as deleting duplicate data, filling in missing data, and unifying data formats. It also carries out processes such as standardization and normalization.

[1434] Input: Sales data and schedule data obtained in Step 1.

[1435] Output: Preprocessed data.

[1436] Specific example: The server deletes duplicate data related to project A and rewrites the scale of project A in a unified format. Additionally, if visit dates and times are missing, it automatically fills in the nearest available date and time.

[1437] Step 3:

[1438] The server collects user sentiment data.

[1439] Specifically, it uses an emotion engine to analyze the user's emotions through voice recognition sensors, facial recognition cameras, text analysis, and other means.

[1440] Input: User's voice tone, facial expressions, and text data.

[1441] Output: Analyzed sentiment data.

[1442] Specific example: The server analyzes the user's voice tone to determine that customer B is very satisfied and saves this as sentiment data.

[1443] Step 4:

[1444] The server trains a machine learning model based on pre-processed sales data and sentiment data.

[1445] In terms of specific operations, the server builds and trains models using algorithms such as logistic regression, random forest, and neural networks.

[1446] Input: Pre-processed sales data and sentiment data.

[1447] Output: Trained machine learning model.

[1448] Specific example: The server learns the correlation between features that lead to high order conversion rates and sentiment data based on past sales data, and builds a highly accurate predictive model.

[1449] Step 5:

[1450] The server uses new sales opportunity data and the latest sentiment data to infer the next action to take.

[1451] Specifically, the process involves inputting new data into a machine learning model and generating an action plan based on the results.

[1452] Input: New sales opportunity data and latest sentiment data.

[1453] Output: Inferred action plan.

[1454] Specific example: The server takes data for "Project B (ID: B456, scale: 20 million yen)" and sentiment data for "Customer B is very excited" and infers the action "immediate follow-up is necessary".

[1455] Step 6:

[1456] The server notifies the sales representative of specific actions to take based on the inference results.

[1457] Specifically, the system generates notification data for the action plan and sends it to the device.

[1458] Input: Inferred action plan.

[1459] Output: Notification data for sales representatives.

[1460] Specific example: The server generates an instruction that says, "The next action is to visit customer C," and sends a notification to the sales representative's smartphone saying, "Visit customer C tomorrow at 10:00."

[1461] (Application Example 2)

[1462] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1463] Traditional sales support systems focused on collecting and analyzing sales data, but lacked action inference that considered customer sentiment data, making efficient sales activities difficult. Similarly, in factory robot management, the inability to integrate robot operation data with staff sentiment data resulted in inadequate maintenance and management. These situations highlight the need for sales activities that consider user sentiment and efficient management of factory robots.

[1464] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1465] In this invention, the server includes means for collecting sales data, means for collecting sentiment data, means for pre-processing the sales data and sentiment data, means for training a machine learning model based on the pre-processed sales data and sentiment data, means for inputting new sales opportunity data and the latest user sentiment data into the machine learning model and inferring the next action to take, means for notifying the sales representative of the inferred action, and means for collecting factory robot operating status data and staff sentiment data and suggesting appropriate actions to the robot manager based on this. As a result, sales activities can be taken efficiently while taking customer sentiment into consideration, and in the management of factory robots, operating status data and sentiment data can be analyzed in an integrated manner to enable appropriate maintenance and management.

[1466] "Sales data" refers to information such as project details and sales activity schedules obtained from customer management systems and schedule management systems.

[1467] "Emotional data" refers to user emotional information analyzed using an emotion engine based on the user's voice tone, facial expressions, text, etc.

[1468] A "machine learning model" is an algorithm trained on collected sales and sentiment data to infer the next action to take.

[1469] A "factory robot" is a machine used to automate manufacturing processes and provides operational status data.

[1470] "Staff" refers to personnel who manage and maintain robots within the factory.

[1471] "Operating status data" refers to information such as the operating time, error rate, and performance of factory robots.

[1472] "Preprocessing" refers to the process of cleaning, standardizing, and normalizing raw data to transform it into a format suitable for machine learning models.

[1473] "Inference" is the process of using a machine learning model to input new data and derive the next action to take.

[1474] "Notification" refers to sending a message to inform sales representatives or factory robot managers of the inferred action.

[1475] The present invention provides a system for collecting and analyzing sales data and sentiment data, inferring the next course of action, and notifying sales representatives and factory robot managers. This system makes sales activities and factory robot management more efficient and effective.

[1476] Hardware and software to be used

[1477] The hardware to be used is as follows:

[1478] 1. Sensor device: A device used to collect data on the operating status of factory robots.

[1479] 2. Camera and microphone: Devices for collecting voice tones and facial expressions of factory staff and users.

[1480] The software to be used is as follows:

[1481] 1. RESTful API: An interface for collecting data from multiple systems.

[1482] 2. Python Libraries: Libraries for data preprocessing and machine learning model training (e.g., scikit-learn).

[1483] 3. HTTP Request Library: A library for sending notifications (e.g., requests).

[1484] Data collection

[1485] The server collects sales data and factory robot operating status data from related systems. Specifically, the server obtains sales opportunity information, visit schedules, robot operating hours, error rates, etc., via APIs.

[1486] Data preprocessing

[1487] The server cleans the collected data and performs standardization and normalization as needed. This makes the data suitable for machine learning models.

[1488] Collection of emotional data

[1489] The server uses an emotion engine to collect user emotion data. For example, it obtains emotional information such as "satisfied" or "excited" from the voice tone and facial expressions of factory staff.

[1490] Model Learning

[1491] The server trains machine learning models based on pre-processed sales data, sentiment data, and factory robot operating status data. These models are trained using methods such as logistic regression, random forests, and neural networks.

[1492] Inference and Action Suggestions

[1493] The server inputs new sales data, the latest sentiment data, and factory robot operating status data into the model to infer the next action to take. For example, it might suggest actions such as "immediate follow-up" for sales or "maintenance required" for factory robots.

[1494] notification

[1495] The server sends notifications to sales representatives and factory robot managers based on the inference results. For example, specific instructions such as "You need to visit the customer as the next action" or "You should perform maintenance on the machine" are sent.

[1496] Specific example

[1497] Data collection: Obtain factory robot operating status data from the API endpoint "http: / / robot_data.api".

[1498] Collection of emotional data: Emotional data obtained from the voice tone and facial expressions of factory staff.

[1499] Inference and Action Suggestions: Infer the next action to take based on new sales opportunity data.

[1500] Notification: A notification is sent to the sales representative stating, "The next action required is to visit the customer."

[1501] Example of a prompt

[1502] "Implement a system that collects robot operating status data from an API endpoint (http: / / robot_data.api), takes into account emotional data collected from voice tone and facial expressions, and suggests the next action to take. Notify the robot administrator of the appropriate action and recommend maintenance if any abnormalities are detected."

[1503] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1504] Step 1:

[1505] The server collects sales data and factory robot operating status data via APIs. Specifically, it obtains sales opportunity information, visit schedules, robot operating hours, error rates, and more. This allows the server to obtain sales data and robot operating status data as input data.

[1506] Step 2:

[1507] The server preprocesses the collected data. Specifically, it performs data cleaning (removing duplicate data, imputing missing values), standardization, and normalization. Using the collected raw data as input, it obtains preprocessed data in a format suitable for machine learning models as output.

[1508] Step 3:

[1509] The server uses an emotion engine to collect user emotion data. Specifically, it analyzes the user's emotions from their voice tone and facial expressions and generates emotion data. This requires sensor devices and audio / video data as input, and the analyzed emotion data is obtained as output.

[1510] Step 4:

[1511] The server trains a machine learning model based on pre-processed sales and sentiment data. Specifically, it uses algorithms such as logistic regression, random forest, and neural networks to learn the relationship between features and successful / unsuccessful deals. It takes pre-processed data as input and produces a trained machine learning model as output.

[1512] Step 5:

[1513] The server inputs new sales opportunity data and the latest sentiment data into a trained machine learning model to infer the next action to take. Specifically, it feeds sales data and sentiment data into the model and obtains actions such as "follow-up required" or "maintenance required" as inference results. New data is used as input, and the inference result is obtained as output.

[1514] Step 6:

[1515] The server sends notifications to sales representatives and factory robot managers based on the inference results. Specifically, it generates a message containing the appropriate action and sends it to the terminal. In this way, the inference results are used as input, and the notification message is sent to the sales representatives and managers as output.

[1516] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1517] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1518] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1519] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1520] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1521] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1522] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1523] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1524] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1525] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1526] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1527] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1528] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1529] 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.

[1530] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1531] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1532] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1533] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1534] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1535] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1536] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[1537] The following is further disclosed regarding the embodiments described above.

[1538] (Claim 1)

[1539] Means of collecting sales data,

[1540] Means for pre-processing the aforementioned sales data,

[1541] A means for training a machine learning model based on the aforementioned pre-processed sales data,

[1542] A means for inputting new sales opportunity data into the aforementioned machine learning model and inferring the next action to take,

[1543] Means for notifying the sales representative of the inferred action

[1544] A system that includes this.

[1545] (Claim 2)

[1546] The system according to claim 1, characterized in that the aforementioned sales data is obtained from a customer management system and a schedule management system.

[1547] (Claim 3)

[1548] The system according to claim 1, further comprising means for learning the characteristics of successful and unsuccessful deals using the machine learning model.

[1549]

[1550] "Example 1"

[1551] (Claim 1)

[1552] Means of collecting sales data,

[1553] Means for pre-processing the aforementioned sales data,

[1554] A means for training a machine learning model based on the aforementioned pre-processed sales data,

[1555] A means for inputting new sales data into the aforementioned machine learning model and inferring the next business action to be taken,

[1556] Means for notifying the person in charge of the business action that has been inferred above

[1557] A system that includes this.

[1558] (Claim 2)

[1559] The system according to claim 1, characterized in that the aforementioned sales data is obtained from a customer management system and a scheduling management system.

[1560] (Claim 3)

[1561] The system according to claim 1, further comprising means for learning the characteristics of closed deals and unclosed deals using the machine learning model.

[1562] "Application Example 1"

[1563] (Claim 1)

[1564] Means of collecting sales data,

[1565] Means for pre-processing the aforementioned sales data,

[1566] A means for training a machine learning model based on the aforementioned pre-processed sales data,

[1567] A means for inputting new sales opportunity data into the aforementioned machine learning model and inferring the next action to take,

[1568] A means of notifying the sales representative of the inferred action,

[1569] A means for collecting and pre-processing machine operation data in a factory and then inferring maintenance actions,

[1570] Means for notifying the inferred maintenance action and

[1571] A system that includes this.

[1572] (Claim 2)

[1573] The system according to claim 1, characterized in that the sales data is obtained from a customer management system and a schedule management system, and the operation data of machinery in the factory is obtained from a sensor device.

[1574] (Claim 3)

[1575] The system according to claim 1, further comprising means for learning the characteristics of successful and unsuccessful orders using the machine learning model, and further comprising means for learning predictive maintenance actions using machine operation data in the factory.

[1576] "Example 2 of combining an emotion engine"

[1577] (Claim 1)

[1578] Means of collecting sales data,

[1579] Means for pre-processing the aforementioned sales data,

[1580] Means for collecting the aforementioned sales data and additional sentiment data,

[1581] A means for training a machine learning model based on the pre-processed sales data and sentiment data,

[1582] A means for inputting new sales opportunity data and the latest sentiment data into the machine learning model and inferring the next action to take,

[1583] Means for notifying the sales representative of the inferred action

[1584] A system that includes this.

[1585] (Claim 2)

[1586] The system according to claim 1, characterized in that sales data is obtained from a customer management system and a schedule management system.

[1587] (Claim 3)

[1588] The system according to claim 1, further comprising means for learning the characteristics of successful and unsuccessful deals using a machine learning model.

[1589] "Application example 2 when combining with an emotional engine"

[1590] (Claim 1)

[1591] Means of collecting sales data,

[1592] Means for pre-processing the aforementioned sales data,

[1593] A means for training a machine learning model based on the aforementioned pre-processed sales data and user sentiment data,

[1594] A means for inputting new sales opportunity data and the latest user sentiment data into the machine learning model and inferring the next action to take,

[1595] A means of notifying the sales representative of the inferred action,

[1596] A means of collecting operational data of factory robots and emotional data of staff, and based on that, suggesting appropriate actions to robot managers,

[1597] A system that includes this.

[1598] (Claim 2)

[1599] The system according to claim 1, characterized in that the aforementioned sales data and factory robot operating status data are acquired from the management system.

[1600] (Claim 3)

[1601] The system according to claim 1, further comprising means for learning the characteristics of ordered projects, lost projects, and robot operating status data using the machine learning model. [Explanation of Symbols]

[1602] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. Means of collecting sales data, Means for pre-processing the aforementioned sales data, A means for training a machine learning model based on the aforementioned pre-processed sales data, A means for inputting new sales opportunity data into the aforementioned machine learning model and inferring the next action to take, Means for notifying the sales representative of the inferred action A system that includes this.

2. The system according to claim 1, characterized in that the aforementioned sales data is obtained from a customer management system and a schedule management system.

3. The system according to claim 1, further comprising means for learning the characteristics of successful and unsuccessful deals using the machine learning model.

Images