Task management device, task management device control method, and task management device control program

Abstract

Properly manage tasks related to tax and accounting work. [Solution] A task management device is provided, which is a device for managing tasks related to a business's tax and accounting work, comprising a task information generation unit that receives instruction information regarding task generation and generates task information, a display control unit that presents the task information to a user, and a task management unit that registers tasks based on the user's approval, wherein the task information generation unit generates the task information using public information including at least tax and accounting laws and regulations and notices, the business's past task information including at least one of the user's comments or revision history, and a generation AI model.

Description

[Technical Field]

[0001] The present invention relates to an apparatus for managing tasks relating to the tax and accounting operations of a business operator. [Background technology]

[0002] Conventionally, a system has been known in which task data relating to taxation is shared and managed between a manager such as an accountant or tax accountant and a person to be managed such as a sole proprietor or a client company (Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2018-128853 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the system described in Patent Document 1 assumes that either the administrator or the person being managed has at least the correct knowledge of tasks related to tax procedures, etc. However, because the tasks related to tax and accounting work are diverse and laws and regulations are frequently revised, it is necessary to always be familiar with the latest specialized knowledge, and task management is a heavy burden for both the administrator and the person being managed.

[0005] Therefore, an object of the present invention is to support the correct management of tasks related to tax and accounting work. [Means for solving the problem]

[0006] In order to solve the above problems, the device of the present invention is a device for managing tasks related to a business's tax and accounting work, and is equipped with a task information generation unit that receives instruction information regarding task generation and generates task information, a display control unit that presents the task information to a user, and a task management unit that registers tasks based on the user's approval, and is characterized in that the task information generation unit generates the task information using public information including at least laws and regulations related to tax and accounting, the business's past task information including at least one of comments or revision history by the user, and a generation AI model. [Effects of the Invention]

[0007] According to the present invention, tasks relating to tax and accounting work can be easily and correctly managed. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a diagram illustrating an example of the configuration of a system according to an embodiment of the present invention. [Figure 2] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 3] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 4] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 5] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 6] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 7] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 8] 1 is a diagram illustrating an example of a hardware configuration of a server 1 according to an embodiment of the present invention. [Figure 9] 2 is an example of a software configuration of a server 1 according to an embodiment of the present invention. [Figure 10] 1 is a diagram illustrating an example of a configuration of task information according to an embodiment of the present invention. [Figure 11] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 12] 1 is an example of a screen configuration according to an embodiment of the present invention. [Figure 13] FIG. 1 is a process flow diagram according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that the examples described below do not unduly limit the content of the present invention described in the claims. Furthermore, not all of the configurations described below are necessarily essential constituent elements of the present invention.

[0010] 1 is a diagram showing an example of the configuration of a task management support system according to one embodiment of the present invention. The task management support system includes at least a server 1 and a user terminal 2.

[0011] The present invention relates to a device for supporting tax and accounting task management for businesses, such as corporations and individuals running sole proprietorships. Businesses encounter a variety of tax and accounting tasks throughout the year. These tasks range from routine tasks that apply to almost all businesses, such as filing corporate tax returns within two months of the closing date, conducting year-end tax adjustments in December, and paying withholding income tax at the designated time, to ad hoc tasks such as submitting advance notification of executive bonuses and processing paperwork related to employee retirement, each of which often has a legal deadline.

[0012] The "user" in the present invention may be the business owner or a supporter such as a tax accountant's office or an accounting firm that supports the business owner. Business owners can use the device of the present invention to thoroughly identify their own tasks and manage them so that they are completed within the deadline, even if they lack specialized knowledge. Supporters such as tax accountant's offices or accounting firms can thoroughly identify the tasks of their clients and support their work so that they are completed within the deadline. Furthermore, the tax and accounting tasks in the present invention typically relate to accounting procedures, filing and reporting to tax offices, etc., but may also include tasks related to social insurance, labor insurance, pensions, human resources, and labor.

[0013] The basic functions of the task management support system according to the present invention are explained using Figures 2 to 7. Figures 2 to 7 show example screen configurations of the task management support system according to the present invention when the user is a supporter at a tax accountant's office, accounting firm, or other similar organization. Figure 2 shows an example of the configuration of the home screen of the application of the task management support system according to the present invention. The task status display field in the upper section displays the task status of all business clients of the supporter (user), and, as an example, displays the total number of unprocessed tasks. "Overdue Tasks" refers to the number of tasks that have exceeded the deadline set when the task was registered in the task management support system. "Tasks of the Month" refers to the number of registered tasks that have a deadline in the current month. "Tax Categories of the Month" refers to the number of tax categories related to tasks that have a deadline in the current month. Tax categories are conceptually representative of task classifications, such as "Corporate Tax Settlement and Filing" and "Consumption Tax Filing." Each tax category contains multiple specific tasks.

[0014] The bottom of the screen has a client task status display column, which shows the task status for each client. As an example, the client name, closing month, number of overdue tasks, number of tasks for this month, and number of tax categories for this month are aggregated and displayed for each client. In the example in Figure 2, for BBB Co., Ltd., there are four overdue tasks, six tasks for this month, and two tax categories for this month: "Corporate Tax Closing and Filing" and "Consumption Tax Filing." For AAA Co., Ltd. and CCC Co., Ltd., there are no overdue tasks or tasks for this month. This screen allows the user, who is the supporter, to grasp the task status of all clients at a glance across the board.

[0015] Figure 3 is an example of the screen layout that displays a list of clients for the user who is a supporter. The user can register the information of business clients in this system and register them as clients. When registering, the user can enter basic information about the client, such as the client's name, whether it is a corporation or individual, the fiscal year end, location, and the representative's name, and can also assign a person in charge on the supporter's side. The user who is a supporter can centrally manage multiple clients on this screen.

[0016] Figure 4 shows an example of a screen displaying a list of tasks. The list display area can display information about tasks, including, but not limited to, the task name, tax category, status, client name, fiscal year end, task assignee, and deadline. The displayed items can be changed as needed and customized by the user. While Figure 4 shows an example of a list displaying tasks for BBB Co., Ltd., tasks for all clients can be displayed in order of the nearest deadline. For example, a filter function can be provided to display only tasks with specific content. By specifying the content of any item in the filter settings field, only the corresponding tasks can be displayed in the list. Figure 4 shows that the status is "Incomplete" and the client is "BBB Co., Ltd.", as entered in the filter settings field.

[0017] Figure 5 shows an example of a screen layout when tasks are displayed in a calendar format. A calendar for a specific month is displayed, and the number of tasks with deadlines set for each day is displayed numerically. When displaying a calendar, you can also use the filter function as described above.

[0018] Next, Figure 6 shows an example screen configuration in which one of the tasks listed for BBB Co., Ltd. in Figure 4 is selected and detailed task information is displayed. For the task name "Preparing to collect the complete set of documents required for consumption tax returns," the task owner, start date, deadline, checklist, and comment field are displayed. However, the displayed details are not limited to these and can be set as needed and may be user-configurable. The checklist may be a list of more specific actions to take when completing the task or points to be aware of when completing the task. A check box may be provided for each check item, allowing the user to check the box upon completion. The comment field is a free-form entry field for the task. For example, a supervisor (e.g., the tax accountant at the head of the office) may enter instructions for the task owner, or the comment field may display notes made by the task owner when the task was completed in the previous period to pass on information for the next period. Furthermore, after completing the task, the task status can be changed from "Incomplete" to "Completed" by clicking the "Task Complete" button in the upper right corner of the screen.

[0019] Figure 7 shows an example of the screen display of the task registration input field when registering a new task. The items to be entered when registering a task may include, but are not limited to, information such as the client, tax category, title, memo, period, and person in charge of the task, and can be changed as needed and may also be customizable by the user.

[0020] The above shows an example of the basic screen configuration of the task management support system according to the present invention, but this is merely an example, and any screen configuration is acceptable as long as it is capable of executing each of the elements of the present invention described below.

[0021] 1 is a diagram illustrating a task management support system according to the present invention. A server 1 functions as a platform that provides the task management support system, and outputs task information to a user terminal 2 connected via a network 3. As will be described in detail later, the server 1 generates tasks and task-related documents while exchanging information with the user terminal 2. At this time, the server 1 generates tasks and task-related documents using a large-scale language model.

[0022] Large-scale language models (LLMs) are a type of generative artificial intelligence (AI) model specialized for natural language processing (NLP). They are language models that realize advanced natural language generation built through deep learning of massive amounts of text data. A language model is a system that predicts the next word after a given word by understanding the patterns of natural language used by humans and analyzing the probability of word occurrence. Examples of large-scale language models include ChatGPT (Chat Generative Pre-trained Transformer) (registered trademark), Gemini (registered trademark), PaLM2 (registered trademark), and LaMDA2 (registered trademark). Words and sentences used as instructions to be input into these large-scale language models are generally referred to as prompts.

[0023] The server 1 may be any information processing device capable of implementing the functions described in each embodiment. The server 1 includes, for example, a server device, a computer (e.g., a desktop, a laptop, a tablet, etc.), etc. Note that while FIG. 1 illustrates the server 1 alone, this is not limiting. That is, each function described as being provided by the server 1 may be implemented by multiple servers, or may be implemented by a single server. Also, a large-scale language model may be provided by another server. Note that the server 1 may be, for example, a distributed server system that cooperates by communicating via a network, or may be a so-called cloud server. That is, the server 1 is not limited to a physical server, and may also include a virtual server implemented by software.

[0024] The user terminal 2 is a communication terminal used by a user who uses the task management support device, and has an application for using the task management support device installed thereon, and is capable of sending and receiving various types of information to and from the server 1. Note that it is not essential to install an application on the user terminal 2; the user may access a web page for using the task management support service provided by the server 1 from the user terminal 2 via a web browser or the like, and send and receive various types of information to and from the server 1.

[0025] The user terminal 2 may be any terminal that can implement the functions described below. For example, the user terminal 2 may be a computer (e.g., a desktop computer, a laptop computer, a tablet, etc.), a smartphone, etc.

[0026] The task management support system may further include a database server (not shown) that stores various information (data) used in the task management support system, such as training data to be input to a large-scale language model, prompts, and user information.

[0027] The network 3 may include a wireless network or a wired network, such as a wireless LAN (WLAN), a wide area network (WAN), integrated service digital networks (ISDNs), wireless LANs, code division multiple access (CDMA), long term evolution (LTE), LTE-Advanced, fourth generation (4G), fifth generation (5G), sixth generation (6G) or later mobile communication systems. The network 3 is not limited to these examples, and may also be a public switched telephone network (PSTN), Bluetooth (registered trademark), optical fiber, a satellite communication network, or a combination thereof.

[0028] Next, the hardware configuration and functional configuration of the server 1 and the user terminal 2 according to one embodiment of the present invention will be described with reference to FIG.

[0029] The server 1 and the user terminal 2 each include at least a processor 10, a memory 11, a storage 12, a transmitting / receiving unit 13, an input / output unit 14, etc., which are electrically connected to each other via a bus 15.

[0030] The processor 10 is a computing device that controls the overall operation of the server 1, controls the transmission and reception of data between each element, and performs information processing necessary for executing applications. For example, the processor 10 is a CPU (Central Processing Unit), and executes programs stored in the storage 12 and deployed in the memory 11 to perform various information processing.

[0031] The memory 11 includes a main memory configured with a volatile storage device such as a DRAM (Dynamic Random Access Memory) and an auxiliary memory configured with a non-volatile storage device such as a flash memory or an HDD (Hard Disc Drive). The memory 11 is used as a work area for the processor 10, and also stores a BIOS (Basic Input / Output System) that is executed when the server 1 starts up, various setting information, etc.

[0032] The storage 12 stores various programs such as application programs. A database that stores data used for each process may be constructed in the storage 12.

[0033] The transmitting / receiving unit 13 connects the server 1 to the network 3. The transmitting / receiving unit 13 may include a short-range communication interface for Bluetooth (registered trademark) and BLE (Bluetooth Low Energy).

[0034] The input / output unit 14 is an information input device such as a keyboard and a mouse, and an output device such as a display, which are used as needed.

[0035] A bus 15 is commonly connected to the above elements and transmits, for example, address signals, data signals and various control signals.

[0036] 9 is a diagram showing an example of the software configuration of the server 1. The server 1 includes a task management unit 21, an acquisition unit 22, a preprocessing unit 23, a task information generation unit 24, a document information generation unit 25, a display control unit 26, a task management information storage unit 31, a publicly known information storage unit 32, and a business operator data storage unit 33.

[0037] The task management unit 21, acquisition unit 22, preprocessing unit 23, task information generation unit 24, document information generation unit 25, and display control unit 26 are realized by the processor 10 of the server 1 reading out programs stored in the storage 12 into the memory 11 and executing the programs. The task management information storage unit 31, publicly known information storage unit 32, and business entity data storage unit 33 are realized as part of the storage area provided by at least one of the memory 11 and the storage 12.

[0038] The task management information storage unit 31 holds task information for each business registered in this task management system. Figure 10 shows an example of the configuration of task information stored in the task management information storage unit 31. Task information may be managed by creating a table for each business, or tasks for each business may be managed by including information identifying the business in the task information. Task information may hold predetermined items (task name, status, start date, deadline, memo, order, required time, etc.) for each tax category, but the types of items are not limited to these.

[0039] The task management unit 21 manages task information. That is, it executes processes such as registering tasks, changing their status, and notifying users. Task registration is performed by recording the task information in the user's task management information storage unit 31 when the user inputs an instruction to approve a task generated by the task information generation unit 24, as will be described later. Status changes are performed by the user executing the task and selecting the task completion button as shown in FIG. 6, thereby updating the status of the task information through user processing, such as changing the status of the task to completed. In addition, if there is a task whose start date or deadline has arrived, the user can be notified of this.

[0040] The database unit may include a public information storage unit 32 that stores various publicly available information related to tax and accounting, and a business data storage unit 33 that stores user-specific business information. The public information stored in the public information storage unit 32 typically includes information on laws, ministerial ordinances, cabinet orders, and other statutes, as well as notices, standards, guides, and procedural summaries published by public institutions such as the national government, prefectures, and municipalities. It may also include legal precedents, news, and commentary articles related to tax and accounting, and the database unit can be constantly updated with the latest information. The business information stored in the business data storage unit 33 may include business-specific information such as the business's various past tax returns, notifications, applications, payment slips, and other tax documents and their attachments, accounting data (journalization data and various ledger data), articles of incorporation, registers, certificates of all historical matters, number of employees, employee joining and leaving information, payroll data, information about the business type, minutes of shareholders' meetings and board of directors meetings, and shareholder lists. The business data storage unit 33 may store business-specific information related only to the business for which the task is to be generated, or may also store business-specific information related to other businesses that are not the business for which the task is to be generated. The business-specific information of other businesses may be selected and included as business-specific information for businesses with similar capital and number of employees, or similar information such as industry. It may also include an external database that can be acquired. An external database may be a database aggregating data on startup companies. If the user is a supporter such as a tax accountant's office, the business-specific information of other business clients may be included. Furthermore, from among other business clients, businesses with similar capital and number of employees, or similar information such as industry, may be selected and included in the database. Various data stored in the database may be mapped by theme as needed. For example, mapping multiple tax and accounting-related themes, such as data on withholding income tax, data on executive compensation, and data on social insurance, can improve the accuracy of data generation using large-scale language models. The publicly known information storage unit 32 and the business entity data storage unit 33 may be separate databases or may be an integrated database, and the format is not important.

[0041] The database unit can also include past task information for the client. That is, it stores past task information registered by the task management support device according to the present invention. The task management information storage unit 31 stores information about tasks to be performed. However, for example, information about completed tasks may be transferred to the business entity data storage unit 33 as needed, so that the task information generation unit 24 can use the information as input to the AI ​​model when generating tasks. The timing of transferring task information to the business entity data storage unit 33 can be set as appropriate, such as transferring tasks for the current fiscal year when the fiscal year ends. The past task information stored in the business entity data storage unit 33 can include information such as the actual start date, completion date, task manager, and comments, as well as information about the revision history of these items. By using past task information as input, the task information generation unit 24 can set the start date earlier than usual if, for example, a comment such as "It would be better to start earlier" is included in the comment field for past task information. Furthermore, when there are multiple tasks related to one tax category, the start order for each task can be adjusted based on the actual start date and completion date of the past task information.

[0042] The acquisition unit 22 acquires a task generation instruction related to tax and accounting from a user. The user inputs the task generation instruction into the application of the task management support system according to the present invention via the input / output unit of the user terminal 2, thereby enabling the acquisition unit 22 to acquire the task generation instruction. The task generation instruction is instruction information related to task generation, and may be in the form of a question about a specific event, such as "An employee is retiring at the end of next month. Please tell me the necessary tasks," or in the form of a question about tax and accounting, such as "Can I pay a bonus to an executive?" In other words, the task generation instruction data does not necessarily have to be a direct instruction to generate a task, but may also be a somewhat vague question about tax and accounting. The task generation instruction may also be input as text data or image data. For example, a task generation instruction may be input by uploading an image of a notice received from the tax office to the application.

[0043] The preprocessing unit 23 preprocesses the task generation instruction data from the user acquired by the acquisition unit 22 as necessary to generate preprocessed task generation instruction data. In the following description, task generation instruction data that is not preprocessed because it is not necessary is also referred to as preprocessed task generation instruction data. As described below, the task information generation unit 24 generates task information using a generation AI model based on the preprocessed task generation instruction data. However, the task information generation unit 24 can process the task generation instruction data to facilitate the functioning of the generation AI model. Preprocessing can include adjusting the format of the task generation instruction data, such as the number of characters, or, if the task generation instruction data is image data, converting it into text data using image analysis technology such as OCR. The preprocessing unit 23 may also generate prompt data that can be input to the generation AI model based on the preprocessed task generation instruction data. Prompt data may be generated by preparing a prompt template that predefines prompts required for task generation and applying the preprocessed task generation instruction data to the prompt template.

[0044] The task information generation unit 24 generates task information based on the preprocessed task generation instruction data using a large-scale language model. In one embodiment of the present invention, the task information generation unit 24 can extract public information and business entity information related to the preprocessed task generation instruction data from the database unit. The task information generation unit 24 can also extract past task information (including comments, revision history, etc.) related to the preprocessed task generation instruction data from the database unit.

[0045] The task information generation unit 24 can generate task information such as tax category, title, start date, deadline, and worker using public information, business operator information, past task information, and a generation AI model. The task information generation unit 24 can generate task information by providing a prompt including public information, business operator information, and preprocessed task generation instruction data to a large-scale language model. The task information generation unit 24 can also generate task information by providing a prompt including public information, business operator information, past task information, and preprocessed task generation instruction data to a large-scale language model. If multiple tasks are required for one tax category, information regarding their order can also be read from the database unit and included in the prompt. For tasks determined by basic information such as each business operator's fiscal year-end, the deadline can be determined by referring to the basic information of each business operator. The start date can be determined, for example, by estimating the work time required for the task and calculating backward from the deadline.

[0046] The display control unit 26 displays on the user terminal the task information generated by the task information generation unit 24. When the user inputs an instruction to approve the task generated by the task information generation unit 24, the task management unit 21 registers the task information in the task management information storage unit 31 of the user.

[0047] Figure 11 is an example of a task generation screen. In Figure 11, the user inputs the text "How do I apply for simplified taxation?" as task generation instruction data. In response, the task information generation unit 24 generates the following text data and displays it to the user: "To apply for simplified taxation, the following steps are required: 1. Check eligibility: Taxable sales for the tax period must be 50 million yen or less: For sole proprietors, taxable sales for the year before last must be 50 million yen or less; for corporations, taxable sales for the business year before last must be 50 million yen or less. Business type: Not all businesses are eligible for the simplified taxation system. Please check with the tax office for the types of businesses that are eligible. 2. Submit a notification form for selecting the simplified taxation system. Where to submit: The tax office with jurisdiction over the tax payment location. Submission deadline: One month before the start of the tax period. 3. Prepare the necessary documents: Notification form for selecting the simplified taxation system for consumption tax: This can be downloaded from the tax office's website. Business-related documents: Documents that confirm the business details, such as a final tax return." Furthermore, a message including a task generation link, "To generate a simplified taxation task, please click the link," is displayed.

[0048] If the user wishes to create a task in response to the message "If you would like to create a simplified taxation task, please click the link," the user clicks the "Create a simplified taxation task" text (task creation link) to transition to the task registration screen shown in Figure 12. The task registration screen displays a task registration format in which the task information generation unit 24 has filled in the appropriate information, allowing the user to register each item required for task registration, such as "tax category," "title," "memo," "deadline," and "task owner." The user can confirm the content of each item, correct it if necessary, and click "Register Task" to add and register a task related to simplified taxation in the task management system. In this way, the task information generation unit 24 can respond to the user's questions with information including explanations about the task and generate a task registration format with information required for task registration. The task information generation unit 24 generates the answers and the information required for task registration using a generative AI model that is fed with information obtained from the database described above, thereby always proposing tax and accounting tasks based on the latest public information. Furthermore, because the database unit contains past user task information and basic information about the business, task information optimized for each user can be generated.

[0049] The task information generation unit 24 can generate the task registration format in the form of a URL with parameters, for example. In the above example, the generated URL with parameters is something like "...client=Test Co., Ltd.&title=Application for Simplified Taxation&memo=Content: Submit the 'Notification of Selection of Simplified Consumption Tax Tax System' to the tax office in charge of the tax payment area by the day before the first day of the tax period...(omitted)," and the URL is attached as a link in a message that reads, "If you want to generate a simplified taxation task, please click the link." The task information generation unit 24 may generate information in a format other than a URL with parameters, as long as it can display the task registration format.

[0050] The task information generation unit 24 can also generate task information by combining a generative AI model and a rule base. For example, a task related to a corporate tax return is a common task for all corporations, and its deadline is set to two months after the closing date. Therefore, at least the information regarding the deadline can be generated rule-based based on the basic information of the business. On the other hand, for items that may vary depending on the business or user (e.g., person in charge, memo, start date, order, etc.), it is more appropriate to generate them using a generative AI model that includes task information from past users in prompts, as this allows for the generation of tasks that more closely reflect the actual circumstances of each business. The task information generation unit 24 can be configured in advance to determine the content of routine task items using a rule base and the rest using a generative AI model.

[0051] The task information generation unit 24 can regenerate task information when the generated task information and the user's task generation instruction data do not match. For example, the task information generation unit 24 can determine the match rate between the generated task information and keywords included in the user's task generation instruction data, and if the match rate is low, it can determine that the quality of the generated task information is poor and regenerate the task information. If necessary, a message such as "Please enter a more specific question" or "Is your question about XX?" can be displayed to the user to prompt them to re-enter the task generation instruction data.

[0052] When a document related to taxation or accounting needs to be generated, the document information generation unit 25 can generate the document using a generative AI model. In the example of FIG. 11, a "Notification of Simplified Consumption Tax System Election" needs to be submitted, so the document information generation unit 25 creates the document. Documents may include, but are not limited to, documents submitted to tax offices, local governments, social insurance offices, etc., such as notifications, tax returns, applications, and payment slips, as well as their attachments; minutes of shareholders' meetings and board of directors meetings; and various internal documents such as shareholder registers. The document information generation unit 25 may generate a document based on instruction information regarding document creation from a user. In the example of FIG. 11, a document can be generated when the user enters additional instruction information to create a document, such as "Create a Notification of Simplified Consumption Tax System Election." Alternatively, text such as "Click the link to create a Notification of Simplified Consumption Tax System Election" may be displayed, and the document may be generated when the user clicks on the link. Furthermore, if the task information includes the generation of a document, the document may be generated automatically, and the generated document may be displayed together with the "answer" as shown in FIG. 11.

[0053] The document information generation unit 25 can generate documents by extracting required fields and entry guidelines for those fields from the public information storage unit 32 and then extracting business information to be entered for those fields from the business data storage unit 33. The business data storage unit 33 contains business information such as business's past submitted documents (e.g., tax returns, applications, and payment slips) and their attachments, minutes (e.g., minutes of general shareholders' meetings and minutes of board meetings), and various internal documents (e.g., shareholder registers). The document information generation unit 25 extracts this business information from the business data storage unit 33, extracts public information (e.g., entry guidelines) from the public information storage unit 32, and provides the extracted public information and business information in prompts to the generation AI model, allowing it to infer where to pull numbers and information from the business data for the required fields included in the document and create a document with information entered into the required fields in a predetermined format defined for each document. The user can review the generated document, accept any necessary corrections, and, once the user approves, the document file can be output in PDF format, Word format, or other formats.

[0054] FIG. 13 shows a basic processing flow for task generation according to the present invention. First, the acquisition unit 22 acquires task generation instruction data from a user (S101). The task generation instruction information may be text data or image data, and any format is acceptable. The task generation instruction information may also be a directive, a question, or an interrogative sentence, and the user can freely input it. The preprocessing unit 23 preprocesses the acquired task generation instruction information as needed (S102). This step S102 is optional. The task generation instruction information may be modified, for example, to have a character length that is easy for the generation AI model to process, or may be converted from image data to text data using image analysis (OCR, etc.).

[0055] The task information generation unit 24 generates task information using the generation AI model (S103). The task information generation unit 24 reads public information, business operator information, and task information from the public information storage unit 32 and the business operator data storage unit 33, and provides the read information and the preprocessed task generation instruction data to the generation AI model to generate a response to the user and task information based on the task generation instruction data. The task information can be presented to the user in a format in which the necessary information is entered into a task registration format, but the format is not limited to this. If the quality of the generated task information does not meet a certain standard (S104=NO), the task information generation unit 24 recreates the task information (S105). For example, S104 can be performed by checking whether keywords match between the task generation instruction data and the generated task information. If the task information is determined to be correct (S104=YES), the display control unit 26 displays a task registration screen such as that shown in FIG. 12 on the user terminal based on the task information (S106). The user can register the task information by modifying the task information as necessary and approving it (S107). Upon receiving an instruction to register task information, the task management unit 21 registers the task information in the user's task management information storage unit 31 and performs task management.

[0056] Next, the flow for generating a document will be described. When document generation is required based on task generation instruction information input by the user, document information generation unit 25 generates document information (S201). Display control unit 26 presents the generated document to the user (S202), and the user corrects the document information as necessary and approves it (S203), thereby outputting the document (S204). As in S104, a step for determining whether the generated document information is correct may be provided, and the document may be recreated if the generated document information does not satisfy predetermined conditions.

[0057] Although an example of a processing flow has been shown above, the order and the like can be changed as appropriate within the scope of achieving the object of the present invention.

[0058] According to the present invention, tasks and documents can be automatically generated using a generative AI model, allowing users to properly manage tax and accounting tasks even if they lack specialized knowledge, while significantly reducing work time. In the present invention, the database is constantly updated, making it possible to respond to frequent legal changes in real time. Furthermore, by including not only publicly available information but also business-specific information such as the user's (or, if the user is a tax accountant or other supporter, the business's) past tax returns, notification documents, accounting data (journal entries and various ledger data), articles of incorporation, registers, and minutes of shareholders' meetings and board meetings, as well as data containing past task information, in the prompts, it is possible to accurately generate tasks and documents tailored to the user's circumstances. Furthermore, while the above explanation is based on the example of the user being a supporter, the user may also be the business itself. A system may be used in which the supporter and the business can share task information, allowing both parties to register and update task information.

[0059] The above-described embodiment is merely an example for facilitating understanding of the present invention, and is not intended to limit the present invention. The present invention can be modified and improved without departing from the spirit thereof, and it goes without saying that the present invention includes equivalents thereof.

[0060] For example, in this embodiment, information read from the database unit is provided to the large-scale language model in a prompt. However, the generative AI model may be fine-tuned. In this case, the generative AI model that generates task information can be fine-tuned using the latest public information each time the latest public information is acquired, or by storing the latest public information and periodically tweaking the generative AI model. The generative AI model can also be fine-tuned using the latest business information each time the latest business information is acquired, or by storing the latest business information and periodically tweaking the generative AI model. The generative AI model can also be fine-tuned based on the task information (including revision history and comments included therein) each time task information is completed, or by storing completed task information and periodically tweaking the generative AI model. The generative AI model used by the document information generation unit 25 can also be fine-tuned using submitted documents such as various notifications, tax returns, applications, and payment slips, as well as their attachments, minutes of shareholders' meetings and board of directors meetings, and various internal documents such as shareholder registers.

[0061] Furthermore, learning by fine tuning may be combined with learning in which data extracted from the database is given to prompts (one-shot / few-shot learning). For example, the generative AI model may be fine-tuned using information up to a certain date and time, the date and time may be stored, information after the date and time may be read from the database, and prompts including the read information may be given to a generative AI model that has been fine-tuned using information before the date and time to generate task information. [Explanation of symbols]

[0062] 1 server 2. User terminal 3 Network

Claims

1. A device for managing tasks related to tax and accounting work of a business, a task information generating unit that receives instruction information related to task generation and generates task information; a display control unit that presents the task information to a user; a task management unit that registers a task based on the approval of the user; The task information generation unit extracts the public information and past task information related to the instruction information from a database that stores public information including at least laws and regulations and notices related to taxation and accounting, and the business operator's past task information including at least one of comments or revision history by the user, and generates the task information by providing a prompt including the extracted public information and past task information and the instruction information to a generation AI model.

2. 2. The task management device according to claim 1, wherein the display control unit presents the task information to the user in a state in which at least some of the required items have been entered in advance in a task information registration format.

3. 2. The task management device according to claim 1, wherein the generative AI model learns the public information and the past task information periodically or irregularly, and the task information generation unit generates the task information by providing the instruction information to the generative AI model.

4. 2. The task management device according to claim 1, wherein the database includes the business entity's unique information, and the task information generation unit extracts the public information related to the instruction information, the past task information, and the unique information and provides them to the generation AI model.

5. 2. The task management device according to claim 1, wherein the database includes registry information regarding the business operator, and the task information generation unit extracts the public information regarding the instruction information, the past task information, and the registry information and provides them to the generation AI model.

6. a document information generating unit that generates document information; business information relating to the business; The task management device according to claim 1 , wherein the document information generation unit generates the document information by providing the public information and the business information to a generation AI model.

7. A method for controlling a task management device for managing tasks related to tax and accounting work of a business, comprising: a task information generating step of receiving instruction information regarding task generation and generating task information; a display control step of presenting the task information to a user; a task management step of registering a task based on the approval by the user; A control method for a task management device, characterized in that the task information generation step extracts the public information and the past task information related to the instruction information from a database that stores public information including at least laws and regulations and notices related to taxation and accounting, and the business operator's past task information including at least one of comments or revision history by the user, and generates the task information by providing a prompt including the extracted public information and the past task information and the instruction information to a generation AI model.

8. A control program for a task management device for managing tasks related to tax and accounting work of a business, the control program being installed on a computer. a task information generation function that receives instruction information regarding task generation and generates task information; a display control function for presenting the task information to a user; a task management function for registering a task based on the user's approval; The task information generation function extracts public information and past task information related to the instruction information from a database that stores public information including at least tax and accounting laws and regulations and notices, and the business operator's past task information including at least one of comments or revision history by the user, and generates the task information by providing a prompt including the extracted public information and past task information and the instruction information to a generation AI model, said control program for a task management device.

Images