Business scenario creation support system and business scenario creation support method

Abstract

This invention provides a business scenario creation support system and a business scenario creation support method that assist in the creation of business scenarios by effectively utilizing information publicly available on the internet. [Solution] The business scenario creation support server 1 collects business information about each company's business that is publicly available on the Internet 101, inputs a prompt containing specific information to identify the company to be supported into the large-scale language model 21, and if the large-scale language model 21 generates business scenario information about the business scenario of the company to be supported based on the collected business information, the server 1 retrieves that business scenario information from the large-scale language model 21 and provides the retrieved business scenario information to the company to be supported.

Description

Technical Field

[0001] The present invention relates to a business scenario creation support system and a business scenario creation support method for assisting a company in creating a business scenario.

Background Art

[0002] Various technologies for assisting in creating future business scenarios in companies have been proposed. For example, Patent Document 1 discloses a business evaluation system that evaluates business value based on predictions of investments required for the progress of a business and revenues obtained from the business when creating a business scenario. This business evaluation system can improve the efficiency of probabilistic business value evaluation arithmetic processing by including a schedule that branch-displays a plurality of events assumed at the planning time and a business progress change criterion that sets criteria for changing the progress of the business depending on the branch of events in the schedule, and can assist in evaluating the effects of measures that can be implemented as the plan progresses.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in recent years, with the progress of information technology, information related to the business of each company has been widely disclosed. In order to efficiently create a business scenario, a technology for effectively utilizing this information is required.

[0005] The present invention has been made in view of such circumstances, and its main object is to provide a business scenario creation support system and a business scenario creation support method that can assist in creating a business scenario by effectively utilizing information publicly available on the Internet. [Means for solving the problem]

[0006] To solve the above problems, a business scenario creation support system according to one aspect of the present invention comprises: a collection unit that collects business information relating to the business of each company that is publicly available on the internet; an input unit that inputs a prompt including specific information that identifies the company to be supported into a large-scale language model; an acquisition unit that, when the large-scale language model that has received the input of the prompt generates business scenario information relating to the business scenario of the company to be supported based on the collected business information, acquires the business scenario information from the large-scale language model; and a provision unit that provides the business scenario information to the company to be supported.

[0007] In the above embodiment, the system may further include an extraction unit that extracts information related to the prompt from the collected business information, the input unit inputs the extracted business information and the prompt into the large-scale language model, and the acquisition unit acquires the business scenario information generated by the large-scale language model based on the extracted business information.

[0008] Furthermore, in the above embodiment, the collection unit may collect the business information, including press release article information of each company's business announcements.

[0009] Furthermore, in the above embodiment, the collection unit may collect the business information, including word-of-mouth information about each company's business.

[0010] Furthermore, in the above embodiment, the business information may be accompanied by company identification information for identifying each company.

[0011] Furthermore, in the above embodiment, the input unit may input the prompt, which includes instructions regarding the identification of business challenges in the company to be supported and the setting of solutions to those challenges, into the large-scale language model, and the acquisition unit may acquire the business scenario information, which includes the business challenges and solutions to those challenges, from the large-scale language model.

[0012] In the above embodiment, the input unit may input the prompt, which includes instructions for selecting companies that can collaborate with the company to be supported, to the large-scale language model, and the acquisition unit may acquire the business scenario information, including the companies that can collaborate, from the large-scale language model.

[0013] In the above embodiment, the input unit may input the prompt, which includes instructions for setting the content of the collaboration with the collaborating companies, to the large-scale language model, and the acquisition unit may acquire the business scenario information, which includes the content of the collaboration, from the large-scale language model.

[0014] A method for supporting the creation of business scenarios according to one aspect of the present invention involves collecting business information relating to the business of each company that is publicly available on the internet, inputting a prompt containing specific information that identifies the company to be supported into a large-scale language model, and when the large-scale language model that has received the prompt input generates business scenario information relating to the business scenario of the company to be supported based on the collected business information, obtaining the business scenario information from the large-scale language model and providing the business scenario information to the company to be supported. [Effects of the Invention]

[0015] According to the present invention, it becomes possible to efficiently create highly feasible business scenarios. [Brief explanation of the drawing]

[0016] [Figure 1] A block diagram showing the configuration of the business scenario creation support system and its communication destinations. [Figure 2] A diagram showing an example layout of company tables that make up a company database. [Figure 3] This diagram shows an example of the layout of a PR article table that makes up a PR article database. [Figure 4]A diagram showing an example of the layout of a review table that constitutes a review database. [Figure 5] A diagram showing an example of the layout of a review table that constitutes a review database. [Figure 6] A flowchart showing an example of the procedure for data registration processing. [Figure 7] A flowchart showing an example of the procedure for business scenario creation support processing. [Figure 8] A diagram showing an example of a prompt. [Figure 9] A diagram showing an example of the output of a response sentence for the prompt shown in FIG. 8. [Figure 10] A diagram showing another example of a prompt. [Figure 11A] A diagram showing an example of the output of a response sentence for the prompt shown in FIG. 10. [Figure 11B] A diagram showing an example of the output of a response sentence for the prompt shown in FIG. 10.

Mode for Carrying Out the Invention

[0017] Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. Note that each of the embodiments shown below exemplifies methods and apparatuses for embodying the technical idea of the present invention, and the technical idea of the present invention is not limited to the following. Various changes can be made to the technical idea of the present invention within the technical scope described in the claims.

[0018] As will be described later, in the present embodiment, by utilizing information on the business of each company publicly available on the Internet and a large language model (hereinafter also referred to as "LLM"), support is provided for creating future business scenarios for each company. Hereinafter, an example of creating support for a business scenario related to "co-creation" in which a plurality of companies create new value by collaborating in business is also shown. In this specification, "company" refers to all organizations and individuals conducting business activities, including corporations (including private and public corporations) and individual business owners.

[0019] (System Configuration) Figure 1 is a block diagram showing the configuration of the business scenario creation support system and its communication destination according to this embodiment. The business scenario creation support system according to this embodiment consists of a business scenario creation support server (hereinafter simply referred to as "support server") 1. This support server 1 is a computer system operated by an operating company that provides business scenario creation support services to each user company, and is connected to the generation AI server 2 and user terminal 3 via the internet 101 so as to be able to communicate.

[0020] The Generative AI Server 2 is a computer system equipped with LLM21. LLM21 is a pre-trained machine learning model built using large amounts of text data, and when it receives a prompt (question) as input, it outputs a response corresponding to that prompt. For LLM21, for example, GPT (Generative Pre-trained Transformer) and BERT (Bidirectional Encoder Representations from Transformers) can be used.

[0021] User terminal 3 is an information terminal provided to each company that uses the business scenario creation support service, and consists of, for example, a personal computer, a tablet device, and a smartphone. Each company is required to register as a user with the service provider in order to use the business scenario creation support service. During this user registration, information about each company is provided to the service provider. This point will be discussed later.

[0022] The following describes the detailed configuration of Support Server 1. Support Server 1 consists of one or more computers equipped with a control unit including a CPU, RAM, and ROM, as well as a storage unit. The control unit executes the processes described later. The storage unit of Support Server 1 contains the following databases: Corporate Database (DB) 11, PR (Press Release) Article Database (DB) 12, Word-of-Mouth Database (DB) 13, and Scenario Database (DB) 14. The details of these databases are described below.

[0023] (A) Company DB11 Corporate DB11 is a database for storing information about each user company. Each company utilizing the business scenario creation support service provides information about itself to the service provider during user registration. Corporate DB11 is built based on the information provided to the service provider at this time. The information stored in Corporate DB11 is updated as needed based on information received from each company after user registration.

[0024] The company database 11 consists of multiple tables. Figure 2 shows an example of the layout of one of these tables, the company table. As shown in Figure 2, the company table stores various information about each company, such as the company name, address, industry, characteristics, challenges, desires, and conditions, linked to a company ID used to identify each company.

[0025] The "characteristics" mentioned above refer to the characteristics of each company as it perceives them, such as information about its business communicated to stakeholders like customers and business partners. "Challenges" refer to the challenges each company perceives. Furthermore, "hopes" are what each company hopes to achieve through co-creation, and "conditions" are the conditions sought in co-creation.

[0026] (B) PR article DB12 The PR article DB12 is a database for storing information on press release articles published on the Internet. The support server 1 repeatedly collects press release articles published on the Internet at appropriate times and stores the information obtained thereby in the PR article DB12.

[0027] The PR article DB12 is composed of a plurality of tables. FIG. 3 is a diagram showing an example of the layout of a PR article table, which is one of them. As shown in FIG. 3, in the PR article table, information such as the "release date", "title", and "content" of each PR article is stored in association with an article ID for identifying each PR article. Here, the "release date", "title", and "content" are the release date, title, and content described in the collected PR article.

[0028] As shown in FIG. 3, the PR article table also stores a company ID defined in the above company table in association with the article ID. This company ID is for identifying the company that is the sender of the PR article and is assigned to the PR article table by the data registration process described later.

[0029] (C) Review DB13 The review DB13 is a database for storing information on reviews for each company published on the Internet. The support server 1 repeatedly collects reviews published on the Internet at appropriate times and stores the information obtained thereby in the review DB13.

[0030] The review DB13 is composed of a plurality of tables. FIG. 4 is a diagram showing an example of the layout of a review table, which is one of them. As shown in FIG. 4, in the review table, information such as the "posting date" and "review content" of each review is stored in association with a review ID for identifying each review.

[0031] As shown in Figure 4, the review table also stores the company ID, defined in the company table above, which is linked to the review ID. This company ID is used to identify the company mentioned in the review and is assigned to the review table through the data registration process described later.

[0032] (D) Scenario DB14 Scenario DB14 is a database for storing information related to business scenarios, and is constructed using the information stored in the aforementioned company DB11, PR article DB12, and word-of-mouth DB13. Support server 1 generates the information to be registered in Scenario DB14 by executing the data registration process described later.

[0033] Scenario DB14 consists of multiple tables. Figure 5 shows an example of the layout of one of these tables, the Scenario Table. As shown in Figure 5, the Scenario Table stores information related to each business scenario, linked to a Scenario ID that identifies each business scenario, and converted into vector values. This information corresponds to the information stored in Company DB11, PR Article DB12, and Word of Mouth DB13.

[0034] As shown in Figure 5, the scenario table also stores the company ID, as defined in the company table above, which is linked to the scenario ID. This company ID is used to identify which company's business scenario it is and is assigned when registering information in the scenario DB14 through the data registration process described later.

[0035] Scenario DB14 is a database used to implement what is known as RAG (Retrieval Augmented Generation). As will be described later, support server 1 searches Scenario DB14, extracts appropriate information, and passes it to generation AI server 2. This allows generation AI server 2 to utilize the information stored in Scenario DB14.

[0036] (System operation) Next, we will explain the operation of the support server 1 configured as described above. The support server 1 performs the following: (1) data collection processing to collect information publicly available on the internet, (2) data registration processing to register the collected information in a database, and (3) business scenario creation support processing to provide information about business scenarios to the user in cooperation with the generation AI server 2 using that database. The details of each of these processes will be explained below.

[0037] (1) Data collection process As described above, support server 1 periodically collects PR articles and reviews published on the internet and stores them in PR article DB 12 and review DB 13. For example, PR articles distributed by press release distribution service providers and reviews posted on various social media platforms are collected by support server 1 and stored in the databases.

[0038] (2) Data registration process Figure 6 is a flowchart showing an example of the data registration process performed by support server 1. This data registration process is a batch process that is repeatedly executed at predetermined times, such as midnight every day.

[0039] Support server 1 first extracts the data to be processed from PR article DB12 and word-of-mouth DB13 (S101), and then performs data cleansing to correct any deficiencies in the data (S102). This data cleansing includes, for example, correcting inconsistencies and misspellings in company names and standardizing notation rules.

[0040] Next, the support server 1 identifies which company's PR article or review the data to be processed belongs to, and assigns the identified company's company ID to the data (S103). At this time, if the identified company is a company whose information is stored in the company DB 11, the company ID defined for that company is assigned. Otherwise, the support server 1 assigns a new company ID and assigns it to the data.

[0041] As described above, once a company ID is assigned to the data, the support server 1 updates the database by storing that company ID in the PR article DB 12 or the word-of-mouth DB 13 (S104).

[0042] Next, the support server 1 converts the data to be processed into a vector value (S105) and registers that vector value in the scenario DB 14 along with the company ID (S106). If a vector value for the same company ID is already stored in the scenario DB 14, the current vector value is registered in association with that vector value. This makes it easy to search for vector values ​​related to the same company.

[0043] By repeatedly executing the above data registration process with different data being processed, company IDs are stored in PR article DB12 and word-of-mouth DB13, and vector values ​​are accumulated in scenario DB. Information stored in company DB11 is also converted to vector values ​​and stored in scenario DB14 in the same manner as above.

[0044] By assigning a company ID as described above, it becomes possible to accurately associate and handle PR articles and reviews related to the same company. In generation AI, the inaccurate recognition of information linked to proper nouns has been a concern, but this embodiment resolves that problem by introducing a company ID.

[0045] (3) Business scenario creation support processing Each user company receives support from support server 1 when creating its own business scenario. The following business scenario creation support process is executed at this time.

[0046] Figure 7 is a flowchart illustrating an example of the procedure for business scenario creation support processing, which is executed through the cooperation of support server 1, generation AI server 2, and user terminal 3. A representative from each company inputs a request for support in creating their company's business scenario into user terminal 3. Upon receiving this request from the user (S201), user terminal 3 transmits it to support server 1 (S202).

[0047] The above-mentioned requests for assistance in creating a document include instructions such as providing information to identify the company (company name, company ID, etc.), creating a summary of customer reviews about the company, identifying challenges in the company's business, defining solutions to those challenges, selecting a co-creation partner company, and creating a business scenario for collaboration with the co-creation partner company (hereinafter referred to as the "co-creation scenario").

[0048] When the support server 1 receives a creation support request from the user terminal 3, it generates a prompt using the information contained in the creation support request (S301). A specific example of this prompt will be described later.

[0049] Next, support server 1 searches scenario DB 14 (S302) and obtains information (vector values) related to the above prompt. Then, support server 1 sends the extracted vector values, the scenario-related information including the company ID associated with those vector values, and the prompt to generating AI server 2 (S303).

[0050] When the generation AI server 2 obtains scenario-related information and prompts from the support server 1, it generates a response text using LLM21 (S401). This response text is generated using the scenario-related information. The generation AI server 2 sends the generated response text to the support server 1 (S402).

[0051] When the support server 1 obtains the response text from the generation AI server 2, it edits the response text into a format that can be presented to the user (S304) and sends the edited response text to the user terminal 3 (S305). The user terminal 3, having received this response text, outputs it to its display unit (S203). A specific example of this output screen will be described later.

[0052] (Examples of prompts and responses) Figure 8 shows an example of a prompt, and Figure 9 shows an example of the output response to that prompt. The prompt example shown in Figure 8 instructs the user to summarize the reviews about their company based on word-of-mouth, identify issues, and propose solutions to those issues. Furthermore, it is stipulated that the solutions must be related to the Sustainable Development Goals (SDGs). The name of the company asking the question is entered in the [Company] field.

[0053] Support server 1 will retrieve vector values ​​of information stored in company DB11, PR article DB12, and word-of-mouth DB13, linked to the company ID of the company in question, from scenario DB14 as information related to the above prompt, and provide them to generation AI server 2.

[0054] As shown in Figure 9, the response to the above prompt will include a "summary of the customer feedback" and "the identified issues and solutions." Furthermore, the issues and solutions will be tailored to the SDGs.

[0055] Figure 10 shows another example of a prompt, and Figures 11A and 11B show examples of output responses to that prompt. The prompt example shown in Figure 10 instructs the system to summarize customer reviews about the company, and then, referencing the summary of those reviews and the "characteristics," "challenges," "desires," and "conditions" in the company table of the company DB11, create a co-creation scenario based on the PR article DB12. In order to create a co-creation scenario, it is necessary to select a co-creation partner company and define the content of the collaboration with that company. Therefore, this prompt example includes instructions for selecting a co-creation partner company and defining the content of the collaboration with that company.

[0056] Support server 1 will retrieve vector values ​​of information stored in company DB 11, PR article DB 12, and word-of-mouth DB 13, linked to the company ID of the company in question and company IDs of other companies different from the company in question, from scenario DB 14 as information related to the above prompt, and provide them to generating AI server 2. The above company IDs of other companies may be company IDs of all companies other than the company in question, or they may be company IDs of companies that have a specific relationship with the company in question, such as being in the same industry.

[0057] The response to the above prompt includes the reference information used in creating the response, as well as the body of the response. Figure 11A shows an example of the output of the reference information, and Figure 11B shows an example of the output of the body of the response.

[0058] In the example shown in Figure 11A, the reference information includes a summary of the reviews, the company's characteristics, challenges, what it wants to achieve through co-creation, the conditions it is seeking, and the content of the referenced PR article. This is the information that the generating AI server 2 referenced when creating the response.

[0059] Furthermore, in the example shown in Figure 11B, the main body of the response includes the solution set based on the summary of word-of-mouth, the company's challenges and the solutions set for those challenges, and the co-creation scenario. The co-creation scenario includes the partner company and the details of the collaboration with that company.

[0060] As described above, in this embodiment, business information tailored to each company is generated using PR articles and word-of-mouth published on the internet and provided to the company. This enables each company to efficiently create business scenarios suited to its own needs. In particular, in this embodiment, information on companies that can collaborate with each company and the details of such collaborations are provided, making it easy to create co-creation scenarios that were previously difficult to create.

[0061] (Other embodiments) In the above embodiment, the support server 1 extracts prompt-related information from the scenario DB 14 and provides it to the LLM 21, but the present invention is not limited to this. For example, the information stored in the scenario DB 14 may be provided to the LLM 21 for additional learning, thereby performing fine tuning to adjust the parameters of the LLM 21. This also allows for obtaining a response that takes the contents of the scenario DB 14 into consideration.

[0062] Furthermore, in the above embodiment, the data in the company DB11, PR article DB12, and word-of-mouth DB13 are used after being vectorized, but this is not the only way to do so. For example, various types of data, such as text data in each database, may be used in their original format. [Explanation of Symbols]

[0063] 1. Business Scenario Creation Support Server 11 Company Databases 12 Article Database 13 User Reviews Database 14 Scenario Database 2. Generation AI Server 21. Large-Scale Language Models (LLMs) 3. User terminals 101 Internet

Claims

1. A collection department that collects business information about each company's business that is publicly available on the internet, An input unit that inputs a prompt containing specific information to identify the company to be supported into a large-scale language model, Based on the collected business information, if the large-scale language model that receives the prompt input generates business scenario information relating to the business scenario of the supported company, an acquisition unit acquires the business scenario information from the large-scale language model, A provision unit that provides the aforementioned business scenario information to the companies to be supported. A business scenario creation support system equipped with the following features.

2. The system further includes an extraction unit that extracts information related to the prompt from the collected business information, The input unit inputs the extracted business information and the prompt into the large-scale language model. The acquisition unit acquires the business scenario information generated by the large-scale language model based on the extracted business information. A business scenario creation support system according to claim 1.

3. The collection unit collects the business information, including press release articles containing business announcements from each company. A business scenario creation support system according to claim 1.

4. The collection unit collects the business information, including word-of-mouth information about each company's business. A business scenario creation support system according to any one of claims 1 to 3.

5. The aforementioned business information is accompanied by company identification information to identify each company. A business scenario creation support system according to any one of claims 1 to 3.

6. The input unit inputs the prompt, which includes identifying business challenges in the company to be supported and instructions for setting solutions to those challenges, into the large-scale language model. The acquisition unit acquires the business scenario information, including the challenges of the business and solutions to those challenges, from the large-scale language model. A business scenario creation support system according to any one of claims 1 to 3.

7. The input unit inputs the prompt, which includes instructions for selecting companies that can collaborate with the company to be supported, into the large-scale language model. The acquisition unit acquires the business scenario information, including the collaborating companies, from the large-scale language model. A business scenario creation support system according to any one of claims 1 to 3.

8. The input unit inputs the prompt, which includes instructions for setting the content of collaboration with the collaborating companies, to the large-scale language model. The acquisition unit acquires the business scenario information, including the details of the collaboration, from the large-scale language model. A business scenario creation support system according to claim 7.

9. We collect business information about each company's operations that is publicly available on the internet. A prompt containing specific information to identify the companies to be supported is input into a large-scale language model. Based on the collected business information, if the large-scale language model that receives the prompt input generates business scenario information regarding the business scenarios of the supported company, the business scenario information is obtained from the large-scale language model. The aforementioned business scenario information is provided to the companies to be supported. Methods for supporting the creation of business scenarios.

Images