Information processing method, program, and information processing device

Abstract

This invention provides an information processing method, program, and information processing device that support the creation of new documents by utilizing various (diverse) documents created and accumulated in the past. [Solution] The method involves creating a template with multiple items based on a document of the same type as the document to be created, using a language model, obtaining information for the multiple items, and then creating a new document based on the created template and the obtained information for the multiple items using a language model.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a technique for confirming that a document is created based on an expected template. In the technique disclosed in Patent Document 1, from the correspondence relationship of the description items in a plurality of documents having a relationship in which a change in one document (source document) also affects other documents (destination documents), based on the description items of the source document, identify the description items to be included in the destination document, and create a document using, as a template for the destination document, a document including a set of the identified description items and identification information for identifying the content of the description items of the source document corresponding to the identified description items. Further, in the technique disclosed in Patent Document 1, for a document to be determined, it is determined using identification information whether the template used for creating the document is a template generated based on the description items of the source document of the document, and when it is determined for all description items that the template is a template generated based on the description items of the source document, it can be specified that the document is created based on an appropriate template.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Companies and other organizations often create new documents by referring to various (diverse) documents that have been created and accumulated in the past. However, in many cases, a large amount of documents are stored but not effectively utilized. The technology disclosed in Patent Document 1 does not mention supporting the creation of new documents by effectively utilizing various (diverse) documents that have been created and accumulated in the past.

[0005] One aspect of this approach is to provide information processing methods that can support the creation of new documents by utilizing various (diverse) documents that have been created and accumulated in the past. [Means for solving the problem]

[0006] One aspect of the information processing method involves a computer creating a template with multiple items based on a document of the same type as the document to be created, using a language model, acquiring information for the multiple items, and then using the language model to create the document to be created based on the created template and the acquired information for the multiple items. [Effects of the Invention]

[0007] One aspect of this is that it can support the creation of new documents by utilizing various (diverse) documents that have been created and accumulated in the past. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram showing an example of the configuration of an information processing device. [Figure 2] This flowchart shows an example of the process for creating a business requirements definition document. [Figure 3] This is an explanatory diagram showing an example of a business requirements definition document. [Figure 4] This is an explanatory diagram showing an example of a business requirements definition document. [Figure 5]It is an explanatory diagram showing an example of a prompt. [Figure 6] It is an explanatory diagram of the process for creating a business requirements definition document. [Figure 7] It is an explanatory diagram of the process for creating a business requirements definition document. [Figure 8] It is an explanatory diagram of the process for creating a business requirements definition document. [Figure 9] It is a flowchart showing an example of the procedure for creating a business requirements definition document in Embodiment 2. [Figure 10] It is an explanatory diagram showing a configuration example of a document DB. [Figure 11] It is a flowchart showing an example of the procedure for creating a business requirements definition document in Embodiment 3. [Figure 12] It is an explanatory diagram showing an example of the display of a hearing sheet. [Figure 13] It is a diagram showing a configuration example of a template change history DB. [Figure 14] It is a diagram showing a configuration example of an operation history information DB. [Figure 15] It is a flowchart showing an example of the procedure for changing a template. [Figure 16] It is an explanatory diagram showing an example of the display of a template. [Figure 17] It is a flowchart showing an example of the procedure for creating a business requirements definition document in Embodiment 4. [Figure 18] It is a flowchart showing an example of the procedure for editing a document. [Figure 19] It is an explanatory diagram showing an example of the display of a document. [Figure 20] It is a flowchart showing an example of the procedure for changing a template. [Figure 21] It is a flowchart showing an example of the procedure for recommending a template. [Figure 22] It is an explanatory diagram showing an example of the display of a template.

Modes for Carrying Out the Invention

[0009] Hereinafter, an information processing method, a program, and an information processing apparatus according to the present disclosure will be described in detail based on the drawings showing embodiments thereof. In the following embodiments, an information processing apparatus that supports the creation of a business requirements document (Business Requirements Document: BRD) in which business goals and business cases to be achieved by a project are described will be described as an example. However, the present disclosure is not limited to the business requirements document, and can be applied to an apparatus that supports the creation of various requirement definition documents such as a functional requirements document, a non-functional requirements document, a user story, and a use case. Further, the present disclosure can be applied to an apparatus that supports the creation of a document having a plurality of essential description items (a document structured according to a plurality of description items), such as a technical specification document, a project plan document, a user manual, a software design document, a test plan document, an operation manual, and a risk management plan document, in addition to the requirement definition document.

[0010] (Embodiment 1) In this embodiment, an information processing apparatus that supports the creation of a new document of the same type by using various (diverse) documents created and accumulated in the past will be described. Specifically, the information processing apparatus of this embodiment supports the creation of a new business requirements document by using the business requirements document created in the past. Note that documents of the same type are documents having the same or similar usage purposes, document titles, or description items, and mean documents having the same attributes or types classified by usage purposes, document titles, or description items. FIG. 1 is a block diagram showing a configuration example of the information processing apparatus. The information processing apparatus 10 is a computer capable of various information processing and transmission / reception of information, and is composed of a personal computer, a server computer, or the like.

[0011] The information processing device 10 includes a control unit 11, a storage unit 12, a communication unit 13, an input unit 14, a display unit 15, a reading unit 16, etc., and these units are connected via a bus. The control unit 11 includes one or more processors (arithmetic processing units) such as a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), a GPU (Graphics Processing Unit), or an AI chip (AI semiconductor). The control unit 11 executes the processing that the information processing device 10 should perform by appropriately executing the program P stored in the storage unit 12. If the control unit 11 includes multiple processors, each process may be executed by the same processor, or each process may be executed by different processors.

[0012] The memory unit 12 includes RAM (Random Access Memory), flash memory, hard disk, SSD (Solid State Drive), etc. The memory unit 12 stores the program P (program product, computer program) executed by the control unit 11 and various data. The memory unit 12 also temporarily stores data generated when the control unit 11 executes program P. The memory unit 12 also stores the language model M. The language model M is a general-purpose large-scale language model (LLM: Large Language Model) constructed by pre-training with a large set of documents, etc. The language model M is trained to, for example, when input data including text data written in natural language is input, perform calculations to generate output data according to the content of the input data based on the input data, and output the generated output data. The language model M can be constructed using algorithms such as GPT (Generative Pre-trained Transformer)-3, GPT-3.5, GPT-4, GPT-4 Turbo, GPT-4o (Omni), GPT-4o mini, Bard, Llama (Large Language Model Meta AI), Gemini, and Claude, and may be constructed by combining multiple algorithms. Note that the language model M is not limited to the transformer-based models described above. The language model M is intended to be used as a program module constituting artificial intelligence software. The language model M performs predetermined operations on input data and outputs the results. The memory unit 12 stores data such as the coefficients and thresholds of the function that defines these operations as the language model M. Instead of storing the language model M in the memory unit 12, the information processing device 10 may access and read the language model M from a language processing server. Furthermore, the language model M may be fine-tuned using various (diverse) documents created and accumulated in the past. In addition to fine-tuning the entire language model M, the parameters of only certain layers may be fine-tuned using techniques such as LoRA (Low-Rank Adaptation).The storage unit 12 may be composed of multiple storage devices, and a part of the storage unit 12 may be other storage devices connected to the information processing device 10, or other storage devices that the information processing device 10 can communicate with.

[0013] The communication unit 13 is a communication module for processing wired or wireless communication, and transmits and receives information with other devices via the network N. The network N may be the Internet or a public telephone network, or it may be a LAN (Local Area Network) built in a facility where the information processing device 10 is installed. The input unit 14 receives operation input from the user and sends control signals corresponding to the operation content to the control unit 11. The display unit 15 is a liquid crystal display or an organic EL (Electro Luminescence) display, etc., and displays various information according to instructions from the control unit 11. The input unit 14 and the display unit 15 may be a touch panel configured as an integrated unit.

[0014] The reading unit 16 reads information stored on a portable storage medium 10a such as a CD (Compact Disc), DVD (Digital Versatile Disc), USB (Universal Serial Bus) memory, SD card, microSD card, or CompactFlash (registered trademark). The program P and various data to be stored in the storage unit 12 may be read by the control unit 11 from the portable storage medium 10a via the reading unit 16 and stored in the storage unit 12. Alternatively, the program P and various data may be written to the storage unit 12 during the manufacturing stage of the information processing device 10, or the control unit 11 may download them from another device via the communication unit 13 and store them in the storage unit 12.

[0015] In this embodiment, the information processing device 10 is not limited to a single computer, but may be a multi-computer system comprising multiple computers, or it may be a virtual machine virtually constructed by software within a single device. Furthermore, the information processing device 10 may be a local server installed in the facility where the information processing device 10 is located, or it may be a cloud server connected via network N. In addition, the program P may be deployed and executed on a single computer or at one site, or it can be distributed across multiple sites and deployed to be executed on multiple computers interconnected via network N. Moreover, the input unit 14 and the display unit 15 of the information processing device 10 are not mandatory, and it may be configured to accept operations via a connected computer, or to output the information to be displayed to an external display device.

[0016] The information processing device 10 of this embodiment creates a template that will serve as the basis for a business requirements specification document, and a hearing sheet that contains the names of several items considered necessary for the business requirements specification document, based on a previously created business requirements specification document (a document of the same type as the document to be created). The information processing device 10 also retrieves information for each item in the hearing sheet and inserts it into the hearing sheet based on the created template and hearing sheet, a document containing information about the user's company that will create the new business requirements specification document, an internal database, a document containing information that will be the target of the business requirements specification document to be created (for example, information on the project name, start date, end date, purpose, background, etc.) (for example, a document containing corporate information, management policies, etc. of each target company), or other various referenceable information sources (hereinafter referred to as reference documents). Furthermore, the information processing device 10 creates a new business requirements specification document based on the created template and the hearing sheet into which the information for each item has been inserted. In this way, the information processing device 10 can create a new business requirements specification document based on a previously created business requirements specification document.

[0017] The following describes a series of processes by which the information processing device 10 of this embodiment creates a new business requirements specification document based on a previously created business requirements specification document. Figure 2 is a flowchart showing an example of the business requirements specification document creation process, Figures 3 and 4 are explanatory diagrams showing an example of a business requirements specification document, Figure 5 is an explanatory diagram showing an example of a prompt, and Figures 6 to 8 are explanatory diagrams of the business requirements specification document creation process.

[0018] A user who wishes to create a new business requirements specification document based on a previously created business requirements specification document using the information processing device 10 prepares a previously created business requirements specification document as a sample document. When preparing multiple sample documents, the sample documents may include multiple business requirements specification documents created in the same format, or they may include multiple business requirements specification documents with different (not unified) formats. The user also prepares a reference document as described above. The user stores the prepared business requirements specification document (sample document) and reference document in the storage unit 12 of the information processing device 10, or in the storage unit of another terminal that the information processing device 10 can communicate with.

[0019] The control unit 11 of the information processing device 10 obtains a previously created business requirements definition document as a sample document (S11). Here, the control unit 11 may, for example, read a sample document stored on the hard disk of the storage unit 12 into RAM, or it may obtain a sample document from another terminal that can communicate and store it in the storage unit 12. In this embodiment, a configuration using two sample documents will be described as an example. Figure 3 shows an example of a business requirements definition document for Company AAA, and Figure 4 shows an example of a business requirements definition document for Company BBB. As shown in Figures 3 and 4, a business requirements definition document is a document in which items that are considered to need to be described are written according to a somewhat fixed format, but the categorization and item names of each item do not have to be exactly the same.

[0020] The control unit 11 creates an information extraction prompt (S12) to instruct the system to extract information on items deemed necessary for the business requirements definition document from the acquired sample document. Figure 5A shows an example of an information extraction prompt, which includes a command statement for the process to be executed and text about the sample document (input document). The text about the sample document may contain the file name of the sample document or the content (body) of the sample document. In the example of Figure 5A, two sample documents are listed as text about the input document, but a command statement may be included for each sample document. For example, the command statement for the first sample document may be "Analyze the following document and extract the items," and the command statements for the second and subsequent sample documents may be "Analyze the following document and extract any items that should be added." A prompt template containing only the command statement in the prompt of Figure 5A is stored in the storage unit 12, and the control unit 11 creates an information extraction prompt by inserting the text about the acquired sample document into the prompt template. Figure 5B shows a modified example of the information extraction prompt. If the items to be included in the business requirements specification are somewhat specified, a prompt containing the specified items may be created in addition to the command statements and sample document text, as shown in Figure 5B. In the example of Figure 5B, in addition to the prompt in Figure 5A, it includes items that are considered necessary for the business requirements specification. In the example of Figure 5B, each item is listed, but a descriptive text for each item may also be included, for example, for the item "Project Overview," a text explaining the content that is likely to be included as the project overview. In the case of the prompt in Figure 5B, a prompt template containing the command statements and pre-specified items is stored in the storage unit 12, and the control unit 11 can create the information extraction prompt in Figure 5B by inserting the text related to the acquired sample document into the prompt template.

[0021] The control unit 11 inputs the created information extraction prompt to the language model M, and the language model M extracts (acquires) information for each item (item name, its value, and attribute of the value) from each sample document (S13). For example, by inputting the information extraction prompt to the language model M, the control unit 11 extracts the item name and value from each sample document, and the language model M infers (estimates) the attribute of the value of each item from the extracted item name and value, thereby acquiring the item name, value, and attribute of each item. The attribute is information that indicates whether the value of each item is the name of a person, organization, project, etc., an amount of money, a due date, a procedure, etc. The control unit 11 acquires the information of each item extracted from each sample document by the language model M and stores the acquired information of each item in a DB (database) prepared in, for example, the storage unit 12 (S14). A DB is prepared for each type of sample document. For example, a database is prepared for each type of document, such as business requirements definition documents, functional requirements definition documents, non-functional requirements definition documents, user stories, use cases, technical specifications, project plans, user manuals, software design documents, test plans, operation manuals, and risk management plans. Information on each item extracted from each sample document is stored in the database corresponding to the type of sample document. Figure 6 shows an example of a database that stores information on each item extracted from a sample business requirements definition document. The database shown in Figure 6 includes major item columns, minor item columns, value columns, and attribute columns, and stores the information (value) and its attributes described in the item, associated with the item name of the major item and the item name of the minor item. Note that some major items do not have minor items and only store the value (information) and attributes. Here, the item names and attributes of major and minor items are information used in common in business requirements definition documents (sometimes referred to as static data below), while the values ​​of each item are different information for each business requirements definition document (sometimes referred to as dynamic data below). Therefore, the control unit 11 divides the information extracted from the sample document into static data and dynamic data and stores them in the storage unit 12 (DB).Furthermore, the control unit 11 may not only extract information for each item from the sample document, but may also be configured to obtain items deemed necessary for the document to be created (for example, a business requirements definition document) from the user (the user creating the new document), for example, in QA (Question and Answer) format.

[0022] The control unit 11 determines whether or not the item information storage process has been completed for all sample documents (S15). If it determines that the process has not been completed (S15: NO), it returns to step S14 and stores the information (item name, value, attribute) of each item extracted from the sample documents (item name, value, attribute) in the DB for the sample documents for which the item information storage process has not been performed (S14).

[0023] When the control unit 11 determines that it has finished storing item information for all sample documents (S15:YES), it creates a template / hearing sheet creation prompt to instruct the creation of a template that will serve as the basis for the business requirements definition document and a hearing sheet in which each item is described, based on the item information extracted from each sample document (S16). Figure 5C shows an example of a template / hearing sheet creation prompt, and in the example of Figure 5C, it includes a command statement and text about the information extracted from the sample documents as reference information. Here, the text about the information extracted from the sample documents may contain the file name of the extracted information (DB), or it may contain the content of the extracted information (content of the DB). In addition, the command statement here may include text that specifies the output format depending on whether each piece of information is static or dynamic data, for example, "Please create a template and hearing sheet that have input fields for the values ​​of each item, with the item names written as they are." In the prompt shown in Figure 5C, a prompt template containing only instruction statements is stored in the memory unit 12, and the control unit 11 can create the prompt shown in Figure 5C by inserting text related to the extracted information of each sample document into the prompt template.

[0024] The control unit 11 inputs the created template and hearing sheet creation prompt to the language model M, and the language model M creates the template and hearing sheet indicated by the prompt (S17). Specifically, the control unit 11 acquires the template and hearing sheet created by the language model M and stores the acquired template and hearing sheet in the storage unit 12. Figure 7 shows an example of a template and hearing sheet created based on the extracted information of the business requirements definition document in Figures 3 and 4. The template contains multiple major and minor items (static data) that are considered necessary for the business requirements definition document, identified based on the extracted information (DB shown in Figure 6) extracted from each sample document, and explanatory text regarding the information (values, dynamic data) to be entered into each item. The explanatory text here may include sentences that explain what should be entered, such as "Insert project name" for the item "Project name". The items described in the template may include all items in each sample document, or similar items may be grouped together as a single item. Therefore, the control unit 11 creates a template based on the extracted information from each sample document and the classification result of whether each piece of information is static or dynamic data. For static data (item names), it records them as they are, and for dynamic data (information for each item), it writes a descriptive text for each piece of information. The descriptive text for each piece of information may be stored in the storage unit 12 in advance, or it may be created using the language model M. The hearing sheet contains each item (major and minor items) described in the template, and descriptive texts (questions) regarding the information to be entered for each item. The hearing sheet may also have input fields for each item, and may be configured to accept input of information for each item through this hearing sheet.

[0025] The control unit 11 inserts corresponding information into each item in the hearing sheet (S18). The information inserted here includes information about the client for whom the newly created business requirements definition document is to be created (in this case, CCC Corporation) and information about the company (our company) that is creating the business requirements definition document. Figure 8 shows an example of how the hearing sheet is displayed. For example, as shown in Figure 8, the control unit 11 displays the hearing sheet on the display unit 15, associating each item with its corresponding input field. The control unit 11 then accepts information input for each input field in the hearing sheet via the input unit 14, and inserts the entered information into each input field for display. Alternatively, the control unit 11 may be configured to provide a pull-down menu in the corresponding input field, such as the project start date, project end date, and project manager items in Figure 8, allowing the user to select any one option from multiple choices, and accepting the selection of any option (information) via the pull-down menu. The control unit 11 may also extract and display options from various documents created in past projects, or from documents (e.g., reference documents) containing information about the client (CCC) for which the business requirements definition document is being created, or information about the company (our company) that created the business requirements definition document. For example, if there is a reference document containing a list of names of employees who could become project managers within our company, the control unit 11 may extract the names of these employees from this reference document and set up a pull-down menu displaying the extracted names in the input field for the project manager. Furthermore, if various documents created in past projects (documents other than the business requirements definition document) contain descriptions of items such as business objectives, business background, and business goals, the control unit 11 may extract the descriptions for each item from these documents and present the extracted descriptions so that they can be selected. In addition, when information is entered into the input fields for each item of the hearing sheet, the control unit 11 may display information to assist in entering information into each item based on the entered information.For example, if "customer account management function" is entered in the "Include" item of the scope in the hearing sheet, information such as "account registration," "login function," and "account settings" may be added to the input field as functions related to "customer account management function." Information about such related functions can be identified, for example, from a previously created business requirements definition document, and by storing the related information in the storage unit 12 in advance, the control unit 11 can present it to the user by displaying other related information in the input field when any of the related information is entered. The user can input information for each item by selecting the appropriate information from the presented information, thereby reducing the burden of input operations on the user. Furthermore, the control unit 11 may obtain information for each item in the hearing sheet from the user (the user creating the new document), for example in QA format. In addition, the control unit 11 may insert information into each item in the hearing sheet using the language model M. For example, the control unit 11 may create prompts to instruct the insertion of information into each input field in the hearing sheet based on the template and hearing sheet created in step S17 and the reference document, and input the created prompts to the language model M, thereby obtaining a hearing sheet in which information has been inserted into each input field by the language model M.

[0026] The control unit 11 creates a document creation prompt (S19) to instruct the creation of a new business requirements definition document (business requirements definition document for CCC Corporation) based on the hearing sheet with information inserted into the input fields for each item and the template created in step S17. Figure 5D shows an example of a document creation prompt, in which the example in Figure 5D includes an instruction statement and, as information to be used, text relating to the template created in step S17 and the hearing sheet with each piece of information inserted. Here, the text relating to the template and hearing sheet may contain the file names of the template and hearing sheet, or it may contain the contents of the template and hearing sheet. A prompt template containing only an instruction statement in the prompt of Figure 5D is stored in the storage unit 12, and the control unit 11 can create the prompt of Figure 5D by inserting the created template and the hearing sheet with each piece of information entered into the prompt template.

[0027] The control unit 11 inputs the created document creation prompt to the language model M, and the language model M creates the document (S20). Specifically, the control unit 11 retrieves the document (business requirements definition document) created by the language model M and stores the retrieved document in the storage unit 12. This creates a business requirements definition document for CCC. Although not shown in the diagram, the business requirements definition document for CCC, like the business requirements definition documents shown in Figures 3 and 4, associates multiple description items such as project information, business objectives, business background, scope, functional requirements, non-functional requirements, stakeholders, budget and resources, and includes information specific to CCC as information related to each item.

[0028] The process described above allows for the creation of a template that serves as a blueprint for a business requirements specification document, using a previously created business requirements specification document. Furthermore, a hearing sheet listing items deemed necessary for inclusion in the business requirements specification document can be created using a previously created business requirements specification document. By using such templates and hearing sheets, the burden of creating a new business requirements specification document of the same type as a previously created one can be reduced. Therefore, it is possible to support the creation of new documents of the same type by utilizing previously created documents. In the process described above, by registering the extracted information from the sample document into a database as shown in Figure 6, the search process can be accelerated by using this database in document searches to find desired documents from a large volume of past documents.

[0029] In the process shown in Figure 2, steps S11-S17 and steps S18-S20 may be executed at different times. For example, by executing steps S11-S17 when a sample document is ready, a template and interview sheet of the same type as the sample document are created based on the sample document and stored in the storage unit 12. Then, when creating a new document, the template and interview sheet are read from the storage unit 12 and steps S18-S20 are executed, making it possible to create a new document using the template and interview sheet that have been created in advance and stored in the storage unit 12. Alternatively, after the template and interview sheet have been created by executing steps S11-S17, the system may be configured to create a new document by executing steps S18-S20 using the created template and interview sheet. For example, if the created template and interview sheet are adopted as the standard format within a department, the system may be configured to create a new document using this template and interview sheet. Alternatively, the system may be configured so that the user specifies the template to be used according to the content of the document to be created.

[0030] In this embodiment, the process of creating templates and hearing sheets from sample documents is not limited to being performed locally by the information processing device 10. For example, a server that performs the above-described process may be provided. In this case, the information processing device 10 is configured to send sample documents to the server and to obtain templates and hearing sheets created on the server.

[0031] In this embodiment, various documents can be used as reference documents to obtain information to be inserted into each item in the hearing sheet. For example, if the document to be created is a consent form for a clinical trial, the research protocol for that clinical trial may be used as the reference document. In this case, the information processing device 10 can create a template and a hearing sheet from a previously created consent form for a clinical trial, extract information for each item in the hearing sheet from the research protocol, and create a new consent form for a clinical trial based on the hearing sheet and template into which the extracted information has been inserted. Furthermore, if the document to be created is a newsletter that provides specialized research results to the general public, the research report on those research results may be used as the reference document to extract information for each item in the hearing sheet. Furthermore, if the document to be created is a user manual (user guide) for a product, the technical specifications for that product may be used as the reference document to extract information for each item in the hearing sheet. Furthermore, if the document to be created is an FAQ (Frequently Asked Question) regarding legal documents such as contracts or terms of service, the legal document may be used as the reference document to extract information for each item in the hearing sheet. Furthermore, if the document to be created is an explanatory material that provides advanced scientific discoveries to the general public or students, information for each item in the interview sheet may be extracted using a paper on the scientific discovery as a reference document. As described above, by using information extracted from various reference documents, it becomes possible to create a document for the general public from a specialized document. In addition, since specialized terminology extracted from the specialized document is used in this process, it becomes possible to create various documents that use specialized terminology.

[0032] (Embodiment 2) This embodiment describes an information processing device that assists in the creation of documents (for example, business requirements specifications) without using sample documents (various documents created and accumulated in the past). The information processing device of this embodiment has the same configuration as the information processing device 10 of Embodiment 1 shown in Figure 1, so a description of the configuration will be omitted.

[0033] The information processing device 10 of this embodiment assists in the creation of documents using the language model M by executing the processes of steps S16 to S20 in Figure 2. In the information processing device 10 of this embodiment, the control unit 11 creates a template / hearing sheet creation prompt to instruct the language model M to create a template that will serve as the basis for the business requirements definition document and a hearing sheet in which each item is described (S16). The prompt here can be a prompt containing an instruction such as, for example, "Identify the items that are considered necessary for the business requirements definition document and the items that are considered desirable to be included in the business requirements definition document, and create a template for the business requirements definition document and a hearing sheet in which the identified items are described." The prompt may also contain explanatory text and definition text for the template and hearing sheet, and may contain text specifying the output format.

[0034] The control unit 11 inputs the created template and hearing sheet creation prompt to the language model M, and the language model M creates a template and hearing sheet for creating a business requirements definition document (S17). Here again, the control unit 11 creates a template and hearing sheet as shown in Figure 7. Subsequently, by executing the processes in steps S18 to S20, the control unit 11 uses the language model M to insert information corresponding to each item in the hearing sheet, creates a document (business requirements definition document) based on the template and the hearing sheet with the information for each item inserted, and stores the created document in the storage unit 12. Here again, a business requirements definition document for CCC Corporation is created.

[0035] Since the language model M has been pre-trained on a large amount of documents written in various formats, including business requirements definitions, it can create templates and interview sheets with a certain degree of accuracy even without using sample documents or information extracted from sample documents. Therefore, it becomes easier to create new documents by using templates and interview sheets created using the language model M. Alternatively, the control unit 11 may be configured to display the created templates and interview sheets on the display unit 15 after processing in step S17, allow the user to edit them, and then perform the processing from step S18 onwards. Alternatively, the control unit 11 may be configured to have the language model M modify the contents (items) of the templates and interview sheets based on the created templates and interview sheets and documents created in the past other than business requirements definitions after processing in step S17, and then perform the processing from step S18 onwards based on the templates and interview sheets modified by the language model M.

[0036] The following describes the process when the processing by the information processing device 10 of Embodiment 1 and the processing by the information processing device 10 of this embodiment are combined. Figure 9 is a flowchart of an example of the business requirements definition document creation process of Embodiment 2. The process shown in Figure 9 is the process shown in Figure 2 with step S31 added before step S11. The same steps as in Figure 2 will not be explained.

[0037] In the process shown in Figure 9, the control unit 11 of the information processing device 10 is configured to receive instructions on whether or not to use a previously created sample document when creating a new document. For example, in the process of creating a new document, the control unit 11 provides an input field on the initial screen displayed on the display unit 15 for inputting (selecting) whether or not to use a sample document, and receives instructions on whether or not to use a sample document through this input field. The control unit 11 determines whether or not to use a sample document through the input field (S31), and if it determines that the use of a sample document has been instructed (S31: YES), it proceeds to step S11, and if it determines that the use of a sample document has not been instructed (S31: NO), it proceeds to step S16. That is, if the user (document creator) instructs to use a sample document, the control unit 11 extracts information on items considered necessary for the business requirements definition document from the sample document and executes a process to create a template and a hearing sheet using the extracted information. On the other hand, if the use of a sample document is not instructed, the control unit 11 executes a process to create a template and a hearing sheet using the language model M without using a sample document. In this way, it is possible to switch whether or not to use sample documents. For example, if a user wants to create a document that reflects the format or corporate culture of documents previously created within the company, they may be instructed to use sample documents.

[0038] The process in Figure 9 is configured to allow the user to choose whether or not to use sample documents, but the system is not limited to this configuration. For example, the system could switch based on whether or not there are a predetermined number of sample documents. For example, in step S31 in Figure 9, the control unit 11 determines whether or not the number of documents prepared as sample documents (business requirements definition documents created in the past) is greater than or equal to a predetermined number. If the control unit 11 determines that the number of sample documents is greater than or equal to a predetermined number, it proceeds to step S11; if it determines that the number of sample documents is less than a predetermined number, it proceeds to step S16. In such a configuration, if a predetermined number or more of sample documents are available, the system extracts information about items considered necessary for the business requirements definition document from the sample documents and uses the extracted information to create a template and a hearing sheet. On the other hand, if a predetermined number of sample documents are not available, the system creates a template and a hearing sheet using the language model M without using sample documents. In this way, the system can switch whether or not to use sample documents depending on the number of sample documents. Furthermore, if dynamic data for specific items for the document to be created is already stored in the DB in the storage unit 12, the control unit 11 may decide not to use sample documents. In this case, the control unit 11 may create a template and a hearing sheet using the language model M without using a sample document, and create a new document using dynamic data previously stored in the DB.

[0039] (Embodiment 3) This embodiment describes an information processing device that suggests information to be entered when inputting information into each item of a hearing sheet. The information processing device of this embodiment has the same configuration as the information processing device 10 of Embodiment 1 shown in Figure 1, so the explanation of the configuration will be omitted. In this embodiment as well, the configuration that supports the creation of a business requirements definition document will be explained as an example.

[0040] In this embodiment, when the information processing device 10 creates a template from a sample document, it stores the information of each item in the template (item information) and the information of each item extracted from the sample document in the document DB prepared in the storage unit 12. Furthermore, after creating a document using the template, the information processing device 10 stores the information of each item in the document DB. Figure 10 is an explanatory diagram showing an example of the document DB configuration. The document DB shown in Figure 10 uses a NoSQL database that can handle JSON (JavaScript® Object Notation) format, which can enumerate and store combinations of item names and values ​​for each item, but other formats may also be used. The document DB can be configured using, for example, a document-oriented NoSQL database such as MongoDB or a graph DB such as Neo4j. In this embodiment, item information for similar documents is stored in one document DB, but for example, a separate document DB may be prepared in the storage unit 12 for each document. Hereinafter, documents created in the past will be referred to as past documents. Also, when the information of each item in the sample document and past documents is used in a new document, they become information source documents for the new document.

[0041] The document database shown in Figure 10 sequentially stores template item information, item information extracted from sample documents when the template was created, and item information from past documents in an append-type event log format. The template item information includes the template ID assigned to the template, a timestamp indicating the creation date and time of the template, and the item name for each item (the value for each item is left blank). The item information extracted from sample documents includes the template ID, a timestamp indicating the creation date and time of the template, and the item name and value for each item. The item information for past documents includes the template ID of the template used to create the past document, a timestamp indicating the creation date and time of the past document, and the item name and value for each item. The item name and value for each item may be input information for the template, i.e., differential information for the template. In the example in Figure 10, the sample document item information is the information for each item extracted from the sample document shown in Figure 3, and the past document item information is the information for each item in the business requirements definition document for CCC created using the hearing sheet shown in Figure 8. The item information for sample documents and past documents is stored by associating each item name with its value. For example, the value of each item is stored corresponding to the item name of the major item "project (project information)" and the sub-items "name (project name)", "start_date (project start date)", "end_date (project end date)", "pm (project manager)", and "key_stakeholders (key stakeholders)". Similarly, the value of each item is stored corresponding to the item name of the major item "business" and the sub-items "objectives (purpose)" and "background (background)". The document database also stores usage history information, indicating that each value was used when other past documents were created, associated with the value of each item.The usage history information includes, for example, the creation date and time (usage date and time) when other past documents were created using each value, the purpose of those other past documents (e.g., the company to which the document was submitted and the submission date), and information about the user (creator) (information about the document). If those other past documents are viewable, the usage history information may also include the file name of those other past documents. The date and time, purpose, and user information included in the usage history may be obtained from other past documents that were created and the property information of those past documents. In addition, the document DB may store information such as the attributes of the values ​​of each item and descriptions of each item. The attributes of the values ​​of each item and descriptions of each item may be created using a language model M based on the information of each item extracted from the document, or they may be entered via the input unit 14 at any time, or they may be stored in the storage unit 12 in advance.

[0042] Figure 11 is a flowchart showing an example of the business requirements definition document creation process in Embodiment 3, and Figure 12 is an explanatory diagram showing an example of the display of the interview sheet. The process shown in Figure 11 is the same as the process shown in Figure 2, but with step S41 added instead of steps S11 to S17, steps S42 to S50 added between steps S18 and S19, and step S51 added after step S20. The same steps as in Figure 2 will not be explained. In the following explanation, it will be assumed that the process shown in Figure 2 has been executed at least once, and that the business requirements definition document template, the interview sheet, and the document DB shown in Figure 10 have been created and stored in the storage unit 12.

[0043] In this embodiment, the control unit 11 of the information processing device 10 reads the business requirements definition document template and the hearing sheet from the storage unit 12 (S41). The control unit 11 may also read the document DB associated with the template ID of the template from the storage unit 12 along with the template and the hearing sheet. The control unit 11 then performs the processing of step S18 on the read hearing sheet and inserts the corresponding information into each item in the hearing sheet.

[0044] After processing in step S18, the control unit 11 generates comments for each item in the hearing sheet based on the contents stored in the document database (S42). For example, the control unit 11 reads the value of each item from the document database and generates a comment that includes the read value and the creation date and creator information of past documents created using the read value. This generates a comment that presents the value of each item and the creation date and creator of past documents that used that value. The creation date and creator information of past documents may be stored in the document database. The control unit 11 may also read the usage history information of each item's value from the document database and generate a comment that includes information such as the creation date and time, purpose, and user included in the usage history. In this case, a comment is generated that presents the creation date, purpose, and creator of other past documents created using the value of each item. For example, the control unit 11 may generate a comment such as "Creation date: X month X day, User: YY, Purpose: ZZZ, Value: ********". If the document database stores item information for multiple past documents, the control unit 11 may generate multiple comments based on the item information of each past document. The control unit 11 may also generate comments suggesting the values ​​of each item based on the stored contents of the document database, based on the frequency of use of each item's value, or on the validity obtained from the usage history and frequency of use. Note that the control unit 11 does not need to generate comments for all items. The items for which comments are generated may be items identified based on the attributes of the item's value, items that have been set in advance as targets for comment generation, or items that are arbitrarily (randomly) selected. For example, by setting items with predetermined attributes (e.g., date or amount) as items for which comments are generated, the items for which comments are generated can be identified based on the attributes of each item. Alternatively, by setting items for which comments are generated for each document type, project budget, document creation department, or stakeholder, the items for which comments are generated can be identified according to each document.Furthermore, the control unit 11 may arbitrarily select past documents to use for generating comments, or it may use item information from past documents that have been pre-configured for comment generation. In addition, if there are past documents that have been previously adopted (used), the control unit 11 may be configured to prioritize the use of item information from those past documents when generating comments.

[0045] The control unit 11 adds the generated comments to the hearing sheet into which the information for each item was inserted in step S18 (S43). The control unit 11 displays the hearing sheet with the added comments on the display unit 15, as shown in Figure 12 (S44). The hearing sheet in Figure 12 has the same items and values ​​as the hearing sheet in Figure 8, and a leader line and comment C1 are added to the input field of the item into which a comment was generated in step S42. In the example in Figure 12, comment C1 is added to three items, but comments may be added to all items, or comments may be added to items identified by a predetermined process. Each comment C1 includes other examples of the value of the corresponding item, as well as the creation date and creator of a past document that used the value of the other example. Note that comment C1 is a comment on the value (input information) of each item, and may also include the file name and purpose of the past document that used the value of the other example. Furthermore, it may be configured to allow setting an instruction not to add comments to any item, in which case a hearing sheet can be created in which no comments are added to the set item.

[0046] Each comment C1 is provided with an "Accept" button to indicate acceptance of the value displayed in the comment, and a "Cancel" button to indicate rejection. The control unit 11 determines whether either of the "Accept" buttons has been operated (S45), and if it determines that it has been operated (S45:YES), it changes the value of the item for which the "Accept" button was operated to the value in the comment (S46). The control unit 11 identifies the past document corresponding to the comment C1 for which the "Accept" button was operated, and stores usage history information in the item information of the identified past document in the document DB, indicating that the value of the item in comment C1 was used (S47). As a result, if the values ​​of each item used in a past document are used in a new document, this usage history is accumulated in the document DB. If the control unit 11 determines that the "Accept" button has not been operated (S45:NO), it skips steps S46 to S47 and determines whether either of the "Cancel" buttons has been operated (S48). If the control unit 11 determines that either of the "Cancel" buttons has been operated (S48:YES), it terminates the display of the comment for which the "Cancel" button was operated (S49). If the control unit 11 determines that the cancel button has not been pressed (S48: NO), it skips step S49 and determines whether the document creation button provided on the screen has been pressed (S50). If the control unit 11 determines that the document creation button has not been pressed (S50: NO), it returns to step S45 and continues to accept input from the accept button or cancel button.

[0047] If the control unit 11 determines that the document creation button has been operated (S50: YES), it proceeds to step S19. In step S19, the control unit 11 creates a document creation prompt using the hearing sheet, whose values ​​have been changed by the operation of the adoption button, and the template read from the storage unit 12 in step S41. The control unit 11 inputs the created document creation prompt into the language model M, and the language model M creates the document (S20). The control unit 11 also generates a generation transaction that includes the template ID and version of the template used to create the document, information on the past document used (source document), and information on the items used from the past document, and adds this to the created document as metadata. This enables traceability, allowing retrospective tracking of which template (template ID and version) was used in the created document, which past document's values ​​were used for which items, etc. In addition to the information described above, the control unit 11 may include in the generated transaction information information of the language model M used to create the document (e.g., model ID and version), information of the prompt for creating the document (e.g., prompt ID), difference information between output versions (editing information or change information for the document), protocol_id as shown in Figure 14, and event_id as shown in Figure 13. After creating the document in step S20, the control unit 11 stores the information of each item of the created document in the document DB of the storage unit 12 (S51). This allows the item information of the document created in step S20 to be stored in the document DB. Through the above process, when creating a new document, comments on each item can be presented based on the content of past documents (document DB), and the user can choose whether or not to adopt the commented content in the new document. Therefore, the creator of the new document can create the new document by referring to the terminology and sentences in past documents, thereby improving the quality of the new document.

[0048] In this embodiment, the control unit 11 may further perform a process to determine whether the inserted value is appropriate for each item in the hearing sheet when a value is inserted into that item. For example, if a value other than a date is inserted into the input field of an item whose attribute stored in the document DB is a due date, the control unit 11 will determine that the inserted value for that item is inappropriate and will generate a comment such as "Please enter a date" and add it to the item. In the example in Figure 12, January 1, 2024 is entered in the input field for the planned project start date, and January 31, 2024 is entered in the input field for the planned project end date. In this case, the control unit 11 may, for example, calculate the period from the planned project start date to the planned project end date, determine whether the period is appropriate, and if it determines that it is inappropriate, generate a comment C2 as shown in Figure 12 and add it to the item. For example, the control unit 11 reads the planned project start date and planned project end date for each past document from the document DB, calculates the respective project period, and calculates the shortest, longest, and average project periods. The control unit 11 then compares the period from the planned project start date to the planned project end date entered in the hearing sheet with the shortest, longest, and average project periods calculated, as shown in Figure 12, to determine whether the period is appropriate. If the period falls within the range of the shortest to longest calculated project periods, the control unit 11 may determine that the period is appropriate. In the example in Figure 12, the shortest project period is 6 months, while the period from the planned project start date to the planned project end date entered in the hearing sheet is 1 month, so comment C2 as shown in Figure 12 is generated. For example, if a period for which setting the project period is prohibited is registered, the control unit 11 may determine whether the period from the planned project start date to the planned project end date entered in the hearing sheet includes the prohibited period, and if it does, it may determine that the period is inappropriate and generate a comment notifying the user to make appropriate input.Furthermore, when a value is inserted into each item of the hearing sheet, the control unit 11 may determine whether the item name into which the value was inserted or the value inserted into each item contains a redundant expression. If it determines that a redundant expression is present, it may generate a comment such as "The item names of item AA and item BB are similar" or "The values ​​of item AA and item BB are similar" and attach it to the corresponding item. The determination of whether the item name of the hearing sheet or the value of each item contains a redundant expression can be performed, for example, using a language model M. For example, a hearing sheet with the values ​​of each item inserted and a prompt including the command "Detect redundant expressions from this hearing sheet" may be input to the language model M, and the language model M may perform the detection of redundant expressions. Furthermore, based on the contents stored in the document DB, the control unit 11 may generate comments that notify the user of suggested values ​​according to the frequency of use of the values ​​of each item used in past documents, or comments that warn that the entered value is inappropriate. Furthermore, the control unit 11 may generate comments that suggest the values ​​of each item used in documents that have been adopted in the past. Furthermore, the control unit 11 may, for items into which numerical values ​​are entered, determine whether the entered value satisfies the set conditions, for example, if upper and lower limits are set in advance, and if a currency unit is set, and generate a warning comment if it does not. In such a configuration, for example, when a user enters an item value into the input field of the hearing sheet via the input unit 14, the appropriateness of the entered value is determined, and a comment is displayed if it is inappropriate. Therefore, it is possible to support the user in the operation of entering values ​​for each item into the hearing sheet, and to suppress input errors and the input of inappropriate values. In addition, if there are items in the hearing sheet for which values ​​have not been entered, the control unit 11 may add a comment to those items to notify them of the missing input. In this case, missing inputs can be suppressed.

[0049] In this embodiment, when values ​​are inserted into each item of the hearing sheet, examples of other values ​​are suggested, and the values ​​of each item can be appropriately changed by referring to the suggested content. Alternatively, for example, after processing in step S20, the same processing as in steps S42 to S49 may be performed on the created document to add comments C1, C2, etc., as shown in Figure 12. In this case, when the document is created (after processing in step S20 in Figure 11), other examples of values ​​for each item are suggested, and the values ​​of each item can be changed. Therefore, even in this case, the values ​​of each item in the new document can be created by referring to terms and sentences in past documents.

[0050] In this embodiment, based on the item information of each document stored in the document DB, feature quantities such as the usage trend of input information (values), input error patterns, input error rate, change history, and acceptance / rejection history of input values ​​may be extracted for each item and trained on the language model. An input error here means a state in which a value with content or format different from the input trend in past documents or the defined format (type) is entered for each item. For example, this includes format errors (input of a value in a format different from the defined format), content validity errors (input of a value that does not match the content and meaning of the item), missing input, ambiguous expression, etc. Such input errors are errors that can be automatically detected using, for example, a language model, and are so-called human errors. In this case, when creating a new template, when inputting values ​​for each item in a hearing sheet, or when changing (editing) the values ​​of each item in a created document, the language model can generate recommended values ​​and candidate items to be made required, and present them to the user. Alternatively, the system may analyze the input history of values ​​for each item in past documents to identify the type or category of each item's value (date, name, amount, identification number, etc.), and then determine whether the input information for each item is of a different type or category than expected when creating a new document. In this case, if the input information for each item is of a different type or category than expected, an alert can be output, such as displaying a warning comment below the input field or notifying the user via a pop-up, thereby enabling error detection at the time of inputting the value for each item and reducing human errors in the input operation for each item.

[0051] Embodiment 1 described above is configured to create a template and a hearing sheet from a sample document, and then create a new document using the template and the hearing sheet. Embodiment 2 is configured to create a template and a hearing sheet using the language model M without using a sample document, and then create a new document using the template and the hearing sheet. Embodiment 3 is configured to create a new document using the created template and hearing sheet. In Embodiments 1 to 3, the creation and use of the hearing sheet is not mandatory. For example, the control unit 11 may collect information for each item and store it in the storage unit 12, and instead of a hearing sheet, it may be configured to write the stored information to the document creation prompt. Even when using such a document creation prompt, the language model M can create a document with values ​​inserted for each item.

[0052] (Embodiment 4) Embodiments 1 to 3 described above are configured to create templates and interview sheets, and then to create new documents using the templates and interview sheets. Embodiment 3 describes a configuration in which, each time a new document is created, the item information of the created document is stored in a document database as shown in Figure 10. Note that the new documents may include documents created without using the information processing device 10 in Embodiments 1 to 3, and documents created using the information processing device 10 that have been modified by the user.

[0053] When a template is generated for a certain type of document (in this embodiment, a business requirements definition document), the generated template can be used to create a new document. If multiple templates are generated, for example, multiple templates may be merged to create a single template. The process of merging multiple templates may be performed using, for example, a language model M. For example, the control unit 11 of the information processing device 10 creates a prompt that includes multiple templates and an instruction statement such as, "From the information of each item in the multiple templates, merge the duplicate items and add items that are included in at least one template to create one standard template." The control unit 11 may then input the created prompt into the language model M and create a template using the language model M. In this case, for example, a template can be created in which the names of each item, the order in which each item appears, and the terms used for the values ​​of each item are standardized (normalized).

[0054] In this embodiment, the template can be modified. The template can be modified, for example, according to a modification instruction entered via the input unit 14 on a screen displaying the template on the display unit 15. When a modification is made to the template, the information processing device 10 prepares a template modification history DB that stores information about the modification and an operation history information DB that stores information about the history of the modification in the storage unit 12, and stores them as modification log information. The template modification history DB and the operation history information DB have an event log recording mechanism in which each piece of information is sequentially recorded in event log format. Figure 13 shows an example of the configuration of the template modification history DB, and Figure 14 shows an example of the configuration of the operation history information DB. The template modification history DB shown in Figure 13 and the operation history information DB shown in Figure 14 use NoSQL databases that can handle JSON format, but formats other than JSON format may also be used.

[0055] The template change history database is stored in the storage unit 12, linked to the template ID of the template. The template change history database shown in Figure 13 stores information about template creation and information about template changes in an append-type event log format. Information about creation includes the user who created the template, the date and time of creation, a description of the template's contents, and the reason for its creation. Information about changes includes the user who modified the template (modifier), the date and time of change, a description of the changes (changed items), a description of the changes (before and after information), the reason for the change, tags indicating the reason for the change, and related information such as related tickets used in project management tools. Related tickets include information about other related tickets (work or tasks) and information indicating the relationship with other tickets. This template change history database records log information about change operations on templates, such as adding, deleting, making items mandatory, merging, reordering, and changing terminology. The operation history information database is also stored in the storage unit 12, linked to the template ID of the template. The operational history information database shown in Figure 14 stores template creation information and template modification information in an append-type event log format, recording the history of template changes. Creation information includes the template ID, the user who created it, and the creation date and time. Modification information stores information such as the template ID, the items changed, the content of the changes, the user who made the changes, the date of the changes, and the reason (basis) for the changes. This operational history information database records log information that shows the history of changes made to templates. By recording such log information, it is possible to record when, by whom, for what reason, and how the template was changed. This allows users other than the person who updated the template to understand the history of the template changes and the basis for the structure of each item in the template. Therefore, it becomes possible to create documents based on the basis. Furthermore, when selecting and evaluating standard templates to be used for document creation, using log information as a basis for judgment makes it possible to make appropriate selections and evaluations that consider whether the history of changes to each item is appropriate.For example, templates that are frequently changed may lack stability, while templates that are rarely updated and have been in use for a long time can be evaluated as highly reliable. Templates whose generated documents contribute to results can be evaluated as suitable for standardization.

[0056] When changes are made to a template, it is possible to either update the template by overwriting it, or create a new modified template separate from the original by saving it as a new file. For example, by making changes to a template according to its intended use, it is possible to create modified templates that are modified for each use. When a modified template is saved as a new file, further changes can be made to that modified template. When an update is made to a modified template, a template change history DB and an operation history information DB are generated for that modified template, and the change log information is stored, just as when the template is updated.

[0057] Figure 15 is a flowchart showing an example of the procedure for modifying a template, and Figure 16 is an explanatory diagram showing an example of how the template is displayed. The control unit 11 of the information processing device 10 performs the following processing when a template is created for a certain type of document. A corresponding hearing sheet may also be created along with the template.

[0058] The control unit 11 reads a template from the storage unit 12 (S61) and displays the read template on the display unit 15 as shown in Figure 16 (S62). The template on the display screen is displayed so that each item and its description can be selected. The control unit 11 determines whether or not a selection for any item has been accepted (S63), and if it determines that it has not been accepted (S63: NO), it waits until it is accepted. If the control unit 11 determines that a selection for any item has been accepted (S63: YES), it displays an input field for accepting changes to the selected item (S64). In the example in Figure 16, the item "Business Objectives" is selected with cursor C3, and when a predetermined operation (e.g., left click) is performed in this state, the input field C4 for changes is displayed to the right of the template. The input field C4 has, for example, an input field for the change content and an input field for the reason for the change, and accepts input of the change content and reason for the change for the selected item. The input field for changes allows users to select one of the following: adding, deleting, making required, merging, or reordering items, or changing terminology. If adding, merging, reordering, or changing terminology is selected, the field accepts input of specific changes.

[0059] The control unit 11 accepts the input of the change content via the input field C4 and displays the entered change content in the input field C4 (S65). The input field C4 is provided with an OK button to reflect the entered change content in the template and a Cancel button to instruct the user to cancel the change process. The control unit 11 determines whether or not the OK button has been pressed (S66), and if it determines that it has been pressed (S66:YES), it changes the displayed template based on the change content entered in the input field C4 (S67). If the control unit 11 determines that the OK button has not been pressed (S66:NO), that is, if it determines that the Cancel button has been pressed, it terminates the display of the input field C4 (S68) and proceeds to step S69.

[0060] The screen in Figure 16 is equipped with an overwrite save button for instructing the user to overwrite the currently displayed template, and a new save button for instructing the user to save the template as a new template (modified template). The control unit 11 determines whether the overwrite save button has been operated (S69), and if it determines that it has been operated (S69: YES), it overwrites the currently displayed template with the template stored in the storage unit 12 (S70). The control unit 11 also stores change log information in the storage unit 12 based on the changes made to the template in step S67 (S71). Here, the control unit 11 stores information regarding the changes to the template in the template change history DB provided in the storage unit 12, and stores information regarding the history of changes to the template in the operation history information DB provided in the storage unit 12. Information regarding the changes and information regarding the history of changes are entered, for example, via the input field C4 in Figure 16.

[0061] If the control unit 11 determines that the overwrite save button has not been operated (S69: NO), it determines whether the new save button has been operated (S72). If the control unit 11 determines that the new save button has not been operated (S72: NO), it returns to step S63 to accept the selection of an item and the input of changes to the selected item. If the control unit 11 determines that the new save button has been operated (S72: YES), it newly stores (newly saves) the currently displayed template as a modified template in the storage unit 12 (S73). For example, the control unit 11 generates a new file name by adding information indicating that it is a modified template to the file name of the template, and stores the modified template in the storage unit 12 with the generated file name. The control unit 11 may also add information (e.g., version) indicating that the modified template is a modification (improvement) of the template as metadata (property information) to the modified template. In this case, the relationship between templates before and after modification can be understood.

[0062] After creating a modification template, the control unit 11 generates a template change history DB and an operation history information DB corresponding to the modification template in the storage unit 12, and stores the change log information in the template change history DB and the operation history information DB (S74). Here, the control unit 11 generates information related to the creation of the modification template and stores it in the template change history DB, and generates information about the creation of the modification template and stores it in the operation history information DB. The information related to the creation of the modification template and the creation information may include the changes from the template and the reasons for the changes. Through the above process, a template can be modified to create and store a modification template that differs from the original template, for example, in some items. In addition, if the control unit 11 has created a modification template, it may also generate a hearing sheet corresponding to the modification template and store it in the storage unit 12.

[0063] In the process described above, the user makes changes to the template via the input unit 14. Therefore, for example, if there are inconsistencies in the terminology used in the template, a template (or modified template) with unified terminology can be created by making changes to unify it. By using such a template, it becomes possible to create a consistent, high-quality document with unified terminology. The information processing device 10 may also use the language model M to extract inconsistencies in terminology in the template, and use the language model M to make changes that eliminate the extracted inconsistencies, thereby creating a modified template. In this embodiment, the modified template can be saved, and it becomes possible to create a document using the modified template. Multiple modified templates may be created and saved, and any of the templates (the original template or any of the modified templates) can be used depending on the document to be created, making it possible to create a document using a template suitable for the purpose.

[0064] Template modifications are not limited to user-driven configurations; for example, they may be performed using a language model M. For instance, the control unit 11 of the information processing device 10 may create a prompt containing the template and a message regarding the modifications to the template, and obtain the modified template by inputting the created prompt into the language model M. For example, the input field C4 in the screen of Figure 16 may be provided with an input field for modification details and a button for instructing the creation of a modified template using the language model M. In this case, the control unit 11 can obtain the modified template by creating a prompt containing the message entered in the input field and inputting it into the language model M. The modifications to the template may be determined based on, for example, the frequency of use (occurrence frequency) in similar documents created in the past, or based on the items and terminology used in the latest document. Furthermore, the control unit 11 may decide to modify items and terminology used in the template if their frequency of use in similar documents created in the past is low. In this case, the control unit 11 may present the items and terminology decided to modify to the user and accept input of the modification details from the user, or modify them to items and terminology with a higher frequency of use. By modifying each item and term in the template based on its frequency of use, it becomes possible to create a more appropriate template.

[0065] The process described above allows for the modification of templates when modifications are needed during actual use, and also allows for the accumulation of modification history (operational history) information. This accumulated modification history information can be used as intellectual property (knowledge) along with the template. For example, when creating a new document, a message such as "The format has been changed due to many input errors in the past. Please be careful when entering data" can be displayed, and various suggestions based on the template's modification history can be made. Furthermore, when modifying templates in the future, it will be possible to propose modification plans based on past modification history. In this way, it becomes possible to create new documents and make further modifications to templates by referring to past modification history.

[0066] As a result of the process described above, if a modified template is created in addition to the template, the user can choose which template to use when creating a new document. Figure 17 is a flowchart of an example of the business requirements definition document creation process in Embodiment 4. The process shown in Figure 17 is the same as the process shown in Figure 2, but with steps S81 to S82 added instead of steps S11 to S17, and step S83 added after step S20. The same process as in Embodiment 1 is performed for the same steps as in Figure 2.

[0067] In this embodiment, templates (including modification templates) are created based on business requirements definition documents created in the past and stored in the storage unit 12. When a new business requirements definition document is created, the control unit 11 of the information processing device 10 accepts the selection of one template from the business requirements definition document templates and modification templates (S81). The selection of a template may be made by the user, for example, via the input unit 14, or it may be made based on information in a document prepared as a reference document. For example, if CCC Company is extracted from the reference document as information of the client for which the business requirements definition document is to be created, the control unit 11 will read the modification template for CCC Company if one is stored in the storage unit 12. If the modification template for CCC Company is not stored in the storage unit 12, the control unit 11 may read any template.

[0068] The control unit 11 reads the selected template and the corresponding hearing sheet from the storage unit 12 (S82). Based on the read template and hearing sheet, the control unit 11 executes steps S18 to S20 to create a document based on the template. The control unit 11 also generates a generation transaction that includes information about the template used in the newly created document (e.g., template ID and version), information about past documents used, and information about items used from past documents, and adds it to the newly created document as metadata. After processing in step S20, the control unit 11 stores the information of each item in the created document in the document DB prepared in the storage unit 12 (S83). In this embodiment, when a template is modified (revised), or when a modified template is created from a template, the reason for the change and the history of the change can be saved. Therefore, the change history and reasons for the change to the template can be checked as needed, and it is possible to understand who changed the template, when, and for what reason, and as a result, the reasons for using each item and the values ​​of each item can be made transparent and inherited in the created document. Furthermore, based on the generation transaction attached as metadata to each document, information about the template used for each document can be managed, and the reasons for changes to the template can be traced back based on the template's change history, making it possible to explain the reasons for the structure of the created document. Since the template used for a document and past documents can be identified, the reproducibility of documents created using the language model M is improved. Also, if a problem occurs in a document, the template used to create that document and past documents can be identified, enabling early resolution of the problem and prevention of recurrence. Since it is possible to create documents according to the past change history of the template, the quality of documents can be maintained and improved, and third-party verification can be performed during document audits, improving third-party auditability.

[0069] (Embodiment 5) Embodiment 3 describes a configuration in which item information of created documents is stored in a document database as shown in Figure 10. Embodiment 4 describes a configuration in which, when a template and a modified template are changed, change log information is stored in a template change history database and an operation history information database corresponding to each template. This embodiment describes an information processing device that can manage traceability in document creation based on a template, a template change history database, an operation history information database, and a document database. The information processing device of this embodiment has the same configuration as the information processing device 10 of Embodiment 1 shown in Figure 1, so the explanation of the configuration is omitted. In this embodiment, the template in the business requirements definition document, the template change history database and operation history information database of the said template, and the document database in which item information of past documents created using the said template are stored are stored in the storage unit 12.

[0070] In this embodiment, when a template is modified, information about the changes is stored in the template change history database, and information about the history of the changes is stored in the operation history information database. Therefore, the template modification process (evolution process) is accumulated in the template change history database and the operation history information database, making it possible to manage the template's evolution process (lifecycle). Furthermore, when a new document is created, a generation transaction containing information about the template used, information about past documents used, and information about items used from past documents is added to the new document as metadata. Therefore, the template used to create the new document and the past documents (information source documents) used can be identified from the generation transaction of the new document. In addition, the template change history database and the operation history information database of the identified template can be used to obtain the template's change history (change log information). Therefore, information about the created document can be traced retrospectively, achieving traceability.

[0071] The following describes the process of displaying the change history of the template used to create a document, as well as item information from other past documents that used the same template, when editing an existing document. Figure 18 is a flowchart showing an example of the document editing process, and Figure 19 is an explanatory diagram showing an example of how the document is displayed.

[0072] The control unit 11 of the information processing device 10, for example via the input unit 14, receives a specification of a document to be edited and reads the specified document from the storage unit 12 (S91). The control unit 11 displays the read document on the display unit 15 (S92). For example, the control unit 11 displays the left half of the screen in Figure 19 as the document. Based on the generation transaction attached as metadata to the document, the control unit 11 identifies the template and past documents (information source documents) used to create the document (S93). The control unit 11 generates comments regarding the identified template and past documents (S94), and adds the generated comments to the displayed document and displays it (S95). The control unit 11 generates a comment for each item in the template. For example, the control unit 11 generates a comment C5 that includes information such as the template ID and creation date of the template used, and information such as the document name, creation date, and creator of the past document used, and displays it as shown in Figure 19.

[0073] In the example in Figure 19, comment C5 is provided with a button to view the template's change history and a button to view other past documents. The control unit 11 determines whether the template's change history viewing button (labeled as the change history viewing button in Figure 18) has been operated (S96). If the control unit 11 determines that the template's change history viewing button has been operated (S96: YES), it reads the template's change history information from the template's change history DB and operation history information DB (S97) and displays the read change history information (S98). The change history information is displayed as shown in comment C6 in Figure 19 and includes the change date, the information before the change, and the reason for the change as part of the template's changes. Comment C6 is provided with a "Revert to Previous State" button and an exit button. If the user wants to revert to the state before the change, considering the information before the change and the reason for the change, they operate the "Revert to Previous State" button. If the "Revert to Previous State" button is operated, the control unit 11 reverts the corresponding item in the currently displayed document to its state before the change, changes the content of comment C5 to the information of the template before the change, and ends the display of comment C6. If the exit button for comment C6 is pressed, the control unit 11 will terminate the display of comment C6 without taking any further action.

[0074] If the control unit 11 determines that the template change history viewing button has not been operated (S96: NO), it skips steps S97 to S98. The control unit 11 determines whether the "View Other Past Documents" button in comment C5 (labeled as the document viewing button in Figure 18) has been operated (S99). If the control unit 11 determines that the "View Other Past Documents" button has been operated (S99: YES), it reads information about other past documents from the document DB corresponding to the template ID of the template in question (S100). The control unit 11 displays the information about the other past documents (S101). The information about the other past documents is displayed as shown in comment C7 in Figure 19, and includes the document name, creation date, creator, and the values ​​of the corresponding items. Comment C7 displays information for two past documents, and an "Adopt" button is provided for each past document. The user operates the "Adopt" button if they want to adopt the values ​​from the other past documents. When the "Adopt" button is pressed, the control unit 11 changes the value of the corresponding item in the currently displayed document to the value of another past document and terminates the display of comment C7. When the "End" button for comment C7 is pressed, the control unit 11 terminates the display of comment C7 without taking any action. In addition to displaying other past documents, the control unit 11 may also output the data to a file or transfer the data to a specified device.

[0075] The control unit 11 determines whether the overwrite save button in the screen shown in Figure 19 has been operated (S102). If it determines that the button has been operated (S102: YES), it overwrites the currently displayed document with the contents of the document stored in the storage unit 12 (S103). The control unit 11 also stores the information of each item in the overwritten document in the document database prepared in the storage unit 12 (S104). As a result, the item information of the overwritten document is accumulated in the document database. If the control unit 11 determines that the overwrite save button has not been operated (S102: NO), i.e., if the cancel button has been operated, it terminates the process.

[0076] Through the process described above, information about the template used to create the document and information about past documents can be identified based on the generation transaction attached as metadata to the document. Furthermore, since template change log information is stored in the template change history DB and operation history information DB, a third party can retrospectively refer to the change history (set of change events) of the template used to create the document. In addition, since the template ID of the template and the item information of past documents created using that template are stored in the document DB, other past documents that used each template can be identified. Therefore, from a template, other past documents that used that template can be identified, and information about each item in those past documents can be referenced. Thus, in this embodiment, it is possible to refer to documents created using a template from a template, and it is also possible to refer to the template used to create that document and past documents from a document, realizing bidirectional referencing. In this way, due to the relationship in which the records on the template side and the records on the document side refer to each other based on identification information, it is possible to retrospectively identify the corresponding template (template ID, version, and change history) from any document, and to identify the set of documents created using that template in a forward direction from the template side. The identification information does not necessarily have to be a unique identifier (ID), as long as it is information that can identify the template or document. For example, it may be configured to use the file name, creation date and time, creator, hash value based on the document content, or a combination of these.

[0077] By adding metadata to each document, including information about the template used and information about past documents (source documents), it becomes possible to track which template each document was generated from and from which values ​​in which source documents. This significantly improves the reproducibility and explainability of the document creation process, reducing the effort required for document quality assurance and auditing. Furthermore, by accumulating template change log information, the change history and reasons for changes during operation are preserved, allowing subjective judgments and empirical corrections to be formalized as knowledge assets and reflected in future template designs. This enables continuous improvement of the consistency and accuracy of created documents. Additionally, by establishing a bidirectional reference relationship between documents and templates, it is possible to automatically analyze which documents are affected when a template is changed, and which template version a particular document is based on. This improves the efficiency of template change management and auditing, and suppresses the misuse of templates or the creation of inconsistent documents.

[0078] The following describes the process of automatically modifying other templates that are affected when one template is changed. Figure 20 is a flowchart of an example of the template modification process. The process shown in Figure 20 is the same as the process shown in Figure 15, but with steps S111 to S112 added after step S71, and steps S113 to S114 added after step S74. The same steps as in Figure 15 are omitted from the explanation. Steps S61 to S68 in Figure 15 are not illustrated in Figure 20.

[0079] In the process shown in Figure 20, after step S71, the control unit 11 identifies other templates affected by the changes made to the template in step S67 (S111). For example, the control unit 11 identifies other templates containing the changed items based on the template change history DB of each template. The control unit 11 then applies the same changes made in step S67 to the identified other templates and updates the change log information of those templates (S112). Here, the control unit 11 stores information about the changes made to the other templates in the template change history DB of those templates, and stores information about the history of the changes made to those templates in the operation history information DB of those templates.

[0080] Furthermore, after processing in step S74, the control unit 11 performs the same processing as in steps S111 to S112 (S113 to S114). This allows other templates affected by the changes in the modified template from the original template to the modified template to be identified, the same changes to the identified other templates to be made, and the change log information of the other templates to be updated. This ensures that when a template is changed, the same changes are made to all other templates affected by this change, reducing the burden of template modification work. For example, if the imperial era changes, and one template is changed from the old era to the new era, the same change can be reflected in all other templates that include the era name item. In addition, information indicating that the changes were made automatically may be added to the change log information of other templates that have been automatically changed in conjunction with a change to a particular template. This allows the system to display that the changes were made automatically and the reason for the changes when a new document is created using this template, corresponding to the automatically changed items. For example, for each item in the hearing sheet shown in Figure 12, the message "Changed in conjunction with the change to template LL" and the wording before and after the change may be displayed as comments. In this case, the system may be configured to allow the user to choose whether or not to accept the automatically modified content. Alternatively, the message "There are changes due to a change in template LL. The changed items are MM and NN" may be added and displayed at the beginning of the hearing sheet shown in Figure 12. Alternatively, a comment indicating that there are changes may be added to the template without actually making any changes to the template itself. Furthermore, when displaying an existing document, a comment indicating that there are changes may be added to the corresponding items.

[0081] As mentioned above, if changing one template automatically changes other templates, it will affect documents created using those other templates. Therefore, when such a document is displayed, comments may be added to the document indicating that the items that were automatically changed in the template were changed and explaining the reason for the change. This will notify the user that the document needs to be changed because other templates have been modified.

[0082] In this embodiment, if improvement suggestions, exception handling, and requests for a template arise during its operation, this information may be stored in the operation history information DB as template change log information (operation history). For example, the control unit 11 of the information processing device 10 receives improvement suggestions, exception handling, and requests for any template via the input unit 14, and stores the received information in the operation history information DB as operation history information. This allows, for example, the template improvement cycle carried out within an organization to be accumulated as explicit knowledge and operational knowledge, making it reusable. In this case, when creating a new document for a new project that is similar in purpose or attributes to each template, it becomes possible to recommend the template to be used and recommend values ​​for each item based on the change history of that template, based on the operation history of each template.

[0083] Figure 21 is a flowchart showing an example of the template recommendation processing procedure, and Figure 22 is an explanatory diagram showing an example of template display. The control unit 11 of the information processing device 10 receives input of the purpose or attributes of a newly created document, for example, via the input unit 14 (S121). The purpose includes the content of the project and the recipient of the document. The attributes include the project's objective, duration, related companies, budget, the affiliation and attributes of the document creator, etc. Based on the received purpose and attributes, the control unit 11 searches the operation history information DB for each template and searches for templates with similar purposes or attributes (S122). In addition, if issues or requests arising in the project are entered as purposes or attributes, the control unit 11 may search for templates in which these issues or requests have been resolved.

[0084] If the control unit 11 finds at least one template, it reads the found template from the storage unit 12 and reads the operation history information for that template from the operation history information DB (S123). If the control unit 11 cannot find a template, it may display a message such as "No similar projects were found" and terminate the process. The control unit 11 displays the read template and operation history information on the display unit 15 (S124). For example, the control unit 11 displays a screen as shown in Figure 22. In the screen of Figure 22, the template is displayed on the left and the operation history information is displayed on the right. In the example of Figure 22, operation history information indicating that there was a request for the template is displayed. Also in the example of Figure 22, a template change history viewing button is provided, and when the template change history viewing button is operated, the change history of the displayed template is displayed, similar to comment C6 in Figure 19. When the template change history is displayed, it is possible to return the template to its state before it was displayed.

[0085] The screen in Figure 22 is equipped with an "Adopt" button to instruct the user to create a document using the currently displayed template, and a "Cancel" button to instruct the user not to create a document. The control unit 11 determines whether the "Adopt" button has been operated (S125). If it determines that the button has been operated (S125: YES), it identifies the currently displayed template as the template to be used (S126). If the control unit 11 determines that the "Adopt" button has not been operated (S125: NO), i.e., if the "Cancel" button has been operated, it terminates the process. Through the above process, when a new project is launched and a new document is created, it is possible to recommend a template used to create a past document created in a past project with a similar situation to the new project. Furthermore, if there have been improvement suggestions, exception handling, or requests related to the template, these can be presented along with the template to show what kind of situation occurred in the past and how it was resolved. Therefore, if any trouble has occurred in the past, the solution to this trouble can be recommended along with the template, and the past trouble-solving method can be reused as operational knowledge. The control unit 11 identifies a template to be used to create a new document, and then executes the processes shown in steps S18 to S20 of Figure 2, for example. This creates a new document based on the identified template.

[0086] Traditionally, documents have been created based on templates. The technology disclosed in Patent Document 1 only verifies whether a document was created based on an appropriate template, and does not include a mechanism for structurally recording and reusing the history of changes and improvements to the template itself during operation, the reasons for those changes, and operator information. Furthermore, the relationship between the generated document and the template is limited to one-way referencing, and it is not possible to retrospectively identify which template, which version, and which item was used from the document side. Therefore, challenges remain in terms of improving the quality of generated documents, ensuring accountability, modifying templates, and reusing modified templates. In such conventional technologies, templates are treated statically, the operational improvements and reasons for changes to templates are not made into assets, and it is impossible to trace the basis for the structure of the created document.

[0087] In contrast, the technology disclosed here stores modification, changes, and reasons for changes during the template's operation as change log information (event log), and manages the created document, the template used to create the document, and its change history in a bidirectional relationship. This enables traceability that allows retrospective tracking of the reasons for the structure of the created document, and allows for continuous improvement of the template. Therefore, by managing dynamically and continuously evolving templates, the quality of documents created using templates can be guaranteed. This improves the consistency, explainability, and reproducibility of quality in document creation, and makes it possible to technically guarantee the auditability and reusability of the entire process in automated document creation using the language model M. Therefore, it becomes possible to ensure accountability for documents, verify and audit by third parties, and automatically optimize the document quality improvement cycle by linking with the language model.

[0088] The technology disclosed herein automates consistency checks during the creation of new documents by managing templates, template change history databases, operation history information databases, and document databases in association with each type of document. Furthermore, since the information is saved append-only to the template change history database and operation history information database, tampering detection and rollback processing of the change history become easier, thereby increasing the reliability of the document creation system.

[0089] The technology disclosed herein allows the template and document database (past documents) used during document creation to be referenced as input constraint information to the language model M. Therefore, if the language model M outputs an item that does not exist in the template or an inappropriate value type (date, amount, name, etc.), it can automatically detect the inconsistency based on the specifications of the document database and template, and regenerate the document or issue a warning. This configuration suppresses the occurrence of hallucination in the language model M and makes it possible to technically ensure the reliability of the output document. The technology disclosed herein enables the management of dynamic evolution of templates, the management of traceability in the document creation process, the appropriateness of input values ​​for each item based on past documents, and the reuse of template change history by linking the management of templates and the management of operational history (change history) in the document creation system. Therefore, it greatly contributes to the quality assurance and continuous improvement of documents in document generation utilizing the language model M.

[0090] In each of the embodiments described above, the documents to be created can be various types of documents. For example, documents that record the background of a decision-making process may be created, such as memos on the decision-making process, memos on opposing opinions and concerns raised in meetings, meeting minutes, and retrospective reports on failed projects. In this case, new documents and a document database can be created using the document templates described above, and a template change history database and an operation history information database can be generated, storing the template change log information. In this case, these can be used as a reference when making similar decisions, and the repetition of the same mistakes can be prevented. Documents related to meeting minutes of verbal exchanges with customers, draft history of proposals to customers, and pre-contract negotiation history with customers may also be created. In this case, the templates generated from the documents, the new documents and document database using the templates, the template change history database and operation history information database can be used as evidence if a dispute arises with a customer. Documents that record changes in business processes may also be created, such as the history of changes in business flows, the reasons for introducing or discontinuing tools and their effects, and the revision history and background of manuals. In this case, the content of business improvements can be accumulated, and the tendency for business improvements to become dependent on individuals can be suppressed. Furthermore, documents that visualize the company's culture and values ​​may be created, such as records of internal events (e.g., photos, impressions, objectives), employee feedback from surveys and interviews, and messages from management regarding their vision and policies. In this case, the transmission of company culture and the sharing of values ​​within the company become possible. Documents recording experiments and verifications, such as logs of pilot measures (hypotheses, methods, and results), records of small-scale PoCs (Proof of Concept), and unadopted ideas, may also be created. In this case, a foundation for retrying and the reuse of ideas become possible. Documents that record the content of verbal instructions within the company, and documents that codify unwritten rules and customs within the company, may also be created. In this case, the locus of responsibility can be clarified, and verification can be conducted at a later date.

[0091] The document creation support processes described in each of the embodiments above are particularly effective in industries where the digitalization of various documents is considered to be lagging. Examples of industries where document digitalization is lagging include the apparel industry, agriculture, forestry and fisheries, education, construction, manufacturing, and wholesale and retail. These industries have structural factors that make the introduction of digitalization difficult, and their DX (Digital Transformation) is considered to be lagging due to a focus on on-site operations and a high degree of reliance on individual expertise. Industries where DX is lagging tend to have a large amount of paper-based work and a large volume of documents, so the technology disclosed here will yield remarkable results. Furthermore, the technology disclosed here is also highly effective in industries with a large variety and volume of documents. Examples of industries with a large volume of documents include the legal and administrative industries, the medical and nursing care industries, the construction and real estate industries, the pharmaceutical and chemical industries, the finance and insurance industries, and the education industry. These industries have a large amount of documents that are legally and institutionally required to be recorded, and the complexity and multi-layered nature of their operations also tend to result in a large volume of documents. Industries with a large volume of documentation will see greater benefits from this disclosed technology, including reduced burdens in document creation and management.

[0092] In the embodiments described above, various processes such as extracting information for each item from a sample document, creating templates and interview sheets, and creating documents were explained using the same language model M. Alternatively, a separate language model may be provided for each process. For example, a language model fine-tuned using diverse documents may be used for each process. Furthermore, metadata regarding the language model M used to create the document may be added to the created document.

[0093] The matters described in each embodiment can be combined with each other. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. In addition, the claims use a form in which claims referencing two or more other claims (multi-claim form), but are not limited to this. A form in which multi-claims referencing at least one multi-claim (multi-multi-claim) may also be used.

[0094] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the invention is indicated by the claims, not in the sense described above, and all modifications within the sense and scope equivalent to the claims are intended. [Explanation of Symbols]

[0095] 10 Information Processing Devices 11 Control Unit 12 Storage section 13 Communications Department 14 Input section 15 Display section 16 Reading Unit M language model

Claims

1. Using a language model from documents of the same type as the document to be created, a template with multiple items based on the said documents of the same type is created. Obtain information on the aforementioned multiple items, Based on the created template and the acquired information on the multiple items, a language model is used to create a new document to be created. An information processing method in which a computer performs the processing.

2. Using a language model, create a template with multiple items that are considered necessary to include in the document to be created. Obtain information on the aforementioned multiple items, Based on the created template and the acquired information on the multiple items, a language model is used to create a new document to be created. An information processing method in which a computer performs the processing.

3. Using a language model from the aforementioned similar documents, a hearing sheet with multiple items is created. Based on the created hearing sheet, information on the multiple items is obtained. Based on the template and the hearing sheet in which information for the multiple items is embedded, a language model is used to create the document to be created. The information processing method according to claim 1, wherein the processing is performed by the computer.

4. Information for the aforementioned multiple items is obtained from a reference document that contains information deemed necessary to be included in the document to be created. The information processing method according to claim 1 or 2, wherein the computer performs the processing.

5. Using a language model, information for the multiple items is obtained from the aforementioned similar documents. Based on the information obtained for the aforementioned multiple items, a language model is used to create a template containing the aforementioned multiple items. The information processing method according to claim 1, wherein the processing is performed by the computer.

6. The information obtained for the aforementioned multiple items is divided into the item name and input information for each item. Based on the acquired information for the multiple items and the classification results for each piece of information, the language model is used to create the template. The information processing method according to claim 5, wherein the computer performs the processing.

7. The memory unit stores information about the aforementioned similar documents and input information for each item used in the aforementioned similar documents. Comments relating to the input information for each item stored in the memory unit are added to each item of the hearing sheet into which the aforementioned information is embedded. The information processing method according to claim 3, wherein the computer performs the processing.

8. The memory unit stores information about the aforementioned similar documents and input information for each item used in the aforementioned similar documents. Comments relating to the input information of each item stored in the storage unit are added to each item of the document to be created. The information processing method according to claim 1, wherein the processing is performed by the computer.

9. The created template and the modified template created based on the template are stored in the memory unit. Using the template created or the modified template, a new document to be created is generated. The information processing method according to claim 1 or 2, wherein the computer performs the processing.

10. Based on the change instructions for the aforementioned template or the aforementioned modified template, the aforementioned template or the aforementioned modified template is modified, The storage unit stores change log information, including the date and time of the change, the person who made the change, the content of the change, and the reason for the change, related to the change to the template or the modified template. The information processing method according to claim 9, wherein the computer performs the processing.

11. The memory unit stores information about documents of the same type as the document to be created, and input information for each item used in said documents of the same type. A generation transaction is generated that includes the identification information and version information of the template used to create the document to be created, as well as the identification information of past documents that serve as the source of input information for each item. The aforementioned generated transaction is added as metadata to the document to be created. The information processing method according to claim 10, wherein the processing is performed by the computer.

12. Based on the generation transaction attached as metadata to the document to be created, the identification information and version information of the template used to create the document to be created, as well as the identification information of past documents that serve as the source of input information for each item, are obtained. Based on the aforementioned generated transaction, identify other documents created using the template, Output the aforementioned other documents together with the document to be created. The information processing method according to claim 11, wherein the processing is performed by the computer.

13. The operation history related to the operation of multiple templates is stored in the storage unit. Based on the purpose and attributes of the document to be created and the operational history of the multiple templates, the system recommends a template to be used to create the document. The information processing method according to claim 11, wherein the processing is performed by the computer.

14. A template containing multiple items based on the document of the same type as the document to be created is stored in the memory unit, and is created using a language model from the same type of document as the document to be created. Obtain information for the multiple items included in the template, Based on the template and the information obtained for the multiple items, a language model is used to create a new document to be created. An information processing method in which a computer performs the processing.

15. The document to be created is generated by inputting a prompt to the language model that includes the template information and the information of the same type of document. The information processing method according to claim 14, wherein the processing is performed by the computer.

16. Using a language model from documents of the same type as the document to be created, a template with multiple items based on the said documents of the same type is created. Obtain information on the aforementioned multiple items, Based on the created template and the acquired information on the multiple items, a language model is used to create a new document to be created. A program that instructs a computer to perform a process.

17. In an information processing device having a control unit, The control unit, Using a language model from documents of the same type as the document to be created, a template with multiple items based on the said documents of the same type is created. Obtain information on the aforementioned multiple items, Based on the created template and the acquired information on the multiple items, a language model is used to create a new document to be created. Information processing device.

Images