Information processing device and information processing method

Abstract

This invention provides an information processing device and information processing method that reduce the effort required for user preparation while supporting audits of audit target files regardless of the format of the audit target files. [Solution] The document audit process extracts text from the audit target file, generates prompts to extract descriptions corresponding to the check items required for the audit from the extracted text and output them as tabular data, obtains the tabular data by inputting the prompts into the language model, inputs the descriptions shown in the obtained tabular data into an audit checklist for the check items required for the audit, and outputs the audit checklist with the descriptions entered.

Description

Technical Field

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

Background Art

[0002] As the background art of this technical field, there is Japanese Patent Application Laid-Open No. 2024-91362 (Patent Document 1). This publication describes that a medical audit for recording a procedure in a beauty clinic and auditing whether the procedure is performed according to a specific procedure has a function of generating a model for each procedure from an image obtained by photographing the procedure, a function of analyzing a newly acquired video, and a function of auditing whether the procedure in the video is performed as described in the above procedure (see the abstract).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The items in the document to be checked in document auditing differ depending on the type (format) of the document. Therefore, in auditing various types of documents, in order for the auditor to extract the descriptions corresponding to the items to be checked, experience and knowledge are required. As a result, there is a risk of the vulgarization of the audit work and the concealment of the know-how of the audit work.

[0005] For example, when mechanizing document auditing using RPA (Robotic Process Automation), it is necessary for the user to specify in advance the position information on the document where the items to be checked are described for each document format.

[0006] Furthermore, the technology described in Patent Document 1 involves feeding video footage of a procedure to a learning model to audit whether the procedure is being performed correctly, but generating such a learning model requires a large amount of training data.

[0007] Thus, conventional technologies require a great deal of preparation by the user to enable document auditing, such as specifying location information for each document format and preparing a large amount of training data.

[0008] Therefore, one aspect of the present invention reduces the effort required for user preparation while providing audit support for audit target files regardless of the format of the audit target files. [Means for solving the problem]

[0009] To solve the above problems, one aspect of the present invention adopts the following configuration. The information processing device comprises a processor and a storage device, the storage device holds an audit target file, item information indicating check items necessary for auditing, and an audit checklist relating to the check items, the processor extracts text from the audit target file, generates a prompt to extract descriptions corresponding to the check items indicated by the item information from the extracted text and output them as tabular data, the generated prompt is input to a language model to obtain the tabular data output from the language model, the descriptions indicated by the obtained tabular data are input to the audit checklist, and the audit checklist with the descriptions entered is output. [Effects of the Invention]

[0010] According to one aspect of the present invention, it is possible to reduce the effort required for user preparation while providing audit support for audit target files regardless of the format of the audit target files.

[0011] Other issues, configurations, and effects not mentioned above will be clarified by the following description of the embodiments. [Brief explanation of the drawing]

[0012] [Figure 1] This is a block diagram showing an example configuration of the document audit support device in Example 1. [Figure 2] This is an explanatory diagram showing an example of an audited file in Example 1. [Figure 3] This is an example of a set table of required items in Example 1. [Figure 4] This figure shows an example of the data structure of the audit checklist in Example 1. [Figure 5] This flowchart shows an example of the document audit process in Example 1. [Figure 6] This is an explanatory diagram showing an example of the screen configuration of the main screen in Example 1. [Figure 7] This figure shows an example of a question sent to the generation AI server in Example 1. [Figure 8] This figure shows an example of a response sent from the generation AI server in Example 1. [Figure 9] This figure shows an example of the data structure of the audit checklist with various confirmation results and judgment results entered in Example 1. [Figure 10] This is an explanatory diagram showing an example of an audited file with a mark in Example 1. [Modes for carrying out the invention]

[0013] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In this embodiment, the same components will be denoted by the same reference numerals in principle, and repeated descriptions will be omitted. It should be noted that this embodiment is merely one example for realizing the present invention and does not limit the technical scope of the present invention. [Examples]

[0014] FIG. 1 is a block diagram showing a configuration example of a document audit support device 100. The document audit support device 100 is configured by, for example, a computer having a CPU (Central Processing Unit) 101, a memory 102, an auxiliary storage device 103, a network I / F (InterFace) 104, an input I / F 105, and an output I / F 106.

[0015] The CPU 101 is an example of a processor and executes a program stored in the memory 102. The memory 102 includes a non-volatile storage element ROM (Read Only Memory) and a volatile storage element RAM (Random Access Memory). The ROM stores invariant programs (e.g., BIOS (Basic Input / Output System)). The RAM is a high-speed and volatile storage element such as DRAM (Dynamic Random Access Memory) and temporarily stores the program executed by the CPU 101 and the data used during program execution.

[0016] The auxiliary storage device 103 is a large-capacity and non-volatile storage device such as a magnetic storage device (HDD (Hard Disk Drive)) or a flash memory (SSD (Solid State Drive)), and stores the program executed by the CPU 101 and the data used during program execution. That is, the program is read from the auxiliary storage device 103, loaded into the memory 102, and executed by the CPU 101.

[0017] The network I / F 104 is an interface that controls communication with other devices according to a predetermined protocol. Further, the network I / F 104 may include a serial interface such as USB (Universal Serial Bus).

[0018] The document audit support device 100 is connected to a network 160, such as the Internet, using a network interface 104. The document audit support device 100 is connected to a generative AI (Artificial Intelligence) server 150 via the network 160.

[0019] The input interface 105 is an interface for connecting the input device 107 and the document audit support device 100. The input device 107 is a device that accepts user input such as a keyboard or mouse. The output interface 106 is an interface for connecting the display device 108 and the document audit support device 100. The display device 108 is a device that outputs the execution results of a program, such as a display or printer, in a format that can be viewed by the user.

[0020] Some or all of the program executed by the CPU 101 may be provided to the document audit support device 100 via the network 160 from an external computer equipped with a non-temporary storage medium such as removable media (CD-ROM, flash memory, etc.) or non-temporary storage device, and stored in a non-volatile auxiliary storage device 103, which is also a non-temporary storage medium. For this reason, the document audit support device 100 may have an interface for reading data from the removable media.

[0021] The document audit support device 100 is a computer system that is configured on a single physical computer or on multiple logically or physically configured computers, and may operate in separate threads on the same computer, or may operate on a virtual computer built on multiple physical computer resources.

[0022] The CPU 101 includes, for example, a text extraction unit 111, a question generation unit 112, an audit check unit 113, and a file editing unit 114. The text extraction unit 111 extracts text from the audit target file, which is a document file containing text. The question generation unit 112 generates question sentences to extract descriptions corresponding to the check items necessary for auditing the audit target file from the text extracted by the text extraction unit 111, and sends the generated question sentences (prompts) to the generation system AI server 150. Hereinafter, the check items necessary for auditing the audit target file will also be referred to as required items.

[0023] The generative AI server 150 has a language model such as an LLM (large language model), and generates an answer to the received question using the language model and transmits it to the document audit support device 100. For example, the document audit support device 100 may have a language model such as an LLM (local LLM), in which case the document audit support device 100 does not need to be connected to the generative AI server 150, and generates the answer by inputting the question into its own language model.

[0024] The audit check unit 113 audits the audit target file according to the response received from the generation AI server 150 and the user input to the input device 107. The file editing unit 114 performs editing on the audit target file, such as marking the text corresponding to the extracted required items.

[0025] For example, the CPU 101 functions as a text extraction unit 111 by operating according to a text extraction program loaded into memory 102, and functions as a question generation unit 112 by operating according to a question generation program loaded into memory 102. The relationship with the program is similar for other functional units included in the CPU 101.

[0026] Furthermore, some or all of the functions of the functional unit of the document audit support device 100 may be implemented by dedicated hardware such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array).

[0027] The auxiliary storage device 103 includes, for example, an audit target file DB (DataBase) 121 and an audit target file definition DB 122. The audit target file DB 121 stores audit target files and audit target files edited by the file editing unit 114 in association with each other. The audit target file definition DB 122 stores information showing a list of items necessary for auditing audit target files and an audit checklist showing items to be confirmed during the audit of audit target files, associated with each type of document for audit target files.

[0028] In addition, some or all of the information stored in the auxiliary storage device 103 in Figure 1 may be stored in the memory 102, or in an external server connected to the document audit support device 100.

[0029] In this embodiment, the information used by the document audit support device 100 is not dependent on a specific data structure and may be represented in any data structure. For example, a data structure appropriately selected from a table, list, database, or queue can store the information.

[0030] Figure 2 is an explanatory diagram showing an example of an audited file. Document 200 is an example of an audited file and is stored in audited file DB121. Document 200 is a document file containing text. The type of document 200 in Figure 2 is "Program Change Proposal," as described at the beginning of document 200.

[0031] Figure 3 shows an example of the required items set table 300. The required items set table 300 is defined for each type of document of the audited file and is stored in the audited file definition DB122. In the example in Figure 3, the required items set table 300 is for a "program change proposal" ("program change proposal" is an example of a type of document of the audited file).

[0032] The required items set table 300 includes, for example, a number column 301 and a required items column 302. The number column 301 indicates a group of required items. The required items column 302 indicates required items belonging to the same group. In the required items set table 300, required items belonging to the same group are related required items.

[0033] Figure 4 shows an example of the data structure of audit checklist 400. Audit checklist 400 is defined for each type of document of the audit target file and is stored in the audit target file definition DB122. In the example in Figure 4, audit checklist 400 is for "Program Change Proposal" ("Program Change Proposal" is an example of a type of document of the audit target file).

[0034] The audit checklist 400 includes, for example, a number column 401, a check content column 402, a confirmation result column 403, a confirmation result column 404, and a judgment result column 405. The number column 401 indicates a group of required items. The check content column 402 indicates the check content for the audited file. The check content indicated in the check content column 402 includes, for example, checking whether the required items (i.e., related required items) belonging to the group indicated by the corresponding number in the audit checklist 400 are described in the audited file.

[0035] The "Confirmation Results" column 403 shows the confirmation result regarding whether the required items belonging to the group indicated by the corresponding number in the audit checklist 400 are included. Specifically, for example, the confirmation results column 403 will contain an excerpt of the entries related to the required items belonging to the group indicated by the corresponding number in the audit checklist 400, along with information indicating whether or not those entries are included.

[0036] The Content Verification Result Column 404 shows the verification results regarding the content of the required items belonging to the group indicated by the corresponding number in the Audit Checklist 400. Specifically, for example, the Content Verification Result Column 404 will contain information indicating other supporting documents necessary to verify the content of the required items belonging to the group indicated by the corresponding number in the Audit Checklist 400, and information indicating the validity of the content, as input as verification results.

[0037] The judgment result column 405 shows the judgment result obtained by combining the confirmation result shown in the presence / absence confirmation result column 403 and the confirmation result shown in the content confirmation result column 404.

[0038] The required items set table 300 and the check contents column 402 of the audit checklist do not contain information such as the location of the required items within the audited file, and the values ​​in each column are defined in natural language. In other words, the document audit support device 100 of this embodiment can perform checks on the required items of the audited file with high accuracy, even without being given information on the location of the check items, that is, regardless of the format of the audited file.

[0039] Figure 5 is a flowchart illustrating an example of the document audit process. Before the document audit process begins, the required items set table 300 is configured. Also, before the document audit process begins, the information in the number column 401, the information in the check content column 402, and the information regarding other required supporting documents in the content confirmation result column 404 of the audit check sheet 400 are configured. Furthermore, before the document audit process begins, the audit target files are registered in the audit target files DB121.

[0040] The text extraction unit 111, for example, displays the main screen 600 (described later) on the display device 108 and retrieves the audit target files from the audit target file DB 121 according to the input from the user to the input device 107 via the main screen 600 (S501).

[0041] Figure 6 is an explanatory diagram showing an example of the screen configuration of the main screen 600. The main screen 600 includes, for example, an input area 601, an input area 602, an input area 603, a display area 604, and a display area 605. The input area 601 is an area for specifying the audit target file to be acquired in step S501. The input area 602 is an area for specifying the save location of the audit checklist 900 in which various confirmation results and judgment results have been entered. The input area 603 is an area for specifying the save location of the marked audit target file, which will be described later.

[0042] Display area 604 is for displaying information indicating the type of document the audited file is. Display area 605 is for displaying the progress of the document audit process, etc.

[0043] Returning to the explanation of Figure 5, the text extraction unit 111 extracts text from the audit target file acquired in step S501 (S502). Specifically, for example, the text extraction unit 111 extracts text by selecting and copying all the text contained in the document that is the audit target file. Alternatively, for example, if the audit target file is image data, the text extraction unit 111 may extract text using a predetermined character recognition method such as OCR (Optical Character Recognition).

[0044] Furthermore, the text extraction unit 111 may add markdown to the extracted text that indicates, for example, the structure of the audited file as a document (for example, the structure of the table if the text is contained within a table, as in document 200 in Figure 2). The text extraction unit 111 may also perform predetermined preprocessing on the extracted text to improve the accuracy of the response by the generation AI server 150.

[0045] The question generation unit 112 obtains the required item set table 300 and the audit checklist 400 corresponding to the document type of the audit target file obtained in step S501 from the audit target file definition DB 122 (S503).

[0046] Furthermore, information indicating the type of document of the audited file is pre-associated with the audited file and stored in the audited file DB121, and the question generation unit 112 identifies the type from the audited file DB121. Alternatively, the question generation unit 112 may identify the type from the text within the audited file (for example, the type may be identified by the description "Program Change Proposal" at the beginning of document 200 in Figure 2, or by characters whose font size is greater than or equal to a predetermined value).

[0047] The question generation unit 112 generates a question sentence (a prompt to be input to the language model of the generation AI server 150) for each group indicated by the required item set table 300 obtained in step S503, to extract the description corresponding to the required item (i.e., related required item) included in that group from the text extracted in step S502 (S504).

[0048] Figure 7 shows an example of a question sentence 700 sent to the generation AI server 150. The question sentence 700 in Figure 7 is an example of a prompt for extracting descriptions corresponding to the required items belonging to the group numbered "1" in the example required item set table 300 in Figure 3. As the question sentence 700 shows, the question generation unit 112 requests that the response from the language model be output as tabular data.

[0049] Returning to the explanation of Figure 5, the question generation unit 112 sends the group-specific question sentences generated in step S504 to the generation AI server 150 (S505). The generation AI server 150 inputs the received question sentences (prompts) into its own language model and sends the answers output from the language model to the document audit support device 100. The question generation unit 112 receives the group-specific answers sent from the generation AI server 150 (S506).

[0050] Figure 8 shows an example of a response 800 sent from the generative AI server 150. The response 800 in Figure 8 is an example of a response output when the question sentence 700 in Figure 7 is input to the language model. In the example in Figure 8, the response 800 is written in JSON (JavaScript® Object Notation) format, which is an example of tabular data.

[0051] Returning to the explanation of Figure 5, the audit check unit 113 enters the entries corresponding to the required items indicated by the response received in step S507 into the entry confirmation result column 403 of the audit check sheet 400 (S507).

[0052] Figure 9 shows an example of the data structure of the audit checklist 900, which has been entered with various confirmation results and judgment results to be output in step S510, described later. The audit checklist 900 in Figure 9 is an example in which the information in the confirmation result column 403 (whether or not it is written), the information in the content validity confirmation result column 404 (content confirmation result column), and the information in the judgment result column 405 have been entered, compared to the example of the audit checklist 400 in Figure 4.

[0053] As shown in Figure 9, in step S507, for example, the audit check unit 113 extracts the entries corresponding to the required items for each group from the responses for each group, and inputs the extracted entries into the entry presence / absence confirmation result column 403 (the entry extraction column) corresponding to the group.

[0054] Furthermore, as shown in Figure 9, in step S507, for example, if the audit check unit 113 has extracted all the necessary items for each group, it inputs information indicating that the necessary items are present in the presence or absence confirmation result field 403 (the presence or absence field) corresponding to that group. Also, as shown in Figure 9, in step S507, for example, if there are necessary items that could not be extracted for each group, the audit check unit 113 inputs information indicating the necessary items that could not be extracted in the presence or absence confirmation result field 403 (the presence or absence field) corresponding to that group.

[0055] Returning to the explanation of Figure 5, the audit check unit 113 performs a check process on the items in the audit check sheet 400 that can be checked mechanically, and inputs the results of the check process into the entry status check result column 403 and / or entry content check result column 404 (S508).

[0056] Specifically, for example, regarding the check item "(6) Are the approval dates in the order (2) ≤ (4) ≤ (5)?" in the audit checklist 900 of Figure 9, since the order of the dates attached to the approval stamps numbered "2", "4", and "5" can be mechanically compared, "Date order OK" can be entered as information indicating the validity of the content in the content confirmation result column 404.

[0057] Furthermore, for example, if the document audit support device 100 has or is connected to employee roster data (including employee names and job titles) of the company to which the user belongs, it can refer to the employee roster data to check whether the "Proposal Manager" and "Development Manager" hold "manager positions or higher" when checking items such as "Is there an approval stamp of a manager or higher in the Proposal Source column" and "Is there an approval stamp of a manager or higher in the IT department?" in the audit checklist 900 in Figure 9. Therefore, it is possible to input "'Proposal Manager' is a manager-equivalent position" and "'Development Manager' is a manager-equivalent position" as information indicating the validity of the content in the content confirmation result column 404.

[0058] If there are no items that can be mechanically verified by the audit check unit 113, the process in step S508 is omitted.

[0059] The audit check unit 113 displays the audit check sheet 400 after the verification process in step S508 on the display device 108, obtains the verification results for items that have not been verified by user input into the input device 107, and inputs the obtained verification results into the presence / absence verification result column 403 and / or content verification result column 404 and judgment result column 405 of the audit check sheet 400 (S509).

[0060] The audit check unit 113 displays the audit check sheet 900, in which the confirmation results from steps S508 and S509 have been entered, on the display device 108 (S510). In step S510, the audit check unit 113 saves the audit check sheet 900, in which the various confirmation results and judgment results from steps S508 and S509 have been entered, to the audit target file DB 121, in association with the audit target file acquired in step S501.

[0061] The file editing unit 114 marks the strings corresponding to the responses received in step S506 (i.e., required items included in the text) in the audit target file acquired in step S501, and saves the audit target file with the marks (hereinafter also called the marked audit target file) as a separate file, associating it with the original audit target file (i.e., the audit target file acquired in step S501) in the audit target file DB 121 (S511). The marks are an example of identification information for identifying the required items included in the audit target file.

[0062] In steps S510 and S511, the original audit target file, the stamped audit target file, and the audit checklist 900 with various confirmation results and judgment results entered are associated and saved in the audit target file DB121. For example, the audit check unit 113 may retrieve the stamped audit target file and the audit checklist 900 with various confirmation results and judgment results entered, which are stored in association with each other, according to user input to the input device 107 and display them on the display device 108. This allows the user to refer to the stamped audit target file and the audit checklist 900 with various confirmation results and judgment results entered at any time, even after the document audit process is completed.

[0063] The file formats of the original audited file and the marked audited file may or may not be the same. For example, if the file editing unit 114 marks an audited file using a specific file format, unless the original audited file's file format is that specific format, it will generate a file by converting the original audited file's file format to that specific format and then perform the marking process on the generated file (if the original audited file's file format is that specific format, then there is no need to convert the file format).

[0064] The file editing unit 114 displays the stamped audit target files on the display device 108 (S512) and terminates the document audit process.

[0065] Figure 10 is an explanatory diagram showing an example of a marked audit file. Document 1000 is an example of a marked audit file output in step S512 and is stored in the audit file DB121, associated with the original audit file. Document 1000 is a version of Document 200 in Figure 2, with markings added to the text corresponding to the required items defined in the required item set table 300 in Figure 3.

[0066] As shown in Figure 10, the text extracted as required items is enclosed in a double-lined rectangle in document 1000, and a callout indicating the group of required items is attached to that rectangle (that is, the same callout is attached to required items (related check items) belonging to the same group, and different callouts are attached to required groups belonging to different groups). In other words, the excerpts corresponding to each number indicated in column 401 of audit checklist 900 correspond to the entries enclosed in rectangles with the corresponding numbered callouts in document 1000. Note that both the rectangle and the callout are examples of markings.

[0067] The rectangles included in Document 1000 (the stamped audited file) allow users to easily verify whether the extracted required items and their content are appropriate. Additionally, the callouts in Document 1000 allow users to easily identify which text within Document 1000 corresponds to the required items. Furthermore, because the original audited file and the stamped audited file are saved in association, users can retrieve both files together and easily compare them.

[0068] As described above, the document audit support device 100 of this embodiment extracts text from the audit target file and requests the generation system AI server 150 to extract the necessary items from the text. This allows the device to extract the necessary items from the audit target file using natural language definitions without requiring the user to set location information within the audit target file. Furthermore, since this embodiment does not use machine learning techniques, there is no need to prepare a large amount of training data (audit target files). Therefore, the document audit support device 100 of this embodiment can support document auditing while reducing the user's preparation effort.

[0069] Furthermore, the document audit support device 100 reduces the user's workload by inputting the extracted required items into the audit checklist 400 and automatically performing mechanical verification processes such as checking whether the required items are filled in and verifying the date.

[0070] Furthermore, the document audit support device 100 outputs not only the audit checklist 900 with various confirmation results entered, but also the audit target file with a stamp, allowing the user to easily check the contents of the necessary items corresponding to the various confirmation results within the audit target file.

[0071] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are described in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described. It is also possible to replace parts of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add configurations from other embodiments to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace parts of the configuration of each embodiment with other configurations.

[0072] Furthermore, each of the above configurations, functions, processing units, and processing means may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. Alternatively, each of the above configurations and functions may be implemented in software by having the processor interpret and execute programs that implement each function. Information such as programs, tables, and files that implement each function can be stored in memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

[0073] Furthermore, the control lines and information lines shown are those deemed necessary for explanatory purposes, and not all control lines and information lines are necessarily shown in the actual product. In reality, it is safe to assume that almost all components are interconnected. [Explanation of Symbols]

[0074] 100 Document audit support device, 101 CPU, 102 Memory, 103 Auxiliary storage device, 104 Network I / F, 105 Input I / F, 106 Output I / F, 107 Input device, 108 Display device, 111 Text extraction unit, 112 Question generation unit, 113 Audit check unit, 114 File editing unit, 121 Audit target file DB, 122 Audit target file definition DB, 150 Generation system AI server, 300 Required item set table, 400 Audit check sheet

Claims

1. An information processing device, Equipped with a processor and memory, The storage device holds the audit target file, item information indicating the check items necessary for the audit, and an audit checklist related to the check items. The aforementioned processor, Extract text from the aforementioned audited file, A prompt is generated to extract the entries corresponding to the check items indicated by the aforementioned item information from the extracted text and output them as tabular data. When the generated prompt is input to the language model, tabular data output from the language model is obtained. The information indicated in the acquired tabular data is entered into the audit checklist. An information processing device that outputs an audit checklist with the above information entered.

2. An information processing apparatus according to claim 1, The aforementioned audit checklist includes the content of the check regarding the validity of the information entered in the audit checklist, The aforementioned processor, The results of the checks regarding the validity checks are entered into the audit checklist. An information processing device that outputs an audit check sheet into which the aforementioned check results and the aforementioned description have been entered.

3. An information processing apparatus according to claim 2, Further comprising an input device and a display device, The aforementioned processor, The audit checklist with the above information entered is output to the display device. The input of the check results for the content of the validation check is received via the input device. An information processing device that inputs the check results received via the input device into the audit check sheet.

4. An information processing apparatus according to claim 2, The content of the validity check includes at least one of the following: a first check regarding the presence or absence of a description corresponding to the check item, and a second check regarding the date and time indicated by the description corresponding to the check item. The aforementioned processor, Based on the information entered in the aforementioned audit checklist, perform at least one of the aforementioned checks. An information processing device that inputs the results of at least one of the checks performed into the audit checklist.

5. An information processing apparatus according to claim 1, The aforementioned item information indicates the group to which each of the aforementioned check items belongs. The check items belonging to the same group are check items that are related to each other. The aforementioned audit checklist shows the check contents for each group, The aforementioned processor, For each of the aforementioned groups, a prompt is generated to extract the descriptions corresponding to the check items belonging to that group from the extracted text and output them as tabular data. When the group-specific prompts are input to the language model, the tabular data for each group output from the language model is obtained. An information processing device that inputs the descriptions shown in the tabular data for each group as check results for the check contents corresponding to that group in the audit checklist.

6. An information processing device according to claim 5, The aforementioned processor, To the audited file, identification information is assigned to identify each of the entries shown in the acquired tabular data, and the identification information indicates the group to which the check item corresponding to the entry belongs. An information processing device that stores the audit target file to which the aforementioned identification information has been added as a separate file from the audit target file before the aforementioned identification information was added to the storage device.

7. An information processing apparatus according to claim 6, The processor is an information processing device that stores in the storage device an audit target file to which the identification information has been assigned and an audit checklist in which the check results have been entered, in association with each other.

8. An information processing apparatus according to claim 7, Further comprising an input device and a display device, The processor is an information processing device that retrieves the audit target file to which the identification information is attached and which is stored in association with the data, and the audit check sheet to which the check results are entered, from the storage device and displays them on the display device, in accordance with the input to the input device.

9. An information processing apparatus according to claim 6, The aforementioned processor, The process of assigning the identification information to the audited file is performed using a specific file format. An information processing device in which the file format of the audited file to which the aforementioned identification information is assigned is the aforementioned specific file format.

10. An information processing method using an information processing device, The aforementioned information processing device has a processor and a storage device, The storage device holds the audit target file, item information indicating the check items necessary for the audit, and an audit checklist related to the check items. The aforementioned information processing method is The processor extracts text from the audited file, The processor generates a prompt to extract the entries corresponding to the check items indicated by the item information from the extracted text and output them as tabular data. The processor acquires tabular data output from the language model when the generated prompt is input to the language model. The processor inputs the description indicated by the acquired tabular data into the audit checklist. An information processing method wherein the processor outputs an audit checklist into which the above description has been entered.

Images