Financial due diligence material creation support device, financial due diligence material creation support method, and financial due diligence material creation support program

Abstract

An object of the present invention is to provide a financial due diligence material creation support device for efficiently creating financial due diligence materials. [Solution] The financial due diligence material creation support device acquires financial statement information including information indicating the financial statements of the target company for each of a plurality of years, acquires text from each of the plurality of financial reports, extracts item information indicating items including a plurality of account items and total items and amount information indicating the amount corresponding to each item from the acquired text, and generates financial statements including information regarding the trends in the amounts of each account item and total item based on the item information and amount information. The financial due diligence material creation support device acquires answer information indicating interviews with the management of the target company or answers to questions, and generates a summary of the answers using a trained model that has been machine-learned to summarize and output input sentences.

Description

[Technical field]

[0001] The present invention relates to a financial due diligence material creation support device, a financial due diligence material creation support method, and a financial due diligence material creation support program. [Background technology]

[0002] In recent years, mergers and acquisitions between companies have become widely known in Japan as M&A (Merger and Acquisitions).

[0003] To make M&A successful, it is necessary to conduct detailed research and analysis of the target company in advance, but such research and analysis require a great deal of effort. In particular, financial due diligence, which investigates and analyzes the financial situation of a company, requires organizing a huge amount of financial statements and interviews with management, etc., and compiling them into financial due diligence materials that allow the parties involved to decide whether or not to proceed with the M&A.

[0004] In relation to this, for example, various OCR (Optical Character Recognition) technologies are known that automatically recognize letters and numbers by analyzing images captured by a scanner of paper documents in order to extract information from documents (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2020-16946 A Summary of the Invention [Problem to be solved by the invention]

[0006] However, simply extracting information from paper documents using OCR technology as described above does not allow you to organize the information necessary to determine whether or not to proceed with an M&A transaction and create financial due diligence documents.

[0007] The present invention has been made in consideration of the above-mentioned circumstances, and aims to provide a financial due diligence document creation support device, a financial due diligence document creation support method, and a financial due diligence document creation support program for efficiently creating financial due diligence documents. [Means for solving the problem]

[0008] The above object can be achieved by the following means.

[0009] The financial due diligence material creation support device includes a financial statement information acquisition unit, a text acquisition unit, an extraction unit, a financial statement generation unit, a response information acquisition unit, a summary generation unit, and an output unit. The financial statement information acquisition unit acquires financial statement information including information indicating the financial statements of the target company for each of a plurality of years. The text acquisition unit acquires text from each of a plurality of financial statements indicated by the financial statement information acquired by the financial statement information acquisition unit. The extraction unit extracts item information indicating items including a plurality of account items and total items, and amount information indicating amounts corresponding to each item, from the text acquired by the text acquisition unit. The financial statement generation unit generates financial statements including information on the trends in amounts in each account item and total item, based on the item information and amount information extracted by the extraction unit. The response information acquisition unit acquires response information indicating an interview with the manager of the target company or responses to questions. The summary generation unit inputs the response information acquired by the response information acquisition unit into a trained model that has been machine-learned to summarize and output the input text, thereby generating a summary sentence. The output section outputs the financial statements generated by the financial statement generation section together with the summary generated by the summary generation section.

[0010] The financial due diligence material creation support method is a computer-executed financial due diligence material creation support method, and includes a financial information acquisition unit, a text acquisition unit, an extraction unit, a financial statement generation unit, a response information acquisition unit, a summary generation unit, and an output unit. The financial information acquisition step acquires financial statement information including information indicating the financial statements of the target company for each of a plurality of years. The text acquisition step acquires text from each of a plurality of financial statements indicated by the financial statement information acquired in the financial information acquisition step. The extraction step extracts item information indicating items including a plurality of account items and total items, and amount information indicating amounts corresponding to each item, from the text acquired in the text acquisition step. The financial statement generation step generates financial statements including information on the trends in amounts in each account item and total item, based on the item information and amount information extracted in the extraction step. The response information acquisition step acquires response information indicating an interview with the manager of the target company or responses to questions. The summary generation step generates a summary sentence by inputting the response information acquired in the response information acquisition step into a trained model that has been machine-learned to summarize and output the input sentence. The output step outputs the financial statements generated in the financial statement generation step together with the summary generated in the summary generation step.

[0011] The financial due diligence material preparation support program is configured to cause a computer to function as the above-mentioned financial due diligence material preparation support device. Effect of the Invention

[0012] According to the financial due diligence material creation support device of the present invention, financial statement information including information showing the financial statements of the target company in each of a plurality of years is acquired, and text is acquired from each of the plurality of financial statements shown by the acquired financial statement information. The financial due diligence material creation support device extracts item information showing items including a plurality of account items and total items, and amount information showing the amount corresponding to each item, from the acquired text. The financial due diligence material creation support device generates financial statements including information on the transition of amounts in each account item and total item based on the extracted item information and amount information. The financial due diligence material creation support device acquires answer information showing an interview with the manager of the target company or answers to questions. The financial due diligence material creation support device inputs the acquired answer information into a trained model that has been machine-learned to summarize and output the input text, thereby generating a summary. The financial due diligence material creation support device outputs the generated financial statements and the generated summary together as financial due diligence materials. This makes it possible to efficiently create financial due diligence materials. [Brief description of the drawings]

[0013] [Figure 1] 1 is a diagram showing a schematic configuration of a system according to an embodiment of the present invention; [Diagram 2] FIG. 2 is a block diagram showing a schematic configuration of a user terminal. [Diagram 3] FIG. 2 is a block diagram showing a schematic configuration of the support device. [Figure 4] FIG. 2 is a block diagram showing a functional configuration of the support device. [Diagram 5] 10 is a flowchart showing a procedure of a process executed in the support device. [Figure 6] 7 is a flowchart of a subroutine showing the procedure of the financial statement generation process shown in step S104 of FIG. 5. [Figure 7] FIG. 2 is a conceptual diagram for explaining a flow of processing executed in the support device. [Figure 8] 11 is a diagram for explaining a position determination process of a total item executed in the support device. FIG. [Figure 9] 11 is a diagram for explaining a major item determination process for bad debt reserves executed in the support device. FIG. [Figure 10] 11 is a diagram for explaining a major item determination process for bad debt reserves executed in the support device. FIG. [Figure 11] 11 is a diagram for explaining a major item determination process for bad debt reserves executed in the support device. FIG. [Figure 12] 13 is a diagram for explaining a major item determination process for bad debt reserves executed in the support device. FIG. [Figure 13A] 11 is a diagram for explaining an account item consolidation process executed in the support device. FIG. [Figure 13B] 11 is a diagram for explaining an account item consolidation process executed in the support device. FIG. [Figure 14] 11 is a diagram for explaining a spelling variation correction process executed in the support device. FIG. [Figure 15] 11 is a diagram for explaining an increase / decrease output process executed in the support device. FIG. [Figure 16] FIG. 11 is a diagram for explaining a verification process executed in the support device. [Figure 17] FIG. 13 is a diagram showing an example of financial due diligence materials output by the support device. [Figure 18] FIG. 13 is a diagram showing an example of financial due diligence materials output by the support device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Hereinafter, an embodiment of the present invention will be described with reference to the attached drawings. In the description of the drawings, the same elements are given the same reference numerals, and duplicated explanations will be omitted. In addition, the dimensional ratios of the drawings are exaggerated for the convenience of explanation, and may differ from the actual ratios.

[0015] <System configuration> FIG. 1 is a diagram showing a schematic configuration of a system according to an embodiment of the present invention.

[0016] As shown in FIG. 1, the system is made up of a user terminal 100 and an assistance device 200, and each component is configured to be able to communicate with each other via a network.

[0017] The user terminal 100 is an information terminal used by a requester, such as a notebook PC, a desktop PC, a smartphone, or a tablet PC.

[0018] The support device 200 is a server provided by a business operator that manages and operates a financial due diligence material creation support service. The support device 200 constitutes a financial due diligence material creation support device.

[0019] Each component will be described in detail below.

[0020] <User terminal 100> FIG. 2 is a block diagram showing a schematic configuration of a user terminal.

[0021] 2, the user terminal 100 has a CPU (Central Processing Unit) 110, a ROM (Read Only Memory) 120, a RAM (Random Access Memory) 130, a storage 140, a communication interface 150, an operation unit 160, and a display unit 170. Each component is connected to each other via a bus 180 so as to be able to communicate with each other.

[0022] The CPU 110 controls the above components and performs various arithmetic processing in accordance with programs recorded in the ROM 120 and the storage 140 .

[0023] The ROM 120 stores various programs and various data.

[0024] The RAM 130 serves as a working area for temporarily storing programs and data.

[0025] The storage 140 stores various programs including an operating system, and various data.

[0026] The communication interface 150 is an interface for communicating with external devices such as the support device 200, and uses standards such as 3G, 4G, and 5G for mobile phone communications, and standards such as Wi-Fi (registered trademark), for example.

[0027] The operation unit 160 includes a pointing device such as a mouse, and a keyboard, and is used to perform various input operations.

[0028] The display unit 170 is, for example, a liquid crystal display, and displays various information. The display unit 170 may function as the operation unit 160 by being configured as a touch panel display.

[0029] <Support device 200> FIG. 3 is a block diagram showing a schematic configuration of the support device.

[0030] 3, the support device 200 has a CPU 210, a ROM 220, a RAM 230, a storage 240, a communication interface 250, an operation unit 260, and a display unit 270. Each component is connected to be able to communicate with each other via a bus 280. Note that each component of the CPU 210, the ROM 220, the RAM 230, the storage 240, the communication interface 250, the operation unit 260, and the display unit 270 has the same functions as the CPU 110, the ROM 120, the RAM 130, the storage 140, the communication interface 150, the operation unit 160, and the display unit 170 of the user terminal 100, and therefore a duplicated description will be omitted.

[0031] Storage 140 stores programs for causing CPU 210 to function as a settlement information acquisition unit, a text acquisition unit, an extraction unit, a financial statement generation unit, a response information acquisition unit, a summary generation unit, an output unit, a determination unit, an identification unit, a classification setting reception unit, a sum setting unit, a sum calculation unit, a table registration unit, a conversion unit, a calculation unit, an increase / decrease output unit, an editing unit, a verification unit, a notification unit, and a comment generation unit. Storage 140 also stores various information required for each of the above units to execute processing.

[0032] <Functions of the Support Device 200> FIG. 4 is a block diagram showing a functional configuration of the support device.

[0033] As shown in FIG. 4, the support device 200 functions as a financial statement information acquisition unit 2101, a text acquisition unit 2102, an extraction unit 2103, a financial statement generation unit 2104, a response information acquisition unit 2105, a summary generation unit 2106, an output unit 2107, a judgment unit 2108, an identification unit 2109, a classification setting reception unit 2110, a sum setting unit 2111, a sum calculation unit 2112, a table registration unit 2113, a conversion unit 2114, a calculation unit 2115, an increase / decrease output unit 2116, an editing unit 2117, a verification unit 2118, a notification unit 2119, and a comment generation unit 2120 by the CPU 210 reading the program stored in the storage 240 and executing the processing.

[0034] The financial statement information acquisition unit 2101 acquires financial statement information including information showing the financial statements of the target company for each of a plurality of years.

[0035] The text acquisition unit 2102 acquires text from each of the multiple financial statements indicated by the financial statement information acquired by the financial statement information acquisition unit 2101 .

[0036] The extraction unit 2103 extracts item information indicating items including multiple account items and total items, and amount information indicating amounts corresponding to each item, from the text acquired by the text acquisition unit 2102. The extraction unit 2103 also extracts position information regarding the position where each item is described in the financial statement.

[0037] The financial statement generating unit 2104 generates financial statements including information on the trends in amounts for each account item and total item based on the item information and amount information extracted by the extracting unit 2103.

[0038] The answer information acquisition unit 2105 acquires answer information indicating answers to interviews or questions for managers of the target company. The answer information acquisition unit 2105 may acquire answer information indicating answers to interviews or questions for each account item.

[0039] The summary generating unit 2106 inputs the answer information acquired by the answer information acquiring unit 2105 into a trained model that has been machine-learned to summarize and output the input sentence, to generate a summary. The trained model is, for example, a large-scale language model. The summary generating unit 2106 may generate a summary for each account item by inputting the answer information for each account item, information on the amount corresponding to the account item, and preset instructions and rules for creating a summary to the trained model.

[0040] The output unit 2107 outputs the financial statements generated by the financial statement generation unit 2104 together with the summary generated by the summary generation unit 2106. The output unit 2107 may output the summary generated for each account item in association with each account item of the financial statements.

[0041] Based on the item information and position information extracted by the extraction unit 2103, the determination unit 2108 compares the position of the account item of accounts receivable or accounts payable with the position of the total item of current assets or current liabilities, and determines whether the total item is written above or below each account item included in the total item.

[0042] The identification unit 2109 identifies whether the account item for the allowance for doubtful accounts is listed as a current asset or as a fixed asset based on the determination result by the determination unit 2108 and the recording position of the account item for the allowance for doubtful accounts extracted by the extraction unit 2103 and the recording position of the total item.

[0043] The classification setting receiving unit 2110 receives the setting of classification information indicating the classification for each account item.

[0044] If the amount of the target account item is less than a specified percentage of the total amount of all account items belonging to the classification set for the account item, the sum setting unit 2111 sets the target account item as a summation target.

[0045] The summation calculation unit 2112 adds up the amounts of each account item set as the object of summation by the summation setting unit 2111 to calculate the total amount to be summed, integrates each account item as other items, and outputs it in association with the calculated total amount to be summed.

[0046] The table registration unit 2113 accepts registration of association table information for converting the names of a plurality of account items corresponding to the same account item into one unified name.

[0047] The conversion unit 2114 converts the names of the account items extracted by the extraction unit 2103 into unified names using the association table information whose registration has been accepted by the table registration unit 2113 .

[0048] The calculation unit 2115 calculates the inter-year increase or decrease in the amount corresponding to each account title and total item for each of a plurality of years, based on the item information and amount information extracted by the extraction unit 2103.

[0049] The increase / decrease output unit 2116 outputs the increase / decrease in the amount between fiscal years calculated by the calculation unit 2115 .

[0050] The editing unit 2117 edits the financial statements generated by the financial statement generating unit 2104 and the summary generated by the summary generating unit 2106 so that they can be output in a predetermined format. The output unit 2107 outputs the financial statements and the summary in a predetermined format.

[0051] The verification unit 2118 judges whether or not the total of the amounts corresponding to the multiple account items extracted by the extraction unit 2103 matches the amount corresponding to the total item.

[0052] If the verification unit 2118 determines that there is no match, the notification unit 2119 notifies the user.

[0053] The comment generating unit 2120 generates a summary comment regarding the amount and the trend of the amount for each account item and total item based on the financial statement generated by the financial statement generating unit 2104. The output unit 2107 outputs the summary comment generated by the comment generating unit 2120 together with the financial statement and the summary text.

[0054] The user terminal 100 and the assistance device 200 may include components other than those described above, or may not include some of the components described above.

[0055] <Processing Overview> FIG. 5 is a flowchart showing the procedure of the process executed in the support device. FIG. 6 is a flowchart of a subroutine showing the procedure of the financial statement generation process shown in step S104 of FIG. 5. FIG. 7 is a conceptual diagram for explaining the flow of the process executed in the support device. FIG. 8 is a diagram for explaining the position determination process of the total item executed in the support device. FIGS. 9 to 12 are diagrams for explaining the major item determination process of the bad debt reserve executed in the support device. FIGS. 13A and 13B are diagrams for explaining the account item consolidation process executed in the support device. FIG. 14 is a diagram for explaining the notation variation correction process executed in the support device. FIG. 15 is a diagram for explaining the increase / decrease output process executed in the support device. FIG. 16 is a diagram for explaining the verification process executed in the support device. FIG. 17 is a diagram showing an example of a financial due diligence material output by the support device. FIG. 18 is a diagram showing an example of a financial due diligence material output by the support device. The processes shown in the flowcharts of FIGS. 5 and 6 are stored as a program in the storage 240 of the support device 200, and are executed by the CPU 210 controlling each unit.

[0056] As shown in FIG. 5, the support device 200 acquires financial statement information including information showing financial statements of a target company for each of a plurality of years (step S101). For example, the user terminal 100 transmits a PDF file or the like showing financial statements of a certain company for each of a plurality of years (years) to the support device 200 via a network as financial statement information. A financial statement is a document prepared, for example, when a company performs settlement processing and declares taxes to a tax office, and includes a profit and loss statement (PL or P / L) and a balance sheet (BS or B / S) of the company for the year (year) for which declaration is to be made. As shown in FIG. 7, the support device 200 acquires information showing financial statements for a plurality of years (years). The financial statement information includes text such as letters, numbers, and symbols as image data (image information). The financial statement information may include such text as text data (text information) instead of image data.

[0057] The support device 200 acquires text from each of the multiple financial statements indicated by the financial statement information acquired in the processing of step S101 (step S102). For example, the support device 200 can acquire various texts included in the financial statement by executing a text recognition process by a known method on image data included in the financial statement information. The text recognition process may be executed by the support device 200, or may be executed in an external device such as a cloud server, and the execution result may be acquired by the support device 200. Alternatively, if various texts are already included as text data in the financial statement information, the support device 200 can acquire various texts from the text data.

[0058] The support device 200 extracts item information indicating items including a plurality of account items and total items, and amount information indicating amounts corresponding to each item, from the text acquired in the process of step S102 (step S103). Specifically, the support device 200 extracts item information indicating account items and total items, and amount information indicating amounts corresponding to each account item and each total item, from the profit and loss statement and balance sheet included in the financial statement. The process of extracting the parts corresponding to the profit and loss statement and balance sheet from the financial statement may be executed by the support device 200 using a known text matching technique or pattern matching technique, or may be executed by receiving an instruction from the user for the relevant part. In addition, the process of extracting account items and total items, and amounts corresponding to each account item and each total item from the profit and loss statement and balance sheet may be executed by the support device 200 using a known text matching technique or pattern matching technique, or may be executed by receiving an instruction from the user.

[0059] The support device 200 generates financial statements including information on the transition of amounts in each account item and total item based on the item information and amount information extracted in the process of step S103 (step S104). The process of generating financial statements will be described in detail later.

[0060] The support device 200 acquires answer information indicating answers to interviews or questions given to the manager of the target company (step S105). The interviews include, for example, a management interview and a financial and tax interview. The management interview is an interview on the overall management of the target company conducted with the manager and executives of the target company. The management interview includes, for example, all contents related to the management of the target company, such as the business situation, personnel situation, performance situation, market situation, competitor situation, risk factors in business operation, competitive strategy with competitors, business issues, and future management plans. The financial and tax interview is an interview on finance and tax conducted with the manager of finance and tax of the target company. The financial and tax interview includes, for example, all contents related to the finance and tax of the target company, such as sales situation, cost situation, labor cost, rent, sales and administrative expenses such as advertising expenses, other income and expenses, various taxes, assets, liabilities, and net assets. The answer information indicating the answers to the interview may be obtained, for example, as text data (file) created by transcribing the contents of questions and answers in an oral interview. Alternatively, if the interview is conducted text-based through online text chat, e-mail, or the like, a record of the exchanged text data may be obtained as the answer information to the interview. The answers to the questions are so-called Q&A, and for example, questions about the P&L or B / S included in the financial statement and answers by the management of the target company are written in a Q&A sheet in a table format or bullet-point format. In the Q&A sheet, the account items of the P&L and B / S that are the subject of each question are clearly stated, so that it is easy to recognize which account items the question and answer relate to. The answer information indicating the answers to the questions may be obtained, for example, as table data (file) indicating the Q&A sheet.

[0061] The support device 200 inputs the answer information acquired in the process of step S105 into a trained model that has been machine-learned to summarize and output the input sentence, and generates a summary sentence. The trained model is, for example, a known large-scale language model (LLM: Large Language Model). The large-scale language model is a machine-learning model that can process various natural language processing tasks. For example, ChatGPT (registered trademark) (Chat Generative Pre-trained Transformer) or the like can be used as the large-scale language model. As shown in FIG. 7, the support device 200 inputs the answer information into the trained model, thereby generating a summary sentence of the answer information for each account item included in the PL and BS of the financial statement. The summary sentence generation process will be described in detail later.

[0062] 7, the support device 200 outputs the financial statements generated in the process of step S104 and the summary generated in the process of step S106 together so that they can be viewed as one document or image (step S107). The output process will be described in detail later.

[0063] <Details of financial statement generation process in step S104> As shown in FIG. 6, the support device 200 executes the processes of steps S1041 to S1046 included in the financial statement generation process of step S104.

[0064] <Position determination / major item determination processing> The support device 200 determines the position of each item described in the financial statements (PL and BS), and determines into which major category a given item is classified (step S1041).

[0065] Specifically, the support device 200 extracts position information on the position where each item is described in the financial statement. The position information is extracted, for example, as coordinate information based on a certain position on the image of the financial statement. The position information may be relative information indicating the positional relationship between the extracted items. As shown in FIG. 8, the support device 200 compares the description position of the account item of accounts receivable or accounts payable with the description position of the total item of current assets or current liabilities based on the item information and position information extracted from the financial statement. The support device 200 judges whether the total item is described above or below each account item included in the total item. As shown as (1) in FIG. 8, if the description position of the total item of current assets or current liabilities is above the description position of the account item of accounts receivable or accounts payable, the support device 200 judges that the description format is such that the total item is described above each account item included in the total item. On the other hand, as shown as (2) in FIG. 8, if the position of the total item for current assets or current liabilities is lower than the position of the account item for accounts receivable or accounts payable, the support device 200 determines that the description format is one in which the total item is written below each account item included in that total item.

[0066] The support device 200 specifies whether the account item of the allowance for doubtful accounts is listed as a current asset or a fixed asset based on the determination result regarding the entry position of the total item and the entry position of the extracted account item of the allowance for doubtful accounts and the entry position of the total item. For example, as in the example of FIG. 9, in a format in which the total item is listed above each account item, if the account item of the allowance for doubtful accounts is located between the total item of current assets and the total item of fixed assets, the allowance for doubtful accounts is specified to be listed as a current asset. Also, as in the example of FIG. 10, in a format in which the total item is listed above each account item, if the account item of the allowance for doubtful accounts is located below the total item of current assets and the total item of fixed assets, the allowance for doubtful accounts is specified to be listed as a fixed asset. Also, as in the example of FIG. 11, in a format in which the total item is listed below each account item, if the account item of the allowance for doubtful accounts is located above the total item of current assets and the total item of fixed assets, the allowance for doubtful accounts is specified to be listed as a current asset. In addition, in the case of a format in which total items are listed below each account item, such as the example in Figure 12, if the account item for the allowance for doubtful accounts is located between the total item for current assets and the total item for fixed assets, the allowance for doubtful accounts is identified as being listed as a fixed asset.

[0067] <Account item summary processing> The support device 200 executes an account item consolidation process that consolidates account items with small amounts among the account items belonging to the same classification as miscellaneous items and sums them up (step S1042).

[0068] Specifically, the support device 200 accepts and registers in advance the setting of classification information indicating the classification for each account item. The classification registered as classification information may be a classification having multiple stages such as major items (major classifications), medium items (medium classifications), and minor items (minor classifications), as shown in Fig. 13A. The classification information set for each account item and the relationship between each account item and each total item (such as which account item each total item includes) as shown in Fig. 13A are registered in advance as an account item master and stored in the storage 240.

[0069] When the amount of the target account item is less than a predetermined ratio with respect to the total value of the amounts of all account items belonging to the classification set for the account item, the support device 200 sets the target account item as the one to be added together. The support device 200 calculates the total amount of the account items to be added together by adding up the amounts of each account item set as the one to be added together, integrates each account item as other items, and outputs the total amount of the account items to be added together in association with the calculated total amount of the account items to be added together. The above process will be described using an example in which the predetermined ratio is set to 5%. For example, as shown in FIG. 13B, if there are five account items 1 to 5 that belong to the same major item, and the total amount of account items 1 to 2 accounts for 95% of the total amount of the major item, the amounts of account items 3 to 5 are each 5% or less of the total amount of the major item. In this case, the support device 200 sets account items 3 to 5 as the ones to be added together. Then, the support device 200 integrates account item 3 to account item 5 as other items, and outputs the total amount of account item 3 to account item 5 as the total amount to be added together in association with the other items.

[0070] <Spelling variation integration processing> The support device 200 converts the names of the account items into unified names based on preregistered association table information (step S1043).

[0071] Specifically, the support device 200 accepts the registration of association table information for converting the names of a plurality of account items corresponding to the same account item into one unified name by associating them with each other. For example, the support device 200 accepts the registration of association table information as shown in Fig. 14. Fig. 14 shows an example of association table information for converting the names (names) of various account items such as cash and deposits, cash deposits, cash, ordinary deposits, deposits, and petty cash into a unified name of cash and deposits by associating them with each other. The support device 200 converts the names of each account item extracted from the financial statement into a unified name by using the association table information as described above.

[0072] <Increase / decrease calculation process> The support device 200 calculates and outputs the increase / decrease value between the years of the amount corresponding to each account item and total item in each of the multiple years based on the item information and amount information extracted in the processing of step S103 (step S1044). Specifically, as shown in FIG. 15, the support device 200 outputs the increase / decrease value between the years of the amount of each account item and each total item based on the amount in multiple years. In the example of FIG. 15, the increase / decrease value of the amount from 2020 to 2021 and the increase / decrease value of the amount from 2021 to 2022 are calculated based on the amounts of each account item and each total item in 2020, 2021, and 2022. The calculated increase / decrease value is output in association with each account item and each total item. The support device 200 generates a table including information on the amount and the transition of the amount of each account item and each total item in multiple years as a financial statement, as shown in FIG. 15.

[0073] <Verification process> The support device 200 verifies whether the extracted amount is correct by determining whether the total of the amounts corresponding to the multiple account items extracted in the process of step S103 matches the amount corresponding to the total item of the classification set for the multiple account items (step S1045). Specifically, as shown in Fig. 16, the support device 200 determines whether the total of the read amounts of each account item (amounts read and extracted from the financial statement) matches the read amount of the total item of the classification set for the account item (amounts read and extracted from the financial statement). If it is determined that the amounts do not match, the support device 200 notifies the user that a discrepancy has occurred.

[0074] For example, two account items, "Capital Surplus" and "Other Capital Surplus," have "Capital Surplus" set as the classification for sub-item 1. For the account item "Capital Surplus," "Sub-item 1" is set in the Total Range column, and "Capital Surplus" is set in the Sub-item 1 column. This indicates that the account item "Capital Surplus" is the sum of the account items for which "Capital Surplus" is set in sub-item 1, that is, "Capital Surplus" and "Other Capital Surplus." Here, the total of the read amounts for "Capital Surplus" and "Other Capital Surplus" is 110,000 yen. However, the read amount for "Capital Surplus," which is the sum of "Capital Reserve" and "Other Capital Surplus," is 300,000 yen. In this case, since the total of the read amounts for each account item does not match the read amount for the total item of the classification set for that account item, the support device 200 notifies the user of the occurrence of a discrepancy by displaying a screen, sending a message, issuing an audio alert, etc.

[0075] <Processing for generating summary comments from transition tables> The support device 200 generates a summary comment on the amount and the trend of the amount for each account item and total item based on the financial statement generated as shown in FIG. 15 (step S1046). The summary comment is created based on a predetermined generation rule, for example. An example of the generation rule is "summarize and describe the breakdown and trend of the three largest amounts of account items in the latest accounting period." For example, for an account item of "cash and deposits," the following summary comment can be generated. The breakdown of "cash and deposits," A to C (for example, names of financial institutions), A1 to A3 (for example, account information of A financial institution), and information indicating the corresponding amounts, etc., are acquired in advance by the support device 200 by reading and extracting the financial statement, and are stored in the storage 240.

[0076] <Example of summary comment for transition table> Cash and deposits The amount for fiscal year YY is XXX million yen, with A being XXX million yen, B being XXX million yen, and C being XXX million yen. The breakdown of A is A1: XXX million yen, A2: XXX million yen, and A3: XXX million yen. The breakdown of B is as follows: B1 is XXX million yen, B2 is XXX million yen, and B3 is XXX million yen. The breakdown of C is as follows: C1 is XXX million yen, C2 is XXX million yen, and C3 is XXX million yen. There was a decrease of XXX million yen from fiscal year YY to fiscal year ZZ, with the largest decrease being for A: XXX million yen, B: XXX million yen, and C: XXX million yen. The large decreases in B are for B1, XXX million yen, B2, XXX million yen, and B3, XXX million yen.

[0077] Summary comments such as those described above are printed together with the financial statements and the summary.

[0078] When the process of step S1046 is completed, the support device 200 ends the process of step S104, and executes the processes of step S105 and thereafter shown in FIG.

[0079] <Details of the generation process of the summary text of S106> The support device 200 can create a summary text of the response information for each accounting item, for example, by giving the following prompt to the large language model based on an instruction from the user or the like.

[0080] <Example of prompt> # Work instruction · You are a professional in M&A. · The provided data is related to the financial and tax situation of the company. · Please summarize the content described for each accounting item. · The output should be the summary content (up to 3 sentences in Japanese), in JSON format (the key is the accounting item, and the value is a list of strings split by sentence). # Accounting item · Cash and deposits · Accounts receivable · Inventories # Rule · For accounting items not described in the data, output "Not described in the data". · Describe the cited data No. in parentheses at the end of the string. · Do not include facts without description. · Output in "is the case". · The length of the list of strings may vary depending on the accounting item.

[0081] The support device 200 can, by means of the above prompt, instruct the generation of a summary text for each accounting item while requesting a summary as an expert. Also, the support device 200 can, by means of the above prompt, generate a summary text for each accounting item by quoting the summary comment generated by the process of step S1046 in FIG. 6.

[0082] The support device 200 can combine (integrate) the abstract text generated as described above and the abstract comment generated by the process of step S1046 in FIG. 6 based on the following rules, and output it as an abstract text for each account item.

[0083] <Example of rules for combining> · Summarize the abstract text for each account item. · The order of description shall be the summary comment of the transition table, the abstract text of Q&A, the abstract text of the financial and tax interview, and the abstract text of the management interview. · If there is no description of the corresponding account item in Q&A or the interview, write "Not described in the data".

[0084] The support device 200 can generate the following abstract text for an account item such as "accounts receivable" by the above-described process.

[0085] "For the fiscal year ending December 2022, it is XXX million yen, A is XXX million yen, B is 22 million yen, and C is 13 million yen. From the fiscal year ending December 2020 to the fiscal year ending December 2022, it has decreased by XXX million yen. The largest decreases are 571 million yen for A, 61 million yen for C, and 47 million yen for D. The balance of accounts receivable has decreased to XXX million yen in the fiscal year ending December 2021, XXX million yen in the fiscal year ending December 2022, and XXX million yen in August 2023. The reason for the decrease in accounts receivable is due to the decrease in sales, especially affected by the decline in sales due to COVID-19. The proportion of accounts receivable for individual installment payments is large, and approximately XXX billion yen is due to installment payments (delinquent accounts). Data for the past two years must be considered, and delinquent accounts that have not been deemed uncollectible in court have accumulated (delinquent accounts). For delinquent accounts, provisions are made at 50% for delays of more than three months and 100% for delays of more than six months (delinquent accounts). Not described in the data." <Details of the output process of S107> The support device 200 edits the financial statements generated by the process of step S104 and the summary sentence generated by the process of step S106 so that they can be output in a predetermined format specified in advance, and outputs them as financial due diligence materials. Specifically, as shown in FIG. 17 and FIG. 18, the support device 200 arranges the financial statements and the summary sentence in a list format and outputs them as financial due diligence materials. FIG. 17 is a financial due diligence material showing an overview of BS, and FIG. 18 is a financial due diligence material showing an overview of PL. As shown in FIG. 17 and FIG. 18, a transition table of BS and PL is output on the left side of the material, and a summary sentence generated for each account item of BS and PL is output in association with each account item on the right side of the material. The support device 200 can output the financial due diligence materials on a web page so that the contents can be easily confirmed and corrected, and can be easily converted into an output format (file format) to be finally used and downloaded.

[0086] As described above, according to the support device 200 of the present embodiment, financial statement information including information showing financial statements of the target company in each of a plurality of years is acquired, and text is acquired from each of the plurality of financial statements shown by the acquired financial statement information. The support device 200 extracts item information showing items including a plurality of account items and total items, and amount information showing amounts corresponding to each item, from the acquired text. The support device 200 generates financial statements including information on the transition of amounts in each account item and total item based on the extracted item information and amount information. The support device 200 acquires answer information showing answers to interviews or questions given to managers of the target company. The support device 200 inputs the acquired answer information into a trained model that has been machine-learned to summarize and output input text, thereby generating a summary sentence. The support device 200 outputs the generated financial statements and the generated summary sentence together as financial due diligence materials. This makes it possible to efficiently create financial due diligence materials. Conventionally, in order to prepare financial due diligence materials, it was necessary to check a huge amount of paper documents such as financial statements for each year, collect and organize the necessary figures, and compile them into financial statements, which required a huge amount of work. In addition, simply preparing financial statements and organizing the figures only allows one to grasp the facts and results such as increases and decreases in figures, but does not allow one to grasp the causes of the increases and decreases in figures. Therefore, it was necessary to separately conduct interviews and Q&As with managers, etc., extract summaries from the huge amount of documents recording the results, and organize them together with the financial statements. According to the support device 200 of this embodiment, financial due diligence materials can be prepared by simply inputting financial statements for each of multiple years and answer information to interviews or questions, so that financial due diligence materials can be prepared efficiently.

[0087] The support device 200 further extracts position information on the position where each item is described in the financial statement. Based on the extracted item information and position information, the support device 200 compares the description position of the account item of accounts receivable or accounts payable with the description position of the total item of current assets or current liabilities. The support device 200 judges whether the total item is described above or below each account item included in the total item. This makes it possible to reliably and easily determine and grasp whether the total item is described in the BS and PL of the financial statement in the description format where it is described above or below. In the financial statement, depending on the taste of the preparer, the specifications of the accounting system, and the like, there are both a description format where the total item is described above and a description format where it is described below. For example, an account item called bad debt reserve may belong to a total item called current assets, and may belong to a total item called fixed assets. In this case, if it is not possible to grasp whether the position where the total item of each account item is described is above or below, a problem occurs in that it is not possible to properly grasp which total item the bad debt reserve belongs to. According to the support device 200 of this embodiment, it is possible to reliably and easily determine and understand whether the total items are written in a format where they are written at the top or at the bottom, so that the total items to which each account item belongs can be properly understood.

[0088] Furthermore, the support device 200 identifies whether the account heading for the allowance for doubtful accounts is listed as a current asset or a fixed asset based on the determination result regarding the entry position of the total item, the entry position of the account heading for the allowance for doubtful accounts, and the entry position of the total item. This makes it possible to reliably identify whether the account heading for the allowance for doubtful accounts is listed as a current asset or a fixed asset, and therefore makes it possible to easily create appropriate financial statements without the need for manual confirmation or instructions by the user.

[0089] In addition, the support device 200 accepts the setting of classification information indicating the classification for each account item. If the amount of the target account item is less than a predetermined ratio with respect to the total value of the amounts of all account items belonging to the classification set for the account item, the support device 200 sets the target account item as a summation target. This makes it possible to set account items with relatively small amounts among account items belonging to the same classification as a summation target.

[0090] In addition, the support device 200 calculates the total amount of the account items set as the objects of addition by adding up the amounts of the account items, integrates the account items as other items, and outputs the calculated total amount of the account items. This makes it possible to integrate the account items with relatively small amounts set as the objects of addition as other items and output the total amount of the other items. Therefore, it is possible to output account items with relatively small amounts integrated as other items, and output account items with relatively large amounts and important separately, so that it is possible to generate effective financial statements that allow the contents of important account items to be easily visually confirmed and understood.

[0091] The support device 200 also accepts registration of association table information for converting the names of multiple account items corresponding to the same account item into one unified name by associating them with each other. The support device 200 converts the names of each extracted account item into a unified name using the association table information. Different names are often used for account items of the same content for different companies. Even for the same company, different names may be used for each year when the preparer, preparation method, or preparation system of the financial statement is changed. According to the support device 200 of this embodiment, financial due diligence materials can be generated after converting account items of the same content described with different names into unified names. Therefore, financial due diligence materials can be easily generated from financial statements of multiple years for various different companies.

[0092] In addition, the support device 200 calculates the increase or decrease between the years for the amounts corresponding to each account item and total item in each of the multiple years based on the extracted item information and amount information, and outputs the calculated increase or decrease between the years. This makes it possible to calculate and output the increase or decrease between multiple years even if only the PL and BS for that year are listed in the financial statement. This makes it easy to understand the changes in the PL and BS, and makes it easy to generate financial due diligence materials that allow the management status of the target company to be more accurately understood.

[0093] Moreover, the support device 200 edits the generated financial statements and the generated summary so that they can be output in a predetermined format specified in advance, and outputs the financial statements and the summary in the predetermined format. For example, the support device 200 can output financial due diligence materials on a web page so that the contents can be easily confirmed and corrected, and the financial due diligence materials can be easily converted into a final output format (file format) to be used and downloaded. Conventionally, it was necessary to create a transition table using a spreadsheet application, copy the created transition table, and paste it into a page of a file of a presentation application to use it as a financial due diligence material, and adjust the layout within the page. According to this embodiment, the generated financial statements and the generated summary can be edited and output so that they can be output in a predetermined format specified in advance, so that financial due diligence materials can be created efficiently.

[0094] In addition, the support device 200 judges whether the total of the amounts corresponding to the extracted multiple account items matches the amount corresponding to the total item. If it is judged that they do not match, the support device 200 notifies the user. This makes it possible to verify the validity of the contents extracted from the financial statement, and if the contents are incorrect, the user can be notified and checked, making it possible to create appropriate financial due diligence materials.

[0095] In addition, the support device 200 generates summary comments regarding amounts and trends of amounts for each account item and total item based on the generated financial statements, and outputs the summary comments together with the financial statements and the summary text. This allows comments summarizing the contents of the financial statements to be created and output as text, making it possible to create financial due diligence materials whose contents are easy to understand.

[0096] In addition, the support device 200 acquires answer information indicating answers to the interview or questions for each account item. The support device 200 generates a summary for each account item by inputting the answer information for each account item, information on the amount corresponding to the account item, and a summary creation instruction and creation rule set in advance into the trained model. The support device 200 outputs the summary generated for each account item in association with each account item of the financial statements. This allows the summary of the answers to the interview or questions to be output for each account item, so that the summary corresponding to the account item can be output in association with each account item of the financial statements. Therefore, appropriate financial due diligence materials can be easily created.

[0097] The present invention is not limited to the above-described embodiment, but can be modified in various ways within the scope of the claims.

[0098] For example, the user terminal 100 and the assistance device 200 may each be configured by multiple devices, or may be configured by a single device.

[0099] In addition, the functions of each component may be realized by other components. For example, some of the functions of the support device 200 may be realized by the user terminal 100. Some of the functions of the support device 200 may be realized by an external component such as a cloud server that can be connected via a network.

[0100] Furthermore, the processing units in the flowcharts in the above embodiments are divided according to the main processing contents in order to facilitate understanding of each process. The method of classifying the processing steps does not limit the present invention. Each process can be further divided into more processing steps. Also, one processing step may execute more processes.

[0101] The means and methods for performing various processes in the system according to the above-mentioned embodiment can be realized by either a dedicated hardware circuit or a programmed computer. The above-mentioned program may be provided by a computer-readable recording medium such as a flexible disk or a CD-ROM, or may be provided online via a network such as the Internet. In this case, the program recorded on the computer-readable recording medium is usually transferred to and stored in a storage unit such as a hard disk. The above-mentioned program may be provided as a standalone application software, or may be incorporated into the software of the device as one function of the system. [Explanation of symbols]

[0102] 100 user terminals, 110 CPUs, 120 ROMs, 130 RAM, 140 storage, 150 communication interface, 160 Operation section, 170 display section, 180 Bus, 200 Support equipment; 210 CPUs, 2101 Financial information acquisition department; 2102 Text Acquisition Unit, 2103 Extraction part, 2104 Financial Statement Preparation Department; 2105 Answer information acquisition department; 2106 Summary generator, 2107 output section, 2108 Judgment Department, 2109 Specific Department; 2110 Classification setting reception unit, 2111 Combined setting unit, 2112 Combined power division, 2113 Table registration unit, 2114 conversion unit, 2115 Calculation unit, 2116 Increase / decrease output section, 2117 Editorial Department, 2118 Verification Department, 2119 Notification Department; 2120 Comment generation unit, 220 ROMs, 230 RAM, 240 Storage, 250 communication interface, 260 operation section, 270 display section, 280 bus.

Claims

1. a financial statement information acquisition unit that acquires financial statement information including information showing financial statements of a target company for each of a plurality of fiscal years; a text acquisition unit that acquires text from each of a plurality of financial statements indicated by the financial statement information acquired by the financial statement information acquisition unit; an extraction unit that extracts, from the text acquired by the text acquisition unit, item information indicating items including a plurality of account items and total items, and amount information indicating amounts corresponding to each item; a financial statement generating unit that generates financial statements including information regarding the trends in amounts of the account items and the total items based on the item information and the amount information extracted by the extracting unit; a response information acquisition unit that acquires response information indicating responses to interviews or questions given to managers of the target companies; a summary generation unit that inputs the answer information acquired by the answer information acquisition unit into a trained model that has been machine-learned to summarize and output an input sentence, thereby generating a summary sentence; an output unit that outputs the financial statements generated by the financial statement generation unit and the summary generated by the summary generation unit together; A financial due diligence material creation support device having the above-mentioned.

2. The extraction unit further extracts location information regarding a location where each of the items is described in the financial statement, The financial due diligence material creation support device of claim 1, further comprising a determination unit that compares the position of an account item for accounts receivable or accounts payable with the position of a total item for current assets or current liabilities based on the item information and the position information extracted by the extraction unit, and determines whether the total item is listed above or below each account item included in the total item.

3. The financial due diligence material creation support device of claim 2, further comprising an identification unit that identifies whether the account item of the bad debt reserve is listed as a current asset or a fixed asset based on the judgment result by the judgment unit and the position of the account item of the bad debt reserve extracted by the extraction unit and the position of the total item.

4. a classification setting receiving unit that receives a setting of classification information indicating a classification for each of the account items; an aggregation setting unit that sets the target account item as an aggregation target when the amount of the target account item is less than a predetermined ratio with respect to the total value of the amounts of all the account items belonging to the classification set for the target account item; The financial due diligence material creation support device according to any one of claims 1 to 3, further comprising:

5. A financial due diligence material creation support device as described in claim 4, further comprising an aggregation calculation unit that adds up the amounts of each account item set as the object of aggregation by the aggregation setting unit to calculate a total amount to be aggregated, integrates each account item as other items, and outputs the calculated total amount to be aggregated in association with the calculated total amount to be aggregated.

6. The financial due diligence material creation support device according to claim 4 , wherein the classification has a plurality of stages.

7. a table registration unit that accepts registration of correspondence table information for converting names of a plurality of account items corresponding to the same account item into a single unified name; a conversion unit that converts the names of the account items extracted by the extraction unit into the unified names using the correspondence table information; The financial due diligence material creation support device according to any one of claims 1 to 3, further comprising:

8. a calculation unit that calculates an increase or decrease between years of an amount corresponding to each account title and total item for each of a plurality of years based on the item information and the amount information extracted by the extraction unit; an increase / decrease output unit that outputs the increase / decrease in the amount between the years calculated by the calculation unit; The financial due diligence material creation support device according to any one of claims 1 to 3, further comprising:

9. an editing unit that edits the financial statements generated by the financial statement generating unit and the summary generated by the summary generating unit so that they can be output in a predetermined format that is specified in advance; 4. The financial due diligence material creation support device according to claim 1, wherein the output unit outputs the financial statements and the summary in the predetermined format.

10. a verification unit that judges whether or not a total of amounts corresponding to the plurality of account items extracted by the extraction unit matches an amount corresponding to the total item; a notification unit that notifies a user when the verification unit determines that the two do not match; The financial due diligence material creation support device according to any one of claims 1 to 3, further comprising:

11. a comment generating unit that generates summary comments regarding amounts and trends of amounts in each of the account items and total items based on the financial statements generated by the financial statement generating unit; 4. The financial due diligence material creation support device according to claim 1, wherein the output unit outputs the summary comment together with the financial statements and the summary.

12. the answer information acquisition unit acquires answer information indicating answers to the interview or the questionnaire for each of the account items, the summary generation unit generates the summary for each account item by inputting the answer information for each account item, information on the amount corresponding to the account item, and preset summary creation instructions and creation rules into the trained model; 4. A financial due diligence material creation support device according to claim 1, wherein the output unit outputs the summary generated for each account item in association with each account item of the financial statements.

13. The financial due diligence material creation support device according to any one of claims 1 to 3, wherein the trained model is a large-scale language model.

14. A method for supporting the preparation of financial due diligence materials executed by a computer, comprising: A financial statement information acquisition step of acquiring financial statement information including information showing financial statements of the target company for each of a plurality of fiscal years; a text acquisition step of acquiring text from each of a plurality of financial statements indicated by the financial statement information acquired in the financial statement information acquisition step; an extraction step of extracting, from the text recognized in the text acquisition step, item information indicating items including a plurality of account items and total items, and amount information indicating amounts corresponding to each item; a financial statement generating step of generating financial statements including information regarding the trends in amounts of the respective account items and the total items based on the item information and the amount information extracted in the extracting step; a response information acquisition step of acquiring response information indicating responses to interviews or questions given to managers of the target companies; a summary generating step of generating a summary sentence from the answer information acquired in the answer information acquiring step by using a trained model that has been machine-learned to summarize and output an input sentence; an output step of outputting the financial statements generated in the financial statement generation step and the summary generated in the summary generation step together; Support methods for preparing financial due diligence materials, including:

15. A financial due diligence material preparation support program for causing a computer to function as the financial due diligence material preparation support device according to any one of claims 1 to 3.

Images