Programs, information processing devices, methods, and systems

Abstract

Automatically generates sales materials optimized for each customer. [Solution] A program that causes one or more processors to execute the following steps: acquiring basic customer information, information related to transactions with the customer, and external customer information; analyzing each acquired piece of information to extract keywords and the context of those keywords to identify the customer's potential problems; selecting the company's services to solve the problems based on the identified problems and selection rules that define the selection criteria for the company's services, and determining proposal requirements that include the company's services and the problems; incorporating the proposal requirements into a prompt template and generating a prompt that instructs a generating AI model to generate text content that constitutes sales materials for the customer; and inputting the prompt into the generating AI model and incorporating the obtained text content into a document template to generate sales materials for the customer.

Description

Technical Field

[0001] The present disclosure relates to programs, information processing apparatuses, methods, and systems.

Background Art

[0002] In recent years, in corporate sales activities, there has been a demand for improving the efficiency and quality of proposal activities for customers. In particular, with the emergence of general-purpose generative AI (Artificial Intelligence) models such as ChatGPT (registered trademark), the movement to use generative AI models for creating drafts such as proposal documents or sales emails has been accelerating. General-purpose generative AI models can generate natural sentences in response to prompts input by users.

[0003] Also, conventionally, customer relationship management (CRM: Customer Relationship Management) systems represented by Salesforce (registered trademark) have been widely used. A CRM system centrally manages basic customer information, negotiation histories, sales prospects, etc., and sales staff conduct daily sales activities by referring to this information.

[0004] Furthermore, technologies for supporting solution proposals to customers by referring to past knowledge have also been proposed. For example, Patent Document 1 discloses a technology for storing past solutions, negotiation information, etc. in a knowledge database and generating contents such as proposal documents or question lists based on input information obtained in new negotiations and similar knowledge. [[ID=二十二]]

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, in order to create high-quality sales materials using conventional general-purpose generative AI models, sales representatives themselves had to create highly accurate prompts that described the customer's situation, background, and characteristics in detail. Therefore, there was a problem in that the quality of the generated sales materials depended heavily on the prompt-creation skills of the sales representatives.

[0007] Furthermore, while conventional CRM systems accumulate vast amounts of customer data, this data is often managed merely as a record. In other words, they are not sufficient as a mechanism to proactively generate optimized proposals for each customer by comprehensively analyzing the transaction history, external information, and the company's own service information accumulated in the system.

[0008] Furthermore, even in the technology disclosed in Patent Document 1, sufficient consideration was given to dynamically combining and analyzing customer transaction information, the latest external information, and the company's own service information to optimize the logical structure of proposals in the content. In particular, it was difficult to adjust the approach or tone and manner of proposals while taking into account qualitative elements (personality) such as the personal character or work attitude of the customer's representative.

[0009] The purpose of this disclosure is to automatically generate sales materials optimized for each customer, regardless of the sales representative's skills. [Means for solving the problem]

[0010] To solve the aforementioned problems, a program according to one aspect of the present disclosure is a program for operating a computer having one or more processors, and causes one or more processors to execute the following steps: acquiring basic customer information, information on transactions with the customer, and external customer information; analyzing text data contained in each of the acquired basic information, transaction information, and external information, and extracting predefined keywords and the context of those keywords to identify the customer's potential problems as structured data; selecting a service to solve the identified problems by referring to a database storing information on multiple services of the company based on the identified problems and predefined selection rules that define the selection criteria for the company's services, and determining proposal requirements, which are structured data including the selected services and problems; incorporating the determined proposal requirements into a prompt template that templates a logical structure corresponding to a predetermined business logic, and generating a prompt that includes an instruction sentence instructing a generating AI model to generate text content constituting sales materials for customers; and inputting the generated prompt into a generating AI model to output text content, and generating sales materials for customers by incorporating the output text content into a predetermined document template. [Effects of the Invention]

[0011] According to this disclosure, it is possible to automatically generate sales materials optimized for each customer, regardless of the skills of the sales representative. [Brief explanation of the drawing]

[0012] [Figure 1] This is a block diagram showing an example of the overall system configuration. [Figure 2] This is a block diagram showing an example of a user terminal's hardware configuration. [Figure 3] This is a block diagram showing the functional components implemented by the control unit of the user terminal. [Figure 4] Block diagram showing an example of a server hardware configuration. [Figure 5] This is a block diagram showing the functional parts implemented by the server's control unit. [Figure 6] This diagram shows an example of the data structure of a customer database. [Figure 7] This diagram shows an example of the data structure of our company's service database. [Figure 8] This figure shows an example of the data structure of a prompt template database. [Figure 9] This figure shows an example of the data structure of a personality database. [Figure 10] This figure shows an example of the data structure of a document template database. [Figure 11] This flowchart shows an example of server operation. [Figure 12] This is a schematic diagram showing an example of a user terminal screen. [Figure 13] This is a schematic diagram showing other examples of user terminal screens. [Modes for carrying out the invention]

[0013] The embodiments of this disclosure will be described below with reference to the drawings. In all the drawings illustrating the embodiments, common components are denoted by the same reference numerals, and repeated explanations are omitted. The following embodiments are not intended to unduly limit the content of this disclosure as described in the claims. Not all components shown in the embodiments are necessarily essential components of this disclosure. Also, each drawing is a schematic diagram and is not necessarily a strict illustration.

[0014] Also, in the following description, a "processor" is one or more processors. The processor may be expressed, for example, as processing circuitry. At least one processor is typically a microprocessor such as a CPU (Central Processing Unit), but may also be other types of processors such as a GPU (Graphics Processing Unit). At least one processor may be single-core or multi-core. Also, at least one processor may be a general-purpose processor or a special-purpose processor.

[0015] Also, at least one processor may be a processor in a broad sense such as a hardware circuit (e.g., FPGA (Field-Programmable Gate Array), ASIC (Application Specific Integrated Circuit)) that performs part or all of the processing.

[0016] Also, in the following description, an expression such as "xxx table" may be used to describe information from which an output is obtained for an input. This information may be data of any structure or a learning model such as a neural network that generates an output for an input. Therefore, "xxx table" can be referred to as "xxx information".

[0017] Also, in the following description, the configuration of each table is an example. One table may be divided into two or more tables, or all or part of two or more tables may be one table.

[0018] The program may be pre-installed on the information processing device described below, or, for example, the information processing device may be on a readable (e.g., non-temporary) recording medium and the program may be installed on the information processing device. Alternatively, the program may be sent from a program distribution server to the information processing device and installed there. Furthermore, in the following description, two or more programs may be implemented as a single program, or one program may be implemented as two or more programs.

[0019] Furthermore, while various types of identification information are used in the following description, the identification information only needs to be information that indicates a predetermined object, and the specific data is not limited to the embodiments. The identification information may be an identification number, or an identifier that includes letters and symbols.

[0020] [Embodiment] <Overview> The system according to this embodiment is a system that integrates and utilizes diverse information accumulated in a CRM system, etc., to automatically generate various customer-oriented sales materials (proposals, sales emails, or role-playing scripts, etc.) that are optimized for each customer.

[0021] Specifically, the system according to this embodiment analyzes information on past customer transactions, external information of customer companies, and information on the company's own services using a natural language processing model or a rule-based algorithm. Based on the analysis results, the system according to this embodiment automatically generates prompts structured according to predetermined business logic and uses a generation AI model to create sales materials with a logical story structure tailored to customer interests.

[0022] This allows sales representatives to quickly create high-quality, personalized proposals for customers, even without advanced prompt engineering skills.

[0023] <Overall System Configuration> Figure 1 is a block diagram showing an example of the overall configuration of System 1. System 1 includes a user terminal 10, a server 20, and an AI system 30. The user terminal 10, server 20, and AI system 30 are connected to each other via a network 80 such as the Internet or a LAN (Local Area Network), using protocols such as TCP / IP to enable communication.

[0024] System 1 further integrates with CRM system 40 and external information DB (database) 50. CRM system 40 is an external system that manages basic customer information, sales history, and contact person information. External information DB 50 is an external database service that stores or provides publicly available information such as corporate IR information, news releases, and industry reports.

[0025] Note that the external systems that System 1 integrates with are not limited to the CRM system 40 and the external information DB 50. For example, System 1 may integrate with a scheduling management system (not shown) such as Google Calendar® or Microsoft Outlook®. This makes it possible to automatically obtain the availability of sales representatives and insert them as suggested dates into the email text when generating sales emails (appointment request emails). System 1 may also integrate with a business chat tool (not shown) such as Slack® or Microsoft Teams®. This makes it possible to send a notification to the sales representative via the chat tool when the generation of sales materials is complete.

[0026] User terminal 10 is a client terminal (computer, information processing device) used by the user. The user is, for example, a sales representative. The user accesses server 20 via a web browser or dedicated application installed on an information terminal such as a personal computer (PC), tablet, or smartphone, and uses various functions provided by system 1. Specifically, on user terminal 10, it is possible to select the type of sales material to be generated, check customer information, and preview and edit the generated material.

[0027] Server 20 is an information processing device that forms the core of System 1 and is an example of a computer and information processing device according to one aspect of this disclosure. Server 20 is typically a cloud server composed of one or more computers. Server 20 receives requests from user terminals 10, collects necessary information from CRM system 40 and external information DB 50, and performs processing to generate sales materials in cooperation with AI system 30.

[0028] The AI ​​system 30 includes a generative AI model and is a system that provides access to said generative AI model. Specifically, for example, the AI ​​system 30 provides access to a large language model (LLM), which is a type of generative AI model.

[0029] A large-scale language model is a natural language model designed to perform multiple tasks in natural language processing. A large-scale language model is an example of a trained model, trained using a large number of parameters (e.g., billions to hundreds of billions) and high-level computational resources. A large-scale language model is a computer program or algorithm designed to perform tasks in natural language processing. For example, in natural language processing, processes such as morphological analysis, syntactic analysis, information extraction, or text generation enable computers to analyze human language (i.e., natural language) and perform predetermined processing. A large-scale language model receives a prompt (instruction) as input and generates output based on the text or image of that prompt.

[0030] Examples of large-scale language models include the GPT series (Generative Pre-Trained Transformer) developed by OPEN AI, StableLM developed by Stability AI, Llama2 developed by Meta, and Palm2® and LamDA2® developed by Google. Note that the generative AI model included in AI system 30 is not limited to large-scale language models; other language models may also be used. Other language models could include, for example, BERT (Bidirectional Encoder Representations from Transformers) developed by Google.

[0031] The AI ​​system 30 analyzes the input information and generates text content based on the instructions, in response to API calls (prompt input) from the server 20. The server 20 may also contain the functions of the AI ​​system 30; that is, the server 20 may have a generating AI model.

[0032] <User terminal hardware configuration> Figure 2 is a block diagram showing an example of the hardware configuration of a user terminal 10. As shown in Figure 2, the user terminal 10 comprises a control unit 101, a storage unit 102, a communication unit 103, an input unit 104, and an output unit 105. Each block included in the user terminal 10 is electrically connected, for example, by a bus.

[0033] The control unit 101 executes various processes by running various programs stored in the memory unit 102. The control unit 101 is, for example, a processor such as a CPU. A processor is hardware for executing instruction sets described in a program. A processor consists of an arithmetic unit, registers, and peripheral circuits.

[0034] The storage unit 102 includes a main memory and an auxiliary memory. The storage unit 102 stores various programs and various information. For example, the storage unit 102 stores an application program 120. The application program 120 includes, for example, a programming language that runs on a web browser application (not shown) stored in the storage unit 102.

[0035] The communication unit 103 performs modulation and demodulation processing for the user terminal 10 to communicate with an external device (e.g., server 20). The communication unit 103 performs transmission processing on the signal generated by the control unit 101 and transmits it to the external device. The communication unit 103 performs reception processing on the signal received from the external device and outputs it to the control unit 101.

[0036] The input unit 104 receives instructions or information input from the user. The input unit 104 is implemented, for example, by a keyboard, mouse, or touch panel. The output unit 105 presents information to the user. The output unit 105 is implemented, for example, by a display. The display displays various information according to the control of the control unit 101.

[0037] <User terminal functional configuration> Figure 3 is a block diagram showing the functional units realized by the control unit 101. The control unit 101 comprises an operation reception unit 131, a transmission / reception unit 132, and a presentation control unit 133 as functional units. Specifically, the control unit 101 realizes each functional unit by reading the application program 120 stored in the storage unit 102 and executing the instructions contained in the application program 120.

[0038] The operation reception unit 131 processes instructions or information input from the input unit 104. For example, the operation reception unit 131 accepts user selection of document type, input and modification of customer information, input of personality information, and editing operations for generated text content.

[0039] The transmitting / receiving unit 132 performs processing to enable the user terminal 10 to send and receive data with an external device according to a communication protocol. Specifically, the transmitting / receiving unit 132 transmits instructions (such as document generation requests) received by the operation reception unit 131 to the server 20. The transmitting / receiving unit 132 also receives customer information and generated sales document data transmitted from the server 20.

[0040] The presentation control unit 133 controls the output unit 105 to present various information to the user. Specifically, based on the information received from the server 20, the presentation control unit 133 generates various screens (GUIs) as shown in Figures 12 and 13, which will be described later, and displays them on the display.

[0041] <Server Hardware Configuration> Figure 4 is a block diagram showing an example of the hardware configuration of server 20. Server 20 comprises a control unit 201, a storage unit 202, a communication unit 203, and an input / output IF 204. The control unit 201 performs various processes by executing various programs stored in the storage unit 202. The control unit 201 is, for example, one or more processors such as a CPU or a GPU.

[0042] The storage unit 202 includes a main memory and an auxiliary memory. The storage unit 202 stores various programs and various information. In this embodiment, the storage unit 202 stores, for example, an application program 220, a customer DB 401, a company service DB 402, a prompt template DB 403, a personality DB 404, and a document template DB 405. In other words, the storage unit 202 stores a database 400 constructed from each of these tables.

[0043] Application program 220 is application software for executing and managing services provided by system 1. Users can access application program 220 on server 20, for example, by using a web browser application installed on user terminal 10.

[0044] The communication unit 203 performs modulation and demodulation processing for the server 20 to communicate with external devices (e.g., user terminal 10, AI system 30, CRM system 40). The input / output IF 204 functions as an interface for an input device (not shown) to receive input operations from the administrator of the server 20, and an output device (not shown) to output information to the administrator.

[0045] <Server Functional Configuration> Figure 5 is a block diagram showing the functional units implemented by the control unit 201. The control unit 201 comprises a communication control unit 211, an information management unit 212, a data acquisition unit 213, an analysis unit 214, a confirmation unit 215, a prompt generation unit 216, and a data generation unit 217 as functional units. Specifically, the control unit 201 implements each functional unit by reading a program (including an application program 220) stored in the storage unit 202 and executing the instructions contained in the program.

[0046] The communication control unit 211 controls data communication between the user terminal 10, the AI ​​system 30, the CRM system 40, and the external information DB 50. The information management unit 212 manages the reading and writing of various data in the database 400.

[0047] The data acquisition unit 213 acquires various information necessary for generating sales materials for customers. Specifically, the data acquisition unit 213 acquires basic customer information and information related to transactions with customers from the CRM system 40.

[0048] "Customer basic information" refers to static or semi-static information used to identify a customer and define their basic attributes. Specifically, if the customer is a corporation, customer basic information includes information such as the company name, address, industry, number of employees, annual sales, name of the representative, name of the person in charge, and the person in charge's job title. If the customer is an individual, it includes name, address, and attribute information.

[0049] "Information regarding customer transactions" refers to dynamic information that shows the history of all interactions that have occurred between our company and customers from the past to the present, including information before the start of negotiations, information during negotiations, and information regarding interactions during projects after an order has been placed. Specifically, information regarding customer transactions includes past and ongoing negotiation history, meeting minutes data from negotiations, quotation submission history, inquiry history, support response records, and past reasons for lost orders. Furthermore, documents created during the progress of a project after an order has been placed, such as minutes of regular meetings, progress reports, issue management sheets, or change requests, are also included in information regarding customer transactions. Including these documents enables accurate proposals that take into account the current status of the project in the context of upselling or cross-selling (additional proposals) to existing customers.

[0050] Furthermore, the data acquisition unit 213 acquires external customer information from the external information DB 50. "External customer information" refers to information about the customer itself or the market environment surrounding the customer, which can be obtained from publicly available sources, independently of the direct business relationship between the company and the customer. Specifically, external customer information includes the customer company's IR information, the latest news releases, market information, reputation on social media, interview articles, and the activities of competitors. In addition, patent information (patent gazettes, published patent gazettes) filed or held by the customer company is also included in external customer information. By analyzing patent information, it becomes possible to understand the technological fields that the customer is currently focusing on and the direction of future business development.

[0051] The analysis unit 214 analyzes the text data contained in the acquired basic information, transaction-related information, and external information to identify potential customer issues. The analysis unit 214 uses, for example, a natural language processing model to extract predefined keywords (e.g., "cost reduction," "enhanced security," "deterioration," "DX," "efficiency improvement") and the context of those keywords from the text data. In this embodiment, the natural language processing model is assumed to be a generative AI model built into the AI ​​system 30.

[0052] Furthermore, the natural language processing model is not limited to the generative AI model embedded in the AI ​​system 30. For example, other deep learning models such as BERT (Bidirectional Encoder Representations from Transformers) and RoBERTa, or machine learning models such as SVM (Support Vector Machine) or CRF (Conditional Random Fields) may be used. In this case, the natural language processing model may be embedded in the server 20 or the AI ​​system 30, or it may be embedded in another AI system (not shown) different from the AI ​​system 30.

[0053] "Context" refers to the surrounding information associated with a keyword in order to identify its semantic position, importance, or nuance. Specifically, context includes the sentences before and after the keyword appears, its dependency relationships, or its sentiment score (degree of positive / negative). For example, if the keyword "cost reduction" is accompanied by sentences such as "we want to achieve this urgently," or a strong positive sentiment score, then the keyword is recognized as an important issue. On the other hand, if the context is "cost reduction is secondary," then it would be a lower priority.

[0054] The analysis unit 214 comprehensively analyzes the keywords and context extracted by the generating AI model (natural language processing model) to identify potential customer issues as structured data. "Structured data" as potential customer issues refers to information obtained by converting the analysis results of unstructured or semi-structured text data (meeting minutes, news articles, support history, etc.) contained in the customer's basic information, information related to transactions with the customer, and external information of the customer into a machine-readable format for subsequent service selection processing. Structured data is written in a format such as JSON and has attributes such as the type of issue, specific content, and severity. For example, a potential customer issue may be identified as "increased maintenance costs due to aging of the core system (severity: high)."

[0055] The confirmation unit 215 selects the most suitable in-house service for solving the problem by referring to the in-house service DB 402, based on the problem identified as structured data and predefined selection rules that define the selection criteria for the in-house service.

[0056] A "selection rule" is the logic used to determine which of the company's services to recommend for a given problem. For example, a selection rule might be: "If the problem is 'cost reduction,' prioritize the low-cost Plan A," or "If the problem is 'security,' prioritize the high-performance Plan B."

[0057] The confirmation unit 215 confirms the proposal requirements, which are structured data including the selected company services and identified issues. "Proposal requirements" refers to data that describes the results of identifying the customer's potential issues and information about the company services selected to solve those issues in a format that can be mechanically interpreted and used by the subsequent prompt generation process (step S504). Specifically, the proposal requirements are information assets composed of key-value pairs that can be substituted into predetermined variables (placeholders) of the prompt template, containing information about the identified issues (e.g., "Aging core system") and the selected company services (e.g., "Service B: Phased migration"). The proposal requirements become the data that defines the core logic (outline of the proposal) of the sales materials described later.

[0058] The prompt generation unit 216 generates prompts for the generated AI model based on the confirmed proposal requirements. Prompts are mainly inquiry statements input to the generated AI model and are used by the prompt generation unit 216 to instruct the generated AI model to process information. Inquiry statements may include, for example, text, strings, images, videos, audio, etc. Prompts may also include gestures. In this embodiment, prompts include instruction statements, reference data, and information for specifying the output format.

[0059] An "instruction statement" is text data that indicates the content of the information processing to be performed by the generation AI model. Specifically, it is text data that instructs the generation AI model to generate text content that constitutes customer-facing sales materials. In this embodiment, this corresponds to the structure definition of a template stored in the prompt template DB403 (e.g., "You are an excellent consultant. Please create a draft proposal based on the following information.").

[0060] "Reference data" refers to data that the generating AI model uses as a reference (for learning) when performing information processing. In this embodiment, the proposed requirements determined by the confirmation unit 215 fall under this category. The prompt generation unit 216 includes the reference data in the prompt by embedding these proposed requirements in a placeholder in the template.

[0061] Information used to specify the output format includes, for example, specifying a JSON schema, XML format, or other structured data format. This information allows the output from the generating AI model to be obtained in a predetermined structured format (e.g., JSON data with titles and body text for each slide). Based on this information, the generating AI model generates output results according to the specified format.

[0062] The prompt generation unit 216 uses prompt templates stored in the prompt template DB 403. A "prompt template" refers to a template of a logical structure (story structure) corresponding to a predetermined business logic. "Business logic" is an algorithm for converting input data into output data in a specific order, and specifically refers to an effective persuasive procedure and structure depending on the type of sales material (e.g., proposal, email). For example, for a proposal, a standard proposal story structure such as "problem presentation (clarifying the current problem) → solution (presenting a solution using the company's services) → effect (presenting the benefits or ROI of implementation)" is templated. In addition, "standard industry practices," that is, the structure, expression style, or set format of materials that are considered common or effective in a particular industry or business scene (e.g., structure based on frameworks such as AIDMA or SPIN), are also reflected.

[0063] The prompt generation unit 216 incorporates the finalized proposal requirements into predetermined slots in the prompt template, thereby completing a prompt that instructs the generation AI model to generate text content that constitutes customer-facing sales materials.

[0064] The document generation unit 217 inputs the generated prompt into the AI ​​model of the AI ​​system 30 to output text content. "Text content" refers to text data such as specific sentences, headings, and bullet points that make up sales materials for customers.

[0065] The document generation unit 217 generates the final customer-facing sales material by incorporating the output text content into a predetermined document template stored in the document template DB 405. A "document template" refers to template data that defines the appearance (format) of the customer-facing sales material. Specifically, a document template is a PowerPoint® slide master or an email format, etc., with a predefined design and layout.

[0066] Furthermore, the types of customer-facing sales materials generated may include proposals, sales emails, or role-playing scripts, as well as, for example, a Q&A section for anticipated questions and answers (FAQs) for business negotiations. The Q&A section is an internal document that predicts potential questions customers may ask about the proposal and compiles answers to those questions. It is generated by the AI ​​system 30 by inferring logical weaknesses and unclear points in the proposal.

[0067] <Data structure> Figures 6 to 10 show examples of the data structure of tables stored in the storage unit 202. The tables shown in Figures 6 to 10 refer to relational databases, which are data sets called tabular tables, structurally defined by rows and columns, and are used to manage and associate these sets with each other. In databases, tables are called tables, the columns of a table are called columns, and the rows of a table are called records. In relational databases, relationships between tables can be established and associated.

[0068] Typically, each table has a primary key column to uniquely identify records, but setting a primary key column is not mandatory. The control unit 201 can instruct the processor to add, delete, or update records in specific tables stored in the storage unit 202 according to various programs.

[0069] Please note that Figures 6 to 10 are merely examples and do not exclude any data not shown. Furthermore, even data listed in the same table may be stored in separate memory areas within the memory unit 202.

[0070] Figure 6 shows the data structure of Customer DB401. Customer DB401 is a database that manages basic customer information. Customer DB401 consists of columns such as Customer ID, Company Information, Contact Person Information, and Linkage ID.

[0071] The "Customer ID" column stores identification information (e.g., "C001") to uniquely identify each customer. The "Company Information" column stores basic attribute information such as the customer's company name (e.g., "XX Corporation"), industry (e.g., "Manufacturing"), location (e.g., "Minato-ku, Tokyo..."), number of employees, and annual sales. The "Contact Person Information" column stores information about the customer's contact person, such as their name, department, position, and contact information. For example, the "Contact Person Information" column also includes the contact person ID (identification information) linked to the personality DB 404 described later. The "Linkage ID" column stores identification information to identify the customer's record in the external CRM system 40. This allows the server 20 to retrieve negotiation history and meeting minutes data from the CRM system 40.

[0072] Figure 7 shows the data structure of the proprietary service DB402. The proprietary service DB402 is a database that manages information on the proprietary services that can be proposed. The proprietary service DB402 consists of columns such as service ID, service overview, evaluation attributes, performance data, and addressing issue tags.

[0073] The "Service ID" column stores identification information (e.g., "S001") to uniquely identify each company service. The "Service Overview" column stores information that outlines the company's services, such as the service name and a detailed description of what is provided. The "Evaluation Attributes" column stores attribute values ​​used for evaluation according to the selection rules. Specifically, for example, the "Evaluation Attributes" column stores "Cost Level" (e.g., "Low," "Medium," "High") indicating the cost of implementation, "Implementation Speed" (e.g., "1 month," "3 months") indicating the time to implementation, and "Security Level" (e.g., "High") indicating the strength of security features. The "Performance Data" column stores quantitative performance values ​​such as the number of past implementations and the number of success stories. "Success stories" refer to implementation results with customers in the same industry or of similar size, or cases where an improvement in ROI (Return on Investment) was confirmed after implementation. The "Issue Tag" column stores a list of issues that each company's service can solve (e.g., ["Cost Reduction," "Business Efficiency Improvement," "DX"]).

[0074] Figure 8 shows the data structure of the prompt template DB403. The prompt template DB403 is a database that manages templates for instructions to the generated AI model. The prompt template DB403 consists of columns such as template ID, document type, and structure definition.

[0075] The "Template ID" column stores identification information (e.g., "PT001") to uniquely identify each template. The "Document Type" column stores information indicating the type of document to be generated (e.g., "Proposal (for Executives)", "Sales Email (Initial Approach)", "Role-Playing Script", "Expected Q&A"). The "Structure Definition" column stores text data representing the body of the prompt. This text data includes a role definition for the generating AI model (e.g., "You are an excellent consultant") and a specification of the output format, as well as placeholders for dynamically filling in values ​​(e.g., {Problem}, {Solution}, {Tone & Manner}).

[0076] Figure 9 shows the data structure of Personality DB404. Personality DB404 is a database that manages personality information of customer representatives. "Personality information" refers to information that includes at least one of the customer representative's decision-making tendencies and their attitude towards work. Specifically, personality information is qualitative attribute information that indicates the tendencies of an individual customer representative in business communication, decision-making, or thinking and behavior in performing their duties. This is distinct from objective "basic customer information" such as company name or number of employees, and is used to optimize the content of proposals (which service to choose) or the method of proposal (what tone to use for explanation). It should be noted that it is not mandatory for Server 20 to have Personality DB404; it is sufficient if, for example, personality information is to be utilized for the purpose of further optimizing proposal requirements. Personality DB404 consists of columns such as Representative ID, Decision-Making Tendencies, Work Attitude, and Communication Style.

[0077] The "Representative ID" column stores identification information linked to the representative information in Customer DB401. The "Decision-Making Tendency" column stores a value indicating what the customer representative prioritizes in decision-making (e.g., "Cost-Focused," "Speed-Focused," "Quality-Focused," "ROI-Focused"). The "Work Attitude" column stores the customer representative's stance towards business (e.g., "Offensive (Innovation / Challenge)," "Defensive (Stability / Risk Avoidance)"). The "Communication Style" column stores the customer representative's preferred way of receiving information (e.g., "Logical / Data-Focused," "Emotional / Vision-Focused").

[0078] Figure 10 shows the data structure of the document template DB405. The document template DB405 is a database that manages the layout files of the final deliverables. The document template DB405 consists of columns such as template ID, file type, and file body.

[0079] The "Template ID" column stores identification information to uniquely identify each document template. The "File Type" column stores the file format of each document template (e.g., "pptx", "docx"). The "File Body" column stores the actual file for each document template with a defined layout, or the reference path to the storage where the file is stored. In the example in Figure 10, the reference path is stored in this column. The file contains, for example, placeholders into which text content is inserted.

[0080] <Operation> The following describes an example of server 20 operation, referring to the flowchart in Figure 11. Figure 11 is a flowchart showing the processing flow when server 20 generates sales materials for customers.

[0081] Before explaining the flowchart, we will describe the prerequisites for this example. The user is assumed to have already logged in to the service provided by the server 20 via the application program 120 or web browser on the user terminal 10, and is in a state where the process of generating sales materials for customers can be started. For example, it is assumed that the user selects a customer for whom materials are to be created from the customer list (not shown) displayed on the user terminal 10's screen, proceeds through the material type selection screen (not shown), and then transitions to a screen like the one shown in Figure 12 or Figure 13, which will be described later.

[0082] First, in step S501 (the acquisition step), the data acquisition unit 213 of the server 20 receives a document generation instruction sent from the user terminal 10. This instruction includes, for example, the target customer ID and the type of document to be created (e.g., a proposal). The data acquisition unit 213 uses the received customer ID as a key to send a query to the customer DB 401 and acquires the customer's basic information. The data acquisition unit 213 also uses the linkage ID obtained from the customer DB 401 to call the API of the CRM system 40 and acquires the most recent sales meeting minutes text data, transaction history information, etc. (information related to transactions with the customer). Furthermore, the data acquisition unit 213 searches the external information DB 50 as needed and acquires the latest news articles, IR information, patent information, etc. (external information of the customer) related to the customer.

[0083] Next, in step S502 (identification step), the analysis unit 214 of the server 20 analyzes the acquired information. Specifically, the analysis unit 214 transmits the acquired information to the AI ​​system 30 and has it execute a natural language processing (NLP) task. The generating AI model (natural language processing model) uses techniques such as named entity recognition (NER) and sentiment analysis to extract keywords related to the customer's potential issues and the context of those keywords from the information. The analysis unit 214 comprehensively analyzes the response (extraction results) from the AI ​​system 30 and identifies the customer's potential issues as structured data such as JSON format.

[0084] Next, in step S503 (the confirmation step), the confirmation unit 215 of the server 20 selects its own services based on the identified issues and selection rules. Specifically, the confirmation unit 215 searches the in-house service DB 402 for services that have "corresponding issue tags (a list of issues that can be solved by the in-house services)" corresponding to the identified issues. Then, the confirmation unit 215 selects, for example, the in-house service with the highest score. Subsequently, the confirmation unit 215 combines information about the selected in-house service (e.g., name, features, and benefits) with the identified issues to confirm the proposed requirements, which are structured data.

[0085] Next, in step S504 (the step of generating a prompt), the prompt generation unit 216 of the server 20 retrieves a record corresponding to the document type specified by the user from the prompt template DB 403. Then, the prompt generation unit 216 generates a prompt by substituting (slot-filling) each element of the proposed requirements confirmed in step S503 into the placeholders (e.g., {issue}, {solution}) in the retrieved prompt template.

[0086] Next, in step S505, the document generation unit 217 of the server 20 sends the generated prompt to the AI ​​system 30 via API. The AI ​​system 30 inputs the received prompt into the generating AI model. The generating AI model performs inference according to the input prompt, generates text content such as the wording of each slide in the proposal or the subject and body of an email, and returns it to the server 20 as a response.

[0087] Finally, in step S506 (the step of generating sales materials for customers), the material generation unit 217 incorporates the text content received from the AI ​​system 30 into the corresponding text boxes of a predetermined material template stored in the material template DB 405. This allows the material generation unit 217 to generate a file of sales materials for customers. The generated file is, for example, stored in a temporary area of ​​the server 20 and provided in response to a download request from the user terminal 10.

[0088] <Screen example> Figure 12 shows an example of a basic information confirmation screen 600 displayed on the user terminal 10. The left-hand side of the screen contains a name display area 601 where the name of the selected customer, the name of the customer's representative, etc., are displayed. Specifically, for example, the name display area 601 initially displays values ​​obtained from the customer DB 401 by the data acquisition unit 213. Edit buttons 603 are also provided next to each item in the name display area 601. When the user presses an edit button 603, the information displayed in the text boxes for each item (e.g., customer name, customer's representative name) becomes editable.

[0089] Furthermore, a personality information display area 602 is provided in the right-hand area of ​​the screen where personality information is displayed. Specifically, for example, the current settings (e.g., cost-conscious, aggressive stance) are displayed as tags or icons in the personality information display area 602, and can be changed by clicking to toggle or by selecting from a pull-down menu.

[0090] In the example shown in Figure 12, personality information is displayed along with the customer's basic information. However, a dedicated screen (not shown) for inputting or editing only personality information may be provided. Furthermore, personality information is not limited to such screen input. For example, the server 20 may analyze past transaction history, email content, etc., to automatically infer the characteristics of the person in charge, and automatically retrieve and reflect the values ​​stored in the database (personality DB 404). Alternatively, the basic information confirmation screen 600 may display only the customer's basic information.

[0091] Figure 13 shows an example of the proposal requirements confirmation screen 700 displayed on the user terminal 10. The left-hand area of ​​the screen (facing the user) contains the target customer's problem recognition / order section 701. This section displays the potential customer problems analyzed and identified by the generation AI model. The right-hand area of ​​the screen (facing the user) contains the problem / solution hypothesis section 702. This section displays one or more proposal requirements confirmed by the confirmation unit 215. The user can check the contents of these sections and, if necessary, press the regenerate button 703 to have the server 20 regenerate the problems and proposal requirements, or manually modify the problems and proposal requirements. It is also possible to upload reference materials for the generation AI model to refer to.

[0092] <Summary> Thus, according to this embodiment, a generating AI model comprehensively analyzes vast amounts of internal and external data, and the server 20 automatically generates optimal prompts based on predetermined business logic. This enables the automatic generation of high-quality sales materials that accurately capture customer challenges, without relying on the individual skills of sales representatives. As a result, sales activities become more efficient and standardized, and the order rate is improved.

[0093] [Variation] The embodiments described above are merely examples, and various modifications are possible without departing from the spirit of this disclosure. Several modifications are described below.

[0094] <First variation> In the embodiments described above, a process for determining the proposed requirements without using personality information was illustrated, but the invention is not limited to this case. The determination unit 215 may, for example, adjust the selection rules using personality information.

[0095] Specifically, for example, Personality DB404 classifies the personality of customer representatives based on well-known behavioral psychology models such as "DiSC theory" or "Social Style theory," and stores this information as parameters. Personality DB404 manages scores or tags indicating, for example, whether the attributes of a customer representative fall under "Driver (Executive)," "Expressive (Intuitive)," "Amiable (Gentle)," or "Analytical (Analytical)."

[0096] The confirmation unit 215 updates the selection rules by changing the priority in the selection rules or by applying a different selection rule associated with the acquired personality information. In other words, the confirmation unit 215 changes the weighting coefficients of the evaluation function in the selection rules based on the aforementioned personality classification. For example, if the customer's representative is of the "analytical" type, the confirmation unit 215 increases the weighting of the evaluation items "number of past successes" and "quantitative basis for return on investment (ROI)" in the selection rules from the usual 1.0x to 1.5x. Conversely, if the customer's representative is of the "expressive" type, the confirmation unit 215 increases the weighting of the evaluation item "novelty of the service" or "newsworthiness." In this way, the confirmation unit 215 dynamically optimizes the selection logic of its own services so that services that match the thinking characteristics of the customer's representative are ranked higher in the scoring.

[0097] Furthermore, the prompt generation unit 216 includes in the prompt information that specifies the logical structure or expression style of the customer-facing sales material, determined according to the acquired personality information. "Expression style" refers to the linguistic style of the generated text content or attributes that define the impression given to the reader. In other words, the prompt generation unit 216 determines the tone and manner specification prompt based on the personality classification described above. "Tone and manner specification prompt" refers to additional instruction sentences prepared in advance to specify the style, tone, or expression style of the text content output by the generating AI model, according to the personality classification. The tone and manner specification prompt may be stored in the prompt template DB 403 separately from the material type, or it may be stored in the personality DB 404 in association with each style.

[0098] The prompt generation unit 216 inserts (or appends to the end of) the determined tone and manner specified prompt into the {tone and manner} placeholder of the basic prompt (structure definition) obtained from the prompt template DB403. Specifically, for example, the prompt generation unit 216 adds a system prompt for the "driver" type that reads, "State the conclusion first, and write concisely using bullet points. Minimize the use of adjectives and use assertive language." On the other hand, for the "amiable" type, the prompt generation unit 216 adds the instruction, "Choose empathetic words and use expressions that emphasize cooperativeness and a sense of security. Use a polite writing style using the 'desu' and 'masu' forms." This allows the style of the generated text to be automatically adjusted to the format most preferred by the reader.

[0099] <Second variation> Server 20 may, for example, generate a re-contact email as sales material for customers if the customer is a dormant customer. Furthermore, in the process of generating this email, Server 20 may further improve the accuracy of acquiring and utilizing the customer's external information.

[0100] A "dormant customer" refers to a customer who has not been contacted for a long period of time. Specifically, a dormant customer is a customer whose past transaction history or sales negotiation history is recorded in the CRM system 40 (or customer DB 401), but for which a predetermined period (e.g., 180 days, 1 year, etc.) has elapsed since the last contact (e.g., sales negotiation, email exchange, support response), and for whom no proactive approach has been taken.

[0101] Specifically, for example, the data acquisition unit 213 periodically visits news sites, press release distribution sites, and personnel change information sites on the internet using crawling technology or external APIs to acquire the latest external information. The "latest external information" referred to here is not limited to information that can be acquired at the time the instruction to generate sales materials for customers is received, but also refers to differential information that has been newly published since the last contact date and time.

[0102] The selection unit 215 selects candidate company services to propose by comparing keywords extracted from the latest information on transactions with customers (e.g., reasons for past lost deals) and the latest external information on customers with selection rules. Specifically, the selection unit 215 extracts trigger events (hereinafter referred to as "trigger events") that trigger sales opportunities from the acquired latest external information as keywords. Trigger events include, for example, the announcement of a medium-term management plan, relocation of headquarters, changes in executives, fundraising, and the launch of a new business. The selection unit 215 compares the type of trigger event detected with the selling points defined in the company's service DB 402.

[0103] A "selling point" is predefined attribute information that indicates which customer situations or trigger events each company's service is particularly effective for (can strongly appeal to)

[0104] The document generation unit 217 generates a follow-up email containing information about the selected candidate company services. Specifically, the document generation unit 217 instructs the generation AI model to use this trigger event as a topic hook (introduction). In particular, the document generation unit 217 instructs the generation AI model to construct a context that begins with a reference to past events, such as, "I spoke with Mr. / Ms. XX (predecessor) of your company last year in X month," followed by, "I saw that your company recently announced △△ (trigger event), and I thought our □□ service might be helpful, so I contacted you." This emphasizes that it is not just a mass-sent sales email, but a proposal tailored to individual circumstances, thereby improving open and reply rates.

[0105] <Third variation> For example, in the process of identifying potential customer issues, server 20 may have a generating AI model (natural language processing model) analyze the meeting minutes data included in the transaction information. Server 20 may then identify potential customer issues as structured data by having the generating AI model extract predefined keywords and the context of those keywords. In other words, in the third modified example, natural language processing technology may be utilized more advancedly.

[0106] Specifically, for example, the analysis unit 214 performs morphological and syntactic analysis on meeting minutes data (text data), which is an example of information related to transactions with customers, and further performs sentiment analysis.

[0107] For example, if the keyword "cost" is associated with negative sentiments such as "it's a problem if it costs too much" or "our budget is tight," the analysis unit 214 extracts "cost" as an issue to be resolved and sets its severity to "high." On the other hand, if the keyword appears in a positive context, such as "there are cost benefits," as stated by a company sales representative (who may be a user or someone other than a user), the analysis unit 214 distinguishes that statement as a proposal element rather than an issue.

[0108] Furthermore, the analysis unit 214 uses a topic model (such as LDA) to analyze the distribution of frequently occurring topics throughout the meeting minutes data, and derives an abstract issue category such as "decreased system performance" from the co-occurrence relationships of related words (e.g., "slow," "heavy," "stops") as well as a single keyword. This makes it possible for the analysis unit 214 to identify potential issues that are not explicitly stated as "the issue is XX" as structured data.

[0109] Furthermore, the analysis unit 214 does not necessarily have to rely solely on a generative AI model (natural language processing model) for natural language processing. The analysis unit 214 may also perform analysis based on a predefined set of rules (rule-based analysis) as a non-AI method. For example, the analysis unit 214 may apply conditional branching logic such as, "If the industry is 'manufacturing' and the number of employees is '1000 or more,' then it is determined that there is a problem with 'supply chain management'." Alternatively, the analysis unit 214 may perform statistical analysis, such as analyzing the correlation between information about transactions with customers and external information about customers (e.g., frequency of occurrence of specific keywords), and then segment (cluster) customer groups that have specific patterns to infer the problems.

[0110] <Fourth variation> Server 20 may, for example, select its own services by employing a multi-objective optimization method. That is, the selection rules may include a ranking rule that defines the priority of the services to be selected, using at least one of the following as an evaluation metric: cost, risk, or the number of past successes.

[0111] Specifically, for example, the confirmation unit 215 has an evaluation function that scores the degree of suitability between the identified issues and each of the company's services. This evaluation function is calculated by a weighted average of multiple evaluation indicators such as "cost suitability," "delivery time suitability," "functional sufficiency," "performance suitability (number of implementation cases in the same industry)," and "risk reduction." Here, "risk" refers to, for example, the risk of business disruption due to the introduction of the company's services, the risk of data loss during system migration, or the risk that operating costs after implementation will exceed the budget.

[0112] The selection rules are defined as a set of weighting coefficients (weight vectors) for the evaluation indicators used in the evaluation function. The determination unit 215 sets a high weighting coefficient for cost fit in the case of a "cost-focused" rule, and sets high weighting coefficients for performance fit and risk reduction in the case of a "risk-focused" rule.

[0113] The confirmation unit 215 calculates a score for all of the company's services using an evaluation function and selects the top services (for example, the top three) whose scores exceed a predetermined threshold as proposed candidates. Here, the confirmation unit 215 may also include logic to deliberately select three plans with different balances of price range and functionality, such as "Pine, Bamboo, and Plum." This makes it possible to determine proposal requirements that allow customers to present a range of options for comparison.

[0114] Furthermore, selection rules are not necessarily limited to weighted average calculation formulas explicitly defined by humans (the evaluation function mentioned above). For example, selection rules may be implemented as machine learning models. Specifically, a decision tree model trained on past similar case data (e.g., customer attributes, challenges, proposed services, and order results) can be used as a selection rule. In this case, the trained decision tree model functions as a set of rules (predefined selection rules) such as "if the challenge is A and the personality is B, select service X." Similarly, recommendation systems (recommendation engines) such as collaborative filtering can also be used as a type of selection rule to select the optimal company service based on past data patterns. This makes it possible to select services with a high probability of success, taking into account complex factors that cannot be captured by rule-based systems alone.

[0115] <Fifth variation> System 1 may, for example, display and edit the generation processes of various information and data using a more interactive UI (User Interface).

[0116] Specifically, for example, server 20 uses bidirectional communication technologies such as WebSocket to push real-time updates to user terminal 10 on the progress of multiple agent processes running in the background (market research agent, analysis agent, writing agent, etc.). This is an example of a substep that displays information indicating the execution status of multiple internal processes related to the generation of text content. User terminal 10 sequentially displays which processes are currently running and what intermediate data (e.g., searched news article headlines, a list of identified issues) has been obtained. This allows the user to transparently observe the thought process of the generating AI model.

[0117] Furthermore, the server 20 may further perform the following processes before incorporating the text content output from the generating AI model into a predetermined document template: displaying it in an editable format and accepting editing input for the displayed text content. That is, the server 20 adds a function to the editing screen of various generation results on the user terminal 10 that allows the user to issue regeneration instructions for each block of the generated text data (e.g., "Background," "Issues," "Solutions," etc.). Specifically, for example, if only the description of "Solutions" is insufficient, the user selects that section on the editing screen and enters correction instructions (refinement prompts) such as "Be more specific" or "Add examples." The document generation unit 217 partially regenerates and replaces only that section based on the correction instructions received via the user terminal 10. The customer sales material is generated when the text content after accepting editing input is incorporated into a predetermined document template. The corrections made by the user are saved as history and may be used as feedback for learning data (examples of Few-Shot prompts) for the next generation.

[0118] While several embodiments of this disclosure have been described above, these embodiments can be implemented in a variety of other forms, and various omissions, substitutions, and modifications are permitted without departing from the spirit of the invention. These embodiments and their variations are included within the scope and spirit of the invention, as well as within the scope of the claims and its equivalents.

[0119] [Note] The details described in each of the above embodiments are noted below.

[0120] <Note 1> A program for operating a computer having one or more processors, which causes one or more processors to execute the following steps: acquiring basic customer information, information on transactions with the customer, and external customer information; analyzing the text data contained in each of the acquired basic information, transaction information, and external information, and extracting predefined keywords and the context of those keywords to identify the customer's potential problems as structured data; selecting a service from the company to solve the identified problems by referring to a database storing information on multiple company services based on the identified problems and predefined selection rules that define the selection criteria for the company's services, and determining proposal requirements, which are structured data including the selected company services and problems; incorporating the determined proposal requirements into a prompt template that templates a logical structure corresponding to a predetermined business logic, and generating a prompt that includes an instruction sentence instructing a generating AI model to generate text content constituting sales materials for the customer; and inputting the generated prompt into the generating AI model to output text content, and generating sales materials for the customer by incorporating the output text content into a predetermined document template.

[0121] <Note 2> The program described in (Appendix 1) involves obtaining personality information, including at least one of the client's representative's decision-making tendencies and their attitude towards work, in the finalization step, and then finalizing the proposal requirements in consideration of this personality information.

[0122] <Note 3> The program described in (Appendix 2) updates the selection rules by changing the priority in the selection rules or by applying a different selection rule associated with the acquired personality information in the confirmation step.

[0123] <Note 4> The program described in (Appendix 2) or (Appendix 3) includes in the step of generating a prompt information that specifies the logical structure or presentation style of customer-facing sales materials, determined according to the acquired personality information, in the prompt.

[0124] <Note 5> If a specified number of days have passed since the last contact date with the customer, the program, as described in any of (Appendix 1) to (Appendix 4), in the confirmation step, matches keywords extracted from the latest information regarding the transaction with the customer and the latest external information about the customer with selection rules to select candidate company services to propose, and in the step of generating sales materials for the customer, generates a follow-up email containing information about the selected candidate company services.

[0125] <Note 6> In the identification step, a natural language processing model analyzes the meeting minutes data contained in the transaction information and extracts predefined keywords and the context of those keywords to identify the issues as structured data, as described in any of the programs in (Appendix 1) to (Appendix 5).

[0126] <Note 7> The selection rules are those described in any of the programs (Appendix 1) to (Appendix 6), including ranking rules that define the priority of the company's services to be selected, using cost, risk, or the number of past success stories as evaluation indicators.

[0127] <Note 8> The program described in any of (Appendix 1) to (Appendix 7) further includes a substep for generating customer sales materials, which includes a substep for displaying information indicating the status of execution of several internal processes related to the generation of text content.

[0128] <Note 9> The program, as described in any of (Appendix 1) to (Appendix 8), further includes a substep of displaying the text content output from the generating AI model in an editable format before incorporating it into a predetermined document template, and a substep of accepting editing input for the displayed text content, wherein the customer sales material is generated when the text content after accepting editing input is incorporated into a predetermined document template.

[0129] <Note 10> An information processing device comprising a processor, wherein the processor executes all steps in any of the programs described in (Appendix 1) to (Appendix 9).

[0130] <Note 11> A method to be performed on a computer equipped with a processor, wherein the processor performs all steps in any of the programs described in (Appendix 1) to (Appendix 9).

[0131] <Note 12> A system comprising one or more processors that execute all steps in any of the programs described in (Appendix 1) to (Appendix 9). [Explanation of Symbols]

[0132] 1... System 10…User terminal 20... Server 30…AI system 40…CRM system 50…External Information Database 101, 201… Control Units 102, 202...Storage section 103, 203... Communications Department 104...Input section 105...Output section 211...Communication Control Unit 212…Information Management Department 213...Data acquisition unit 214…Analysis Department 215...Confirmation part 216... Prompt generation unit 217…Data generation department 400... Database 401…Customer DB 402…Company's own service database 403…Prompt Template DB 404...Personality Database 405...Document Template Database 600...Basic Information Confirmation Screen 700... Proposal Requirements Confirmation Screen

Claims

1. A program for operating a computer having one or more processors, The steps include obtaining basic customer information, information regarding transactions with the customer, and external information of the customer, The steps include: analyzing the text data contained in the acquired basic information, transaction-related information, and external information, and extracting predefined keywords and the context of those keywords to identify the customer's potential issues as structured data; Based on the identified problem and predefined selection rules that define the selection criteria for the company's services, the company selects a service to solve the identified problem by referring to a database that stores information on multiple company services, and then determines the proposed requirements, which are structured data including the selected company service and the problem. The steps include: generating a prompt that includes an instruction to the generating AI model to generate text content constituting customer sales materials, by incorporating the finalized proposal requirements into a prompt template that templatees a logical structure corresponding to a predetermined business logic; The steps include: inputting the generated prompt into the generating AI model to output the text content, and incorporating the output text content into a predetermined document template to generate the customer-facing sales document; A program that causes one or more of the aforementioned processors to execute it.

2. The program according to claim 1, wherein in the step of determining the requirements, personality information is obtained, including at least one of the decision-making tendencies and attitude toward work of the customer's representative, and the proposal requirements are determined taking into account the personality information.

3. The program according to claim 2, wherein in the confirmation step, the selection rules are updated by changing the priority in the selection rules or by applying another selection rule associated with the acquired personality information.

4. The program according to claim 2, wherein the step of generating the prompt includes in the prompt information specifying the logical structure or presentation manner of the customer sales material determined according to the acquired personality information.

5. If a predetermined number of days have passed since the last contact date with the aforementioned customer, In the aforementioned confirmation step, keywords extracted from the latest information regarding the transaction with the customer and the latest external information about the customer are compared with the selection rules to select candidates for the company's services to be proposed. The program according to claim 1, wherein the step of generating sales materials for the customer generates a follow-up email message containing information about the selected candidate services of the company.

6. The program according to claim 1, wherein in the step of identification, a natural language processing model analyzes the minutes data contained in the information relating to the transaction and identifies the problem as structured data by extracting predefined keywords and the context of those keywords.

7. The program according to claim 1, wherein the selection rules include a ranking rule that defines the priority of the company's services to be selected, with cost, risk, or the number of past success stories as the evaluation indicator.

8. The program according to claim 1, wherein the step of generating customer sales materials further includes a substep of displaying information indicating the execution status of a plurality of internal processes related to the generation of the text content.

9. In the step of generating the aforementioned sales materials for customers, A substep in which the text content output from the generating AI model is displayed in an editable format before being incorporated into the predetermined document template, The method further includes a substep for receiving editing input for the displayed text content, The program according to claim 1, wherein the customer sales material is generated by incorporating the text content after receiving the edit input into the predetermined document template.

10. An information processing device comprising a processor, wherein the processor executes all steps in any one of claims 1 to 9.

11. A method to be performed on a computer having a processor, wherein the processor performs all steps of a program according to any one of claims 1 to 9.

12. A system comprising one or more of the processors that perform all steps in the program according to any one of claims 1 to 9.

Images