Information processing device, information processing method, and information processing program

Abstract

Provided are an information processing device, an information processing method, and an information processing program capable of reducing the number of man-hours required to create a deliverable. The information processing device: acquires past deliverables created in the past in waterfall-type system development including a plurality of development steps; analyzes the format of the acquired past deliverables; provides an input screen for inputting input items required to create a new deliverable; generates a prompt for automatically generating the new deliverable on the basis of the analyzed format of the past deliverables and the input items input to the input screen; provides the generated prompt to a large-scale language model; acquires, from the large-scale language model, a new deliverable generated in accordance with the provided prompt; and provides the acquired new deliverable.

Description

Information processing device, information processing method, and information processing program

[0001] The present invention relates to an information processing device, an information processing method, and an information processing program.

[0002] In recent years, project management tools utilizing IT (Information Technology) have been used to manage projects such as system development. A technique for automatically generating deliverables to be created in a project is known for project management tools. For example, Patent Document 1 describes a technique for obtaining a record linked to a project number included in a specified record of project data specified by an operator from purchase voucher data, expense voucher data, and daily work report data, obtaining a commission rate linked to a customer included in the specified record from a customer commission rate master, calculating the unit price and sales price of purchases and the sales price amount of labor costs based on the obtained record and commission rate, and creating project detail data based on the obtained record and calculated amounts.

[0003] Japanese Patent Application Laid-Open No. 2022-079589

[0004] However, in waterfall-style system development involving multiple development steps, the deliverables created in each development step are interrelated. Therefore, when creating a new deliverable, it is necessary to check its consistency with deliverables created in the past, and this consistency check work can sometimes increase the man-hours required to create the deliverable.

[0005] Furthermore, when modifying descriptions in the creation of a deliverable, in waterfall-type system development, it is necessary to modify the descriptions of each related deliverable, and the modification work can sometimes increase the amount of work required to create the deliverable.

[0006] The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an information processing device, an information processing method, and an information processing program that can reduce the number of steps required to create a product.

[0007] In order to solve the above problems, the information processing device of the embodiment includes a past product acquisition unit that acquires past products created in the past in waterfall-type system development including multiple development steps, a past product analysis unit that analyzes the format of the past products acquired by the past product acquisition unit, an input screen providing unit that provides an input screen for inputting input items necessary to create a new product, a prompt generation unit that generates a prompt for automatically generating a new product based on the format of the past product analyzed by the past product analysis unit and the input items entered on the input screen provided by the input screen providing unit, a prompt providing unit that provides the prompt generated by the prompt generation unit to a large-scale language model, a new product acquisition unit that acquires from the large-scale language model new products generated in response to the prompt provided by the prompt providing unit, and a new product providing unit that provides the new product acquired by the new product acquisition unit.

[0008] According to one embodiment of the present invention, in waterfall-type system development including multiple development steps, past deliverables created in the past are acquired, the format of the acquired past deliverables is analyzed, an input screen is provided for inputting the input items necessary to create a new deliverable, a prompt for automatically generating a new deliverable is generated based on the analyzed format of the past deliverable and the input items entered on the input screen, the generated prompt is provided to a large-scale language model, a new deliverable generated in response to the provided prompt is acquired from the large-scale language model, and the acquired new deliverable is provided, thereby reducing the labor required to create the deliverable.

[0009] FIG. 1 is a diagram showing an example of an overview of an information processing device according to an embodiment. FIG. 2 is a diagram showing an example of the positioning of each step (process) in an embodiment and the connection between deliverables in each step (process). FIG. 3 is a diagram showing an example of the configuration of an information processing device according to an embodiment. FIG. 4 is a diagram showing a first example of an input screen related to automatic generation and updating of a deliverable provided by an information processing device according to an embodiment. FIG. 5 is a diagram showing a second example of an input screen related to automatic generation and updating of a deliverable provided by an information processing device according to an embodiment. FIG. 6 is a diagram showing a sixth example of an input screen related to automatic generation and updating of a deliverable provided by an information processing device according to an embodiment. FIG. 7 is a diagram showing a seventh example of an input screen related to automatic generation and updating of a deliverable provided by an information processing device according to an embodiment. FIG. 8 is a diagram showing an eighth example of an input screen related to automatic generation and updating of a deliverable provided by an information processing device according to an embodiment. FIG. 9 is a block diagram showing an example of the hardware configuration of an information processing device according to an embodiment. FIG. 10 is a diagram showing an example of a management screen for centralized management of deliverables provided to a user terminal by an information processing device according to an embodiment. 10 is a diagram for explaining a case where an information processing device in an embodiment creates program code as a deliverable. FIG. 11 is a flowchart showing a first example of the operation of the information processing device in an embodiment. FIG. 12 is a flowchart showing a second example of the operation of the information processing device in an embodiment. FIG. 13 is a flowchart showing a third example of the operation of the information processing device in an embodiment. FIG. 14 is a flowchart showing a fourth example of the operation of the information processing device in an embodiment. FIG. 15 is a diagram showing a ninth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 16 is a diagram showing a tenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 17 is a diagram showing an eleventh example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment.10 is a diagram showing a twelfth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 11 is a diagram showing a thirteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 12 is a diagram showing a thirteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 13 is a diagram showing a fourteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 14 is a diagram showing a fifteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 15 is a diagram showing a sixteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 16 is a diagram showing a seventeenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 17 is a diagram showing an eighteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 18 is a diagram showing a nineteenth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 19 is a diagram showing a twentieth example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 20 is a diagram showing a twenty-first example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment. FIG. 21 is a diagram showing a twenty-second example of an input screen related to automatic generation and updating of a deliverable provided by the information processing device in the second embodiment.

[0010] An information processing device, an information processing method, and an information processing program according to an embodiment of the present invention will be described in detail below with reference to the drawings. Note that the contents included in multiple figures may be omitted to avoid duplication. First, an overview of the information processing device will be described using Figure 1. Figure 1 is a diagram showing an example of the overview of the information processing device according to an embodiment.

[0011] In FIG. 1 , information processing device 1 is a device that provides a newly generated deliverable to a user (hereinafter referred to as "user") who creates a deliverable related to waterfall-type system development. In this embodiment, system development refers to the development of a system using technology related to IT or DX (Digital Transformation (X)) (hereinafter, "IT or DX" may be referred to as "IT / DX"). DX refers, for example, to the digitization of information or the transformation of management using digitized information. Waterfall-type system development is a development method in which development proceeds sequentially from upstream to downstream development steps. Development in downstream processes is based on the development results of development steps upstream of the current step (the step of interest), and development proceeds on the premise that there will be no reversal from downstream to upstream. Waterfall-type system development is suitable for system development in which changes in development results in development steps upstream of the current step are unlikely to occur. The term "deliverables" refers not only to materials (regardless of file format or format), but also to program code generated by coding, source code, test scenarios, test results, etc., and also to materials that compile these. Furthermore, deliverables may include objects, information, etc. that are produced during or as a result of actions with a goal or purpose.

[0012] In this embodiment, a past work product is a work product created in the past. A new work product is a work product that is newly created (generated). Past work products include work products of the same type as the work product newly created in the same development step as the new work product, and work products created in the past in a development step upstream of the current step of the new work product (the step of interest). If the past work product is of the same type as the new work product, the past work product is used as a sample when creating the new work product. On the other hand, if the past work product is a work product created in the past in a development step upstream of the current step of the new work product, the new work product becomes an interrelated work product created in a development step downstream of the past work product. In waterfall-type system development, a new work product related to downstream development is created based on a past work product created in a development step upstream of the current step. The created new work product is treated as a past work product in a further downstream development step, and a new work product is then created.

[0013] Here, referring to FIG. 2, a specific example of deliverables created in each development step (process) of waterfall-type system development will be illustrated. FIG. 2 shows an example of the positioning of each step (process) and the connections between deliverables in each step (process). Each deliverable is created based on a previous deliverable created in an upstream development step. Note that the following explanation illustrates a case where the development steps (processes) are performed in the order of planning → requirements definition → design → development → testing → release. However, this embodiment is not limited to these development steps (processes). For example, other development steps (processes) may be included, or some development steps (processes) may be omitted.

[0014] Process Level 1: Planning Process Process Level 2: Planning ・Mid-term management plan (for example, digital strategy document): Describes the organization's mid-term business goals and the plan to use digital technology to support them. It includes strategies for improving business efficiency through digitalization, developing a new customer base, and promoting innovation. ・Project plan: Comprehensively describes the project's objectives, scope, schedule, release timing, budget, risk management plan, etc. It provides basic guidelines for project management and execution.

[0015] Level 1: Requirements Definition Level 2: Requirements Definition ・List of Issues / Proposed Solutions: List issues related to the project or business improvement and propose specific solutions for each issue. You may also include information such as the prioritization or impact of the issues. ・Problem Cause Analysis Diagram: A diagram that systematically analyzes and visually displays the causes of problems when they occur. Problem cause analysis uses methods such as fishbone diagrams (cause and effect diagrams) or 5-Why analysis. ・Requirements List: Lists the business requests or requirements for the project or system in a list format. Details of each request, related tasks, importance, and priority of the request are included. ・Requirements Structure Diagram: A diagram showing the relationships and structure between requirements. It visually represents the dependency relationships or hierarchical structure of requirements and is used to efficiently manage overall requirements. Level 2: Requirements Definition ・Systemization Task List (Requirements Organization Table): Lists the tasks to be systemized and includes the basic functional or non-functional requirements, evaluation, priority, and evaluation reasons for each business request or requirement.・Requirements specification: A detailed description of the specific requirements for a system or project. It covers functional requirements, non-functional requirements, and user requirements, and serves as the basis for subsequent design or development steps. It may also include a timeline or estimated costs.

[0016] Process Level 1: Design process Process Level 2: Basic design ・UI design document: Describes the overview of the user interface, screen transition diagram, or wireframe, etc. ・System architecture design document: Defines the overall system structure, or the main components and their relationships. ・Data model design document: Describes the database schema, ER (Entity Relationship) diagram, or data dictionary, etc. ・Security design document: Describes the design to meet security requirements, or measures and policies in detail. ・Basic design document: Describes the system overview, main functions and data structure, or interface definitions, etc. Process Level 2: Detailed design ・Screen specifications: Defines the detailed layout, operation methods, or processing logic of each screen, etc. ・API (Application Programming Interface) design document: Describes the specifications of the API used to interface with external systems or other modules. Database detailed design document: Describes specific database designs such as table definitions, index designs, or relational integrity constraints. Detailed design document: Describes detailed process flows, algorithms, data structures, or interface designs for each function.

[0017] Level 1 process: Development process Level 2 process: Programming Source code: Text written in a programming language that specifically specifies the behavior of the software. Specific code for implementing the system's functions is written here.

[0018] Process Level 1: Testing process Process Level 2: Unit testing ・Unit test plan: Describes the test purpose, environment, schedule, person in charge, test procedures for each function, input data, and expected results. ・Unit test report: Describes an overview of the test results, problems, improvement suggestions, a list of discovered bugs and their status and priority. Process Level 2: Integration test ・Integration test design document: Describes the test purpose, environment, schedule, person in charge, test procedures for interfaces and integration parts between modules, input data, and expected results. ・Integration test report: Describes an overview of the test results, problems, improvement suggestions, a list of discovered bugs and their status and priority. Process Level 2: Comprehensive testing ・Comprehensive test design document: Describes the test purpose, environment, schedule, person in charge, test procedures that cover the functions of the entire system, input data, and expected results. ・Comprehensive test report: Describes an overview of the test results, problems, improvement suggestions, a list of discovered bugs and their status and priority. Process Level 2: User Acceptance Testing ・Acceptance test plan: Describes the test objectives, environment, schedule, person in charge, test procedures based on scenarios used by actual users, input data, and expected results. ・Acceptance test report: Describes an overview of the test results, problems, improvement proposals, a list of discovered bugs and their status and priority.

[0019] Process Level 1: Release process Process Level 2: Manual creation ・User manual: User documentation including how to use the system, troubleshooting, and FAQs. ・Operation manual: How to operate the system, maintenance plan, and emergency response procedures. Process Level 2: Migration ・Migration plan: Migration procedures for the production environment, schedule, risk management, and backup plan. ・Migration test report: Test results in the production environment, recording of issues, and final approval. ・Project completion report: Summary of project results, level of achievement, unmet challenges, and future proposals.

[0020] The above deliverables are only a partial example of the deliverables created in waterfall system development, and in actual system development, dozens of different deliverables may be created. Because deliverables are created based on deliverables created in upstream development steps, in waterfall system development, if a deliverable in a downstream development step is modified, it will require re-examination (rework) in the upstream development process, and it will also be necessary to check the impact that the modified deliverable has on the contents of other deliverables.

[0021] Next, we will use a specific project as an example to explain the evolution of deliverables created in downstream development steps (processes) based on business requirements acquired in upstream development steps (processes) as they proceed in the order of planning, requirements definition, design, development, testing, and release. We will assume that this specific project proceeds under the following assumptions. Business requirements refer to the content and conditions of the business to be systemized. (Assumptions) A person working in the information systems department of a manufacturing company was assigned as the PM (project manager) for a new supply chain management system implementation project that the company was considering implementing. During the upstream process (requirements definition phase) of this project, the following statements (business requirements) were confirmed from the business department: "With our current system, we cannot monitor the progress of the production line in real time. This often delays preparations for the next process. If we could determine in real time whether a specific part is causing a bottleneck or whether there are other issues, we could respond more quickly." The following shows how the contents of this interview (business requirements) will be expanded into what deliverables in the subsequent process, along with the development steps (processes).

[0022] Process Level 1: Requirement definition process Process Level 2: Requirement definition Based on direct feedback (business requirements) from the manufacturing department, document the need to check the progress of the production line in real time. Typical deliverable: Requirement definition document

[0023] Process Level 2: Requirements definition The business requirement of "making it possible to check the progress of the production line in real time" is converted into specific system requirements, and the functions that the system must fulfill are documented. The system requirements expected for the above business requirements are as follows: 1. Real-time progress display of the production line Displays the progress status of each process on the production line in real time. 2. Bottleneck detection function Automatically detects and notifies of bottlenecks that occur during the production process. 3. Rapid problem response support When a problem occurs, a response guide is provided on the system to support rapid problem resolution. Typical deliverables: Requirements definition document

[0024] Level 1: Design process Level 2: Basic design Based on the above requirements, the overall system structure, data model, and interface with external systems are defined. Also, a rough method for meeting the requirements is determined. Typical deliverable: Basic design document

[0025] Process Level 2: Detailed Design Based on the content determined in the basic design, specific program structure, database schema, detailed communication methods with external systems, etc. are designed. Typical deliverables: Detailed design document

[0026] Level 1: Development process Level 2: Programming Create a system program based on the detailed design. At this stage, code is written and the actual system is constructed. Typical deliverables: source code, development documentation

[0027] Process Level 2: Unit and integration testing Test whether the developed functions work correctly individually, and then conduct integration testing between related functions. This confirms whether the individual components work together correctly. Typical deliverables: test specifications, test reports

[0028] Process Level 2: Comprehensive Acceptance Testing (UAT) Testing to verify that the entire system meets requirements and functions properly in the end user's environment. This includes end-user acceptance testing. In addition, the following are examples of UAT (comprehensive acceptance testing) items for each system requirement organized in the requirements definition phase. 1. Real-time production line progress display - Confirm that production line progress is accurately displayed in real time - Confirm that the progress status of each process (completed, in progress, on hold, etc.) is accurately reflected - Confirm that the system display is updated without delay - Confirm that multiple production lines can be monitored simultaneously 2. Bottleneck detection function - Confirm that notifications are sent immediately when a bottleneck occurs - Confirm that the location of the bottleneck (in which process it occurs) can be accurately identified - Confirm that detailed information about the bottleneck that has occurred (cause, scope of impact, etc.) is provided - Confirm that the system returns to normal after the bottleneck is resolved 3. Rapid problem response support - Confirm that the system automatically provides specific response guides when a problem occurs - Confirm that the response guides contain specific content that is actually useful for solving the problem - Confirm that the problem is quickly resolved as a result of implementing the solution provided by the system - Confirm that end users can easily access and understand the problem response guides Typical deliverables: UAT test plan, UAT test report

[0029] Process Level 1: Release Process Process Level 2: Implementation Deploy (deploy, deploy) the system into the actual operating environment and carry out the necessary data migration and user training. Typical deliverables: deployment plan, training materials

[0030] Process Level 2: Operation and Maintenance After the system is operational, any problems that arise will be corrected, and the system will be updated and improved as needed. Typical deliverables: Operation manual, maintenance log, improvement proposal

[0031] 1, an information processing device 1 is communicably connected to a user terminal 2, a management information storage DB 3, and an LLM (Large Language Model) 4. The user terminal 2 is a terminal used by a user.

[0032] The information processing device 1 provides an input screen to the user terminal 2. The input screen is a screen for inputting input items necessary to create a new deliverable, and is displayed on the user terminal 2. The input screen may include an input section for specifying a previous deliverable created in the past in waterfall-type system development. The input screen may also include an input section for inputting additional input items, screen specifications, screen transition diagrams, etc.

[0033] The information processing device 1 acquires the designation of the past work product input on the input screen. The designation of the past work product can be performed, for example, by designating a path to a destination where the file of the past work product is saved, designating a file name, or sending the file itself. The information processing device 1 may also acquire additional input items, screen specifications, screen transition diagrams, etc. input on the input screen. Details of the input screen will be described later.

[0034] The management information storage DB3 stores management information. Management information is information managed by the user, and is, for example, past works created by the user in the past. The management information may include a format for creating a work. The management information may also be information managed by an administrator, as described below. The information processing device 1 instructs the management information storage DB3 to search for rights information based on the specification of the past work acquired from the user terminal 3. The information processing device 1 acquires management information in accordance with the search instruction.

[0035] The management information storage DB3 can use, for example, a Retrieval Augmented Generation (RAG) mechanism. RAG is a mechanism for answering questions about content not included in the LLM's learning data, such as internal company information. RAG is a technology that improves the accuracy of answers by retrieving learning data required for the answer from an internal company database or the like and including it in a prompt to the LLM. A prompt is an instruction or question entered by a user of the LLM4 in a dialogue with the LLM4 or in an interactive system such as a command line interface (CLI).

[0036] The information processing device 1 analyzes the format of the past deliverables contained in the acquired management information. Format analysis refers to the analysis of the description items contained in the past deliverables and the expression format of the description items. As described above, in waterfall-type system development, deliverables in downstream development steps are created based on the description contents written in past deliverables created in upstream development steps. Therefore, to create a new deliverable, the description items of past deliverables created in upstream development steps and format information that serves as a model for creating a new deliverable are required. By analyzing the format of the past deliverables contained in the acquired management information, the information processing device 1 is able to obtain the information necessary to create a new deliverable.

[0037] The information processing device 1 generates a prompt for automatically generating a new deliverable based on the format of the acquired past deliverable and the input items entered on the input screen, and provides the prompt to the LLM 4 .

[0038] The information processing device 1 acquires the new deliverable generated in the LLM 4 in response to the prompt provided to the LLM 4. The information processing device 1 provides the new deliverable acquired from the LLM 4 to the user terminal 2.

[0039] LLM4 is a type of generative artificial intelligence (AI) specialized in generating text, and uses large amounts of text data as training data. The training data learned by LLM4 is obtained from various information sources, such as information on the web, and may contain inaccurate information. Answers generated based on inaccurate information may be inaccurate. LLM4 may also generate inaccurate answers by inferring unknown information that has not been learned from the training data. Meanwhile, management information is stored in management information storage DB3. Management information refers to information managed by an administrator. The administrator can organize, add, modify, or delete the management information. The administrator may be, for example, the administrator of the information processing device 1 or the administrator of management information storage DB3. The administrator of management information storage DB3 may also be a user. The content of the management information is managed by the administrator. The content of the information may be, for example, the accuracy, freshness, detail, or usefulness of the information. Only management information whose contents are managed can be stored in the management information storage DB 3. By generating a prompt including management information whose contents are managed and providing it to the LLM 4, it becomes possible for the LLM 4 to generate accurate deliverables that are useful for system development.

[0040] In the present embodiment, the information processing device 1 generates a new deliverable for the LLM 4, but the information processing device 1 may provide a prompt to a configuration other than the LLM 4. For example, the information processing device 1 may provide a prompt to an internal processing unit (not shown) to acquire a new deliverable. Furthermore, the information processing device 1 may provide a prompt to multiple external devices (not shown) to acquire new deliverables from each of the external devices.

[0041] Next, the configuration of the information processing device 1 will be described using FIG. 3 . FIG. 3 is a diagram showing an example of the configuration of the information processing device 1 according to an embodiment. In FIG. 3 , the information processing device 1 is communicably connected to a user terminal 2, a management information storage DB 3, and an LLM 4 via a network 9. The information processing device 1 has the following functional units: an input screen providing unit 11, a past deliverable acquisition unit 12, a past deliverable analysis unit 13, a prompt generating unit 14, a prompt providing unit 15, a new deliverable acquisition unit 16, a new deliverable provision unit 17, a defect determination unit 18, a consistency determination unit 19, a description correction unit 20, a relevance recording unit 21, and a change reflection unit 22. The above-mentioned functional units of the information processing device 1 according to this embodiment will be described as functional modules implemented by an information processing program (software) according to this embodiment.

[0042] The input screen providing unit 11 provides an input screen for inputting the input items required to create a new deliverable. The input screen will be described in detail later.

[0043] The past deliverable acquisition unit 12 acquires past deliverables created in the past in system development including multiple development steps (sometimes referred to as "development processes"). The past deliverables can be acquired from the management information storage DB 3 or the user terminal 2. The past deliverable acquisition unit 12 acquires past deliverables created in any development step in waterfall-type system development, and the input screen providing unit 11 may provide, based on the past deliverables acquired by the past deliverable acquisition unit 12, an input screen for creating a new deliverable in a development step downstream from the development step in which the acquired past deliverables were created. In other words, the past deliverable acquisition unit 12 can acquire past deliverables created in development steps upstream from the new deliverable. The input screen for creating a new deliverable is a screen for inputting information necessary for creating a new deliverable.

[0044] The past work product analysis unit 13 analyzes the format of the past work product acquired by the past work product acquisition unit 12. The past work product analyzed by the past work product analysis unit 13 may be a work product created in any development step. By analyzing the format of the past work product, it becomes possible to apply the format of the past work product to a new work product.

[0045] The prompt generation unit 14 generates a prompt for automatically generating a new deliverable based on the format of the past deliverable analyzed by the past deliverable analysis unit 13 and the input items entered into the input screen provided by the input screen providing unit 11. The prompt may include image data in addition to text data.

[0046] The prompt providing unit 15 provides the prompt generated by the prompt generating unit 14 to the LLM 4. The new deliverable acquiring unit 16 acquires, from the LLM 4, a new deliverable generated in response to the prompt provided by the prompt providing unit 15. In addition, the new deliverable providing unit 17 provides the new deliverable acquired by the new deliverable acquiring unit 16 to the user terminal 2.

[0047] The defect determination unit 18 determines defects in the input items entered on the input screen provided by the input screen providing unit 11 based on the new deliverable acquired by the new deliverable acquisition unit 16. Defects in the input items include missing input items or incorrect input. If there is a defect in the input items included in the input prompt, the LLM 4 may respond with an error message indicating that the new deliverable cannot be generated. The defect determination unit 18 can prompt the user to correct the input items by determining whether an error message is included in the acquired new deliverable.

[0048] The consistency determination unit 19 determines consistency between multiple entries in the new deliverable acquired by the new deliverable acquisition unit 16. The entry correction unit 20 corrects the multiple entries in the new deliverable based on the consistency determined by the consistency determination unit 19. Consistency between multiple entries in the new deliverable refers to, for example, whether there is an inconsistency between the multiple entries, or whether there is variation in expression (inconsistency in wording), etc.

[0049] The relevance recording unit 21 records the relevance between new deliverables generated in each of the multiple development steps acquired by the new deliverable acquisition unit 16. The change reflecting unit 22 reflects changes to a new deliverable generated in one development step in new deliverables generated in other development steps based on the relevance recorded in the relevance recording unit 21. The change reflecting unit 22 may also be configured to identify new deliverables generated in other development steps that are affected by changes to a new deliverable generated in one development step based on the relevance recorded in the relevance recording unit 21, and reflect the changes in the new deliverables generated in the other development steps. In waterfall system development, as described above, the descriptions of deliverables generated in upstream development steps are reflected in deliverables generated in downstream development steps. Therefore, when multiple deliverables are generated in multiple development steps, it may be unclear which deliverables (downstream steps) were generated based on which deliverables (upstream steps). By recording the relevance between new deliverables, the relevance recording unit 21 makes it possible to clarify the relevance between deliverables generated in waterfall system development. Based on the recorded associations, the change reflecting unit 22 can reflect changes to a new deliverable created in one development step in a new deliverable created in another development step.

[0050] Note that the above-described functional units of the information processing device 1 are merely examples of the functional units of the information processing device 1 and do not limit the functions of the information processing device 1. For example, the information processing device 1 does not need to have all of the above-described functional units, and may have only some of the functional units. Furthermore, the information processing device 1 may have functions other than those described above. For example, the information processing device 1 may have an input function for inputting information, or an output function for notifying the operating status of the device using an LED lamp or the like.

[0051] As described above, the functional units of the information processing device 1 are implemented by software. However, at least one of the functional units may be implemented by hardware.

[0052] Furthermore, any of the above functional units may be implemented by dividing one functional unit into multiple functional units. Furthermore, any two or more of the above functional units may be implemented by integrating them into one functional unit. Furthermore, the above description expresses the functions of the information processing device 1 using functional blocks, and does not indicate, for example, that each functional unit is configured by a separate program file or the like.

[0053] Furthermore, the information processing device 1 may be a device realized by a single housing, or a system realized by multiple devices connected via a network, etc. For example, the information processing device 1 may realize some or all of its functions by a virtual device, such as a cloud service provided by a cloud computing system. That is, the information processing device 1 may realize at least one or more of the above-mentioned functional units in another device. Furthermore, the information processing device 1 may be a general-purpose computer, such as a desktop PC, or a dedicated device with limited functions.

[0054] Next, the input screen provided by the input screen providing unit 11 will be described with reference to Fig. 4 to Fig. 11. Fig. 4 to Fig. 11 are diagrams showing first to eighth examples of input screens related to automatic generation and update of deliverables provided by the information processing device in the embodiment.

[0055] 4, the input screen 1000 has a task bar on the left side of the screen. The task bar has a "Create Deliverable" button, and the screen transitions described below are displayed when the "Create Deliverable" button is operated.

[0056] The "Draft Creation History" button allows you to select a draft (template) for creating a deliverable created in the past. By selecting a draft, you can reuse a draft created in the past.

[0057] The "Deliverable Format" button displays a pull-down menu that allows the user to select the format of the deliverable created in the past. For example, the "Deliverable Format" button allows the user to select a file format for a specific application.

[0058] In addition, when editing an image (correcting, changing, deleting, etc.) on the input screen 1100 of FIG. 5 , "User Response" is selected by default. When an image is used as an input item, the generated deliverable is often an image, and it is often difficult to determine consistency with other deliverables. For this reason, selecting "User Response" by default can prompt the user to determine consistency. Furthermore, when editing an image (correcting, changing, deleting, etc.) on the input screen 1100, the user may be allowed to select not only "User Response" as the response policy, but also "AI Response," in which AI (LLM4) responds on behalf of the user.

[0059] The "additional response content" on the input screen 1100 is reflected in the prompt generated by the prompt generation unit 14. In other words, it is possible to instruct the LLM 4 to generate a new deliverable using the past deliverable edited on the input screen 1100. By operating the "Execute" button (see FIG. 5), a prompt is generated and provided to the LLM 4. The content of the generated new deliverable is highlighted in the correction location 1002 in FIG. 4.

[0060] FIG. 6 is a diagram illustrating the operation of the AI ​​review of a deliverable on the input screen 1200. When the "Create Draft" button is operated in FIG. 4, an "AI Review" tab and an "Edit" tab are displayed. The "AI Review" tab allows input for a new deliverable to be reviewed by AI (LLM4). The "Edit" tab also allows input for a user to review the new deliverable. The download button is a button for displaying the download history, a button for obtaining files of past deliverables that have already been downloaded, and a button for obtaining files of a deliverable currently being created (in the process of being created).

[0061] The review perspective button is a button for registering the review perspective that you want AI (Artificial Intelligence) to review. A variety of requests can be included in the prompt generation, but inputting the requests each time a prompt is generated is time-consuming. By registering the review perspective that you want AI to review in advance, it is possible to reduce the time required to input the requests.

[0062] The input screen 1200 displays the name of the material entered to generate the prompt and a display area for confirming its contents. The input screen illustrates an example in which a requirements definition document is generated as a new deliverable.

[0063] A number is assigned to each review viewpoint registered using the review viewpoint button and displayed on the input screen 1200. The highlighted portion in the figure is the review viewpoint with the number selected by the user in FIG.

[0064] The input screen 1200 displays the review content and total number of reviews created by AI. The review content (review results) is classified as "add," "edit," or "delete." "Add" indicates that there are insufficient input items and additional input items are required. "Edit" indicates that input items need to be edited. "Delete" indicates that input items need to be deleted.

[0065] When the user wishes to modify the review content, the user operates the "Modify" button, which displays an input screen for modifying the review content shown in FIG.

[0066] 7, the review content is displayed on an input screen 1300. A highlighter 1301 is displayed on the portion corresponding to the review content. In FIG. 7, "There is an ambiguous description" is displayed as the review content. The input screen 1300 also displays the content of the review viewpoint registered in the review viewpoint button (shown as "XXX" in the drawing for convenience). The input screen 1300 also displays a proposed revision to the review content (proposed revision to the requirements definition document).

[0067] If the user determines that no corrections are necessary, he or she checks the "Complete without corrections" box. On the other hand, if the user wishes to make corrections, he or she operates the "Edit" button. If it is determined that the corrections need to be made consistent with other deliverables, the input screen shown in Figure 8 is displayed.

[0068] In FIG. 8 , the input screen 1400 displays the review content in an editable manner. The user can make revisions based on the review content to ensure consistency with other deliverables. FIG. 8 illustrates that when a revision to add functional requirements is made to a requirements definition document, the screen image in other deliverables must be changed. Note that the input screen 1400 in FIG. 8 is the same as the input screen 1100 in FIG. 5 in the AI ​​review function, and both screen images fulfill the same role. Furthermore, as with the input screen 1100 in FIG. 5 , when editing an image (e.g., correcting, changing, or deleting), the input screen 1400 allows the user to select not only "User Response" but also "AI Response," in which the AI ​​(LLM4) responds on behalf of the user. The user can modify the screen image based on the review content, making it easier to ensure consistency between past deliverables and new deliverables. The input items are modified by operating the "Execute" button. When a requirements definition document, which is a new deliverable, is modified for the review content, the color of the badge for the review content is changed and a check mark is added to the right of the review content on the input screen 1200 in Figure 6. By operating the "AI Review" button on the input screen 1200, the above-mentioned AI review can be performed at any time based on the review perspective saved with the "Review Perspective" button. An example of the saved review perspective will be described with reference to Figure 9.

[0069] In FIG. 9 , the input screen 1500 displays a list of saved review perspectives. Each review perspective may be directly editable by the user. Alternatively, the review perspectives may be saved in advance in the initial state of the information processing device 1, and the review perspectives may be modified for each user or each company. In the LLM4, the more specific the prompts for the LLM4, the more accurate the responses from the LLM4. By displaying a list of deliverable reviews in an editable manner, the user can easily improve the accuracy of their reviews of the deliverables. The review perspectives are saved by operating the "OK" button.

[0070] Fig. 10 shows an input screen 1600 when the "Edit" tab is displayed. In Fig. 10, the input screen 1600 allows the user to edit a requirements definition document, which is an example of a deliverable.

[0071] In FIG. 10 , a user selects a sentence to edit from the items in the requirements definition document displayed on the input screen 1600. The selected sentence is highlighted 1601. When the sentence to edit is selected, a comment input field 1602 appears to the right of the selected sentence, allowing the user to enter any comment. Pressing the trash can icon 1603 deletes the comment frame 1610 including the comment input field 1602. After pressing the input button 1604, the AI ​​(LLM4) modifies the deliverable based on the editing request entered by the user in the comment input field 1602 (e.g., "Please correct XXX"). Pressing the consistency check button 1611 pops up an input screen 1700 for ensuring consistency within the deliverable, and then pressing the execute button 1608 reflects the modifications in the deliverable. The deliverable is highlighted wherever the modifications are reflected. The content added and edited by the AI ​​(LLM4) is displayed as an AI edit comment 1609. The additional action policy 1605 for ensuring consistency within the deliverable can be selected from three options: Reflect, User Action, and Ignore, from a pull-down menu (not shown). When the additional action content is a screen image change 1606, the action policy is automatically selected as User Action. However, this is not limited to this, and the user may be allowed to select "AI Action," in which the AI ​​(LLM4) takes action on behalf of the user. The additional action content 1607 is input to the AI ​​(LLM4) as a prompt. The additional action content 1607 can be edited by the user. When it is necessary to check the consistency within or between deliverables, an input screen 1700 for ensuring consistency shown in FIG. 11 is displayed by operating the consistency check button 1611. In FIG. 11, the input screen 1700 has the same function as the input screen 1400 described in FIG. 8. In other words, when the "Edit" tab is selected, the user can directly edit the deliverable without performing an AI review. In this case, the consistency between the previous deliverables and the new deliverables is checked, and the new deliverables are displayed in an editable format, allowing the user to easily create new deliverables.

[0072] Next, the hardware configuration of the information processing device 1 will be described with reference to Fig. 12. Fig. 12 is a block diagram showing an example of the hardware configuration of the information processing device according to the embodiment.

[0073] The information processing device 1 includes a CPU (Central Processing Unit) 101 , a RAM (Random Access Memory) 102 , a ROM (Read Only Memory) 103 , an I / O device 104 , and a communication I / F (Interface) 105 .

[0074] The CPU 101 controls the information processing device 1 by executing an information processing program stored in the RAM 102 or the ROM 103. The information processing program is obtained, for example, from a recording medium on which the program is recorded or from a program distribution server via a network, and is installed in the ROM 103, and is read out and executed by the CPU 101.

[0075] The I / O device 104 has an operation input function and a display function (operation display function). The I / O device 104 is, for example, a touch panel. The touch panel allows the user of the information processing device 1 to input operations using a fingertip, a touch pen, or the like. In this embodiment, the I / O device 104 is described as using a touch panel with an operation display function. However, the I / O device 104 may also have a separate display device with a display function and an operation input device with an operation input function. In this case, the display screen of the touch panel can be implemented as the display screen of the display device, and operation of the touch panel can be implemented as operation of the operation input device. The I / O device 104 may also be implemented in various forms, such as a head-mounted, eyeglass-type, or wristwatch-type display.

[0076] The communication I / F 105 is an I / F for communication. The communication I / F 105 executes short-range wireless communication such as wireless LAN, wired LAN, infrared, etc. Although the figure shows only the communication I / F 105 as the communication I / F, the information processing device 1 may have communication I / Fs for multiple communication methods.

[0077] Next, the centralized deliverable management function of the information processing device 1 will be described with reference to FIG. 13 . FIG. 13 is a diagram showing an example of a management screen 1800 for centralized deliverable management provided by the information processing device 1 to the user terminal 2. The information processing device 1 (input screen providing unit 11) provides the management screen 1800, which displays deliverables created in multiple development steps included in waterfall-type system development for each development step. That is, the information processing device 1 (input screen providing unit 11) provides the management screen 1800 to the user terminal 2, and displays a list of deliverables created in multiple development steps (processes) included in waterfall-type system development on the management screen 1800. The management screen 1800 displays a waterfall-type development step 1801, created deliverables 1802, uncreated deliverables 1803, and an input direction 1804 to the next deliverable. The input direction 1804 to the next deliverable indicates the direction of connection between the deliverables. For example, the input direction 1804 to the next deliverable shown in FIG. 13 indicates that the requirements definition document is created based on the project plan, the requirements definition document is created based on the requirements definition document, the detailed design document and user manual are created based on the requirements definition document, the development code is created based on the detailed design document, and the test specification document is created based on the detailed design document and the development code. In other words, the input direction 1804 to the next deliverable indicates the destination (flow direction) of information contained in a deliverable to be reused in the next deliverable. Note that information regarding the connections between deliverables is pre-registered in the information processing device 1 in its initial state, and when the information processing device 1 is first operated, the generated deliverables are associated based on this information. During operation of the information processing device 1, the user changes the existing registered information regarding the connections between deliverables according to the user's preferences or needs. Note that the deliverables shown in FIG. 13 are merely examples and are not limited thereto; various other deliverables can be applied.

[0078] Next, referring to FIG. 14 , a case where the information processing device 1 creates program code as a deliverable will be described. FIG. 14 is a diagram for explaining a case where the information processing device 1 creates program code as a deliverable. The information processing device 1 (new deliverable acquisition unit 16) acquires, as a new deliverable, at least one of materials and programming code generated in response to a prompt generated by the prompt generation unit 14 from a large-scale language model. FIG. 14( a) shows an example of a screen transition diagram 1901 loaded into the AI ​​(LLM4) using multimodal technology. Multimodal technology refers to enabling the AI ​​(LLM4) to understand a combination of two or more different types of data, such as text, audio, images, and video. FIG. 14( b) shows an example of a screen detail diagram 1902 loaded into the AI ​​(LLM4) using multimodal technology.

[0079] The user checks the screen transition diagram 1901 shown in FIG. 14(a) and the screen detail diagram 1902 shown in FIG. 14(b) and instructs the AI ​​(LLM4) to read text, audio, or additional materials as necessary and make corrections. The user then presses the Create Code button 1903, causing the prompt generation unit 14 to generate a prompt for the AI ​​(LLM4) to proceed with coding. The prompt generated by the prompt generation unit 14 is provided to the AI ​​(LLM4) by the prompt providing unit 15, and the AI ​​(LLM4) generates programming code based on the provided prompt. The generated program code is displayed in the new code display area 1905 shown in FIG. 14(c). After the program code is generated, the user can instruct code consistency between this function and other functions by pressing the Consistency Check button 1906. The information processing device 1 may create deliverables for unit testing, integration testing, comprehensive testing, and user acceptance testing (UAT), which are development steps (processes) downstream from programming, based on the generated programming code. Furthermore, based on all the deliverables created in unit testing, integration testing, comprehensive testing, and acceptance testing (UAT), AI may test the program code generated as a deliverable in the test environment provided by the information processing device 1, and may even provide a research report on the results of the test to the user.

[0080] Next, the operation of the information processing device 1 will be described with reference to Fig. 15 to Fig. 18. Fig. 15 to Fig. 18 are flowcharts showing an example of the operation of the information processing device 1 in the embodiment. Note that, although the operation in the following flowcharts will be described as being performed by the information processing device 1, the operation may also be performed by each functional unit of the information processing device 1 described in Fig. 3.

[0081] 15, the information processing device 1 provides an input screen (step S11). After executing the process of step S11, the information processing device 1 determines whether input to the input screen is complete (step S12). If it determines that input to the input screen is not complete (step S12: NO), the information processing device 1 repeats the process of step S12 and waits for input to the input screen to be completed.

[0082] On the other hand, if it is determined that input to the input screen is complete (step S12: YES), the information processing device 1 acquires the past deliverable specified on the input screen (step S13). After executing the process of step S13, the information processing device 1 analyzes the format of the acquired past deliverable (step S14). After executing the process of step S14, the information processing device 1 generates a prompt (step S15) and provides the prompt to the LLM 4 (step S16). After executing the process of step S16, the information processing device 1 determines whether a new deliverable has been acquired from the LLM 4 (step S17). If it is determined that a new deliverable has not been acquired (step S17: NO), the information processing device 1 repeats the process of step S17 and waits for the acquisition of a new deliverable.

[0083] On the other hand, if it is determined that a new deliverable has been acquired (step S17: YES), the information processing device 1 provides the acquired new deliverable to the user terminal 2 (step S18). The provision of the new deliverable can be executed, for example, by displaying it in an editable manner on the input screen described above.

[0084] FIG. 16 illustrates the determination of the presence or absence of defects when the new deliverable described in FIG. 15 is acquired. In FIG. 16, the information processing device 1 acquires a new deliverable (step S21) and determines whether to determine whether there are any defects in the acquired new deliverable (step S22). The determination of defects is performed, for example, based on an explicit instruction from the user. If it is determined not to determine whether there are any defects (step S22: NO), the information processing device 1 repeats the processing of step S22 and waits for an instruction to determine whether there are any defects. On the other hand, if it is determined to determine whether there are any defects (step S22: YES), the information processing device 1 provides an input screen on which the determination result and the defective location can be edited (step S23).

[0085] 17 illustrates the determination of whether or not there is an inconsistency when the new deliverable described in FIG. 15 is acquired. In FIG. 17, the information processing device 1 acquires a new deliverable (step S31) and determines whether or not to determine whether or not there is an inconsistency in the acquired new deliverable (step S32). The determination of an inconsistency is performed, for example, based on an explicit instruction from the user. If it is determined not to determine an inconsistency (step S32: NO), the information processing device 1 repeats the processing of step S32 and waits for an instruction to determine an inconsistency. On the other hand, if it is determined to determine an inconsistency (step S32: YES), the information processing device 1 provides an input screen on which the determination result and the defective portion can be edited (step S33).

[0086] FIG. 18 illustrates the modification of a previous work product related to a new work product described in FIG. 15 when the new work product is acquired. In FIG. 18, the information processing device 1 acquires the new work product (step S41) and records the relationship between the new work product and the previous work product (step S42). For example, if the new work product is generated based on a previous work product created in an upstream development step, the relationship between the new work product and the previous work product that served as the basis for the new work product is recorded. The relationship record may include, for example, recording which descriptions of the work products are related. After executing the process of step S42, the information processing device 1 determines whether an instruction to reflect changes to the descriptions of the new work product in the previous work product has been issued (step S43). If it is determined that there is no instruction to reflect the changes (step S43: NO), the information processing device 1 repeats the process of step S43 and waits for an instruction to reflect the changes.

[0087] On the other hand, if it is determined that an instruction to reflect the changes has been given (step S43: YES), the information processing device 1 reflects the changes in the related past deliverables (step S44). Note that the operations shown in the above flowchart are merely examples of the operations of the information processing device 1 and do not limit the order of processing. For example, specific processing may be performed in an order or at a timing different from the processing order in the flowchart.

[0088] (Regarding the Information Processing Device 10 According to the Second Embodiment) With reference to FIGS. 19 to 32 , the information processing device 10 according to the second embodiment of the present invention will be described in detail. In these drawings, identical or corresponding parts are designated by the same reference numerals, and redundant description will be omitted. Furthermore, in all drawings, only components for explaining the present invention are illustrated, and other components may be omitted. Furthermore, the present invention is not limited to the second embodiment described below. Contents included in multiple drawings may be omitted to avoid duplication. The information processing device 10 according to the second embodiment differs in that the input screens provided during automatic product generation and update are the ninth to twenty-second examples shown in FIGS. 19 to 32 , instead of the first to eighth examples shown in FIGS. 4 to 11 of the information processing device 1 according to the above-described embodiment. Hereinafter, the information processing device 10 according to the second embodiment will be described only with respect to parts that differ from the information processing device 1 according to the above-described embodiment. The same parts as those of the information processing device 1 according to the above-described embodiment will be designated by the same reference numerals, and their description will be omitted. The description of the information processing device 1 according to the above-described embodiment will be used. 19 to 32 will be used to describe ninth to twenty-second examples of input screens provided by the input screen providing unit 11. Figures 19 to 32 are diagrams showing ninth to twenty-second examples of input screens related to automatic generation and update of deliverables provided by the information processing device 10 in the second embodiment.

[0089] There are two main use cases for the automatic generation and updating of deliverables by the information processing device 10. In the first use case, the information processing device 10 generates a draft (initial draft) of a deliverable (AI draft), edits the generated draft of the deliverable (AI editing), and reviews the quality of the edited deliverable (AI quality review). In the second use case, the information processing device 10 reviews a deliverable created and edited by a user (AI quality review). Therefore, in the second use case, the information processing device 10 does not perform AI drafting or AI editing as in the first use case. Note that the AI ​​draft refers to the information processing device 10 generating a draft of a deliverable (initial draft), the AI ​​editing refers to the information processing device 10 editing the draft of a deliverable, and the AI ​​quality review refers to the information processing device 10 reviewing the quality of the deliverable. The review may involve checking and evaluating the quality of the deliverable and identifying areas for improvement.

[0090] In the first use case, the AI ​​draft mainly involves (a) pre-registration of a template specified by the user of the information processing device 10, and (b) analysis of the input of the information processing device 10, and generation of a deliverable according to the template. Next, in the first use case, the AI ​​editing mainly involves (c) presentation of the parts of the deliverable created by the information processing device 10 and the reasons for this (version management is also performed), and (d) automatic correction of the deliverable of the information processing device 10 through dialogue between the user and the information processing device 10. Next, in the first use case, the AI ​​quality review mainly involves (e) reviewing the entire deliverable from a perspective registered by the information processing device 10 and pointing out any omissions in consideration of requirements. In the second use case, the AI ​​quality review mainly involves (f) uploading materials to the information processing device 10 by the user, reviewing the entire deliverable from a perspective registered by the information processing device 10, and pointing out any omissions in consideration of requirements.

[0091] (First Use Case) (a) Pre-registration of a Template Designated by a User of the Information Processing Device 10 Pre-registration of a template designated by a user of the information processing device 10 will be described with reference to FIG. 19 . FIG. 19 is a diagram illustrating a ninth example of an input screen related to automatic product generation and updating provided by the information processing device (input screen providing unit 11) 10. The user terminal 2 displays an input screen 2000 provided by the input screen providing unit 11. In FIG. 19 , the input screen 2000 has a task bar shown on the left side of the screen. The task bar includes a "Settings" button, and the input screen 2000 is displayed when the "Settings" button is operated. The input screen 2000 is used to pre-register a template designated by the user. A template is a model of a product to be generated and is a predetermined product format. The AI ​​(LLM 4) generates a product in accordance with the format determined by the template. The user enters basic information of the template to be pre-registered in a basic information input area 2001 of the input screen 2000. For example, the user inputs the format name of the template to be pre-registered in the format name field and the file format name of the template to be pre-registered in the file format field in the basic information input area 2001. The format name refers to the name of the specific format or configuration of the template to be pre-registered, and the file format name refers to the name of the recording format used when saving data to a file, such as PowerPoint (registered trademark) format.

[0092] Next, the user enters detailed information about the template to be pre-registered in the detailed information input area 2002 of the input screen 2000. The detailed information includes, for example, a table of contents for the generated deliverable, a specific description method (content image) for each table of contents, and content creation guidelines (format, output volume, points to note, etc.). The content creation guidelines may be prompts or backprompts given to the AI ​​(LLM4). A prompt is an instruction or input given to the AI ​​(LLM4) to generate a specific response. A backprompt is an internal setting that defines instructions and rules for the AI ​​(LLM4) to operate, and plays a role in determining the AI ​​(LLM4)'s behavior, output style, prohibited actions, etc. The user can set prompts or backprompts for each table of contents for the generated deliverable, thereby obtaining a deliverable that faithfully reproduces the user's intentions or requests. Furthermore, by pre-registering a template designated by the user, the user can avoid the need to provide a prompt or back prompt to the AI ​​(LLM4) each time a deliverable is generated. The user can pre-register a template designated by the user by pressing the update button 2003 on the input screen 2000. After this, the user can select and specify the pre-registered template when generating a draft of a deliverable, thereby obtaining a deliverable generated in accordance with the format determined by the template.

[0093] (b) Analysis of input by information processing device 10 and generation of deliverables according to templates Analysis of input by information processing device 10 and generation of deliverables according to templates will be described with reference to Figures 20 to 26. Figures 20 to 26 are diagrams showing examples of input screens Nos. 10 to 16 relating to automatic generation and updating of deliverables provided by information processing device (input screen providing unit 11) 10. Note that input refers to information used as a basis when AI (LLM4) generates a deliverable, and may also be called input information, input data, or input materials.

[0094] First, referring to FIG. 20 , the generation of a draft table of contents for a deliverable will be described. The user terminal 2 displays an input screen 2010 provided by the input screen providing unit 11. In FIG. 20 , the input screen 2010 has a task bar on the left side of the screen. The task bar includes a "Generate Deliverable" button, and the input screen 2010 is displayed when the "Generate Deliverable" button is pressed. The input screen 2010 is used by the user to generate a draft table of contents for a deliverable. The user selects the type of deliverable to be generated from among the deliverable types pre-registered in a pull-down menu in a deliverable type selection field 2011. The deliverable type may be a deliverable created in each development step (process) of waterfall-type system development, such as a proposal, requirements definition, basic design, detailed design, programming, unit test plan, integrated test design, acceptance test plan, user manual, or migration plan. Next, the user enters the name of the deliverable to be generated in free text in a deliverable document name input field 2012. The user can arbitrarily set a deliverable name. Next, the user selects a template that will be the format of the deliverable they want to generate from the deliverable templates pre-registered in the pull-down menu of the format selection field 2013. The deliverable template may include a template pre-registered by the user in (a) above, a template registered in the initial state (default) of the information processing device 10, and the like.

[0095] Next, the user selects information that the AI ​​(LLM4) will use as a reference when generating a deliverable from among the inputs pre-registered in the pull-down menu of the input selection field 2014. The inputs displayed in the pull-down menu of the input selection field 2014 may be files in a folder in a terminal specified by the user. Next, the user selects information that the AI ​​(LLM4) will use as a reference when generating a deliverable from among the reference project deliverables pre-registered in the pull-down menu of the reference deliverable selection field 2015. The reference project deliverable may be a deliverable from a past project similar to the project of the deliverable to be generated. The user can reuse a deliverable generated in a past project when the AI ​​(LLM4) generates a deliverable. The user can have the AI ​​(LLM4) generate a draft of the product by pressing the generate button 2016 and display it in the window on the lower left side of the input screen 2010.

[0096] Next, with reference to Figures 21 to 24, we will explain how to generate a draft of a deliverable by specifying a table of contents. The user terminal 2 displays an input screen 2020 provided by the input screen providing unit 11. The input screen 2020 is used when the user generates a draft of a deliverable by specifying a table of contents. The user sets the deliverable material name and selects the deliverable type, format, input, and reference deliverables in the same manner as in the case of generating a draft of the entire table of contents of the deliverable described above. The user selects the table of contents of the deliverable that he or she wants to generate from the table of contents of the deliverables pre-registered in the pull-down menu of the table of contents selection field 2021. The table of contents of the deliverable pre-registered in the pull-down menu is set by selecting the format of the deliverable to be generated.

[0097] When the user checks the checkbox 2022, the AI ​​(LLM4) designs the topics to compensate for the lack of input before generating a draft of the deliverable. Topics refer to the specific content of questions to be asked to the user to compensate for the lack of input, and topic design refers to the AI ​​(LLM4) generating the specific content of questions to be asked to the user. When the user presses the draft generation button 2023, the AI ​​(LLM4) analyzes the selected input data and, referring to the creation guide for the selected table of contents content, designs the topics for the initial table of contents. At the same time as designing the topics, the AI ​​(LLM4) may also design proposed answers corresponding to the topics and present the proposed answers to the user along with the topics. The AI ​​(LLM4) may design three to five topics for one table of contents, and when designing proposed answers, it may design the same number of proposed answers as the number of topics to be designed. The answer plan designed by AI (LLM4) may be editable by the user, and may further include references 2041c (see FIG. 22(b)) cited by the answer plan. The issues, answer plan, and references designed by AI (LLM4) may be displayed in a window on the lower left side of the input screen 2020. Furthermore, the window on the lower right side of the input screen 2020 may display the content that the user should take note of when answering the issues as an "issue answering process."

[0098] The procedure for a user to respond to the issues will be described with reference to FIG. 22 . The user terminal 2 displays input screens 2030-2060 provided by the input screen provider 11, transitioning according to the user's progress in responding to the issues. When the user selects the first table of contents (Chapter 1, 1. Project Background) in the table of contents response status 2031 (see FIG. 22(a)) displayed on the input screen 2030, the issues and proposed answers designed by the AI ​​(LLM4) are displayed in the proposed issue response display area 2041 on the input screen 2040 (see FIG. 22(b)). The table of contents response status 2031 indicates the progress of the user's responses to the issues. Each time the user selects an table of contents, the AI ​​(LLM4) generates a prompt based on the issues in the table of contents prior to the selected table of contents and the user's responses, and designs the issues and proposed answers for the table of contents selected by the user. This is because important information may have been added in tables of contents prior to the selected table of contents. If the issues and proposed answers of the selected table of contents are not designed taking this into account, there is a risk of asking the same question or losing consistency between the tables of contents. The issue and proposed answer display area 2041 displays, for example, the first issue 2041a, the proposed answer to the first issue 2041b, and cited materials 2041c, the second issue 2042a and the proposed answer to the second issue 2042b, and the third issue 2043a and the proposed answer to the third issue 2043b. The cited materials 2041c are cited materials for the proposed answer to the first issue 2041b. The user may answer the issue by modifying the proposed answer displayed in the issue and proposed answer display area 2041 as needed, or may delete the proposed answer and enter a new answer. The user can answer by pressing the answer button 2044 displayed at the bottom of the input screen 2040. In response to the user pressing the answer button 2044, the table of contents answer status 2031 advances the progress of the user's answer by one step (see input screen 2050 in FIG. 22(c)).

[0099] Next, when the user selects the second table of contents (Chapter 1, 2. Project Objectives) in the table of contents answer status 2031 (see FIG. 22(c)) displayed on the input screen 2050, the issues and proposed answers of the second table of contents designed by the AI ​​(LLM4) are displayed in a display area (not shown) for proposed issues and answers on the input screen 2050. When the user answers the issues of the second table of contents and presses the answer button (not shown), the progress indicated by the issue answer status 2061 moves forward by one step (see FIG. 22(d)).

[0100] Referring to Figure 23, a case where a user responds to issues after a deliverable draft has been generated will be described. The generated deliverable draft is displayed on a display screen 2070 of the user terminal 2, as shown in Figure 23(a). Of the generated deliverable drafts, a table of contents (uncreated table of contents 2071) for which no content has been created is displayed with a "+" button at the beginning, as shown in Figure 23(a). When the user clicks the "+" button, a pop-up is displayed that allows the user to select the table of contents for which the deliverable draft is to be generated (see input screen 2080 in Figure 23(b)).

[0101] When the user selects the table of contents for which they want to generate a deliverable draft from the pop-up on the input screen 2080, if there are any issues designed by the AI ​​(LLM4), the input screen 2090 is displayed on the user terminal 2, and if there are no issues designed by the AI ​​(LLM4), the input screen 2100 is displayed on the user terminal 2. The AI ​​(LLM4) creates the issues and proposed answers for the table of contents selected by the user based on the content creation guide for the template selected when generating the deliverable draft, the contents of the table of contents that have already been created, the inputs up to now, and the inputs that have been added. The user answers the issues on the input screen 2090 displayed on the user terminal 2, and the AI ​​(LLM4) generates a draft of the selected table of contents in a manner that maintains consistency with the answers. If AI (LLM4) does not design the issues, AI (LLM4) generates a draft of the table of contents of the deliverable selected by the user from the pop-up on input screen 2080, and displays the generated draft 2101 on input screen 2100.

[0102] The consistency determination unit 19 of the information processing device 10 checks the consistency between data for each table of contents of the generated deliverable draft, creates a table 2111, and displays the table 2111 on the input screen 2110. The table 2111 shows the contradictions contained in the contents of each table of contents, and for each contradiction, it includes a classification column 2112, a response / additional information column 2113 to the issue, a contradiction content column 2114, a correction method column 2115, and a correspondence column 2116. The classification column 2112 stores three types of contradiction classification (requirement contradiction, business contradiction, data contradiction) determined for each contradiction by the consistency determination unit 19. The response / additional information column 2113 to the issue that caused the contradiction or additional information is stored. The contradiction content column 2114 stores an explanation of where the response / additional information to the issue specifically causes the contradiction. The column 2115 for modification method stores the AI ​​(LLM4)'s proposal for the content to be modified, taking into account the entire requirements definition document. The column 2116 for response stores the response policy selected by the user. The user selects the response policy from three options: reflect, postpone, or ignore the modification method proposed by AI (LLM4). If the user selects reflect, the modification method proposed by AI (LLM4) is reflected in the deliverable draft. If the user selects postpone, the modification method proposed by AI (LLM4) is stored in the management information storage DB3 as history information. If the user wants to change the modification method proposed by AI (LLM4), they may press the AI ​​consultation button 2117. When the AI ​​consultation button 2117 is pressed, an input screen 2120 is displayed on the user terminal 2. The input screen 2120 provides a chat-style dialogue in which the user can have a natural written conversation with the AI ​​(LLM4) and receive replies, and the user can change the correction method by responding to the points generated by the AI ​​(LLM4).

[0103] Referring to FIG. 25 , input analysis, which a user performs as preparation before generating a deliverable draft, will be described. The user terminal 2 transitions between input screens 2130 to 2160 provided by the input screen providing unit 11 and displays them in accordance with the progress of the user's input analysis. Input analysis refers to an analysis performed by the AI ​​(LLM4) to determine whether the AI ​​(LLM4) has the information necessary to generate a deliverable draft based on the format selected by the user. When the user clicks the i mark 2131 on the input screen 2130, the user terminal 2 transitions the display to the input screen 2140. The input screen 2140 displays a list 2141 containing the input information required for each table of contents of the deliverable draft and specific examples of the information. When the user presses the input analysis button 2142 on the input screen 2140, the AI ​​(LLM4) analyzes whether the existing input is insufficient or insufficient. The AI ​​(LLM4) displays a list 2152 on the input screen 2150, which is the list 2141 plus a column 2151 for missing data. Missing data 2153 is highlighted in list 2152. When the user uploads the missing data 2153 in list 2152 to information processing device 10 and the missing state of information necessary for generating the deliverable draft is resolved, list 2152 is updated to a status of no missing data. When the user presses data registration button 2161 on input screen 2160, the additional data added by the user is registered in input entry field 2132 on input screen 2130.

[0104] (c) Presenting the parts of the deliverable created by AI (LLM4) and the reasons for their creation (version management is also performed). Referring to FIG. 26, the presentation of the parts of the deliverable created by AI (LLM4) and the reasons for their creation will be described. The user terminal 2 transitions between and displays input screens 2170 to 2190 provided by the input screen providing unit 11. As shown in FIG. 26(a), the input screen 2170 displays the deliverable draft generated by AI (LLM4). When the user presses the version management button 2171 on the input screen 2170, the display on the user terminal 2 switches to the input screen 2180 (see FIG. 26(b)). When the user selects "Check update history" 2181 on the input screen 2180 and presses the confirm button 2182, the input screen 2190 is displayed on the user terminal 2 (see FIG. 26(c)). The input screen 2190 includes a generated draft display area 2191 and an update content list display area 2192. The generated draft display area 2191 is an area that displays the deliverable draft generated by AI (LLM4). The update content list display area 2192 displays a list of the updates to the deliverable draft made by AI (LLM4). The generated draft display area 2191 and the update content list display area 2192 distinguish and highlight an AI (LLM4) created part 2193 and an input citation part 2194. The AI ​​(LLM) created part 2193 is the part written by AI (LLM4), and the input citation part 2194 is the part written by AI (LLM4) citing the input.

[0105] With reference to FIG. 27 , version management of deliverables will be described. Version management of deliverables refers to a function for recording the status (state) and change history of a deliverable. The user terminal 2 transitions between input screens 2200 to 2190 provided by the input screen providing unit 11 and displays them. As shown in FIG. 27( a), the input screen 2200 displays a deliverable draft generated by the AI ​​(LLM4). When the user presses the version management button 2201 on the input screen 2200, the display on the user terminal 2 switches to the input screen 2210 (see FIG. 27( b)). When the user selects "Upgrade and register update history" 2211 on the input screen 2210 and presses the confirm button 2212, the input screen 2220 is displayed on the user terminal 2 (see FIG. 27( c)). When the user selects "Upgrade and register update history" 2211 and presses the confirm button 2212, changes from the previous draft of the deliverable are registered as an update history, which can be viewed at any time. The input screen 2220 includes a previous version display area 2230, a latest version display area 2240, an update history display area 2250, and a version bar 2260. The latest version display area 2240 displays the latest version of the deliverable generated in AI (LLM4). In the latest version display area 2240, changes from one version are highlighted. The previous version display area 2230 displays the deliverable one version before the one displayed in the latest version display area 2240. The update history display area 2250 displays a table 2251 with columns for No., update date, updated area, update content, updater, and version (Ver.). In the table 2251, rows other than the latest version are highlighted (gray) 2252, and rows for the latest version are highlighted (white) 2253. Double-clicking a row for the latest version highlights that row (light blue) 2254, and the corresponding changes in the latest version display area 2240 are also highlighted. Version bar 2260 includes buttons for selecting the version of the deliverable to be displayed in latest version display area 2240 from among the latest version and versions up to 10 generations before.By selecting and pressing a button, the user can switch and view the version displayed in latest version display area 2240. Note that the contents of previous version display area 2230 and update history display area 2250 are changed in conjunction with the version displayed in latest version display area 2240.

[0106] (d) Automatic Correction of the Deliverable of the Information Processing Device 10 through Interaction Between the User and the Information Processing Device 10 With reference to FIG. 28 , automatic correction of the deliverable of the information processing device 10 through interaction between the user and the information processing device 10 will be described. The user terminal 2 transitions between and displays input screens 2270 to 2300 provided by the input screen providing unit 11. As shown in FIG. 28( a), the input screen 2270 displays a deliverable draft generated by the AI ​​(LLM4). When the user selects the portion 2271 to be corrected and presses the AI ​​consultation button 2272, the screen transitions to the input screen 2280, where the portion 2271 that the user wants to correct is extracted and displayed. The user presses the correction proposal button 2281 on the input screen 2280 to select the correction proposal mode, inputs the correction request for the portion 2271 to be corrected, and presses the paper airplane button 2282, whereupon the correction content generated by the AI ​​(LLM4) is displayed on the input screen 2290 (see FIG. 28( c)). If the user has no problems with the corrections displayed on the input screen 2290, the user can press the reflect button 2291, and if the user determines that there are problems with the corrections, the user can edit directly. When the user presses the reflect button 2291 on the input screen 2290, the corrections are reflected in the deliverable generated by the AI ​​(LLM4), the deliverable reflecting the corrections is displayed on the input screen 2300, and a chat history 2301 between the user and the AI ​​(LLM4) is registered in the management information storage DB3 (see FIG. 28(d)).

[0107] Referring to FIG. 29 , we will explain how the user improves the deliverable by answering questions from AI (LLM4). The user terminal 2 transitions between input screens 2310 to 2330 provided by the input screen provider 11. The input screen 2310 displays the deliverable generated by AI (LLM4). The user selects the edit guide tab 2311 displayed on the input screen 2310 and presses the guide generation button 2312. This causes AI (LLM4) to analyze missing information for each table of contents of the deliverable, generate questions for the user, and display them in the display window 2313 (see FIG. 29( a)). The user selects the first table of contents in the display window 2313, which causes AI (LLM4) to send an information request 2321 (see input screen 2320 in FIG. 29( c)). The user responds to AI (LLM4)'s information request 2321 via chat-style input 2322. The user can also provide information by attaching documents or photos along with their response. Based on the information provided by the user, the AI ​​(LLM4) generates a revision proposal 2331 for the relevant table of contents. The changes are highlighted using colored text or the like (see FIG. 29(c)). The user can also edit directly. The user can select from among the following options for the revision proposal 2331 generated by the AI ​​(LLM4): Skip Reflection, Ask AI Question Again, and Reflection using the selection button 2332. By selecting the Reflect button, the revision proposal 2331 can be reflected in the deliverable generated by the AI ​​(LLM4).

[0108] Referring to FIG. 30 , we will explain how AI (LLM4) identifies and batch-corrects deliverables by linking minutes and other information. The user terminal 2 transitions between and displays input screens 2340 and 2350 provided by the input screen provider 11. The input screen 2340 displays the deliverables generated by AI (LLM4). Pressing the batch edit button 2341 on the input screen 2340 displays a pop-up 2342 showing the table of contents of the deliverables. When the user checks the table of contents they want to edit and uploads comments or materials related to the table of contents, AI (LLM4) takes the comments or materials into consideration to create a revision proposal. The contents of the table of contents can be confirmed by clicking the ▽ button on the pop-up 2342, and the × button closes the pop-up 2342. The input screen 2350 displays the current version of the deliverable, the revised version of the deliverable reviewed by AI (LLM4) based on the input information, and the reasons for the revisions. Changes in the revised version of the deliverable are highlighted in blue or other colors. The user can directly edit the content of the revised version of the deliverable. The reason for revision displayed on the input screen 2350 includes the reason for revision by the AI ​​(LLM4) and the description of the quoted input information. When the user presses the edit button 2351, the deliverables are revised all at once.

[0109] (e) The information processing device 10 reviews the entire deliverable from a pre-registered perspective and identifies any missing requirements. Referring to FIG. 31 , the information processing device 10 reviews the entire deliverable from a pre-registered perspective and identifies any missing requirements. The user terminal 2 transitions between input screens 2370 to 2380 provided by the input screen provider 11. As shown in FIG. 31 ( a ), the input screen 2370 displays a deliverable draft generated by the AI ​​(LLM4). The user selects the quality review tab 2731 on the input screen 2370 to enter quality review mode. The user can register their own review perspective by pressing the review perspective registration button 2372. For example, the user can register an RFP (Request for Proposal) that they created. The user can press the review button 2373, causing the AI ​​(LLM4) to begin reviewing the deliverable. When the user clicks on review card 2374 containing the review generated by AI (LLM4), the display on user terminal 2 transitions to a details screen 2380. The user can check AI (LLM4)'s revision proposal 2381 displayed on details screen 2380 and edit it directly as necessary, and when the user presses revision button 2382, the revision proposal 2381 generated by AI (LLM4) is reflected in the deliverable. When checkbox 2383 is checked, AI (LLM4)'s revision proposal 2381 is not reflected and the response is completed without any revision.

[0110] (Second Use Case) (f) A user uploads materials to the information processing device 10, reviews the entire deliverable from a viewpoint registered in the information processing device 10, and identifies any missing requirements. Referring to FIG. 32 , a case will be described in which a user uploads materials to the information processing device 10, reviews the entire deliverable from a viewpoint registered in the information processing device 10, and identifies any missing requirements. The user terminal 2 transitions and displays an input screen 2390 provided by the input screen providing unit 11. As shown in FIG. 32 , the input screen 2390 is displayed on the user terminal 2 by selecting the deliverable creation button 2391 on the sidebar. The user clicks the upload button 2392 and selects, for example, a file stored on the user terminal 2 (e.g., a Word® file of an RFP (request for proposal)). The user selects a business requirements definition document from the pull-down menu in the deliverable type input field 2393 and presses the upload material button to start uploading the materials, which is completed in a few minutes. The uploaded file is displayed in a display window 2395. A file (for example, a request for proposal (RFP)) uploaded to the information processing device 10 can be subject to review by the information processing device 10 using the function (e) above.

[0111] The above describes an embodiment of the present invention with reference to the drawings, but the specific configuration is not limited to this embodiment, and various modifications are also included within the scope that does not deviate from the spirit of the present invention.

[0112] REFERENCE SIGNS LIST 1 Information processing device 2 User terminal 2a Display screen 2b Process level 1 2c Created deliverable 2d Uncreated deliverable 2e Input to next deliverable 2f Create button 2g System department name selection button 2h 3 Management information storage DB 4 LLM 9 Network 10 Information processing device (second embodiment) 11 Input screen providing unit 12 Past deliverable acquisition unit 13 Past deliverable analysis diagram 14 Prompt generation unit 15 Prompt providing unit 16 New deliverable acquisition unit 17 New deliverable provision unit 18 Defect determination unit 19 Consistency determination unit 20 Description item correction unit 21 Relevance recording unit 22 Change reflection unit 101 CPU 102 RAM 103 ROM 104 Touch panel 105 Communication I / F 1000 Input screen 1001 Patch 1002 Correction area 1100 Input screen 1200 Input screen 1300 Input screen 1301 Marker display 1400 Input screen 1500 Input screen 1600 Input screen 1601 Marker display 1602 Comment field 1603 Trash can icon 1604 Input button 1605 Response policy selection field 1606 Change image 1607 Additional response content (prompt) 1608 Execute button 1609 Additional editing comment for AI 1610 Comment frame 1611 Consistency check button 1700 Input screen 1800 Management screen 1900 Program code update screen 1901 Screen transition diagram 1902 Screen details diagram 1903 Code creation button 1904 Coding screen 1905 New code display area 1906 Consistency check button 2000 Input screen 2001 Basic information input area 2002 Detailed information input area 2003 Update button 2010 Input screen 2011 Select field for type of deliverable 2012 Input field for deliverable document name 2013 Format selection field 2014 Input selection field 2015 Reference deliverable selection field 2016 Draft generation button 2020 Input screen 2021 Table of contents selection field2022 Check box 2023 Draft generation button 2030 Input screen 2031 Table of contents response status 2040 Input screen 2041 Display area for proposed responses to issues 2041a First issue 2041b Proposed response to first issue 2041c Cited materials 2042a Second issue 2042b Proposed response to second issue 2043a Third issue 2043b Proposed response to third issue 2044 Response button 2050 Input screen 2060 Input screen 2061 Issue response status 2070 Input screen 2071 Uncreated table of contents 2080 Input screen 2090 Input screen 2100 Input screen 2101 Draft 2110 Input screen 2111 Table 2112 Classification column 2113 Response to issue / additional information column 2114 Column of contradiction details 2115 Column of correction methods 2116 Column of correspondence 2117 AI consultation button 2120 Input screen 2130 Input screen 2131 i mark 2132 Input entry field 2140 Input screen 2141 List 2142 Input analysis button 2150 Input screen 2151 Column of missing data 2152 List 2153 Missing data 2160 Input screen 2161 Data registration button 2170 Input screen 2171 Version management button 2180 Input screen 2181 Check update history 2182 Confirm button 2190 Input screen 2191 Generated draft display area 2192 Update content list display area 2193 AI (LLM) created section 2194 Input citation section 2200 Input screen 2201 Version management button 2210 Input screen 2211 Upgrade and register update history 2212 Confirm button 2220 Input screen 2230 Previous version display area 2240 Latest version display area 2250 Update history display area 2251 Table 2252 Highlight (gray) 2253 Highlight (white) 2254 Highlight (light blue) 2260 Version bar 2270 Input screen 2271 Area to be corrected 2272 AI consultation button 2280 Input screen 2281 Submit revision proposal button 2282 Paper airplane button2290 Input screen 2291 Reflect button 2300 Input screen 2301 Chat history 2310 Input screen 2311 Edit guide tab 2312 Guide generation button 2313 Display window 2320 Input screen 2321 Information request 2322 Chat-style input 2323 Changes 2330 Input screen 2331 Changes 2331 Proposed revision 2332 Select button 2340 Input screen 2341 Bulk edit button 2342 Pop-up 2350 Input screen 2351 Revise button 2360 Input screen 2370 Input screen 2371 Quality review tab 2372 Review viewpoint registration button 2373 Review button 2374 Review card 2380 Details screen 2381 Proposed revision 2382 Revise button 2383 Check box 2390 Input screen 2391 Generate deliverable button 2392 Upload button 2393 Deliverable type input field 2394 Upload material button 2395 Display window

Claims

1. An information processing device comprising: a past product acquisition unit that acquires past products created in the past in waterfall-type system development including multiple development steps; a past product analysis unit that analyzes the format of the past products acquired by the past product acquisition unit; an input screen providing unit that provides an input screen for inputting input items necessary to create a new product; a prompt generation unit that generates a prompt for automatically generating a new product based on the format of the past product analyzed by the past product analysis unit and the input items entered on the input screen provided by the input screen providing unit; a prompt providing unit that provides the prompt generated by the prompt generation unit to a large-scale language model; a new product acquisition unit that acquires from the large-scale language model new products generated in response to the prompt provided by the prompt providing unit; and a new product providing unit that provides the new product acquired by the new product acquisition unit.

2. The information processing device described in claim 1, wherein the past product acquisition unit acquires past products created in any development step in waterfall-type system development, and the input screen providing unit provides an input screen for creating a new product in a development step downstream of the development step in which the acquired past product was created, based on the past product acquired by the past product acquisition unit.

3. An information processing device as described in claim 1 or 2, further comprising a defect judgment unit that judges defects in input items entered into the input screen provided by the input screen providing unit based on the new deliverable acquired by the new deliverable acquisition unit, wherein the input screen providing unit provides an input screen for inputting additional input items corresponding to the defect judged by the defect judgment unit, and the prompt generation unit generates a prompt for correcting the defect judged by the defect judgment unit based on the additional input items entered into the input screen.

4. An information processing device as described in claim 1 or 2, further comprising: a consistency determination unit that determines consistency between multiple entries within the new deliverable acquired by the new deliverable acquisition unit; and an entry correction unit that corrects the entries based on the consistency determined by the consistency determination unit.

5. An information processing device as described in claim 1 or 2, further comprising: an association recording unit that records the association between new deliverables generated in each of the multiple development steps acquired by the new deliverable acquisition unit; and a change reflection unit that reflects changes to new deliverables generated in any of the development steps in new deliverables generated in other development steps based on the associations recorded in the association recording unit.

6. An information processing device as described in claim 1 or 2, wherein the input screen providing unit provides an input screen for further acquiring screen specifications for a display screen to be displayed in the system and a screen transition diagram for the display screen, and the prompt generating unit generates a prompt for automatically generating a program for displaying the display screen based on the screen specifications and screen transition diagram acquired by the input screen providing unit.

7. The information processing device described in claim 1 or 2, characterized in that the input screen providing unit provides a management screen that displays the deliverables created in multiple development steps included in waterfall-type system development for each development step, and the deliverables are displayed on the management screen together with the direction of input to the next deliverable.

8. The information processing device according to claim 1, wherein the new product acquisition unit acquires, as a new product, at least one of materials and programming code generated in response to the prompt from the large-scale language model.

9. An information processing method that causes a computer to execute the following steps: a past work product acquisition step that acquires past work products created in the past in waterfall-type system development including multiple development steps; a past work product analysis step that analyzes the format of the past work products acquired in the past work product acquisition step; an input screen providing step that provides an input screen for inputting input items necessary to create a new work product; a prompt generation step that generates a prompt for automatically generating a new work product based on the format of the past work product analyzed in the past work product analysis step and the input items entered on the input screen provided in the input screen providing step; a prompt providing step that provides the prompt generated in the prompt generation step to a large-scale language model; a new work product acquisition step that acquires from the large-scale language model a new work product generated in response to the prompt provided in the prompt providing step; and a new work product providing step that provides the new work product acquired in the new work product acquisition step.

10. An information processing program for causing a computer to perform the following operations: a past product acquisition process for acquiring past products created in the past in waterfall-type system development including multiple development steps; a past product analysis process for analyzing the format of the past product acquired in the past product acquisition process; an input screen provision process for providing an input screen for inputting input items necessary to create a new product; a prompt generation process for generating a prompt for automatically generating a new product based on the format of the past product analyzed in the past product analysis process and the input items entered on the input screen provided in the input screen provision process; a prompt provision process for providing the prompt generated in the prompt generation process to a large-scale language model; a new product acquisition process for acquiring from the large-scale language model a new product generated in response to the prompt provided in the prompt provision process; and a new product provision process for providing the new product acquired in the new product acquisition process.

Images