Information processing device, program, and method

WO2026182141A1PCT designated stage Publication Date: 2026-09-03LEGALON TECHNOLOGIES INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2026/007087
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-02-26
Filing Date
2026-02-26
Publication Date
2026-09-03

Smart Images

  • Figure JP2026007087_03092026_PF_FP_ABST
    Figure JP2026007087_03092026_PF_FP_ABST
Patent Text Reader

Abstract

The present invention is for assisting a user in the creation of a contract. A document creation assistance device 101 comprises: an acquisition unit 1013 that acquires document information and task information related to the document information and based on a user's behavior; a search unit 1014 that, on the basis of the document information and the task information, acquires information related to a task indicated by the task information, such acquisition being from a related information database which stores related information related to the document information; and an auxiliary information acquisition unit 1018 that acquires auxiliary information obtained by inputting the task information and the information related to the task into a language model.
Need to check novelty before this filing date? Find Prior Art

Description

Information processing apparatus, program, and method

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

[0002] When a user creates a document, it has been practiced to assist document creation by causing a language model to be used to generate answers to questions that arise during document creation and sentences to be used in the document. Documents to be created have various purposes; for example, Patent Document 1 describes an apparatus that, when a financial institution prepares a business evaluation sheet for evaluating a company, uses a language model to obtain an answer to a question acquired through an interactive UI and outputs the answer.

[0003] Japanese Patent No. 7579554

[0004] An object of an embodiment of the present invention is to provide an information processing apparatus, a program, and a method capable of appropriately answering a user's inquiry.

[0005] An information processing apparatus according to an aspect of the present disclosure comprises: an acquisition unit that acquires document information and issue information related to the document information based on a user's behavior; a search unit that acquires information related to an issue indicated by the issue information from a related information database storing related information associated with the document information, based on the document information and the issue information; and an auxiliary information acquisition unit that acquires auxiliary information obtained by inputting the issue information and the information related to the issue into a language model.

[0006] This is a schematic diagram of the document creation support system. This diagram illustrates the general process of processing in the document creation support system. This is a block diagram of the document creation support device. This diagram illustrates an example of processing in the document creation support system. This is a diagram showing an example of an editing screen. This is a diagram showing an example of an editing screen. This diagram illustrates another example of processing in the document creation support system. This is a diagram showing an example of an editing screen. This is an example of information stored in the related information database. This diagram illustrates the general process of processing using metadata in the document creation support system. This is a diagram showing an example of an editing screen. This diagram illustrates the general process of processing using the editable clause in the document creation support system. This is a diagram showing an example of an editing screen. This diagram illustrates the general process of processing using user attributes and behavioral information in the document creation support system. This is a diagram showing an example of an editing screen. This is a diagram showing an example of the hardware configuration of the document creation support device according to this embodiment.

[0007] An embodiment of the present invention will be described with reference to the attached drawings. In each drawing, components denoted by the same reference numerals have the same or similar configuration.

[0008] Figure 1 shows a block diagram illustrating the outline of the document creation support system 10 according to this embodiment. The document creation support system 10 comprises a document creation support device 101, a user terminal 201, and a language model 301. In the document creation support system 10, the document creation support device 101, the user terminal 201, and the language model 301 are connected to communicate via a network N. In this embodiment, document creation is an example of a task performed by the user, and the document creation support device is an example of an information processing device. A contract is an example of document information.

[0009] The document creation support device 101 is a server device having a computer that performs predetermined processing by executing a predetermined program stored in the document creation support device 101.

[0010] The user terminal 201 is, for example, an information processing device used by the user to create a contract using the document creation support device 101, and is, for example, a personal computer, a tablet terminal, or a smartphone.

[0011] The language model 301 is stored in an information processing device, such as a server, which is connected to the document creation support device 101 via a network. The language model 301 is a model trained to perform natural language processing to generate a response sentence in response to an input sentence (prompt). Specifically, it is a Large Language Model (LLM) training model. Examples of large language models that can be used include GPT-4®, PaLM2, BERT, etc. The language model 301 and the server on which the language model is stored may be operated by an entity separate from the vendor operating the document creation support system. Users or the vendor of the document creation support system can access the language model 301 via the network.

[0012] In the document creation support system 10, for example, a user operating the user terminal 201 performs document editing using the cloud-based document editing function provided by the document creation support device 101. For example, the user accesses the website provided by the document creation support device 101 to create a contract as a document, and uses the document creation support device 101 with a browser. The document creation support device 101 can also create and edit documents other than contracts.

[0013] The document creation support device 101 of this embodiment receives questions from the user, identifies the issues the user is having trouble with, and searches for materials to present to the user for assistance in document creation, as well as information from other contracts. The document creation support device 101 of this embodiment inputs the questions from the user, the issues the user is having trouble with, and the information obtained through the search into the language model 301, and obtains auxiliary information from the language model 301, including answers to the questions and advice on the issues the user is having trouble with in document creation. The obtained auxiliary information is displayed on the user terminal 201 as answer information.

[0014] Referring to Figure 2, an overview of the processing in the document creation support system 10 of this embodiment will be described. In step (1A), the document creation support device 101 acquires document information of the contract D that the user will edit. Specifically, the acquisition unit 1013, which will be described later, acquires the document information.

[0015] In step (1B), the document creation support device 101 acquires problem information, which includes information indicating questions from the user or problems identified by the document creation support device 101. Specifically, the acquisition unit 1013, described later, acquires the problem information. Here, the problem information includes problems based on user input, that is, questions which are active problems for the user, and problems identified by the document creation support device 101, that is, passive problems which the user is not clearly aware of.

[0016] When reviewing contracts or handling projects, users need to obtain information (reference information) about the following issues to resolve their questions. Furthermore, finding, comprehensively discovering, and arriving at answers from multiple documents requires time and effort. Information about the issues that those responsible for contract reviews and project handling might use includes past contracts, past project handling history, legal knowledge sources such as books and laws, and internal rules such as company templates. Issue information is the information used to search for this information about the issues. For example, if a question is entered as text, the issue information is the problem the user is trying to solve, as indicated by the question.

[0017] In step (1C), the document creation support device 101 searches the related information database 10121 of the document creation support device 101. Specifically, the search unit 1014 searches the related information database 10121, which stores related information related to the document information, based on the document information and the task information. The search is performed, for example, based on the information in which the task information has been interpreted by the language model.

[0018] In step (1D), the document creation support device 101 obtains information related to the task (information related to the task) from the related information database 10121. Related information F is, for example, a file that serves as a reference for creating a contract. Related information F is, for example, content related to the contract (slides, diagrams, videos, meeting minutes, emails, etc.), contracts for other projects, information for other projects, etc., and other information stored in the related information database 10121. The information related to the task is used as input to the language model 301 described later, and functions as information to ensure the accuracy of the response by the language model 301, and further functions as the basis for the response presented to the user.

[0019] In step (1E), the document creation support device 101 requests the language model 301 to input the task information and related information and generate an answer.

[0020] In the document creation support device 101, user input can be in colloquial natural language, such as, "What terms need to be defined in this case?". The document creation support device 101 interprets the user input through natural language processing, and based on the interpretation result, a question sentence is generated and input to the language model 301.

[0021] In step (1F), the document creation support device 101 obtains auxiliary information from the language model 301 as a response to assist the user in creating a contract. The auxiliary information may include, for example, answers to questions from the user and advice to the user. The auxiliary information may also include reference information that indicates related information. Here, reference information is information for accessing related information, such as file link information or website URL information.

[0022] In step (1G), the document creation support device 101 transmits auxiliary information to the user terminal 201. The user terminal 201 displays auxiliary information A, and the user can create the contract while checking the auxiliary information A and related information F that can be accessed based on the reference information.

[0023] Referring to Figure 2, the various parts of the document creation support device 101 will be described.

[0024] The document creation support device 101 includes a communication unit 1011, a storage unit 1012, an acquisition unit 1013, a search unit 1014, an editing status identification unit 1015, an attribute information acquisition unit 1016, an action information acquisition unit 1017, an auxiliary information acquisition unit 1018, and a presentation unit 1019.

[0025] The communication unit 1011 controls communication between the document creation support device 101 and the server where the user terminal 201 and language model 301 are stored.

[0026] The storage unit 1012 stores various types of information used in processing by the document creation support device 101. The storage unit 1012 has a related information database 10121. Details of the related information database 10121 will be described later.

[0027] The acquisition unit 1013 acquires document information of the contract that the user is editing. The document information includes, for example, metadata associated with the contract.

[0028] Metadata includes information such as the contract type of the contract, case management information (case information) associated with the contract, information of the contract drafter, and the status of the contract. The contract type can be either a typical or atypical contract. Contract types include various types of contracts that use contracts, such as employment contracts, sales contracts, confidentiality agreements, joint research agreements, and license agreements. Case information is information used to identify legal tasks (legal cases) that should be performed for a certain purpose. The status of the contract is information assigned according to the degree of completion of the contract, such as the contract being in draft, the initial draft being completed and under review, awaiting revisions by the other party, or the contract content being finalized.

[0029] Detailed examples of contract metadata include, for example, file name, contract title, contract type or document type, space (information that manages access permissions such as folders), version, contract status (information indicating whether it has been concluded or not), company name / client name, contract creator, creation date / update date, language, and other labels assigned by the company. Detailed examples of case information include, for example, case management number, case name, status, creation date / update date, deadline, requester / person in charge (secondary person in charge), requesting department, space (information that manages access permissions such as folders), case type (classification such as legal consultation or contract drafting request), etc.

[0030] Furthermore, the acquisition unit 1013 acquires issue information used to generate auxiliary information that assists in the creation of the contract. Auxiliary information includes, for example, answers to questions from the user and advice to the user. Issue information includes active issues, which are questions based on input from the user and are issues that the user actively recognizes, or passive issues, which are problems that the user is presumed to be facing when creating the contract, identified based on the user's operations and user attributes, and are issues that the user passively recognizes.

[0031] Furthermore, the acquisition unit 1013 may generate task information (passive tasks) based on the user or document status. Editing status refers to information acquired by the document creation support device 101 to understand the user's status during editing work, such as the type of clause in the contract being edited by the user, the input content, and the position of the editing cursor in the contract. For example, if the user's editing work related to a certain clause takes longer than a predetermined time, the acquisition unit 1013 may presume that the user is having difficulty making a judgment about the content of the clause and generate task information for creating the content of the clause. Alternatively, the acquisition unit 1013 may generate task information for checking whether the definition of a word defined by the user in the contract is the same as the definition in other contracts belonging to the same case. The generation of task information by the acquisition unit 1013 can be done by a rule-based method, by using a learning model that has been trained using data on the user's status and subsequent actions, or by acquiring data on the user's status and subsequent actions and inputting prompts containing instructions regarding that data into a large-scale language model.

[0032] The task information may, for example, be the presentation of reference information about the entire case. Specifically, when a user opens the case details page in the document creation support device 101, the acquisition unit 1013 may interpret the content of the case, for example using a language model 301, and the task may be to "automatically recommend appropriate reference information." At this time, the system may learn the user's preferred response patterns and, depending on the user, display a button that allows the user to request auxiliary information with, for example, a single click.

[0033] Alternatively, the task information may be "to review the contract" when the contract renewal deadline is approaching. In this case, the auxiliary information will be information for displaying the contract to the user after summarizing the key points of the contract to be reviewed. Specifically, when the user opens the case details page in the document creation support device 101, information prompting the user to review the contract can be displayed as auxiliary information. In this case, the auxiliary information may include a draft of the memorandum necessary for contract renewal generated by the language model 301. Furthermore, after the user has reviewed the auxiliary information, the document creation support device 101 may save the draft file of the memorandum and act as an intermediary to contact the person in charge of the case.

[0034] Furthermore, in the case of a particular project, if, after the contract is concluded, the contract matches a pattern commonly used by the company, the task may be to "save that contract as a company template." Specifically, when a user opens the project details page in the document creation support device 101, information prompting the user to save the contract as a company template can be displayed as supplementary information. The contract is saved as a company template when the user approves the recommendation to save it.

[0035] The search unit 1014 searches the related information database 10121, which stores related information related to the contract, based on the document information and the issue information, and obtains information related to the issue (search results). Here, information related to the issue refers to various types of information stored in the related information database 10121, such as files of contracts other than the contract being drafted, files of contract types that are the same as or related to the contract being drafted, files other than contracts in the case to which the contract being drafted is associated (explanatory materials, emails, etc.), legal guideline files, contract template files, etc. Furthermore, information related to the issue does not necessarily have to be a document file.

[0036] The search unit 1014 performs natural language processing to interpret the question text if the task information is question information that includes a question text entered by the user and attached files. The natural language processing is performed by a trained machine learning model. The trained machine learning model may be a language model stored on a server outside the user device, the document creation support device 101, or the document creation support system 10 as shown in this embodiment. Based on the interpretation information of the question text obtained by natural language processing and the document information, the search unit 1014 acquires information related to the task.

[0037] Furthermore, if metadata is associated with the document information, the search unit 1014 may acquire information related to the issue based on the metadata and the issue information. In this case, for example, the search unit 1014 may acquire as information related to the issue files in the related information database 10121 if the metadata associated with the files matches or is synonymous with the metadata of the contract being edited.

[0038] Furthermore, the search unit 1014 may obtain information related to issues concerning the clauses to be edited, based on the clauses to be edited by the user in the contract and the document information.

[0039] Furthermore, the search unit 1014 may acquire information related to the issue based on user attributes, document information, and issue information. Here, user attributes include, for example, information indicating the user's job title or level of experience in contract drafting, and information indicating the user's proficiency in contract drafting. User attributes may also include information such as the number and frequency of similar cases handled, the number and frequency of reviews of similar contracts, the degree of workload the user is currently under (whether or not they have many cases), and information regarding the lead time from the occurrence to the completion of a case (such as the average time required). In addition, the rarity, or difficulty, of the cases and contract types handled by the user may be associated with this information.

[0040] Furthermore, the search unit 1014 may acquire information related to the issue based on user behavior information related to editing the contract, document information, and issue information. Here, user behavior information related to editing the contract refers to information that shows the history of user actions, such as information on other files viewed by the user while editing, and questions entered by the user. In addition, the behavior information may include information indicating that the user transitioned to reference information pages such as other past contracts, books, or laws and guidelines while referring to or editing a case or contract, information indicating that a search action was taken while referring to or editing a case or contract, information indicating that internal comments were added to the contract, or information indicating that metadata such as the person in charge was changed.

[0041] The editing status identification unit 1015 acquires user input for the contract and identifies the editing status based on the input. For example, if the user opens the contract to be edited, the editing status 1015 identifies it as "editing started". Also, if the user's editing cursor has not moved for a predetermined amount of time in relation to a certain clause, the unit identifies it as the editing work taking longer than the predetermined amount of time.

[0042] The attribute information acquisition unit 1016 acquires the user's user attributes. The attribute information acquisition unit 1016 may acquire user attributes that are pre-stored in the document creation support device 101, or it may acquire user attributes through input from the user.

[0043] The behavior information acquisition unit 1017 acquires user behavior information. Specifically, it acquires the user's file viewing history and other information as behavior information in the cloud service provided by the document creation support device 101.

[0044] The auxiliary information acquisition unit 1018 inputs task information and information related to the task into the language model 301, and acquires auxiliary information from the language model 301. At this time, the auxiliary information acquisition unit 1018 may rewrite the task information acquired or generated by the acquisition unit 1013 into a prompt suitable for the language model 301, and then input the task information and the information related to the task into the language model 301. At this time, the auxiliary information acquisition unit 1018 may input a prompt that requests the language model 301 to generate auxiliary information based on the related information included in the information related to the task.

[0045] The auxiliary information may be, for example, information related to the task obtained by the search unit 1014 searching the related information database 10121, that is, information itself such as a contract, a case or a template. Further, the auxiliary information may be facts that can be found from information related to the task, or explanations, opinions, advices, specific examples, sentence examples, or improvement proposals obtained by combining these facts with matters recorded as facts in the language model 301 in advance.

[0046] For example, when a user inputs a question "Please tell me the status of the contract with Company X" as task information during document creation, the auxiliary information may be presented as follows: "Company X has entered into contract a and contract b, and *** exists as rights and obligations. If you intend to enter into a **** contract with another company, caution is required. However, the renewal deadline for contract b is approaching soon, so it is necessary to judge whether to renew it and take corresponding measures. In addition, although contract c was concluded with Company X in the past, it expired on YYYY / MM / DD. Below are links to each contract: Contract a, Contract b (expiring soon), Contract c (expired)". The user can proceed with document creation by referring to the auxiliary information.

[0047] The presentation unit 1019 presents the auxiliary information in a viewable manner for the user. Specifically, the presentation unit 1019 transmits the auxiliary information to the user terminal 201. The method of presenting auxiliary information is not necessarily limited to simply presenting to the user. For example, the presentation unit 1019 may automatically input text into a text box of a contract or a case being edited.

[0048] Furthermore, in the document creation support apparatus 101, when feedback that the generated text is transmitted via chat, email or the like, the status of the matter or the person in charge is updated, or the contract type is incorrect is received from a user, a process of automatically correcting and amending the metadata of the contract may be performed.

[0049] Referring to FIG. 4, an example of processing in the document creation support system 10 when a question is input from a user will be described.

[0050] In step S401, a request to display a contract to be edited is transmitted from the user terminal 201 to the document creation support apparatus 101. This is performed, for example, when the user selects a link to the contract via a web page provided by the document creation support apparatus 101.

[0051] In step S402, information for displaying the contract is transmitted from the document creation support apparatus 101 to the user terminal 201.

[0052] In step S403, the user starts editing the contract via the user terminal 201 using the online editor function provided by the document creation support system.

[0053] A screen for editing will be described with reference to FIG. 5. FIG. 5 shows an editing screen 501. The editing screen 501 includes a contract display area 502, an auxiliary information display area 503, a question input field 504, and a transmission button 505. The contract to be edited is displayed in the contract display area 502. Auxiliary information presented by the document creation support apparatus 101 is displayed in the auxiliary information display area 503.

[0054] In step S404, the user inputs a question into the question input field 504. When the user who has input the question selects the transmission button 505, in step S405, question information is transmitted from the user terminal 201 to the document creation support apparatus 101.

[0055] In step S406, the search unit 1014 interprets the query information. The interpretation of the query information is performed, for example, by inputting the query information into the language model 301. At this time, in order to improve the accuracy of the interpretation, information about the document being edited by the user may be transmitted along with the query information.

[0056] In step S407, the search unit 1014 retrieves information that may contain the answer to the question from the related information database 10121.

[0057] In step S408, the auxiliary information acquisition unit 1018 transmits the task information and the answer basis information to the language model 301.

[0058] In step S409, the language model 301 generates auxiliary information based on the task information and the response justification information.

[0059] In step S410, the auxiliary information acquisition unit 1018 acquires auxiliary information from the language model 301.

[0060] In step S411, the display unit 1019 transmits auxiliary information to the user terminal 201.

[0061] In step S412, auxiliary information is displayed on the user terminal 201.

[0062] An example of displaying supplementary information and information related to the issue will be explained with reference to Figure 6. Figure 6 shows the editing screen 601. The editing screen 601 has a contract display area 602, a supplementary information display area 603, a question input field 604, and a send button 605.

[0063] The auxiliary information display area 603 has a problem display area 6031 and an auxiliary information display area 6032. The problem display area 6031 displays the question entered by the user. In the example in Figure 6, it is shown that a user editing a confidentiality agreement entered the question "What is the definition of confidential information?" in order to ask about the definition of confidential information.

[0064] The auxiliary information display area 6032 displays auxiliary information generated by the language model 301. In the example in Figure 6, the answer is displayed as, "Examples of the definition of confidential information are as follows: ...". File selection buttons 60321, 60322, and 60323 are also displayed to display files containing information related to the task, which are the basis for the answer, on the user terminal 201. The file selection buttons 60321, 60322, and 60323 are buttons corresponding to links that display "Confidentiality Agreement (Company Z)", "Model Agreement", and "Template Confidentiality Agreement" depending on the user's selection. The information displayed on the file selection buttons 60321, 60322, and 60323 is reference information, and the information that can be viewed by selecting the file selection buttons 60321, 60322, and 60323 is related information. This allows the user to obtain the information necessary to create an agreement. Furthermore, the document creation support device 101 can process the display of example sentences in the contract display area 602 in accordance with the user's selection of example sentences displayed in the auxiliary information display area 6032.

[0065] Furthermore, the auxiliary information display area may also display, as auxiliary information, the reply text used when a user (legal staff member) receives an inquiry or request. The reply text is generated based on past case data and emails stored in the related information database 10121. In addition, relevant documents themselves may be made available for viewing as auxiliary information. Furthermore, the auxiliary information may also include proofreading results of the text written in the document or translation results of the text.

[0066] Next, referring to Figure 7, we will describe the processing in the document creation support system 10 when advice for the user's problem identified by the document creation support device 101 is displayed.

[0067] In step S701, a request to display the contract to be edited is sent from the user terminal 201 to the document creation support device 101. In step S702, information for displaying the contract is sent from the document creation support device 101 to the user terminal 201.

[0068] In step S703, the user begins editing the contract through the user terminal 201.

[0069] The editing screen will be explained with reference to Figure 8. Figure 8 shows the editing screen 801. The editing screen 801 has a contract display area 802, an auxiliary information display area 803, a question input field 804, and a send button 805. The contract to be edited is displayed in the contract display area 802. The auxiliary information display area 803 displays auxiliary information presented by the document creation support device 101. The editing screen 801 differs from the editing screen 501 in Figure 5 in that no question content is entered in the question input field 804.

[0070] In step S704, the user performs an editing operation on the contract.

[0071] In step S705, the editing state identification unit 1015 acquires operation input from the user terminal 201. In the example in Figure 8, it is assumed that information including cursor C information indicating the user's editing position and the time since the start of editing are acquired.

[0072] In step S706, the editing state identification unit 1015 identifies the editing state based on the operation input. For example, if the cursor C is repeatedly rewriting a certain clause, it identifies that the editing work is taking longer than a predetermined amount of time. Note that the identification of the editing state may be performed by the document creation support device 101, or it may be performed by processing on the browser of the user terminal 201, which is on the edge side.

[0073] In step S707, the acquisition unit 1013 generates task information based on the editing status. For example, if the editing status indicates that the editing work is taking longer than a predetermined amount of time, the unit estimates that the user is having difficulty making a judgment about the content of the clause and generates task information for creating the content of the clause.

[0074] In step S708, the search unit 1014 interprets the task information. The interpretation of the task information is performed, for example, by inputting the task information into the language model 301. At this time, in order to improve the accuracy of the interpretation, information about the document that the user is editing may be transmitted along with the question information.

[0075] In step S709, the search unit 1014 searches the related information database 10121 based on the interpretation result of the problem information and obtains information related to the problem. For example, if the interpretation result of the problem information is that the user is having trouble with Article 1 of the confidentiality agreement, the search unit performs a keyword search and a search using tags stored in the related information database 10121 to obtain information related to the problem concerning Article 1 of the confidentiality agreement.

[0076] In step S710, the auxiliary information acquisition unit 1018 transmits task information and information related to the task to the language model 301.

[0077] In step S711, the language model 301 generates auxiliary information based on information related to the task.

[0078] In step S712, the auxiliary information acquisition unit 1018 acquires auxiliary information from the language model 301.

[0079] In step S713, the presentation unit 1019 transmits auxiliary information and information related to the task to the user terminal 201.

[0080] In step S714, auxiliary information and information related to the task are displayed on the user terminal 201.

[0081] An example of displaying supplementary information and information related to the issue will be explained with reference to Figure 9. Figure 9 shows the editing screen 901. The editing screen 901 has a contract display area 902, a supplementary information display area 903, a question input field 904, and a send button 905.

[0082] The auxiliary information display area 903 has a problem display area 9031 and an auxiliary information display area 9032. The problem display area 9031 displays text based on the problem generated by the acquisition unit 1013. In the example in Figure 9, the user editing the confidentiality agreement is shown the message, "Are you having trouble defining confidential information?"

[0083] The auxiliary information display area 9032 displays auxiliary information generated by the language model 301. In the example in Figure 9, the answer is displayed as, "Examples of the definition of confidential information are as follows: ...". In addition, file selection buttons 90321, 90322, and 90323 are displayed to show the file that serves as the basis for the answer on the user terminal 201, based on the information related to the task. This allows the user to obtain the information necessary to create the contract without having to input the questions themselves.

[0084] In this way, the document creation support device 101 provides the user with information to solve the user's problems in creating contracts, and makes it possible to efficiently assist the user in creating contracts.

[0085] Referring to Figure 10, the data stored in the related information database 10121 will be explained. The related information database 10121 stores files such as files of contracts other than the contract being created, files of contracts of the same or related contract types as the contract being created, files other than contracts in the case to which the contract being created is associated (explanatory materials, emails, etc.), files of legal guidelines, and contract template files. Note that the related information database 10121 is not limited to databases inside the document creation support device 101, but may also include external databases that are connected to the document creation support device 101 in a communicative manner.

[0086] The information stored in the related information database 10121 will be explained using the contract database 101211 and the case file database 101212 as examples.

[0087] The contract database 101211 stores information about contracts. For example, the information about a contract may include the following items: "Contract ID," "Contract Type," "Case Information," "Clause Information," "Clause Label," "Person in Charge," and "Experience Level." The information about a contract may also include other information such as the contracting party, and each piece of information may have an ID to identify it.

[0088] The "Contract ID" field stores information that identifies the contract. The "Contract Type" field stores the type of contract to which the contract applies. The "Case Information" field stores information that identifies the case in which the contract is used. For example, in the example in Figure 10, the contract type of the contract with contract ID "D001" is stored as "Confidentiality Agreement," and the case information is stored as "Case C." Similarly, the contract type of the contract with contract ID "D002" is stored as "Joint Research Agreement," and the case information is also stored as "Case C."

[0089] The "Article Information" section stores information about the articles of the contract, organized by article, paragraph, and subparagraph (not shown). The "Article Label" section stores information indicating the purpose or nature of each article. For example, in the example in Figure 10, the contract with contract ID "D001" stores that "Article 1, Paragraph 1" has the article label "Definition," and "Article 1, Paragraph 2" has the article label "Exception Clause." Similarly, the contract with contract ID "D002" stores that "Article 1" has the article label "Purpose," and "Article 1, Paragraph 1" and "Article 1, Paragraph 2" have the article labels "Definition."

[0090] The "Contact Person" field stores information (ID, name, etc.) about the person responsible for drafting the contract. The "Experience Level" field stores a level indicating the contact person's proficiency in contract drafting. The level is set such that, for example, a higher number indicates greater proficiency.

[0091] In the example in Figure 10, it is stored that the creator of the contract with contract ID "D001" is "C" and the experience level is "Level 2". Also, it is stored that the creator of the contract with contract ID "D002" is "D" and the experience level is "Level 3".

[0092] The case file database 101212 stores information about files other than contracts, associated with each case.

[0093] The case file database 101212 stores information such as "file ID," "case information," and "file label."

[0094] The "File ID" stores information to identify each file. The "Case Information" stores information to identify the case in which each file is used. The "Case Information" uses information common to the related information database 10121, which includes the contract database 101211. The "File Label" stores information that indicates the nature of the file, which is assigned to the file. In the example in Figure 10, the file with File ID "F001" is a file related to "Case A," and the file label is stored as "Internal Mail." Also, the file with File ID "F002" is a file related to "Case A," and the file label is stored as "Term Sheet."

[0095] Furthermore, the related information database 10121 may store pre-created templates, explanatory articles related to legal documents, user-created templates, public content such as laws and guidelines, books, and uploaded files other than contracts.

[0096] Referring to Figure 11, we will now explain how to search the related information database 10121 using metadata. Figure 11 shows an example where the metadata "Contract Type: Confidentiality Agreement, Case Information: Case A, Creator: A" is associated with contract D being edited by the user.

[0097] In step (11A), the document creation support device 101 acquires metadata and task information. At this time, the task information may be generated by the document creation support device 101 based on user operations.

[0098] In step (11B), the search unit 1014 searches the related information database 10121 based on the metadata and the issue information. For example, the search unit 1014 searches for contracts and files that have metadata indicating that the contract type is a "non-disclosure agreement". In step (11C), the search unit 1014 retrieves information related to the issue from the related information database 10121. If the editing status of the contract is included as metadata, the generation of issue information and auxiliary information may be performed according to the status of the contract.

[0099] In step (11D), the document creation support device 101 inputs metadata, task information, and task-related information into the language model 301. In step (11E), the document creation support device 101 obtains auxiliary information from the language model 301. Subsequently, the auxiliary information is presented.

[0100] Figure 12 shows an example of displaying auxiliary information based on information related to the issue obtained through a search using metadata. Figure 12 shows the editing screen 1201. The editing screen 1201 has a contract display area 1202, an auxiliary information display area 1203, a question input field 1204, and a submit button 1205.

[0101] The auxiliary information display area 1203 has a task display area 12031 and an auxiliary information display area 12032. The task display area 12031 displays text based on the task generated by the acquisition unit 1013. In the example in Figure 12, a message is displayed to the user editing the confidentiality agreement indicating that auxiliary information related to the case is being presented, such as "Regarding the definition of data related to this case."

[0102] The auxiliary information display area 12032 displays auxiliary information generated by the language model 301. In the example in Figure 12, the answer is displayed as follows: "Regarding 'data,' the following definition is found in other contracts related to this case..." In addition, based on the information related to the task, file selection buttons 120321 and 120322 are displayed to show the file that serves as the basis for the answer on the user terminal 201. This allows the user to check auxiliary information based on information from cases common to or related to the contract being edited.

[0103] Referring to Figure 13, a search of the related information database 10121 using the clause to be edited will be explained. Figure 13 shows an example where the clause to be edited is identified during the editing process of contract D. The clause to be edited is identified by the editing status identification unit 1015 based on the operation input.

[0104] In step (13A), the document creation support device 101 acquires information for identifying the article to be edited and task information. At this time, the article to be edited and task information are identified or generated by the document creation support device 101 based on user operations.

[0105] In step (13B), the search unit 1014 searches the related information database 10121 based on the clause to be edited and the issue information. For example, the search unit 1014 searches for contracts or files in which the clause to be edited is "Article 1" of a "Confidentiality Agreement" to which a "clause label" has been assigned. In step (13C), the search unit 1014 retrieves information related to the issue from the related information database 10121.

[0106] In step (13D), the document creation support device 101 inputs the clause to be edited, the task information, and information related to the task into the language model 301. In step (13E), the document creation support device 101 obtains auxiliary information from the language model 301. Subsequently, the auxiliary information is presented.

[0107] Figure 14 shows an example of displaying auxiliary information based on information related to the issue obtained through a search using metadata. Figure 14 shows the editing screen 1401. The editing screen 1401 has a contract display area 1402, an auxiliary information display area 1403, a question input field 1404, and a submit button 1405.

[0108] The auxiliary information display area 1403 has a problem display area 14031 and an auxiliary information display area 14032. The problem display area 14031 displays text based on the problem generated by the acquisition unit 1013. In the example in Figure 14, a message is displayed to the user editing a confidentiality agreement, such as "Are you having trouble with Article 1 of the confidentiality agreement?", indicating that auxiliary information related to the clause being edited in the agreement is available.

[0109] The auxiliary information display area 14032 displays auxiliary information generated by the language model 301. In the example in Figure 14, the answer is displayed as follows: "In a confidentiality agreement, Article 1 defines 'confidential information' as follows...". In addition, based on the information related to the task, file selection buttons 140321 and 140322 are displayed to show the file that forms the basis of the answer on the user terminal 201. This allows the user to check auxiliary information that is common to the contract being edited or based on information of the contract type.

[0110] Referring to Figure 15, a search of the related information database 10121 using user attributes and behavioral information will be explained. Here, the case where both user attributes and behavioral information are used will be explained, but either user attributes or behavioral information may be used. Figure 15 is an example in which the user attributes and behavioral information of a user editing contract D are acquired by the document creation support device 101.

[0111] In step (15A), the document creation support device 101 acquires user attributes, behavioral information, and task information. At this time, the task information may be identified or generated by the document creation support device 101 based on user operations.

[0112] In step (15B), the search unit 1014 searches the related information database 10121 based on user attributes, behavioral information, and task information. For example, the search unit 1014 searches for contracts created by users with similar "experience level" to the user currently editing, or files suitable for the user currently editing, based on the user's behavioral information. The search unit 1014 also searches for contracts and files that the user has already reviewed to help them understand the files. In step (15C), the search unit 1014 retrieves task-related information from the related information database 10121.

[0113] In step (15D), the document creation support device 101 inputs user attributes, behavioral information, task information, and task-related information into the language model 301. In step (15E), the document creation support device 101 obtains auxiliary information from the language model 301. Subsequently, the auxiliary information is presented.

[0114] Figure 16 shows an example of displaying supplementary information based on information related to the issue obtained through a search using user attributes and behavioral information. Figure 16 shows the editing screen 1601. The editing screen 1601 has a contract display area 1602, a supplementary information display area 1603, a question input field 1604, and a submit button 1605.

[0115] The auxiliary information display area 1603 has a task display area 16031 and an auxiliary information display area 16032. The task display area 16031 displays text based on the task generated by the acquisition unit 1013. In the example in Figure 16, a message is displayed to the user editing the non-disclosure agreement, such as "Are you having trouble with the basic creation method?", indicating that auxiliary information to assist in creating the agreement is provided according to the user's level of proficiency in editing the agreement.

[0116] The auxiliary information display area 16032 displays auxiliary information generated by the language model 301. In the example in Figure 16, the response displayed takes into account the user's attributes, such as, "For general information on how to write a non-disclosure agreement, please refer to the following file..." Also, the response displayed takes into account the user's behavior information, such as, "The file "xxx" that you previously viewed was created based on the following template." Furthermore, based on information related to the task, file selection buttons 160321 and 160322 are displayed to show the file that forms the basis of the response on the user terminal 201. This allows the user to check auxiliary information based on their proficiency in creating agreements and auxiliary information according to their behavior history.

[0117] Next, with reference to Figure 17, an example of a hardware configuration when the document creation support system 10 is implemented using the computer 1700 will be described. Figure 17 is a diagram showing an example of the hardware configuration of the computer 1700.

[0118] As shown in Figure 17, the computer 1700 includes, for example, a processor 1701, memory 1702, storage device 1703, input I / F unit 1704, data I / F unit 1705, communication I / F unit 1706, and display device 1707. The computer 1700 includes, for example, a server computer and a personal computer (e.g., a desktop, laptop, tablet, etc.).

[0119] The processor 1701 is a control unit that controls various processes in the computer 1700 by executing programs stored in the memory 1702.

[0120] Memory 1702 is a storage medium such as RAM (Random Access Memory). Memory 1702 temporarily stores the program code of the program executed by the processor 1701, as well as data required during program execution.

[0121] The storage device 1703 is a non-volatile storage medium such as a hard disk drive (HDD) or flash memory. The storage device 1703 stores the operating system and various programs for realizing the above configurations.

[0122] The input interface unit 1704 is a device for receiving input from the user. The input interface unit 1704 can be, for example, a keyboard, mouse, touch panel, various sensors, or a wearable device. The input interface unit 1704 may be connected to the computer 1700 via an interface such as USB (Universal Serial Bus).

[0123] The data I / F unit 1705 is a device for inputting data from outside the computer 1700. The data I / F unit 1705 is, for example, a drive device for reading data stored on various storage media. The data I / F unit 1705 may be provided outside the computer 1700. If the data I / F unit 1705 is provided outside the computer 1700, it is connected to the computer 1700 via an interface such as USB.

[0124] The communication interface unit 1706 is a device for performing data communication with external devices of the computer 1700 via a network such as the Internet, either by wired or wireless connection. The communication interface unit 1706 may be located outside the computer 1700. If the communication interface unit 1706 is located outside the computer 1700, it is connected to the computer 1700 via an interface such as USB.

[0125] The display device 1707 is a device for displaying various types of information. The display device 1707 may be, for example, a liquid crystal display, an organic EL (Electro-Luminescence) display, or a display for a wearable device. The display device 1707 may be located outside the computer 1700. If the display device 1707 is located outside the computer 1700, it is connected to the computer 1700, for example, via a display cable. Furthermore, if a touch panel is used as the input I / F unit 1704, the display device 1707 may be integrated with the input I / F unit 1704.

[0126] The embodiment of this product has been described above. The information processing device according to this embodiment includes: an acquisition unit 1013 that acquires document information and issue information related to the document information based on the user's actions; a search unit 1014 that acquires information related to the issue indicated by the issue information from a related information database in which related information related to the document information is stored, based on the document information and the issue information; and an auxiliary information acquisition unit 1018 that acquires auxiliary information obtained by inputting the issue information and the information related to the issue into a language model.

[0127] This allows information relevant to a task, suitable for document information and designed to assist the user in drafting a contract, to be acquired, and supplementary information to be created and presented to the user. The user can access related information through the reference information included in the supplementary information, enabling them to understand the information used in drafting a contract. Therefore, it becomes possible to assist the user in drafting a contract.

[0128] In the above embodiment, the acquisition unit 1013 acquires question information input by the user as task information, the search unit 1014 performs natural language processing with the question information as input and interpreted information obtained by interpreting the question information as output, and acquires information related to the task based on the document information and the interpreted information, and the auxiliary information acquisition unit 1018 inputs the interpreted information and the information related to the task into a language model and acquires auxiliary information.

[0129] This allows users to obtain supplementary information based on interpretation, even when they input questions verbally, enabling them to obtain supplementary information for contract creation in a simpler way.

[0130] Furthermore, in the above embodiment, the acquisition unit 1013 may generate task information based on the editing status of the document indicated by the document information provided by the user.

[0131] This will generate auxiliary information tailored to the user's editing status, thereby improving user convenience.

[0132] Furthermore, in the above embodiment, an editing state identification unit 1015 may be further provided, which acquires user operation input on a document and identifies the editing state based on the operation input.

[0133] This will generate auxiliary information in response to user actions, thereby improving user convenience.

[0134] Furthermore, in the above embodiment, the document information includes metadata of the document indicated by the document information, and the related information database stores information of documents different from the document, associated with the metadata of the different documents, and the search unit 1014 may acquire information related to the issue based on the metadata and the issue information. The metadata may also include the document type of the document. The metadata may also include case information associated with the document, and the search unit 1014 may acquire information related to the issue based on the case information. This improves user convenience because information related to the document edited by the user can be acquired as information related to the issue.

[0135] Furthermore, in the above embodiment, the document is a document containing at least one clause, the issue information includes an editable clause which is the clause that the user is editing in the document, the related information database stores the clauses contained in the document labeled on a clause-by-clause basis, and the search unit 1014 may acquire information related to an issue related to the editable clause based on the document information and the editable clause.

[0136] This improves user convenience because it allows relevant information at the clause level in contracts edited by users to be retrieved as information relevant to the issue.

[0137] In the above embodiment, the system may further include an attribute information acquisition unit 1016 that acquires the user's user attributes, a search unit 1014 that acquires information related to the task based on the user attributes, document information and task information, and an auxiliary information acquisition unit 1018 that acquires auxiliary information based on the user attributes, task information and information related to the task. In the above embodiment, the system may further include an action information acquisition unit 1017 that acquires user action information related to editing the document indicated by the document information, a search unit 1014 that acquires information related to the task based on the action information, document information and task information, and an auxiliary information acquisition unit 1018 that acquires auxiliary information based on the action information, task information and information related to the task.

[0138] This will allow users to view supplementary information based on their proficiency in contract drafting and their activity history, thereby improving user convenience.

[0139] 10...Document creation support system, 101...Document creation support device, 201...User terminal, 301...Language model, 1013...Acquisition unit, 1014...Search unit, 1015...Editing status identification unit, 1016...Attribute information acquisition unit, 1017...Behavior information acquisition unit, 1018...Auxiliary information acquisition unit, 1019...Presentation unit

Claims

1. An information processing device comprising: an acquisition unit that acquires document information and issue information related to the document information based on user behavior; a search unit that acquires information related to the issue indicated by the issue information from a related information database in which related information related to the document information is stored, based on the document information and the issue information; and an auxiliary information acquisition unit that acquires auxiliary information obtained by inputting the issue information and the information related to the issue into a language model.

2. An information processing apparatus according to claim 1, wherein the acquisition unit acquires question information input from the user as the task information; the search unit performs natural language processing with the question information as input and interpreted information obtained by interpreting the question information as output, and acquires information related to the task based on the document information and the interpreted information; and the auxiliary information acquisition unit inputs the interpreted information and information related to the task into the language model and acquires the auxiliary information.

3. An information processing device according to claim 1, wherein the acquisition unit generates the task information based on the editing state of the document indicated by the document information by the user.

4. An information processing apparatus according to claim 3, further comprising an editing state identification unit that acquires the user's operation input to the document and identifies the editing state based on the operation input.

5. An information processing device according to claim 1, wherein the document information includes metadata of the document indicated by the document information, the related information database stores information of documents different from the document in association with the metadata of the different documents, and the search unit acquires information related to the problem based on the metadata and the problem information.

6. An information processing device according to claim 5, wherein the metadata includes the document type of the document.

7. An information processing device according to claim 5, wherein the metadata includes case information associated with the document, and the search unit acquires information related to the issue based on the case information.

8. An information processing apparatus according to claim 1, wherein the document information refers to a document containing at least one article, the issue information includes an editable article which is an article in the document that the user is editing, the related information database stores articles contained in the document labeled on an article-by-article basis, and the search unit acquires information related to the issue related to the editable article based on the document information and the editable article.

9. An information processing apparatus according to claim 1, further comprising: an attribute information acquisition unit for acquiring user attributes of the user; the search unit for acquiring information related to the issue based on the user attributes, the document information, and the issue information; and the auxiliary information acquisition unit for acquiring auxiliary information based on the user attributes, the issue information, and the information related to the issue.

10. An information processing apparatus according to claim 1, further comprising: an action information acquisition unit that acquires user action information related to editing a document indicated by the document information, the search unit acquires information related to the task based on the action information, the document information, and the task information, and the auxiliary information acquisition unit acquires the auxiliary information based on the action information, the task information, and the information related to the task.

11. A method comprising: a computer acquiring document information and issue information relating to the document information based on user behavior; acquiring information relating to the issue indicated by the issue information from an associated information database storing associated information relating to the document information, based on the document information and the issue information; and acquiring auxiliary information obtained by inputting the issue information and the information relating to the issue into a language model.

12. A program that causes a computer to perform the following actions: to acquire document information and issue information related to the document information based on the user's actions; to acquire information related to the issue indicated by the issue information from an issue information database in which related information related to the document information is stored, based on the document information and the issue information; and to acquire auxiliary information obtained by inputting the issue information and the information related to the issue into a language model.