Processing system, processing method, and processing program

Abstract

It can support efficient and high-quality business execution in government agencies. [Solution] A processing system 10 selects government agency business data required for a target business data creation process and a processing prompt that indicates the processing procedure for the business data creation process from a government agency business database (DB) 11. Based on the selected processing prompt, the processing system 10 sets at least the roles of personnel responsible for creating business data in the target business data creation process in a generation model. In accordance with the selected processing prompt, the processing system 10 loads a guardrail policy and detailed business content into a generation AI 161, and causes the generation AI 161 to perform predetermined processing for each role based on the loaded guardrail policy and detailed business content and the government agency business data selected by a selection unit, thereby creating the target business data.

Description

[Technical Field]

[0001] The present invention relates to a processing system, a processing method, and a processing program. [Background technology]

[0002] Government agencies are facing increasingly complex, difficult, and sophisticated policy challenges, while facing severe financial circumstances, which require simpler and more efficient systems. In light of this, the Ministry of Internal Affairs and Communications has published a proposal stating that it is important to promote efficient administrative operations and the provision of high-quality administrative services to the public. [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Ministry of Internal Affairs and Communications, Approach to Reform of National Administrative Operations, [Retrieved July 1, 2025], Internet<https: / / www.soumu.go.jp / main_content / 000370120.pdf> Summary of the Invention [Problem to be solved by the invention]

[0004] However, certain tasks can only be handled by specific personnel, so the workload tends to be concentrated. In the case of tasks that are basically performed by people or are complex, there are problems such as variations in work quality depending on the experience and ability of individuals, and differences in review perspectives, making it difficult to perform efficient and high-quality work.

[0005] Furthermore, because government agencies have a workforce that is heavily composed of seconded employees, personnel turnover occurs every two to three years, making it difficult to accumulate know-how and carrying out work efficiently and with high quality.

[0006] The present invention has been made in view of the above, and aims to provide a processing system, a processing method, and a processing program that can support efficient and high-quality business execution in government agencies. [Means for solving the problem]

[0007] In order to solve the above-mentioned problems and achieve the object, the processing system of the present invention is characterized by having a selection unit that selects government agency business data required for the business data creation process to be performed and a processing prompt that indicates the processing procedure of the business data creation process from a government agency business database that stores government agency business data that indicates the organization and business content of a government agency and processing prompts used in the business data creation process; a setting unit that sets at least the role of the person in charge of creating business data in the business data creation process to be performed in a generative model based on the processing prompt selected by the selection unit; and a processing execution unit that reads the guardrail policy and detailed business content into the generative model from a guardrail policy database that stores guardrail policies that indicate decision-making priorities in the government agency and a detailed business content database that stores detailed business content including an organizational chart and business content of all government agencies, based on the read guardrail policy and detailed business content and the government agency business data selected by the selection unit, and performs a predetermined process for each role to create the business data to be created.

[0008] In addition, the processing method of the present invention is a processing method executed by a processing system, and includes selecting government agency business data required for the business data creation process to be performed and a processing prompt that instructs the processing procedure of the business data creation process from a government agency business database that stores government agency business data that indicates the organization and business content of a government agency and processing prompts used in the business data creation process, setting at least the role of the person in charge of creating business data in the business data creation process to be performed in a generative model based on the selected processing prompt, reading the guardrail policy and the detailed business content into the generative model in accordance with the selected processing prompt from a guardrail policy database that stores guardrail policies that indicate decision-making priorities in the government agency and a detailed business content database that stores detailed business content including an organizational chart and business content of all government agencies, and performing a predetermined process for each role based on the read guardrail policy and the detailed business content and the selected government agency business data, thereby creating the business data to be created.

[0009] In addition, the processing program of the present invention selects government agency business data required for the business data creation process to be performed and a processing prompt that instructs the processing procedure of the business data creation process from a government agency business database that stores government agency business data that indicates the organization and business content of a government agency and processing prompts used in the business data creation process, and sets at least the role of the person in charge of creating business data in the business data creation process to be performed in a generative model based on the selected processing prompt, and reads the guardrail policy and detailed business content into the generative model from a guardrail policy database that stores guardrail policies that indicate decision-making priorities in the government agency and a detailed business content database that stores detailed business content including an organizational chart and business content of all government agencies in accordance with the selected processing prompt, and causes the generative model to perform predetermined processing for each role based on the read guardrail policy and detailed business content and the selected government agency business data, thereby creating the business data to be created. [Effects of the Invention]

[0010] According to the present invention, it is possible to support efficient and high-quality business execution in government offices. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram illustrating an example of a configuration of a processing system according to an embodiment. [Figure 2] FIG. 2 is a diagram for explaining the configuration of the government agency business database (DB) shown in FIG. [Figure 3] FIG. 3 is a diagram for explaining the configuration of the detailed business content DB shown in FIG. [Figure 4] FIG. 4 is a diagram for explaining the configuration of the new business plan / prompt DB shown in FIG. [Figure 5] FIG. 5 is a diagram showing an outline of the process (business data creation execution process) in the processing system shown in FIG. [Figure 6]FIG. 6 is a sequence diagram showing the processing procedure of the processing method according to the embodiment. [Figure 7] FIG. 7 is a diagram showing an outline of the process (new business data creation process) in the processing system shown in FIG. [Figure 8] FIG. 8 is a sequence diagram showing the processing procedure of the processing method according to the embodiment. [Figure 9-1] FIG. 9A is a diagram illustrating an example of a prompt for starting a process. [Figure 9-2] FIG. 9B is a diagram illustrating an example of a prompt for starting a process. [Figure 10-1] FIG. 10-1 is a diagram showing an example of output. [Figure 10-2] FIG. 10-2 is a diagram showing an example of output. [Figure 10-3] FIG. 10-3 is a diagram showing an example of output. [Figure 11-1] FIG. 11-1 is a diagram illustrating an example of a guardrail policy. [Figure 11-2] FIG. 11-2 is a diagram illustrating an example of a guardrail policy. [Figure 11-3] FIG. 11-3 is a diagram illustrating an example of a guardrail policy. [Figure 12-1] FIG. 12-1 is a diagram illustrating an example of a guardrail policy. [Figure 12-2] FIG. 12-2 is a diagram illustrating an example of a guardrail policy. [Figure 12-3] FIG. 12-3 is a diagram illustrating an example of a guardrail policy. [Figure 13-1] FIG. 13-1 is a diagram showing an example of an organizational chart. [Figure 13-2] FIG. 13-2 is a diagram showing an example of an organizational chart. [Figure 13-3] FIG. 13-3 is a diagram showing an example of an organizational chart. [Figure 14] FIG. 14 is a diagram showing an example of a processing prompt. [Figure 15-1] FIG. 15-1 is a diagram showing an example of output from a generating AI (Artificial Intelligence). [Figure 15-2] FIG. 15-2 is a diagram showing an example of output from the generation AI. [Figure 15-3] FIG. 15-3 is a diagram showing an example of output from the generation AI. [Figure 16] FIG. 16 is a diagram showing an example of a processing prompt. [Figure 17-1] FIG. 17-1 is a diagram showing an example of output from the generation AI. [Figure 17-2] FIG. 17-2 is a diagram showing an example of output from the generation AI. [Figure 17-3] FIG. 17-3 is a diagram showing an example of output from the generation AI. [Figure 18] FIG. 18 is a diagram illustrating a use case. [Figure 19] FIG. 19 is a diagram illustrating a list of points to be checked in a sales contract. [Figure 20-1] FIG. 20-1 is a diagram showing an example of a past sales contract. [Figure 20-2] FIG. 20-2 is a diagram showing an example of a past sales contract. [Figure 21-1] FIG. 21-1 is a diagram illustrating an example of a processing prompt. [Figure 21-2] FIG. 21-2 is a diagram illustrating an example of a processing prompt. [Figure 21-3] FIG. 21-3 is a diagram illustrating an example of a processing prompt. [Figure 21-4] FIG. 21-4 is a diagram illustrating an example of a processing prompt. [Figure 21-5] FIG. 21-5 is a diagram showing an example of a processing prompt. [Figure 21-6] FIG. 21-6 is a diagram showing an example of a processing prompt. [Figure 21-7] FIG. 21-7 is a diagram showing an example of a processing prompt. [Figure 22-1] FIG. 22-1 is a diagram showing an example of output. [Figure 22-2] FIG. 22-2 is a diagram showing an example of output. [Figure 22-3] FIG. 22-3 is a diagram showing an example of output. [Figure 22-4] FIG. 22-4 is a diagram showing an example of output. [Figure 22-5] FIG. 22-5 is a diagram showing an example of output. [Figure 22-6] FIG. 22-6 is a diagram showing an example of output. [Figure 22-7] FIG. 22-7 is a diagram showing an example of output. [Figure 22-8] FIG. 22-8 is a diagram showing an example of output. [Figure 22-9] FIG. 22-9 is a diagram showing an example of output. [Figure 22-10] FIG. 22-10 is a diagram showing an example of output. [Figure 22-11] FIG. 22-11 is a diagram showing an example of output. [Figure 22-12] FIG. 22-12 is a diagram showing an example of output. [Figure 22-13] FIG. 22-13 is a diagram showing an example of output. [Figure 23] FIG. 23 is a diagram illustrating a use case. [Figure 24-1] Figure 24-1 shows an example of a digital organizational chart, which is an input document. [Figure 24-2] Figure 24-2 shows an example of a digital organizational chart, which is an input document. [Figure 24-3] Figure 24-3 shows an example of a digital organizational chart, which is an input document. [Figure 25-1] FIG. 25-1 is a diagram illustrating an example of a processing prompt. [Figure 25-2] FIG. 25-2 is a diagram illustrating an example of a processing prompt. [Figure 25-3] FIG. 25-3 is a diagram illustrating an example of a processing prompt. [Figure 26-1] FIG. 26-1 is a diagram showing an example of output. [Figure 26-2] FIG. 26-2 is a diagram showing an example of output. [Figure 26-3] FIG. 26-3 is a diagram showing an example of output. [Figure 26-4]FIG. 26-4 is a diagram showing an example of output. [Figure 26-5] FIG. 26-5 is a diagram showing an example of output. [Figure 26-6] FIG. 26-6 is a diagram showing an example of output. [Figure 26-7] FIG. 26-7 is a diagram showing an example of output. [Figure 26-8] FIG. 26-8 is a diagram showing an example of output. [Figure 26-9] FIG. 26-9 is a diagram showing an example of output. [Figure 27-1] FIG. 27-1 is a diagram illustrating an example of a processing prompt. [Figure 27-2] FIG. 27-2 is a diagram illustrating an example of a processing prompt. [Figure 27-3] FIG. 27-3 is a diagram illustrating an example of a processing prompt. [Figure 28-1] FIG. 28-1 is a diagram showing an example of output. [Figure 28-2] FIG. 28-2 is a diagram showing an example of output. [Figure 28-3] FIG. 28-3 is a diagram showing an example of output. [Figure 28-4] FIG. 28-4 is a diagram showing an example of output. [Figure 28-5] FIG. 28-5 is a diagram showing an example of output. [Figure 28-6] FIG. 28-6 is a diagram showing an example of output. [Figure 28-7] FIG. 28-7 is a diagram showing an example of output. [Figure 28-8] FIG. 28-8 is a diagram showing an example of output. [Figure 28-9] FIG. 28-9 is a diagram showing an example of output. [Figure 29-1] FIG. 29-1 is a diagram illustrating an example of a processing prompt. [Figure 29-2] FIG. 29-2 is a diagram illustrating an example of a processing prompt. [Figure 29-3] FIG. 29-3 is a diagram illustrating an example of a processing prompt. [Figure 30-1] FIG. 30-1 is a diagram showing an example of output. [Figure 30-2] FIG. 30-2 is a diagram showing an example of output. [Figure 30-3] FIG. 30-3 is a diagram showing an example of output. [Figure 30-4] FIG. 30-4 is a diagram showing an example of output. [Figure 30-5] FIG. 30-5 is a diagram showing an example of output. [Figure 30-6] FIG. 30-6 is a diagram showing an example of output. [Figure 30-7] FIG. 30-7 is a diagram showing an example of output. [Figure 30-8] FIG. 30-8 is a diagram showing an example of output. [Figure 30-9] FIG. 30-9 is a diagram showing an example of output. [Figure 30-10] FIG. 30-10 is a diagram showing an example of output. [Figure 30-11] FIG. 30-11 is a diagram showing an example of output. [Figure 30-12] FIG. 30-12 is a diagram showing an example of output. [Figure 30-13] FIG. 30-13 is a diagram showing an example of output. [Figure 31] FIG. 31 is a diagram illustrating a use case. [Figure 32-1] FIG. 32-1 is a diagram illustrating an example of a processing prompt. [Figure 32-2] FIG. 32-2 is a diagram illustrating an example of a processing prompt. [Figure 32-3] FIG. 32-3 is a diagram illustrating an example of a processing prompt. [Figure 32-4] FIG. 32-4 is a diagram showing an example of a processing prompt. [Figure 32-5] FIG. 32-5 is a diagram showing an example of a processing prompt. [Figure 32-6] FIG. 32-6 is a diagram showing an example of a processing prompt. [Figure 32-7] FIG. 32-7 is a diagram showing an example of a processing prompt. [Figure 33-1] FIG. 33-1 is a diagram showing an example of an output result of a sales contract check operation stored in the detailed operation content DB. [Figure 33-2] FIG. 33-2 is a diagram showing an example of an output result of the sales contract check work stored in the detailed work content DB. [Figure 33-3] FIG. 33-3 is a diagram showing an example of an output result of the sales contract check work stored in the detailed work content DB. [Figure 33-4] FIG. 33-4 is a diagram showing an example of the output result of the sales contract check work stored in the detailed work content DB. [Figure 33-5] FIG. 33-5 is a diagram showing an example of the output result of the sales contract check work stored in the detailed work content DB. [Figure 33-6] FIG. 33-6 is a diagram showing an example of the output result of the sales contract check work stored in the detailed work content DB. [Figure 33-7] FIG. 33-7 is a diagram showing an example of the output result of the sales contract check work stored in the detailed work content DB. [Figure 33-8] FIG. 33-8 is a diagram showing an example of the output result of the sales contract check work stored in the detailed work content DB. [Figure 33-9] FIG. 33-9 is a diagram showing an example of an output result of the sales contract check work stored in the detailed work content DB. [Figure 34-1] FIG. 34-1 is a diagram showing an example of output. [Figure 34-2] FIG. 34-2 is a diagram showing an example of output. [Figure 34-3] FIG. 34-3 is a diagram showing an example of output. [Figure 34-4] FIG. 34-4 is a diagram showing an example of output. [Figure 34-5] FIG. 34-5 is a diagram showing an example of output. [Figure 34-6] FIG. 34-6 is a diagram showing an example of output. [Figure 34-7] FIG. 34-7 is a diagram showing an example of output. [Figure 34-8] FIG. 34-8 is a diagram showing an example of output. [Figure 34-9] FIG. 34-9 is a diagram showing an example of output. [Figure 34-10] FIG. 34-10 is a diagram showing an example of output. [Figure 34-11] FIG. 34-11 is a diagram showing an example of output. [Figure 34-12] FIG. 34-12 is a diagram showing an example of output. [Figure 34-13] FIG. 34-13 is a diagram showing an example of output. [Figure 35-1] FIG. 35-1 is a diagram illustrating an example of a creation prompt. [Figure 35-2] FIG. 35-2 is a diagram illustrating an example of a creation prompt. [Figure 35-3] FIG. 35-3 is a diagram illustrating an example of a creation prompt. [Figure 35-4] FIG. 35-4 is a diagram illustrating an example of a creation prompt. [Figure 36-1] FIG. 36-1 is a diagram showing an example of output from an agent-type AI. [Figure 36-2] FIG. 36-2 is a diagram showing an example of output from an agent-type AI. [Figure 36-3] FIG. 36-3 is a diagram showing an example of output from an agent-type AI. [Figure 37] FIG. 37 is a diagram illustrating an example of a computer that implements a processing system by executing a program. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of a processing system, a processing method, and a processing program according to the present application will be described in detail with reference to the accompanying drawings. Note that the processing system, the processing method, and the processing program according to the present application are not limited to these embodiments.

[0013] [Embodiment Mode] An embodiment will be described. In the embodiment, for example, a case will be described in which a generating AI is used to create business data to be used in government agencies, thereby supporting efficient and high-quality business execution in government agencies. Examples of support processes for government agencies include support processes for checking contracts, support processes for budget formulation work, and support processes for manual creation.

[0014] Fig. 1 is a diagram showing an example of the configuration of a processing system according to an embodiment. As shown in Fig. 1, the processing system 10 is capable of communicating between a business information DB 20 and a user terminal 30. The business information DB 20 stores business information related to businesses in which government agencies are involved. The user terminal 30 is, for example, a terminal device operated by an employee of the government agency.

[0015] The processing system 10 is realized by loading a predetermined program into a computer or the like including a ROM (Read Only Memory), RAM (Random Access Memory), CPU (Central Processing Unit), etc., and having the CPU execute the predetermined program. The processing system 10 also has a communication interface for transmitting and receiving various information to and from other devices connected via a network, etc. For example, the processing system 10 has a NIC (Network Interface Card), etc., and communicates with other devices via telecommunication lines such as a LAN (Local Area Network) or the Internet.

[0016] The processing system 10 has a government agency business DB 11, a guardrail policy DB 12, a detailed business content DB 13, a processing start determination unit 14 (start determination unit), a processing start control unit 15 (selection unit), a processing unit 16, a new business proposal / prompt DB 17, and a new business data creation unit 18.

[0017] The government agency business DB11 stores government agency business data showing organizational charts and business contents (tentative) of all government agencies, as well as processing prompts used in business data creation processing. Figure 2 is a diagram explaining the configuration of the government agency business DB11 shown in Figure 1. As shown in Figure 2, the government agency business DB11 has data areas for all government agencies. As shown in Figure 2, the government agency business DB11 also has an organizational chart folder F11, a business content folder F12, a prompt folder F13, and an input / output yalm folder F14.

[0018] The organizational chart folder F11 stores data indicating the role and organizational structure of the relevant government agency. The business content folder F12 stores data indicating the business content of the relevant government agency. The prompt folder F13 stores processing prompts linked to the business of the relevant government agency. The input / output yalm folder F14 stores input / output yalm, which indicates a data format for writing in a format that is easy for humans to read. Note that in Figure 2, the "folders" for the organizational chart folder F11, business content folder F12, prompt folder F13, and input / output yalm folder F14 have been omitted.

[0019] The processing prompts stored in the prompt folder F13 are used for business data creation processing in the relevant government agency, and contain specific instructions for various business data creation processing to the generation AI 161 (described later). The processing prompts instruct the generation AI 161 on the government agency business data required for the business data creation processing to be started and the processing procedure for the business data creation processing.

[0020] The processing prompt includes a description for setting roles for the target business, including at least the person in charge of creating business data within the agency, and instructing the creation of business data, in the generation AI 161. The roles include the person in charge of creating the business data within the agency, as well as the person's superior and / or related parties.

[0021] The processing prompt also includes content that instructs the generation AI 161 to determine the consistency of the business data created by the processing unit 16 (described later) with past business data of the same type and / or internal information of government agencies in accordance with predetermined determination conditions. The processing prompt includes a description that instructs the generation AI 161 to compare the processing content by the processing execution unit 163 with the processing content when a human performs the same task as the processing performed by the processing execution unit 163 without using the generation AI 161, and to quantitatively and / or qualitatively measure the effectiveness of the task performed by the processing execution unit 163.

[0022] The prompt folder F13 stores, for example, prompts P11 for answering questions in the Diet and prompts P12 for processing contracts, which correspond to simple tasks; problem extraction prompts P21, problem solving prompts P22, and budget content formulation prompts P23, which correspond to complex tasks; and manual creation prompts P31, which correspond to manual creation.

[0023] The prompt P11 for Diet responses sets the role of the person in charge of creating the check data in the generation AI 161, and instructs the AI ​​to check the contents of the Diet responses and create check data that shows the check results.The prompt P12 for contract processing sets the role of the person in charge of creating the check data in the generation AI 161, and instructs the AI ​​to check the contract and create data that shows the check results.

[0024] The issue extraction prompt P21 sets the role of the person in charge of extracting issues for budget content formulation in the generation AI 161, and instructs the AI ​​to extract issues for budget content formulation. The issue resolution prompt P22 sets the role of the person in charge of creating solutions to the extracted issues in the generation AI 161, and instructs the AI ​​to create solutions to the extracted issues. The budget content formulation prompt P23 sets the role of the person in charge of formulating the budget content for implementing the solutions to the issues in the generation AI 161, and instructs the AI ​​to formulate the budget content for implementing the solutions to the issues.

[0025] The manual creation prompt P31 sets the role of the person in charge of creating the manual in the generation AI 161 and instructs the creation of the manual.

[0026] The guardrail policy DB12 stores guardrail policies that indicate the decision-making priorities of government agencies, as well as current policy / budget materials (government information). Guardrail policies set human decision points, processing stopping points, and prioritization of viewpoints. Guardrail policies are created in advance and stored in the guardrail policy DB12.

[0027] The detailed work content DB13 stores detailed work content, including organizational charts and work content, of all government agencies. Figure 3 is a diagram explaining the configuration of the detailed work content DB13 shown in Figure 1. A detailed work content DB13 is prepared for each government agency, and each DB has an organizational chart folder F131, work content F132, work-related document folder F133, and output result folder F134. The organizational chart folder F131 stores an organizational chart showing the confirmed roles and organizational structure of the corresponding government agency. The work content F132 stores the confirmed work content of the corresponding government agency. The work-related document folder F133 stores work-related documents, and the output result folder F134 stores past output results from the processing system 10.

[0028] The process start determination unit 14 identifies the business data creation process to be performed based on whether or not the business information stored in the business information DB 20 has increased or decreased, or whether or not the person in charge has issued a process start instruction, and determines whether or not the business data creation process needs to be started. The process start determination unit 14 has an agent-type AI 141 that makes the determination.

[0029] When the agent AI 141 detects that a new file has been stored in a specific folder in the business information DB 20, it identifies the business to be performed and determines whether or not it is necessary to start business data creation processing for this business to be performed.

[0030] For example, if a folder link to the business information DB 20 is specified in advance, when a new file is stored, the agent-type AI 141 is activated and determines whether to start the business data creation process. In this way, the processing system 10 automatically starts the business data creation process of the generation AI 161 simply by having the user in charge store the file. Therefore, the processing system 10 can further improve the efficiency of tasks that are simple and have a fixed workflow, such as checking the contents of contracts and accounting processing.

[0031] In addition, the processing start determination unit 14 may identify the task to be performed based on whether or not a processing start instruction has been received from a user, such as an office employee, sent from the user terminal 30, and determine whether or not it is necessary to start processing to create business data for the task to be performed.

[0032] The processing start control unit 15 controls the start of the business data creation process. The processing start control unit 15 uses an agent-type AI (not shown) to read a processing start prompt. The processing start prompt is a prompt that describes the data to be selected to perform the business data creation process and the overall processing flow. For example, a processing start prompt is provided in advance for each business data creation process to be performed, and describes the overall processing flow of the corresponding business data creation process to be performed. When the processing start determination unit 14 determines that the business data generation process should be started for the business to be performed, the processing start control unit 15 reads a processing start prompt corresponding to the business data generation process that has been determined to be started.

[0033] Then, in accordance with the read processing start prompt, the processing start control unit 15 selects government agency business data required for the business data creation processing to be performed and a processing prompt that instructs the business data creation processing from the government agency business DB 11. The processing start control unit 15 outputs the selected government agency business data and the processing prompt to the processing unit 16, causing the processing unit 16 to start the target business data creation processing.

[0034] The government agency business DB 11, for example, sets the type of business data creation process and stores a set of government agency business data and a processing prompt for each type. In this case, the processing start control unit 15 selects a set of government agency business data and a processing prompt corresponding to the type of business data creation process to be performed from the government agency business DB 11. In this way, according to the processing system 10, the user does not need to give detailed instructions to the generation AI 161, and the processing system 10 automatically and smoothly creates business data.

[0035] The processing unit 16 includes a generating AI 161 , a role setting unit 162 , a processing execution unit 163 , a consistency determination unit 164 , an effect measurement unit 165 , and a result output unit 166 .

[0036] The generation AI 161 is a generation model that generates business data by referring to at least the government business data output from the process start control unit 15 in accordance with the process prompt selected by the process start control unit 15 .

[0037] The role setting unit 162 sets a plurality of roles including at least the person in charge of creating business data in the generation AI 161 based on the processing prompt selected by the processing start control unit 15. The roles are the person in charge of creating business data, the person in charge's superior (person who checks the business data), and / or related parties.

[0038] The processing execution unit 163 causes the generation AI 161 to read a guardrail policy from the guardrail policy DB 12 in accordance with the processing prompt selected by the processing start control unit 15. The processing execution unit 163 causes the generation AI 161 to read detailed business content corresponding to the business data creation processing to be performed from the detailed business content DB 13 in accordance with the processing prompt selected by the processing start control unit 15.

[0039] Then, the processing execution unit 163, in accordance with the processing prompt selected by the processing start control unit 15, causes the generation AI 161 to perform a predetermined process (e.g., data creation, discussion) for each role set by the role setting unit 162, based on the loaded guardrail policy data and detailed business content and the government agency business data selected by the processing start control unit 15. In this way, the processing system 10 causes the generation AI 161 to create the business data to be created.

[0040] For example, when the role is a person in charge, the processing execution unit 163 causes the generation AI 161 to create business data. When the role is a superior of the person in charge, the processing execution unit 163 causes the generation AI 161 to check the business data created when the person in charge was in charge's role. When the role is a related party, the processing execution unit 163 causes the generation AI 161 to add comments to the business data created when the person in charge was in charge's role from the related party's perspective.

[0041] Here, the processing prompt includes content that instructs the generation AI 161 to determine the consistency of the business data created by the processing execution unit 163 with past business data of the same type and / or internal information of government agencies in accordance with specified judgment conditions.

[0042] The consistency determination unit 164 uses the generated AI 161 in accordance with the processing prompt to determine the consistency between the business data created by the processing execution unit 163 and past business data of the same type and / or internal information of government agencies.

[0043] If the consistency determination unit 164 determines that the data is consistent, it stores the business data created by the processing execution unit 163 in the detailed business content DB. In this case, the business data created by the processing execution unit 163 is used as input data in subsequent processing. Furthermore, the business data determined to be consistent by the consistency determination unit 164 may be output from the result output unit 166 to the user terminal 30.

[0044] If the consistency determination unit 164 determines that there is no consistency, the processing execution unit 163 creates the business data again and repeats the process until a predetermined termination condition is met. The processing execution unit 163 continues creating the business data, for example, until the consistency determination unit 164 determines that there is consistency.

[0045] The processing prompt also includes a description that instructs the generation AI 161 to compare the processing content performed by the processing execution unit 163 with the processing content when a human performs the same task as the processing performed by the processing execution unit 163 without using the generation AI 161, and to quantitatively and / or qualitatively measure the effectiveness of the task performed by the processing execution unit 163.

[0046] The effect measurement unit 165 measures the effect of the work using the generation AI 161 in accordance with the processing prompt. The effect measurement unit 165 compares the processing content by the processing execution unit 163 with the processing content when a person performs the same work as the processing performed by the processing execution unit 163 without using the generation AI 161, and measures the effect of the work performed by the processing execution unit 163 quantitatively and / or qualitatively. The result output unit 166 outputs the business data created by the processing execution unit 163 and the effect of the work measured by the effect measurement unit 165.

[0047] For example, the processing prompt includes a description that instructs the processing execution unit 163 to calculate the time required to execute the processing and the amount of work (number of people and time) that would occur if a person were to execute the same work as the processing executed by the processing execution unit 163 without using the generation AI 161, and to measure the effect of reducing the work load by the processing execution unit 163. In this case, the effect measurement unit 165 calculates the time required to execute the processing by the processing execution unit 163 and the amount of work (number of people and time) that would occur if a person were to execute the same work as the processing executed by the processing execution unit 163 without using the generation AI 161, and quantitatively measures the effect of reducing the work load by the processing execution unit 163.

[0048] The new task plan / prompt DB 17 stores new task plans and new processing procedures created by the new task data creating unit 18, as well as new processing prompt plans.

[0049] Figure 4 is a diagram explaining the configuration of the new business plan / prompt DB 17 shown in Figure 1. The new business plan / prompt DB 17 has a new business plan F171 in which new business plan are saved, a business processing procedure folder F172 in which processing procedures for new business are saved, and a generated AI execution prompt plan folder F173 in which new processing prompt plans are saved.

[0050] The new task data creation unit 18 has an agent-type AI 181. The new task data creation unit 18 performs new task data creation processing using the agent-type AI 181. The new task data creation unit 18 organizes tasks and creates a task list at a predetermined timing based on the data stored in the detailed task content DB 13, in accordance with a creation prompt that instructs the creation of new task data.

[0051] The new task data creation unit 18 compares the old and new task lists and identifies a new task plan indicating the task content of the newly created task and a new processing procedure indicating the processing procedure of the newly created task.The new task data creation unit 18 then creates a new processing prompt plan that instructs the creation of task data related to the new task plan.The new task plan, the processing procedure of the new task, and the new processing prompt plan are stored in the new task plan / prompt DB 17.

[0052] The new business data creation unit 18 transmits the new business proposal, the new processing procedure, and the new processing prompt proposal to the user terminal 30. When the user instructs the new business data creation unit 18 to store the new business proposal, the new processing procedure, and the new processing prompt proposal in the government agency business DB 11 via the user terminal 30, the new business data creation unit 18 stores the new business proposal, the new processing procedure, and the new processing prompt proposal in the government agency business DB 11.

[0053] [Processing Overview] [Implementation process] Fig. 5 is a diagram showing an outline of the process (business data creation execution process) in the processing system 10 shown in Fig. 1. Fig. 6 is a sequence diagram showing the processing procedure of the processing method according to the embodiment.

[0054] In the processing system 10, the processing start determination unit 14 determines whether or not to start business data creation processing based on reference to the business information DB 20 (step S21-1 in FIG. 6) or a processing start instruction from the user (step S21-2 in FIG. 6) ((1) in FIG. 5, step S22 in FIG. 6). The processing start determination unit 14 automatically determines to start processing when the agent-type AI 141 detects a trigger such as a sudden increase or decrease in the business information stored in the business information DB 20. The agent-type AI 141 is also capable of multiple loop processing ((1-1) in FIG. 5). The processing start determination unit 14 may automatically determine to start processing, or may determine to start processing when a person in charge manually kicks it off ((1-2) in FIG. 5).

[0055] If the process start determination unit 14 determines not to start the business data creation process (step S22 in FIG. 6: No), the process returns to the determination of step S22. On the other hand, if the process start determination unit 14 determines to start the business data creation process (step S22 in FIG. 6: Yes), the process start determination unit 14 sends a process start instruction to the process start control unit 15 (step S23 in FIG. 6).

[0056] The process start control unit 15 reads the process start prompt ((2) in FIG. 5, step S24 in FIG. 6). The process start control unit 15 refers to the government agency business DB 11 in accordance with the process start prompt (step S25 in FIG. 6) and selects information that matches the required information (business type) ((3) in FIG. 5). The government agency business DB 11 pre-registers all government agency business in addition to the listed products ((3-1) in FIG. 5). The process start control unit 15 selects, in accordance with the process start prompt, from the government agency business DB 11 the government agency business data required for the business data creation process to be started and a process prompt that instructs the business data creation process (step S26 in FIG. 6).

[0057] The process start control unit 15 outputs the selected government agency business data and a processing prompt to the processing unit 16 (step S27 in FIG. 6), and starts the target business data creation process.

[0058] In the processing unit 16, the role setting unit 162 sets a role including at least the person in charge of creating business data in the generation AI 161 based on the processing prompt selected by the processing start control unit 15 ((4) in Figure 5, step S28 in Figure 6), and the processing execution unit 163 reads the data required for creating business data.

[0059] Specifically, the processing execution unit 163 refers to the guardrail policy DB12 (step S29 in FIG. 6) and causes the generation AI 161 to read the guardrail policy and current policy / budget documents ((5) in FIG. 5, step S30 in FIG. 6). The guardrail policy sets human decision points, processing stopping points, and perspective prioritization ((5-1) in FIG. 5). The processing execution unit 163 refers to the detailed work content DB13 (step S31 in FIG. 6) and causes the generation AI 161 to read the detailed work content corresponding to the work to be processed ((6) in FIG. 5, step S32 in FIG. 6).

[0060] The processing execution unit 163, in accordance with the processing prompt selected by the processing start control unit 15, causes the generation AI 161 to carry out a predetermined process (e.g., data creation, discussion, etc.) for each role set by the role setting unit 162 based on the loaded guardrail policy data, current policy / budget documents and detailed business content, and the government agency business data selected by the processing start control unit 15 ((7) in Figure 5, step S33 in Figure 6).

[0061] For example, as a process divided by role, if the role is a person in charge, process A is performed to have the generation AI 161 create business data. If the role is the person in charge's superior, process B is performed to have the generation AI 161 check the business data created when playing the role of the person in charge. If the role is a related party, process C is performed to have the generation AI 161 add comments to the business data created when playing the role of the person in charge from the perspective of the related party.

[0062] The consistency determination unit 164 uses the generation AI 161 in accordance with the processing prompt to determine the consistency between the business data created by the processing execution unit 163 and past business data of the same type and / or internal information of the government agency ((8) in FIG. 5, step S34 in FIG. 6). In addition to the consistency determination result, the consistency determination unit 164 creates and outputs an explanation to the superior.

[0063] If the consistency determination unit 164 determines that there is consistency (consistency determination (8-1) in FIG. 5: OK, step S35 in FIG. 6: Yes), it stores the business data created by the processing execution unit 163 in the detailed business content DB ((8-2) in FIG. 5, step S36 in FIG. 6). In this case, the business data created by the processing execution unit 163 is used as input data in subsequent processing. The business data determined to be consistent by the consistency determination unit 164 may be output from the result output unit 166 to the user terminal 30.

[0064] If the consistency determination unit 164 determines that the data are inconsistent (consistency determination at (8-3) in FIG. 5: NG, step S35 in FIG. 6: No), the processing execution unit 163 creates business data again (step S33). The processing execution unit 163 repeatedly creates business data and the consistency determination unit 164 repeatedly performs consistency determination ((8-4) in FIG. 5) until the consistency determination unit 164 determines that the data are consistent (the consistency determination returns OK).

[0065] The effect measurement unit 165 measures the effect in accordance with the processing prompt ((9) in FIG. 5, step S37 in FIG. 6). The effect measurement unit 165 compares the processing content by the processing execution unit 163 with the processing content when a person performs the same work as the processing performed by the processing execution unit 163 without using the generated AI 161, and measures the effect of the work performed by the processing execution unit 163 quantitatively and / or qualitatively.

[0066] The result output unit 166 outputs the effect of the work measured by the effect measurement unit 165 to the user terminal 30 ((10) in FIG. 5, step S38 in FIG. 6). The result output unit 166 may also share information with other systems as needed ((10-1) in FIG. 5).

[0067] [New business data creation process] Fig. 7 is a diagram showing an outline of the process (new business data creation process) in the processing system 10 shown in Fig. 1. Fig. 8 is a sequence diagram showing the processing procedure of the processing method according to the embodiment.

[0068] In the processing system 10, the agent-type AI 181 of the new business data creation unit 18 determines whether it is time to create new business data in accordance with a creation prompt that instructs creation of new business data (step S41 in FIG. 8). The agent-type AI 181 operates at a predetermined timing, such as weekly or monthly. If it is not time to create new business data (step S41 in FIG. 8: No), the agent-type AI 181 performs the determination in step S41 again.

[0069] If it is time to create a task (step S41 in FIG. 8: Yes), the agent AI 181 refers to the detailed task content DB 13 in accordance with the creation prompt (step S42 in FIG. 8). Then, the agent AI 181 organizes tasks and creates a task list based on the data stored in the detailed task content DB 13 in accordance with the creation prompt ((11-1) in FIG. 7, step S43 in FIG. 8).

[0070] The agent-type AI 181 compares the old and new task lists in accordance with the creation prompt, identifies new task proposals and new processing procedures based on the differences between the old and new task lists, and creates new processing prompt proposals ((11-2) in FIG. 7, step S44 in FIG. 8). In addition to the new task proposals and new processing procedures, the agent-type AI 181 may further identify task proposals whose content has been updated (updated tasks) and tasks that are not currently being performed (tasks that are candidates for deletion). The agent-type AI 181 may create new processing prompt proposals for updated tasks.

[0071] The agent-type AI 181 stores the new business plan, the processing procedure of the new business, and the new processing prompt plan in the new business plan / prompt DB 17 ((11-3) in FIG. 7, step S45 in FIG. 8).

[0072] Next, the new task data creating unit 18 transmits the new task proposal, the new processing procedure, and the new processing prompt proposal to the user terminal 30, and requests confirmation by the user (step S46 in FIG. 8).

[0073] When a user issues an instruction to add data to the government agency business DB 11 via the user terminal 30 (step S47 in FIG. 8), the new business data creation unit 18 acquires the data to be added from the new business plan / prompt DB 17 (step S48 in FIG. 8). The new business data creation unit 18 stores the new business plan and new processing procedure to be added, and the new processing prompt plan in the government agency business DB 11 (step S49 in FIG. 8).

[0074] The new business data creation unit 18 stores the new business proposal and new processing procedure in the business content folder F12 of the government agency business DB11, and stores the new processing prompt for the generated AI161 in the prompt folder F13 of the government agency business DB11 that is linked to the relevant business.

[0075] In this way, the new business proposals and new processing procedures created in the new business data creation process, as well as the new processing prompt proposals, are visually checked, and if there are no problems, they are manually added to the government agency business DB11 ((11-4) in Figure 7).

[0076] [Effects of the embodiment] In this way, the processing system 10 according to the embodiment uses the generation AI 161 to automatically create business data to be used in government offices, thereby supporting efficient business execution in government offices.

[0077] In particular, the processing system 10 defines the roles required for discussion in the processing prompt obtained from the government agency business database 11 as the person in charge of creating the business data, the person in charge's superior (the person who checks the business data), and / or related parties, and by conducting discussions between the generation AIs 161, it becomes possible to conduct realistic and multifaceted considerations based on the expected opinions of stakeholders.

[0078] Therefore, the processing system 10 can provide high-quality business data to government agencies.

[0079] Furthermore, the processing system 10 determines the consistency between the created business data and the policy / budget document information, and even if it determines that there is no consistency, it executes processing again by the processing execution unit 163 and repeats the business data creation process until it determines that there is consistency.

[0080] This allows the processing system 10 to create business data that is consistent with policy / budget document information, and to provide accurate business data to government agencies. For example, even when performing complex processing such as policy planning within budgetary constraints, the processing system 10 divides the processing content into multiple stages and performs loop processing, making it possible to output optimal business data while maintaining output quality and output volume.

[0081] The processing system 10 then compares the processing content of the processing execution unit 163 with the processing content when a person performs the same work as the processing performed by the processing execution unit 163 without using the generated AI 161, and quantitatively and / or qualitatively measures the effectiveness of the work performed by the processing execution unit 163. This allows the government agency to accurately recognize the degree to which the use of the processing system 10 has improved the efficiency of its operations.

[0082] The processing system 10 then executes new task data creation processing at a predetermined timing. Therefore, with the processing system 10, the person in charge does not need to organize the tasks, compare the old and new task lists, confirm the new task content, or identify the processing procedures for the newly created tasks, thereby realizing quick and accurate task organization while reducing the workload on the person in charge. Furthermore, with the processing system 10, the person in charge does not need to create a new processing prompt proposal that instructs the creation of task data related to the new task proposal, thereby realizing quick and appropriate processing prompt proposals while reducing the workload on the person in charge.

[0083] Below, we will explain examples of data, prompts, and output that are actually used, as well as use cases.

[0084] [Prompt to start processing] 9-1 and 9-2 are diagrams showing examples of a process start prompt. The process start prompt instructs the agent-type AI of the process start control unit 15 to search for and execute the stored prompt according to the user's input, and to carry out the task using the input information described in the prompt (FIG. 9-1).

[0085] The prompt to start the process instructs the user to understand the desired business content and search for related prompts in STEP 1 (Figure 9-2).

[0086] The prompt to start the process instructs the user to execute the business process in accordance with the contents of the related system prompt in STEP 2 (Figure 9-2).

[0087] For example, when the user inputs "Perform Diet response work" from the user terminal 30, the agent-type AI of the processing start control unit 15 will search for the corresponding "processing prompt for Diet response" and search for the related materials listed in the "processing prompt for Diet response" and carry out the work processing.

[0088] [output] 10-1 to 10-3 are diagrams showing output examples. The agent-type AI of the processing start control unit 15 understands the business content desired by the user instructed in STEP 1 and outputs the search results of related prompts (FIG. 10-1). For example, it searches for "20250115_Diet Answer Business" as a processing prompt, and searches for "Question Notification Metadata" as government agency business data required for the business data creation process to be performed, and outputs them to the processing unit 16. In this case, the processing start control unit 15 causes the processing unit 16 to execute the "Completed Answer" creation process based on the processing prompt "20250115_Diet Answer Business" and the "Question Notification Metadata."

[0089] The agent-type AI of the processing start control unit 15 outputs the execution details of the first sub-step “2.1 Acceptance of question notification and metadata registration” among the business processing instructed in STEP 2 (Figures 10-2 and 10-3).

[0090] In this way, the processing start control unit 15 selects government agency business data required for the business data creation process to be performed and a processing prompt that instructs the business data creation process in accordance with the processing start prompt. By using the processing start prompt, the processing start control unit 15 can accurately select the processing prompt and government agency business data corresponding to the business data creation process to be performed by the processing unit 16 from among the multiple business data creation processes. The processing start control unit 15 outputs the selected government agency business data and the processing prompt to the processing unit 16, causing the processing unit 16 to perform the business data creation process to be performed. In this way, the processing start control unit 15 can cause the processing unit 16 to appropriately perform the business data creation process to be performed from among the multiple business data creation processes.

[0091] [An example of a guardrail policy] 11-1 to 11-3 are diagrams showing examples of guardrail policies. The guardrail policies are common data that are input in all cases.

[0092] The guardrail policy describes the scope and purpose (Figure 11-1). The guardrail policy also describes content filtering, such as immediate blocking, replacement / conversion, image processing, citation with source, and user notification, for each risk category, such as illegal / harmful information, discrimination, and privacy, as well as factuality and citation rules (Figure 11-2).

[0093] The guard race policy also includes precautions to be taken when outputting business data, such as notifying the user as a supplementary message if the content to be output is judged to be significantly different from reality, and when considering policy ideas, etc., automatically determining the point at which to stop the consideration in line with the business context while considering the degree of difference from reality and notifying the user of the result of the judgment, as well as the priority of the judgment (Figure 11-3).

[0094] The processing execution unit 163 executes the business data creation process in accordance with the description of the guardrail policy, and therefore can create business data that is backed up by appropriate decision-making in government agencies.

[0095] [An example of a guardrail policy] 12-1 to 12-3 are diagrams showing examples of guardrail policies. The guardrail policies are common data that are input in all cases.

[0096] The guardrail policy describes the key points in government decision-making, such as fairness and transparency (Figure 12-1), and the key points in contract conclusion (Figures 12-2 and 12-3). Key points in contract conclusion include legal and institutional frameworks, contract methods and selection criteria, policy objectives and contracts, ethics and compliance, contract management and audits, and priorities.

[0097] [Organizational Chart] 13-1 to 13-3 are diagrams showing examples of organizational charts. As an example, Figures 13-1 to 13-3 show organizational charts of the Ministry of Land, Infrastructure, Transport and Tourism, but organizational charts are appropriately selected depending on the government agency that performs the work.

[0098] The organizational chart shows the confirmed roles and organizational structure of the relevant government agency. For example, it includes the mission and main roles of the government agency, major policy issues and recent trends (Figure 13-1), organizational structure (Figure 13-2), and an overview of the work of each department (Figure 13-3).

[0099] [Consistency judgment] Next, the processing prompt used by the consistency determination unit 164 will be described. Fig. 14 is a diagram showing an example of the processing prompt. As shown in Fig. 14, the processing prompt instructs the generation AI 161 to determine the consistency between the output (business data) output by the generation AI 161 as a result of processing performed by the processing execution unit 163 and similar output content in the past and / or internal information of the government agency (white papers, budgets, etc.). If the output contradicts the internal information, the processing prompt considers a correction plan and instructs the processing execution unit 163 to execute the business data creation process again.

[0100] Figures 15-1 to 15-3 are diagrams showing examples of output from the generation AI 161. The generation AI 161 outputs the results of judgments for each corresponding item regarding consistency with laws and regulations, consistency with policies, white papers, and budgets, and consistency with public document management standards, as well as a summary of the results (Figures 15-1 to 15-3).

[0101] In this way, the consistency determination unit 164 determines the consistency between the business data created by the processing execution unit 163 and the policy / budget document information, and causes the processing execution unit 163 to repeatedly perform the business data creation process until consistent business data is created, thereby making it possible to provide accurate business data that is consistent with the policy / budget document information to government agencies.

[0102] [Effectiveness measurement] Next, the processing prompt used by the effect measurement unit 165 will be described. Fig. 16 is a diagram showing an example of the processing prompt. As shown in Fig. 16, the processing prompt instructs the generation AI 161 to calculate the time required to execute each step (e.g., steps 1 to 4) in the processing performed by the processing execution unit 163. The processing prompt instructs the generation AI 161 to calculate the amount of work (number of people and time) that would be required if the same work in each step were performed by a human without using the generation AI 161, and to specifically output the effect of reducing the work load in each step by utilizing the generation AI 161 in both quantitative and qualitative terms.

[0103] Figures 17-1 to 17-3 are diagrams showing example outputs from the generation AI 161. The generation AI 161 outputs the results of a comparison of the operations required for steps 1 to 4 (Figure 17-1). In the example of Figure 17-1, the generation AI 161 outputs that in addition to a 69% reduction in work time, there were also advantages in terms of performance effects such as ensuring accuracy, consistency, and compliance. The generation AI 161 then outputs the measurement results for each step, divided into quantitative impact (Figure 17-2) and qualitative impact (Figure 17-3).

[0104] The effectiveness measurement by the effectiveness measurement unit 165 allows the government agency to accurately recognize the degree to which the use of the processing system 10 has improved the efficiency of its operations.

[0105] [Use case 1] A specific example of actual business data creation processing will be described. Use case 1 will be described using a contract check operation as an example of a relatively simple operation. FIG. 18 is a diagram for explaining use case 1.

[0106] As shown in Figure 18 (A), currently, (1) the contract is prepared by the business person, (2) checked by the supervisor, (3) revised by the business person after the supervisor's review, (4) checked by the supervisor again, and (5) revised by the business person again. This process involves multiple reviews with the supervisor before (6) the contract is approved. Because contract administrative work can only be handled by specific personnel, the workload tends to be concentrated. Furthermore, because contract checking is primarily done by people, the burden on both the business person and the supervisor is heavy, and review perspectives differ depending on the supervisor, which can lead to missed checks.

[0107] 18(B), by introducing the processing system 10 (introduction of the generation AI 161), for example, (1) the first review after the contract is created by the business person, (2) checking the contract, and (3) the processing of corrections are substituted by the generation AI 161. The generation AI 161 checks whether the created contract satisfies the points described in the manual, etc., and corrects the contract so that the points are satisfied.

[0108] Then, a secondary review is conducted by the superior ((4) confirmation by the superior)), and (5) the work staff makes further corrections. In this case, the primary review based on certain standards by the generation AI 161 reduces the burden on the superior and the work staff, minimizes oversights, and (6) shortens the time until approval. Of course, the generation AI 161 may also be used to conduct the secondary review.

[0109] The processing system 10 selects a list of points to be checked in a sales contract from the government agency business DB11, selects past sales contracts (past output results) from the detailed business content DB13, and selects a processing prompt for sales contract check processing from the government agency business DB11.

[0110] [Point list] Figure 19 is a diagram illustrating a list of points to be checked in a sales contract. As shown in Figure 19, the list of points to be checked in a sales contract lists 10 points, such as the parties and the display of the property being sold.

[0111] Figures 20-1 and 20-2 are examples of past sales contracts. Past sales contracts include the land's physical condition, rights and legal restrictions (Figure 20-1), and contract details (Figure 20-2) for each item, such as boundaries, land area, roads, contract termination, compensation for damages, and penalties.

[0112] Processing Prompt Figures 21-1 to 21-7 are diagrams showing examples of processing prompts. In each use case below, we will explain an example in which the assumed person, input data, role, processing for each role, output content, discussion and results of assigning other roles to the generation AI 161, consistency assessment, and effectiveness measurement are all shown within a single processing prompt. The processing prompt assumes that the generation AI 161 is a civil servant and a professional with over 20 years of experience in contract procedures with private businesses, and describes the constraints and input data (Figure 21-1).

[0113] The processing prompt instructs the user to read the contract and input data as STEP 1 (Figure 21-2).

[0114] The processing prompt instructs, in STEP 2, to divide the roles into three: "contract administrator," "contract administrator's supervisor," and "contractor," and to execute processing according to the role (Figure 21-3). In STEP 2-1, the processing prompt instructs to read the guardrail policy and execute processing according to the role.

[0115] The processing prompts indicate the processing content for each role (Figure 21-4). For example, in STEP 2-2, the contract administrator is instructed to check whether the "contract" satisfies the "contract check items," and if not, to output a revision proposal that satisfies the items.

[0116] STEP 2-3 instructs the superior and the contracting party to check the content output in STEP 2-2, to hold a discussion in which the superior and the contracting party comment from their respective positions, and to output a draft contract that is a compromise between the two parties.

[0117] In STEP 2-4, the contract administrator instructs the contract draft output in STEP 2-3 to be rechecked based on the contract items.

[0118] The processing prompts in STEPs 3 and 4 instruct the user to confirm the output contents and to save the output contents (FIG. 21-5).

[0119] The processing prompt then instructs the execution of consistency assessment in STEP 5 (Figure 21-6). STEP 5 instructs the system to correctly assess the consistency of the output with similar past output content and internal organizational information (white papers, budgets, etc.).

[0120] The processing prompt, SETP6, instructs the implementation of an effectiveness measurement (Figure 21-7). STEP6 calculates the time required to perform STEPs 1 to 4. Next, it calculates the workload (number of people and time) required when the same tasks in STEPs 1 to 4 are performed by humans without using the generation AI, and specifically outputs the effect of reducing the workload in STEPs 1 to 4 by using the generation AI.

[0121] [output] Next, a description will be given of examples of output from the processing system 10. Figures 22-1 to 22-13 are diagrams showing examples of output.

[0122] The processing system 10 outputs the results of reading the input material instructed in STEP 1 (FIG. 22-1). The processing system 10 outputs the results of reading the input material instructed in STEP 2-1 and the role setting results (FIG. 22-2, FIG. 22-3).

[0123] The processing system 10 outputs the check results of the superior and contracting party instructed in STEP2-2, comments from the discussion, and a compromise between the two parties (Figures 22-4 to 22-8). As shown in Figure 22-4, the output includes an overview of STEP2-2, including items that met the check items and items that require correction. As shown in Figures 22-5 and 22-6, the output also includes a checklist and the results of the policy compliance check.

[0124] As shown in Figure 22-7, the output includes comments from the superior (box W11) and comments from the contracting party (box W12) based on the discussion. As a result of the discussion, the output includes a compromise draft clause (box W13) as shown in Figure 22-8.

[0125] Next, the output will include the results of the final check by the contract administrator, as shown in Figure 22-9. If the user passes the final check (Figure 22-10), a record of the entire process will be generated (Figure 22-11).

[0126] Next, as shown in Figure 22-12, the output shows the results of the consistency assessment instructed in STEP 5. In the example in Figure 22-12, it was determined that consistency was met for all laws and regulations, organizational guidelines, budget policies, and similar past cases.

[0127] As shown in Figure 22-13, the output shows the results of the effectiveness measurement instructed in STEP 6. In the example of Figure 22-13, it shows that the operation volume was reduced by 70% or more for all of STEPs 1 to 4.

[0128] As shown in the results of measuring the effectiveness of the output, the contract check support process by the processing system 10 enabled the quick execution of checks by the business staff, check results by superiors and contracting parties, comments from discussions, compromises between the two parties, and rechecks by the business staff. Furthermore, as shown in the results of judging the consistency of the output, it was shown that the processing by the processing system 10 enabled the proper execution of contract checks.

[0129] In this way, the processing system 10 can standardize contract administration work by introducing a mechanism that allows for easy contract checks. Furthermore, the processing system 10 can prevent oversights by having the generation AI 161 perform a primary check before a human check. Furthermore, the processing system 10 can eliminate variations in personal checks and improve work quality by performing contract work according to a predetermined check rate.

[0130] [Use case 2] Use Case 2 will be explained using a budget formulation task as an example of a complex task. FIG. 23 is a diagram for explaining Use Case 2.

[0131] As shown in Figure 23 (A), currently, (1) the person in charge of the budget reads multiple documents, (2) analyzes and proposes (identifying issues, creating solutions to those issues, and creating budget details), (3) holds internal discussions, (4) modifies the proposal, and (5) it is approved. Since budget formulation work can only be handled by specific personnel, the workload tends to be concentrated. Furthermore, because budget formulation work is complex, analysis and proposals take time and effort, and the quality of the work can vary depending on the experience and ability of each individual, and differences in perspectives can lead to oversights.

[0132] By introducing the processing system 10 (introduction of the generation AI 161), this work can be replaced by the generation AI 161, for example, by the work staff member (1) reading multiple documents and (2) analyzing and proposing ideas. The generation AI 161 reads multiple documents, analyzes them, and then creates a discussion proposal. This reduces the burden on the work staff member and aims to standardize and streamline work.

[0133] For example, in the case of budget formulation work for the Digital Agency, the processing system 10 selects the Digital Agency's organizational chart and processing prompts (problem extraction prompt P21, problem-solving prompt P22, budget content formulation prompt P23) corresponding to each role from the government agency work DB 11. According to these prompts, three roles are set for the generation AI 161, namely, AI researcher, AI strategist, and veteran coordinator, who are in charge of the work, and each performs processing (e.g., analysis, proposal creation, discussion) for each role.

[0134] Figures 24-1 to 24-3 are examples of organizational charts for digital use, which are input materials. The organizational charts include the Digital Agency's mission and main roles, main policy issues and recent trends (Figure 24-1), organizational structure (Figure 24-2), and an overview of the work of each department (Figure 24-3).

[0135] [AI Researcher] [Prompt1] First, we will explain the processing prompts when assigning the role of an AI researcher, who is in charge of the business, to the generation AI 161 and having it organize social issues.

[0136] 25-1 to 25-3 are diagrams showing examples of processing prompts. The processing prompt (Prompt1) in Fig. 25-1 assumes that the generated AI 161 is an AI researcher belonging to the policy team of the Digital Agency, instructs it to organize social issues, and describes input data.

[0137] The processing prompt instructs the following steps (Figure 25-2): creating a "list of issues" (STEP 1), evaluation and comments by three experts (STEP 2), recording the subsequent discussion in minutes (STEP 3), summarizing the ideas (STEP 4), presenting the calculation method steps (STEP 5), creating a table with the final ranking (priority) (STEP 6), checking (STEP 7), determining consistency (STEP 8), and measuring the effectiveness (STEP 9).The processing prompt then instructs the generation AI 161 to create a frame for outputting the results of the processing performed (Figure 25-3).

[0138] [output] An example of output from the generated AI 161 when the role is AI researcher will be described below. Figures 26-1 to 26-9 are diagrams showing example outputs.

[0139] The generation AI 161 generates a list of six social issues instructed by STEP 1 (Figure 26-1). The generation AI 161 outputs the evaluations and comments from three experts A to C instructed by STEP 2 (Figure 26-2), the minutes recording the discussion instructed by STEP 3 (Figure 26-3), a summary of the thinking memos instructed by STEP 4 (Figure 26-4), the calculation method steps instructed by STEP 5 (Figure 26-5), and a summary and prioritized table of STEP 6 instructions (Figure 26-6). In this way, the generation AI 161 outputs six social issues organized from the perspective of an AI researcher.

[0140] Next, the generation AI 161 outputs the check results of the STEP 7 instructions (FIG. 26-7), performs a consistency check on the STEP 8 instructions and outputs the check results (FIG. 26-8), and measures the effectiveness of the STEP 9 instructions and outputs the measurement results (FIG. 26-9).

[0141] In the example of Figure 26-8, the consistency assessment determined that the issue priorities matched the Digital Agency's key areas, the KPIs matched the official dashboard values, and consistency was met for both the guardrail policies.

[0142] As shown in Figure 26-9, the results of measuring the effectiveness of STEPs 1 to 8 showed that operating hours were reduced by 99.4% and labor costs were reduced by 120,000 yen.

[0143] [AI Strategist] [Prompt2] Next, a processing prompt when the generation AI 161 is given the role of an AI strategist and is made to create a solution will be described.

[0144] Figures 27-1 to 27-3 are diagrams showing examples of processing prompts. The processing prompt (Prompt2) in Figures 27-1 to 27-3 assumes that the generating AI 161 is an AI strategist, receives the priority of Prompt1 (Figure 26-6), and instructs the generating AI 161 to create a solution (Figure 27-1).

[0145] The processing prompt instructs the generation AI 161 to organize the issues according to the specified items (STEP 1), have experts review the results of the STEP 1 organization (STEP 2), record them in the minutes (STEP 3), memo the ideas and write down the calculation steps (STEP 4), present an easy-to-implement integration plan based on the guard race policy (STEP 5), check whether it complies with the law and is realistic (STEP 6), determine consistency (STEP 7), and measure the effects (STEP 8) (Figure 27-2).The processing prompt then instructs the generation AI 161 to create a frame for outputting the results of the processing performed (Figure 27-3).

[0146] As described above, the processing prompt causes the generation AI 161 to output multiple solution proposals for a problem, prioritize each of the multiple solution proposals, and output an integrated proposal that integrates the multiple solution proposals. The processing prompt may also cause the generation AI 161 to create a text-based diagram for each solution proposal. The processing prompt may also cause the generation AI 161 to create a project proposal (solution) corresponding to a combination of policy content and budget amount set by the government (public agency). In this case, the processing prompt may include an instruction to cause the generation AI 161 to prioritize creating a solution proposal for an issue whose budget amount meets a threshold and whose policy content falls under a predetermined item. The processing prompt may also cause the generation AI 161 to create a solution proposal based on the issue of the policy content and the current situation.

[0147] The processing prompt may also cause the generation AI 161 to predict the content of policies corresponding to the government's future budget based on past budgets (information combining the content of policies and the budget amount). The processing prompt may also specify a way of thinking when predicting the content of policies. The processing prompt may then appropriately apply a description of a prompt for planning a project related to a budget for which the evaluation results of the budget satisfy the conditions. The processing prompt evaluates the budget by causing the generation AI 161 to calculate scores according to the policies, such as a score for the degree of agreement between the policy content and social trends, a score for the attractiveness of the market related to the policy content, and a score for the urgency of the policy content.

[0148] [output] An example of output from the generated AI 161 when the role is AI strategist will be described below. Figures 28-1 to 28-9 are diagrams showing example outputs.

[0149] The generation AI 161 generates the results of organizing the tasks instructed in STEP 1 (Figures 28-1 and 28-2). Figures 28-1 and 28-2 show how the AI ​​strategist has organized each task (Figure 26-1) created by the AI ​​researcher.

[0150] Next, the generation AI 161 outputs the review by three experts A to C instructed in STEP 2 (Figure 28-3), the minutes instructed in STEP 3 (Figure 28-4), the memo of the thinking instructed in STEP 4, the calculation method steps (Figure 28-5), the integrated plans 1 to 4 instructed in STEP 5 (Figure 28-6), and the check results (STEP 6) (Figure 28-7). In this way, the generation AI 161 creates and checks four solutions created from the perspective of an AI strategist.

[0151] Next, the generation AI 161 judges the consistency of the STEP7 instructions and outputs the judgement result (FIG. 28-8), measures the effect of the STEP8 instructions and outputs the measurement result (FIG. 28-9).

[0152] In the example of Figure 28-8, the consistency assessment showed that the costs matched the priority plan, met the six elements of the Guard Race policy, there was no overlap with past case examples, and the KPIs matched the official dashboard, so it was determined to be consistent.

[0153] As shown in Figure 28-9, the results of measuring the effectiveness showed that operating hours were reduced by 99.4%, resulting in significant savings in personnel and labor costs.

[0154] [Veteran Coordinator] [Prompt3] Next, a processing prompt when assigning the veteran coordinator role to the generation AI 161 will be described.

[0155] Figures 29-1 to 29-3 are diagrams showing examples of processing prompts. The processing prompt (Prompt3) assumes that the generating AI 161 is a veteran coordinator, and instructs it to receive the integration plan of Prompt2 (Figure 28-6) and turn it into a plan that can be implemented immediately (Figure 29-1).

[0156] The processing prompts instruct the generation AI 161 to create a project card for each policy (STEP 1), create a table that combines all projects in STEP 2 and presents annual expenditures and the remaining budget (STEP 2), prepare comments from three experts (STEP 3), record the full discussion in the minutes (STEP 4), create a summary showing agreements and changes and consider priorities based on the Guard Race policy (STEP 5), create a final package including the project cards and financial table (STEP 6), check whether the results are within budget and whether the KPIs are appropriate (STEP 7), determine consistency (STEP 8), and measure effectiveness (STEP 9) (Figure 29-2).The processing prompts then instruct the generation AI 161 to create a frame for outputting the results of the processing performed (Figure 29-3).

[0157] [output] An example of output from the generation AI 161 when the role is veteran coordinator will be described below. Figures 30-1 to 30-13 are diagrams showing example outputs.

[0158] The generation AI 161, acting as a veteran coordinator, generates business cards for each policy instructed in STEP 1 (Figures 30-1 to 30-6). Figures 30-1 to 30-6 show the business cards for the policies listed in the integration plan (Figure 28-6) created by the AI ​​strategist (Figure 27-1).

[0159] Next, the generation AI 161 outputs the STEP 3 instructions showing expenditures by year and remaining budget (Figure 30-7), the STEP 4 instructions showing comments from three experts A to C (Figure 30-8), the STEP 4 instructions meeting minutes (Figure 30-9), the STEP 5 instructions summary (Figure 30-10), and the STEP 6 instructions final package (Figure 30-11). In this way, the generation AI 161 creates a final package adjusted from the standpoint of the AI ​​strategist and checks it.

[0160] Next, the generation AI 161 judges the consistency of the STEP 8 instructions and outputs the judgement result (FIG. 30-12), measures the effect of the STEP 9 instructions and outputs the measurement result (FIG. 30-13).

[0161] In the example of Figure 30-12, the consistency assessment determined that the costs were consistent with the priority plan and the estimated requested amount, and were in compliance with the Accounting Act and the Public Works Bidding and Contracting Fairness Act, and therefore met the consistency criteria.

[0162] As shown in Figure 30-13, the results of measuring the effectiveness showed that operating hours were reduced by 99.1%, resulting in significant savings in personnel and labor costs.

[0163] As in use case 2, the processing system 10 assigns the roles of AI researcher, AI strategist, and veteran coordinator to the generation AI 161, and causes the generation AI 161 to perform processing sequentially for each role according to the processing prompts corresponding to each role. Then, in use case 2, the processing system 10 causes the generation AI 161 to use the output of processing performed in the previous role in processing for the subsequent role to hold a discussion. In this way, the processing system 10 enables the generation AI 161 to extract issues, create solutions to the issues, and create a final package including a budget proposal.

[0164] As shown in the consistency determination results and the effectiveness measurement results, the processing system 10 can improve the optimization and efficiency of processing.

[0165] Therefore, the processing system 10 can ease restrictions on the number of personnel responsible for budget formulation work, and can level out the work. As in Use Case 2, the processing system 10 can provide practical output that takes into account the expected opinions of stakeholders and predetermined evaluation criteria, thereby improving the quality of budget formulation work.

[0166] [Use case 3] Use Case 3 will be explained using an example of manual creation and know-how sharing operations. FIG. 31 is a diagram explaining Use Case 3.

[0167] As shown in Figure 31 (A), currently, (1) a specific person performs the analysis, (2) if there is no manual flow, the specific person creates a manual, (3) an internal discussion takes place, (4) after revisions are made, (5) the manual is approved. Because government agencies have a workforce structure in which many employees are seconded, personnel are replaced every 2 to 3 years, making it difficult to accumulate know-how and carry out work efficiently and with high quality.

[0168] By introducing the processing system 10 (introduction of the generation AI 161), for example, (1) analysis and (2) manualization are replaced by the generation AI 161 as shown in FIG. 31(B).

[0169] Then, the manual created by the generation AI 161 is (3) discussed within the department, (4) revised, and (5) approved. The generation AI 161 prevents work from becoming dependent on individuals by creating a manual that describes the process flow and output results.

[0170] The processing system 10 selects processing prompts for manual creation processing from the government business database 11.

[0171] Processing Prompt Figures 32-1 to 32-7 are diagrams showing examples of processing prompts. The processing prompt in Figure 32-1 indicates that the generated AI 161 is assumed to be a bureaucrat belonging to a Japanese government ministry. The input data includes a complete record of sales contract checks, 10 checkpoints, an organizational chart, priorities, etc.

[0172] The process prompt states that STEP 1 is to confirm the process to be manualized (Figure 32-2).

[0173] The processing prompt instructs, in STEP 2, to extract the output result of the relevant task (for example, contract check task) from the detailed task content DB 13 (FIG. 32-3). FIGS. 33-1 to 33-9 are diagrams showing examples of the output result of the sales contract check task stored in the detailed task content DB 13.

[0174] As shown in Figure 33-1, the output assumes that the generation AI 161 is a bureaucrat and includes a list of input materials and instructions for the document reading results (summary) (STEP 1) (Figure 33-2). The output then includes the setting of guardrail policies and roles (STEP 2-1) (Figure 33-3), the results of the contract administrator's check from 10 perspectives (STEP 2-2) (Figure 33-4), a summary of the tripartite consultation and compromise proposal (STEP 2-3) (Figure 33-5), the final check results (STEP 2-4) (Figure 33-6), and the compromise clause proposal. The output then includes the user's final confirmation results (STEP 2-2) (Figure 33-7), the consistency assessment results (Figure 33-8), and the effectiveness measurement results (Figure 33-9).

[0175] Returning to Figure 32-4, the processing prompt instructs the user to create a manual in STEP 3 and to check the manual contents in STEP 4 (Figure 32-5). The processing prompt instructs the user to perform a consistency assessment in STEP 5 (Figure 32-6). The processing prompt instructs the user to perform an effectiveness measurement in STEP 6 (Figure 32-7).

[0176] [output] Next, a description will be given of examples of output from the processing system 10. Figures 34-1 to 34-13 are diagrams showing examples of output.

[0177] The processing system 10 outputs the confirmation result of the manualization process instructed in STEP 1 (FIG. 34-1).

[0178] The processing system 10 presents the main information of the contract check work extracted from the detailed work content DB 13 instructed in STEP 2 (FIGS. 34-2 to 34-4).

[0179] The processing system 10 outputs the manual specified in STEP 3 (Figures 34-5 to 34-12). The manual includes the files to be used as a prerequisite (Figure 34-5), a business flow chart (Figure 34-6), and detailed procedures such as reading materials (STEP 1), setting guardrail policies (STEP 2-1) (Figure 34-7), checking using a 10-item checklist (STEP 2-2) (Figure 34-8), tripartite discussions and output of compromise drafts (STEP 2-3), execution of final checks (STEP 2-4) (Figure 34-9), final user confirmation (STEP 3), documentation and electronic storage (STEP 4) (Figure 34-10), and effectiveness measurement (STEP 5) (Figure 34-11).

[0180] In this way, the processing system 10 can prevent work from becoming dependent on a specific person by using the generation AI 161 to create a manual for the processing flow and output results. Furthermore, the processing system 10 can quickly create a manual using the generation AI 161, which can help facilitate smooth handover between personnel even in workplaces where personnel changes occur frequently.

[0181] [Example of processing prompt and output for new business data creation process] [Composition prompt] A description will be given of a creation prompt and an output example for new business data creation processing. Figures 35-1 to 35-4 are diagrams showing examples of the creation prompt.

[0182] The creation prompt assumes that the agent-type AI 181 is a team that handles projects and routine tasks in the government sector. The creation prompt instructs the agent-type AI 181 to launch at a set time, either weekly or monthly (Figure 35-1).

[0183] The creation prompt instructs the agent AI 181 to automatically read files stored in the detailed task content DB 13. The creation prompt instructs the agent AI 181 to create an updated task list report and detect differences from existing data. The creation prompt instructs the agent AI 181 to propose a new task plan that reflects the detected differences and a new processing procedure that indicates the processing procedure for the newly created task (Figure 35-1).

[0184] The creation prompt instructs the agent-type AI 181 to output the list of tasks as a report in a specified format (Figure 35-2). The main points of the report are an overview summary, a list by department, a summary of differences, important topics, next actions, and a file update proposal (which corresponds to a proposal for a new task).

[0185] The creation prompt instructs the agent-type AI 181 on the processing procedures for creating output (Figure 35-3). Specifically, the creation prompt instructs the processing procedures to determine the file to be read, read the document, detect differences (new, update, and delete tasks), create a list of tasks, propose file updates and generate system prompts for task processing, and generate a report.

[0186] The creation prompt instructs the agent-type AI 181 on style guides, error handling, and personal information redaction (Figure 35-4).

[0187] [output] An example of output from the agent-type AI 181 will be described below. Figures 36-1 to 36-3 are diagrams showing an example of output from the agent-type AI 181.

[0188] The agent-type AI 181 outputs a summary, a list of new, updated, and deleted tasks by department, a summary of differences, and important topics in accordance with a prescribed format (Figure 36-1).The summary detects one new task (the Digital Division's online resident registration issuance project) and one updated task (the Crisis Management Office's revision of the disaster response manual).

[0189] The agent-type AI 181 outputs a system prompt (processing prompt) for the new task "Resident Registration Online Issuance Project" (Figure 36-2). The new processing prompt specifically describes the purpose, deliverables, scope and tasks, milestones, risks, and reference materials.

[0190] The agent-type AI 181 outputs a system prompt (processing prompt) for the update task "Revising the Disaster Prevention Manual for the Crisis Management Office" (Figure 36-3). The new processing prompt specifically describes the change, required actions, deadlines, risks and points to note, and reference materials.

[0191] Since the processing system 10 executes the new business data creation process at a predetermined timing, the person in charge does not need to perform the complicated and burdensome process of creating a new proposed processing prompt. Therefore, the processing system 10 can realize the creation of a quick and appropriate proposed processing prompt while reducing the workload of the person in charge.

[0192] [System configuration of the embodiment] The components of each server in the processing system 10 are conceptual functional components and do not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of the functions of each server in the processing system 10 is not limited to that shown in the figure, and all or part of them can be functionally or physically distributed or integrated in any unit depending on various loads, usage conditions, etc.

[0193] Furthermore, all or any part of the processes performed in each server of the processing system 10 may be realized by a CPU, a GPU (Graphics Processing Unit), and a program analyzed and executed by the CPU and the GPU. Furthermore, each process performed in each server of the processing system 10 may be realized as hardware using wired logic.

[0194] Furthermore, among the processes described in the embodiments, all or part of the processes described as being performed automatically can be performed manually. Alternatively, all or part of the processes described as being performed manually can be performed automatically using a known method. In addition, the processing procedures, control procedures, specific names, and information including various data and parameters described above and illustrated can be changed as appropriate unless otherwise specified.

[0195] [program] 37 is a diagram showing an example of a computer in which a program is executed to realize the processing system 10. The computer 1000 has, for example, a memory 1010 and a CPU 1020. The computer 1000 also has a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These components are connected by a bus 1080.

[0196] The memory 1010 includes a ROM 1011 and a RAM 1012. The ROM 1011 stores a boot program such as a BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to a hard disk drive 1090. The disk drive interface 1040 is connected to a disk drive 1100. A removable storage medium such as a magnetic disk or optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to a mouse 1110 and a keyboard 1120, for example. The video adapter 1060 is connected to a display 1130, for example.

[0197] The hard disk drive 1090 stores, for example, an OS (Operating System) 1091, an application program 1092, a program module 1093, and program data 1094. That is, a program that defines each process of each server in the processing system 10 is implemented as a program module 1093 in which code executable by the computer 1000 is written. The program module 1093 is stored, for example, in the hard disk drive 1090. For example, a program module 1093 for executing the same process as the functional configuration of each server in the processing system 10 is stored in the hard disk drive 1090. Note that the hard disk drive 1090 may be replaced with an SSD (Solid State Drive).

[0198] Furthermore, setting data used in the processing of the above-described embodiment is stored as program data 1094, for example, in memory 1010 or hard disk drive 1090. Then, CPU 1020 reads program module 1093 and program data 1094 stored in memory 1010 or hard disk drive 1090 into RAM 1012 as necessary and executes them.

[0199] The program module 1093 and program data 1094 are not limited to being stored in the hard disk drive 1090, but may also be stored in, for example, a removable storage medium and read by the CPU 1020 via the disk drive 1100 or the like. Alternatively, the program module 1093 and program data 1094 may be stored in another computer connected via a network (such as a LAN (Local Area Network) or WAN (Wide Area Network)). The program module 1093 and program data 1094 may then be read by the CPU 1020 from the other computer via the network interface 1070.

[0200] The functions of the elements disclosed herein may be implemented using circuitry or processing circuitry including general-purpose processors, special-purpose processors, integrated circuits, ASICs (Application Specific Integrated Circuits), FPGAs (Field Programmable Gate Arrays), conventional circuits, and / or combinations thereof that are programmed using one or more programs stored in one or more memories or otherwise configured to perform the disclosed functions. A processor is considered processing circuitry or circuitry because it includes transistors and other circuitry. A processor may also be a programmed processor that executes a program stored in a memory. In this disclosure, a circuit, unit, or means is hardware that performs the recited functions or hardware that is programmed to perform the recited functions. The hardware may be any hardware disclosed herein that is programmed or configured to perform the recited functions.

[0201] There is a memory that stores a computer program including computer instructions. These computer instructions provide the logic and routines that enable the hardware (e.g., processing circuitry or circuitry) to perform the methods disclosed herein. The computer program can be implemented in commonly known forms as a computer-readable storage medium, a computer program product, a memory device, a recording medium such as a CD-ROM or DVD, and / or the memory of an FPGA or ASIC.

[0202] Although the present invention has been described above as an embodiment, the present invention is not limited to the descriptions and drawings that form part of the disclosure of the present invention. In other words, other embodiments, examples, and operational techniques that can be made by those skilled in the art based on the present invention are all included in the scope of the present invention. [Explanation of symbols]

[0203] 10 Processing System 11 Government Agency Business Database 12 Guardrail Policy DB 13 Detailed work content DB 14 Processing start determination unit 15 Processing start control section 16 Processing section 17 New Business Proposal Prompt DB 18 New Business Data Creation Department 20 Business Information DB 30 User terminals 141,181 Agent-based AI 161 Generation AI 162 Role Setting Section 163 Processing Department 164 Consistency judgment part 165 Effectiveness Measurement Department 166 Result output section

Claims

1. A start determination unit that identifies the business data creation process to be performed based on whether or not there is an increase or decrease in business information stored in a business information database that stores business information of government agencies, or whether or not there is an instruction to start processing from a person in charge of creating business data in the business data creation process to be performed, and determines whether or not it is necessary to start the business data creation process; a selection unit that, when the start determination unit determines that the business data generation process to be performed is to be started, reads a process start prompt that describes data to be selected for the business data generation and an overall process flow, and selects, in accordance with the read process start prompt, government agency business data required for the business data generation process to be performed and a processing prompt that indicates a processing procedure for the business data generation process to be performed from a government agency business database that stores government agency business data that indicates the organization and business content of the government agency and processing prompts used in the business data generation process; a setting unit that sets at least the role of the person in charge in a generative model based on the processing prompt selected by the selection unit; a processing execution unit that, in accordance with the processing prompt selected by the selection unit, causes the generation model to read the guardrail policies and detailed job contents from a guardrail policy database that stores guardrail policies indicating decision-making priorities in the government agency and a detailed job contents database that stores detailed job contents including organizational charts and job contents of all government agencies, and performs predetermined processing for each role based on the read guardrail policies and detailed job contents and the government agency business data selected by the selection unit to create the business data to be created; A processing system comprising:

2. the processing prompt includes an instruction to determine the consistency between the business data created by the processing execution unit and past business data of the same type and / or internal information of a government agency in accordance with a predetermined determination condition; a consistency determination unit that determines consistency between the business data created by the processing execution unit and past business data of the same type and / or internal information of a government agency using the generative model in accordance with the processing prompt; When the consistency determination unit determines that the business data created by the processing execution unit is consistent with past business data of the same type and / or internal information of a government agency, the consistency determination unit stores the business data created by the processing execution unit in the detailed business content database; The processing system according to claim 1, characterized in that if the consistency determination unit determines that the business data created by the processing unit is inconsistent with the past business data of the same type and / or internal information of a government agency, the processing unit creates the business data again.

3. an effect measurement unit that compares the processing content of the processing execution unit with the processing content when a person performs the same work as the processing performed by the processing execution unit without using the generative model, and quantitatively and / or qualitatively measures the effect of the work performed by the processing execution unit; an output unit that outputs the business data created by the processing execution unit and the effect of the work measured by the effect measurement unit; 10. The processing system of claim 1, further comprising:

4. The role is a person in charge of creating the business data, the person in charge's superior, and / or a related person, If the role is a person in charge, the generation model is caused to create the business data; If the role is a superior of the person in charge, the generative model checks the business data created during the role of the person in charge; The processing system according to claim 1, characterized in that, when the role is the person in question, the generative model is caused to add comments to the business data created during the role of the person in question from the perspective of the person in question.

5. a new business data creation unit that, at a predetermined timing, causes the generative model to organize the business based on the data stored in the detailed business content database, and creates a new business proposal indicating the business content of the newly generated business, a new processing procedure indicating the processing procedure of the newly generated business, and a new processing prompt proposal instructing the creation of business data related to the newly generated business; The processing system according to claim 1, characterized in that the new business data creation unit stores the new business proposal and new processing procedure that have been instructed to be stored in the government agency business database, and the new processing prompt proposal, in the government agency business database.

6. A processing method performed by a processing system, comprising: Identifying the business data creation process to be performed based on whether or not there is an increase or decrease in business information stored in a business information database that stores business information of government agencies, or whether or not there is an instruction to start processing from a person in charge of creating business data in the business data creation process to be performed, and determining whether or not it is necessary to start the business data creation process; If it is determined in the determination of whether or not to start that the business data generation process to be performed is to be started, a processing start prompt is read that describes data to be selected for the business data generation and the overall flow of the processing, and in accordance with the read processing start prompt, government agency business data required for the business data generation process to be performed and a processing prompt that indicates a processing procedure for the business data generation process to be performed are selected from a government agency business database that stores government agency business data that indicates the structure and business content of the government agency and processing prompts used in the business data generation process; Based on the selected processing prompt, set at least the role of the person in the generative model; According to the selected processing prompt, the generative model is caused to read the guardrail policies and the detailed job contents from a guardrail policy database that stores guardrail policies indicating decision-making priorities in the government agency and a detailed job contents database that stores detailed job contents including organizational charts and job contents of all government agencies; and causing the generation model to perform a predetermined process for each role based on the loaded guardrail policy and detailed business content and the selected government agency business data, thereby creating business data to be created. A processing method comprising:

7. Identifying the business data creation process to be performed based on whether or not there is an increase or decrease in business information stored in a business information database that stores business information of government agencies, or whether or not there is an instruction to start processing from a person in charge of creating business data in the business data creation process to be performed, and determining whether or not it is necessary to start the business data creation process; If it is determined in the determination of whether or not to start that the business data generation process to be performed is to be started, a processing start prompt is read that describes data to be selected for the business data generation and the overall flow of the processing, and in accordance with the read processing start prompt, government agency business data required for the business data generation process to be performed and a processing prompt that indicates a processing procedure for the business data generation process to be performed are selected from a government agency business database that stores government agency business data that indicates the structure and business content of the government agency and processing prompts used in the business data generation process; Based on the selected processing prompt, set at least the role of the person in the generative model; According to the selected processing prompt, the generative model is caused to read the guardrail policies and the detailed job contents from a guardrail policy database that stores guardrail policies indicating decision-making priorities in the government agency and a detailed job contents database that stores detailed job contents including organizational charts and job contents of all government agencies; and causing the generation model to perform a predetermined process for each role based on the loaded guardrail policy and detailed business content and the selected government agency business data, thereby creating business data to be created. A processing program that causes a computer to execute the above.

Images