Information processing system, information processing method and program

Abstract

To provide a mechanism for creating a template of a target document, even in a situation in which the document has not been created in the past.SOLUTION: An information processing system includes: title reception means for receiving input of a document title; prompt generation means for generating a prompt for instructing generation of a document template corresponding to the document title, on the basis of the received document title; and template acquisition means for inputting the generated prompt to a generative AI, and acquiring the document template corresponding to the document title.SELECTED DRAWING: Figure 12

Description

[Technical Field]

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

[0002] Japanese Patent No. 6985322 (Patent Document 1) is a document disclosing the background art of this technical field. This publication states that "in the workflow system 1 according to this embodiment, a template (hereinafter also referred to as a 'request template') in which required fields to be filled in for each request for approval are defined is created in advance using AI (Artificial Intelligence) technology such as machine learning, using requests for approval that have been previously drafted and approved" (see paragraph 0011). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6985322 Summary of the Invention [Problem to be solved by the invention]

[0004] In the invention described in this publication, a request form template is created using a request form that was previously drafted. Therefore, in order to create a request form template, the invention requires that a request form has been created in the past. If a request form has not been created in the past, there is a problem in that a request form template cannot be created. The present invention has been made in view of the above circumstances, and provides a mechanism for creating a template for a target document even when the target document has not been created in the past. [Means for solving the problem]

[0005] In order to solve the above problems, for example, the configurations described in the claims are adopted. The present application includes multiple means for solving the above-mentioned problems, and one example is an information processing system comprising a title receiving means for receiving input of a document title, a prompt generating means for generating a prompt that instructs the user to generate a document template corresponding to the document title based on the received document title, and a template acquisition means for inputting the generated prompt into a generation AI and acquiring a document template corresponding to the document title. [Effects of the Invention]

[0006] According to the present invention, it is possible to provide a mechanism for creating a template of a target document even when the target document has not been created in the past. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 shows an example of the configuration of a document generation system 100. [Figure 2] FIG. 2 shows an example of the configuration of the document generation server 101. [Figure 3] FIG. 3 shows an example of the configuration of the document management server 102. [Figure 4] FIG. 4 shows an example of the configuration of the generation AI server 103. [Figure 5] FIG. 5 shows an example of the configuration of the user terminal 104. [Figure 6] FIG. 6 shows an example of an employee DB 600 . [Figure 7] FIG. 7 shows an example of a template generation sequence. [Figure 8] FIG. 8 shows a continuation of the template generation sequence shown in FIG. [Figure 9] FIG. 9 shows an example of a search processing flow. [Figure 10] FIG. 10 shows an example of a document creation sequence. [Figure 11] FIG. 11 shows an example of a document template creation screen. [Figure 12]FIG. 12 shows an example of a document template creation screen. [Figure 13] FIG. 13 shows an example of a document template. [Figure 14] FIG. 14 shows an example of a document template. [Figure 15] FIG. 15 shows an example of a document template. [Figure 16] FIG. 16 shows an example of a document creation screen. [Figure 17] FIG. 17 shows an example of the completed document. [Figure 18] FIG. 18 shows an example of a template generation sequence according to a modified example. [Figure 19] FIG. 19 shows an example of a template generation sequence according to a modified example. DETAILED DESCRIPTION OF THE INVENTION

[0008] 1. Example Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0009] 1-1.Configuration FIG. 1 shows an example of the configuration of a document generation system 100. The document generation system 100 is an information processing system comprising a document generation server 101, a document management server 102, a generation AI server 103, and multiple user terminals 104, and each device is connected via a network. The network may be wired or wireless, and each device can send and receive information via the network.

[0010] Here, the document generation server 101 is a server that provides a document generation service. The document management server 102 is a server that manages the document template data generated by the document generation server 101 . The generation AI server 103 is a server that provides services using the generation AI. The user terminal 104 is a terminal of a user who uses the document generation service provided by the document generation server 101 .

[0011] Each device constituting the document generation system 100 may be, for example, a portable terminal (mobile terminal) such as a smartphone, tablet, mobile phone, or personal digital assistant (PDA), or may be a wearable terminal such as glasses, a wristwatch, or clothing. Each device may also be a stationary or portable computer, or a server located on the cloud or a network. Each device may also function as a VR (Virtual Reality) terminal, an AR (Augmented Reality) terminal, or an MR (Mixed Reality) terminal. Alternatively, a combination of multiple of these terminals may be used. For example, a combination of one smartphone and one wearable terminal may logically function as a single terminal. Other information processing terminals may also be used.

[0012] Each device constituting the document generation system 100 includes a processor that executes an operating system, applications, programs, etc., a main storage device such as a RAM (Random Access Memory), an auxiliary storage device such as an IC card, hard disk drive, SSD (Solid State Drive), or flash memory, a communication control unit such as a network card, wireless communication module, or mobile communication module, an input device such as a touch panel, keyboard, mouse, voice input, or input based on motion detection captured by a camera unit, and an output device such as a monitor or display. Note that the output device may also be a device or terminal that transmits information to be output to an external monitor, display, printer, or other device.

[0013] The main memory stores various programs, applications, etc. (modules), and the processor executes these programs and applications to realize the various functional elements of the overall system. These modules may be implemented in hardware, such as by integration. Each module may be an independent program or application, or may be implemented as a subprogram or function within a single integrated program or application.

[0014] In this specification, each module is described as a subject that performs processing, but in reality, the processing is performed by a processor that processes various programs, applications, etc. (modules). Various databases (DBs) are stored in the auxiliary storage device. A "database" is a functional element (storage unit) that stores a set of data so that it can accommodate any data manipulation (e.g., extraction, addition, deletion, overwriting, etc.) from a processor or an external computer. There are no limitations on how the database is implemented; for example, it can be a database management system, spreadsheet software, or a text file such as XML or JSON.

[0015] 1-1-1. Document Generation Server 101 FIG. 2 shows an example of the configuration of the document generation server 101. The document generation server 101 is configured, for example, by one or more servers arranged on a cloud. The server's main memory device 201 stores programs and applications such as a UI module 210, a reference template acquisition module 211, a template generation module 212, a save module 213, a template acquisition module 214, an attribute acquisition module 215, and a document generation module 216. These programs and applications are executed by the processor 203 to realize each functional element of the document generation server 101. Each module will be described below.

[0016] The UI module 210 provides various screens to the user and accepts user operations through the provided screens. For example, the module provides the user with a document template creation screen and accepts input of a document title from the user through the screen. The module also provides the user with a document creation screen and accepts designation of the user and document template to be entered in the document to be created.

[0017] The reference template acquisition module 211 acquires similar documents whose similarity to the document title received from the user satisfies a predetermined condition from the document management server 102. A method for acquiring similar documents (corresponding to reference templates to be described later) will be described later.

[0018] The template generation module 212 generates a prompt that instructs the user to generate a document template corresponding to the document title based on the document title received from the user. The module then inputs the generated prompt to the generation AI to obtain a document template corresponding to the document title. The module then provides the obtained document template to the user.

[0019] When generating a prompt, the module acquires input items in tag format and includes them in the prompt. Furthermore, the module includes in the prompt an instruction for outputting predetermined input items as input items in the tag format in the generated document template. Note that the predetermined input items here are terms that indicate where user attribute information is to be entered. Furthermore, when a similar document is acquired by the reference template acquisition module 211, the module includes in the prompt an instruction for generating a document template by referring to the acquired similar document. The predetermined input items and tag format input items will be described later.

[0020] The saving module 213 saves the document template instructed to be saved by the user in the document management server 102 .

[0021] The template acquisition module 214 acquires a document template designated by the user from the document management server 102 .

[0022] The attribute acquisition module 215 acquires attribute information about a user designated by the user to be included in the document to be created.

[0023] The document generation module 216 completes the document by replacing tag-format input items in the document template acquired by the template acquisition module 214 with user attribute information acquired by the attribute acquisition module 215. The module then provides the completed document to the user.

[0024] Next, the auxiliary storage device 202 of the document generation server 101 will be described. The auxiliary storage device 202 stores screen data 220 and an employee DB 600. Of these, the screen data 220 includes data for a document template creation screen and data for a document creation screen. On the other hand, the employee DB 600 is a database of employees of the organization to which the user of the user terminal 104 belongs.

[0025] 6 shows an example of an employee DB 600. The employee DB 600 shown in the figure is made up of records each having fields of employee ID, last name, first name, email address, affiliated organization, and job title.

[0026] 1-1-2. Document management server 102 FIG. 3 shows an example of the configuration of the document management server 102. The document management server 102 is configured, for example, by one or more servers arranged on a cloud. The server's main memory device 301 stores programs and applications such as a reference template search module 310, a storage module 311, and a template search module 312. These programs and applications are executed by a processor 303 to realize the various functional elements of the document management server 102. Each module will be described below.

[0027] The reference template search module 310 receives a request from the document generation server 101 and acquires similar documents whose similarity to the document title received from the user satisfies a predetermined condition. The module acquires the similar documents from the auxiliary storage device 302 and provides the acquired similar documents to the document generation server 101. Note that a specific method for acquiring similar documents (corresponding to reference templates described later) will be described later.

[0028] The saving module 311 receives a request from the document generation server 101 and saves in the auxiliary storage device 302 the document template instructed to be saved by the user.

[0029] The template search module 312 receives a request from the document generation server 101 and acquires the document template specified by the user from the auxiliary storage device 302. The module then provides the acquired document template to the document generation server 101.

[0030] Next, the auxiliary storage device 302 of the document management server 102 will be described. The auxiliary storage device 302 stores document template information 320, an inverted index 321, a distributed representation model 322, and distributed representation information 323. Of these, the document template information 320 is a collection of document templates generated by the document generation server 101.

[0031] The inverted index 321 is an index of the document template information 320, and is an index composed of a combination of each word (token) included in the document template and the ID of the document template in which the word appears. By referring to this inverted index 321, it is possible to quickly identify a document template that includes the word to be searched.

[0032] The distributed representation model 322 is a trained model that has been trained using text information published on the web or the like as training data, and that takes character strings as input and outputs distributed representations. For example, a method called word2vec is used to train this model. This method is a method of model training based on the co-occurrence between words in each sentence. Another method is, for example, Sentence BERT. This method is a method of model training that provides similarities between text pairs as training data in addition to the co-occurrence between words.

[0033] The distributed representation information 323 is a collection of distributed representations of each document template that constitutes the document template information 320. Each distributed representation is generated by inputting the document template into the distributed representation model 322. By referring to this distributed representation information 323, it is possible to identify the distributed representation of the target document template.

[0034] 1-1-3.Generation AI Server 103 FIG. 4 shows an example of the configuration of the generation AI server 103. The generation AI server 103 is configured, for example, by one or more servers placed on the cloud. The server's main memory device 401 stores programs and applications such as a generation processing module 410. The processor 403 executes these programs and applications to realize each functional element of the generation AI server 103.

[0035] The generation processing module 410 of the server inputs the prompt sent from the document generation server 101 into the generation AI and acquires the document template. Then, the module returns the acquired document template to the document generation server 101.

[0036] Here we will explain the generation AI. Generative AI (or generative AI) is a deep learning model that has been trained on a large dataset. In other words, the generative AI is a large-scale language model (LLM). Examples of large-scale language models include GPT-3, GPT-4 ("GPT" is a registered trademark), and BERT. Since large-scale language models are created by training deep learning models in advance using large datasets, they can be used without training data or additional training. In particular, this generation AI is a document generation AI that outputs text in response to input text.

[0037] Next, the auxiliary storage device 402 of the generation AI server 103 will be described. The auxiliary storage device 402 temporarily stores a prompt 420 and document template information 421. Of these, the prompt 420 is a prompt sent from the document generation server 101. On the other hand, the document template information 421 is a document template generated as a result of the generation process.

[0038] 1-1-4. User terminal 104 FIG. 5 shows an example of the configuration of the user terminal 104. The user terminal 104 is, for example, a smartphone, a tablet terminal, a notebook PC, a desktop PC, or the like. The terminal's main memory device 501 stores programs and applications such as a browser module 510. The processor 503 executes these programs and applications to realize the various functional elements of the user terminal 104.

[0039] The browser module 510 of the terminal communicates with the document generation server 101, acquires a screen corresponding to the user's operation, and displays it on the display. The auxiliary storage device 502 of the terminal stores the screen data 520 sent from the document generation server 101 .

[0040] 1-2.Operation 1-2-1. Template generation process Next, the template generation process executed by the document generation system 100 will be described with reference to FIGS. 7 and 8 show an example of a template generation sequence. A sequence 700 shown in these figures shows the flow of processing for generating a document template using a generation AI.

[0041] First, the user of the user terminal 104 performs an operation to display a document template creation screen. In response to this operation, the browser module 510 requests the document template creation screen from the document generation server 101 (step 701). In response to this request, the UI module 210 of the document generation server 101 returns the document template creation screen (step 702). The browser module 510 displays the received document template creation screen on the display (step 703).

[0042] Fig. 11 shows an example of a document template creation screen displayed on a display. Screen 1100 shown in the figure is a screen for creating a document template. This screen includes a document generation button 1101, a template name input field 1102, an editing area 1103, and an editing button group 1104. Each display element will be explained below.

[0043] First, a document generation button 1101 is a button for instructing the generation of a document template. The template name input field 1102 is a field for inputting the title of the document template to be generated. The editing area 1103 is an area for displaying a document template in an editable manner. The edit button group 1104 is a group of buttons for editing the document template displayed in the edit area 1103. The edit button group 1104 includes a pull-down menu 1105. The pull-down menu 1105 is a menu that allows the user to manually select and insert input items in tag format, which will be described later.

[0044] On this screen, the user inputs the title of the document template to be generated (for example, "Notice of Transfer") in the template name input field 1102, and selects the document generation button 1101 (step 704). In response to the selection of this document generation button 1101, the browser module 510 transmits a template generation request to the document generation server 101 (step 705). This template generation request includes the template title input by the user.

[0045] Upon receiving this request, the document generation server 101 generates a document template. Specifically, first, the reference template acquisition module 211 acquires a reference template (step 706). At that time, the module transmits a search request for the reference template to the document management server 102. This search request includes the template title entered by the user.

[0046] Upon receiving this request, the reference template search module 310 of the document management server 102 searches for a reference template (step 707). This search process will be described with reference to FIG. 9 shows an example of a search process flow. A flow 900 shown in the figure shows the flow of processing for searching the auxiliary storage device 302 for a reference template.

[0047] The reference template search module 310 first divides the template title entered by the user into words (tokens) (step 901). For example, the module divides the template title "Transfer Notice" into "Transfer" and "Notice."

[0048] Next, the module refers to the inverted index 321 to identify a document template that includes all of the words generated by the division (step 902).

[0049] Next, the module converts the template title into a distributed representation using the distributed representation model 322 (step 903). The module then refers to the distributed representation information 323, calculates the similarity between the distributed representation generated by the conversion and the distributed representations of all document templates created in the past, and identifies document templates with high similarity (step 904).

[0050] The similarity here is an index that indicates how closely two distributed representations (vectors) are oriented in the same direction, such as cosine similarity. Methods for identifying document templates include identifying the top n similarities and identifying all document templates whose similarity is greater than a threshold. The above is the description of the search processing flow 900.

[0051] Returning to the description of the sequence 700 shown in FIG. If the reference template search module 310 identifies a document template in the above steps 902 and 904, it returns the identified document template to the document generation server 101 as a reference template (step 708). If the module cannot identify a document template, it returns a message to that effect to the document generation server 101 (not shown).

[0052] Next, the template generation module 212 of the document generation server 101 generates a prompt (also called a "command statement") to be sent to the generation AI server 103 (step 709). The generated prompt includes the following sentence:

[0053] (1) A statement that instructs the user to generate a document template corresponding to the template title entered by the user. (2) A statement that instructs the document template to be generated in HTML format. (3) Tag format input items (4) A statement that instructs the generated document template to output specified input items as tag-format input items. (5) Text of the reference template (6) A statement instructing the user to generate a document template using the reference template.

[0054] The above-mentioned predetermined input items are specifically "document creation date," "last name," "first name," "personal email address," "organization name," and "job title." These input items are entered when using the document template. These input items are output in the specified display style by the directive (4) above. In addition, the notation of input items other than "Document Creation Date" is changed as shown below.

[0055] <Before change> <After change> "Name Last Name" "Basic Information > Last Name" "First name" "Basic information > First name" "Personal email address" "Basic information > Personal email address" "Name of organization" "Affiliation>Organization (chief)" "Position" "Affiliation > Position (Chief)"

[0056] The tag-formatted input items are input items written using HTML tags, which define the type and display style of the input item.

[0057] The template generation module 212 sends the prompt generated as described above to the generation AI server 103 (step 710).

[0058] The generation processing module 410 of the generation AI server 103 inputs the prompt sent from the document generation server 101 into the generation AI and acquires the document template (step 711). Then, the module returns the acquired document template to the document generation server 101 (step 712). As a result, the template generation module 212 acquires the document template.

[0059] Next, the UI module 210 of the document generation server 101 reflects the acquired document template on the document template creation screen (step 713). Fig. 12 shows an example of the document template creation screen on which the document template is reflected. In the screen 1200 shown in the figure, a document template 1201 is displayed in the editing area 1103.

[0060] Document template 1201 is a template for a transfer notice. This document template 1201 includes the following input fields: "Document Creation Date" 1202, "Basic Information > Last Name" 1203, "Basic Information > First Name" 1204, "Basic Information > Personal Email Address" 1205, "Affiliation > Organization (Chief Duty)" 1206, and "Affiliation > Position (Chief Duty)" 1207. These input fields are fields that are entered when using the document template. These input fields are converted and output from the input fields described above, and are each surrounded by a box and shaded. By looking at these character decorations, the user can recognize that these input fields are fields that should be entered when using document template 1201. At the same time, the user can recognize that these input fields will be automatically converted in the document creation process described below.

[0061] 13 shows another example of a document template, which is a grade notification template 1300. 14 shows yet another example of a document template. The document template 1400 shown in the figure is a template for a working conditions notice. 15 shows an example of yet another document template. The document template 1500 shown in the figure is a template for a joining notice form. The document templates 1300, 1400, and 1500 shown in these figures also include input items such as "Document creation date" 1202, "Basic information > Last name" 1203, and "Basic information > First name" 1204. The document templates 1300 and 1400 also include input items such as "Basic information > Personal email address" 1205, "Affiliation > Organization (chief)" 1206, and "Affiliation > Position (chief)" 1207.

[0062] 12, the user uses the edit button group 1104 to make necessary edits to the document template 1201 (step 714). When the document template is completed as a result of editing, the user selects a save button (not shown) (step 715). In response to this operation, the browser module 510 transmits a template save request to the document generation server 101 (step 716). This template save request includes the document template completed by the user.

[0063] Upon receiving this request, the document generation server 101 saves the document template. Specifically, the saving module 213 transmits the document template instructed to be saved by the user to the document management server 102 (step 717).

[0064] The storage module 311 of the document management server 102 stores the document template sent from the document generation server 101 in the auxiliary storage device 302 (step 718). In addition, the module divides the document template into words (tokens) and updates the inverted index 321 so that each word corresponds to the ID of the document template (step 719). Furthermore, the module converts the document template into distributed representations using the distributed representation model 322 and stores them in the auxiliary storage device 302 (step 720). The above is the description of the template generation sequence 700.

[0065] According to the above-described sequence 700, the user can generate a desired document template simply by inputting the template title. Therefore, unlike the prior art, there is no need to store document templates of the same type in order to generate a document template. In addition, the generative AI can be used to generate document templates. As mentioned above, the generative AI can be used without training data or additional learning. Therefore, it can generate document templates that comply with the latest legal amendments.

[0066] Furthermore, a document template in which predetermined input items have been converted into input items in tag format can be acquired without any special operation on the user's side. Furthermore, if a similar document has been created in the past, a document template can be acquired without the user having to perform any special operations. For example, a user can acquire an employment contract template (a template divided into chapters such as Article XX, Clause XX) that was created by referring to a contract.

[0067] 1-2-2. Document creation process Next, the document creation process executed by the document creation system 100 will be described with reference to FIG. An example of a document creation sequence is shown in Fig. 10. The sequence 1000 shown in the figure shows the flow of processing for creating a document using a previously created document template.

[0068] First, the user of the user terminal 104 performs an operation to display a document creation screen. In response to this operation, the browser module 510 requests the document creation screen from the document creation server 101 (step 1001). In response to this request, the UI module 210 of the document creation server 101 returns the document creation screen (step 1002). The browser module 510 displays the received document creation screen on the display (step 1003).

[0069] 16 shows an example of a document creation screen displayed on a display. A screen 1600 shown in the figure includes a template list 1601, radio buttons 1602, and a pull-down menu 1603. Each display element will be described below.

[0070] First, the template list 1601 is a list of document templates that the user has created in the past. The document templates included in the list are searched for and identified by the template search module 312 of the document management server 102. Each document template has a check box 1604 and a confirmation button 1605. Of these, the check box 1604 is a GUI for selecting the document template to be used, while the confirmation button 1605 is a button for displaying the contents of the document template in a pop-up.

[0071] Radio buttons 1602 are buttons for selecting a method for specifying the target employee. There are two methods for specifying: selecting from the employee names displayed in a pull-down menu 1603, or entering the employee ID in an input field (not shown) displayed in place of the pull-down menu. A pull-down menu 1603 is a menu for selecting the name of the target employee.

[0072] On this screen, the user selects the document template to be used and specifies the target employee (step 1004). Then, the user selects a create button (not shown). In response to the selection of this create button, the browser module 510 transmits a document creation request to the document generation server 101 (step 1005). This document creation request includes the ID of the document template selected by the user and the employee ID.

[0073] Upon receiving this request, the document generation server 101 generates the desired document. Specifically, first, the template acquisition module 214 acquires the document template selected by the user from the document management server 102 (step 1006). Next, the attribute acquisition module 215 acquires the attribute information of the employee specified by the user from the employee DB 600 (step 1007).

[0074] Then, the document generation module 216 completes the document by replacing the input items with appropriate information in the document template acquired in step 1006 (step 1008). Specifically, the module replaces "document creation date" with the execution date of this process (e.g., February 22, 2024) based on the tag. The module also replaces "basic information > last name," "basic information > first name," "basic information > personal email address," "affiliation > organization (chief duty)," and "affiliation > position (chief duty)" with the employee attribute information acquired in step 1007 based on the tag.

[0075] The module then displays the completed document on the display of the user terminal 104 (step 1009). FIG. 17 shows an example of a completed document displayed on the display. The completed document 1700 shown in the figure is a transfer notice. This completed document 1700 includes a document creation date 1701 and employee attribute information. The employee attribute information includes a last name 1702, a first name 1703, a personal email address 1704, an affiliated organization 1705, and a job title 1706.

[0076] The user may download and save this completed document 1700 to his / her own terminal, may send it to the terminal of the target employee, or may print it out on a printer. The document creation sequence 1000 has been described above.

[0077] According to the sequence 1000 described above, the user can obtain a document with the document creation date and employee attribute information entered simply by selecting the document template to be used and specifying the target employee.

[0078] 2. Variations The above embodiment may be modified as follows: The following modifications may be combined with each other.

[0079] 2-1. How to determine tag-format input items (Part 1) In the above embodiment, the tag-format input items included in the prompt are fixed. However, depending on the user, the corresponding attribute information may not be registered in the database, and the document template may not be able to be converted to the corresponding attribute information when used. Therefore, the tag-format input items to be used may be changed depending on the user. This allows different tag-format input items to be used for each user.

[0080] In this case, the template generation module 212 of the document generation server 101 identifies tag-format input items according to the document creator and includes them in the prompt in step 709 of FIG. 7. The input items identified by this module may be identified, for example, by referring to a table stored in advance in the document generation server 101. This table may be, for example, a table that associates user IDs with tag-format input items to be used. Alternatively, the tag-format input items may be identified by referring to the employee DB 600. Specifically, tag-format input items for which a corresponding column exists in the employee DB 600 may be identified as the input items to be used. For example, if the employee DB 600 has columns for "last name" and "first name," the input items to be used may be "basic information > last name" and "basic information > first name." However, if the employee DB 600 does not have a column for "personal email address," the input items to be used may not be "basic information > personal email address." Note that even if a column exists in the employee DB 600, if no data is stored in the column and the column is unused, the corresponding tag-format input item may be unused.

[0081] As another method, a list of input items in tag format may be displayed in step 703 of FIG. 7, and an input item selected by the user from the list may be used.

[0082] 2-2. Excluding input items with unusable tag formats In the above modification (1), usable tag-format input fields are specified in advance, and a document template is generated using only the specified input fields. Instead of this method, a document template may be generated using fixed input fields, and unusable input fields in the generated document template may be replaced with alternative text. This method also allows different tag-format input fields to be used for each user.

[0083] In this case, the document generation server 101: an input item identification module 217 (not shown) that identifies input items in an unusable tag format according to the creator of the document; a correction module 218 (not shown) that converts, when an input item in tag format corresponds to an input item in the specified unusable tag format in the document template acquired by the template generation module 212, the input item in tag format into an output format that does not use the tag; Further provided are:

[0084] Among these modules, the input item identification module 217 identifies input items in unusable tag formats by, for example, referring to a table stored in advance in the document generation server 101. This table may be, for example, a table associating user IDs with input items in usable tag formats, or a table associating user IDs with input items in unusable tag formats. Alternatively, unusable input items in tag formats may be identified by referring to the employee DB 600. Specifically, tag input items for which no corresponding column exists in the employee DB 600 may be identified as unusable. For example, if the employee DB 600 does not have a column for "personal email address," it may determine that "basic information > personal email address" is unusable. Note that even if a column exists in the employee DB 600, if no data is stored in the column and the column is unused, the corresponding tag input item may be determined to be unusable.

[0085] Alternatively, a list of input items in tag format may be displayed in step 703 of FIG. 7, and input items selected by the user as unavailable or not selected as available from the list may be determined to be unavailable.

[0086] On the other hand, the correction module 218 converts input items in unusable tag format into an output format that does not use the tag. For example, the module converts the tag-format input items "Basic Information > Last Name" and "Basic Information > First Name" into alternative wording that does not use the tag (for example, "Yamada Taro" or "Please enter your name"). This prevents character decoration specific to tag-format input items, allowing the user to recognize that the alternative wording will not be automatically converted in the document creation process described above.

[0087] 2-3. How to determine tag-format input items (part 2) In the above embodiment, the tag-format input items included in the prompt are fixed. However, the tag-format input items inserted into the document template differ depending on the type of document template. Therefore, the tag-format input items to be used may be changed depending on the title of the document template to be generated. This allows different tag-format input items to be used for each document template.

[0088] In this case, the template generation module 212 of the document generation server 101 identifies input items in tag format based on the document title and includes them in the prompt in step 709 of Fig. 7. At this time, the module acquires the corresponding input items from the document management server 102.

[0089] The document management server 102 further includes an input item identification module 313 (not shown) that receives a request from the document generation server 101 and identifies input items in tag format whose similarity to the document title satisfies a predetermined condition. The document management server 102 also stores in advance distributed representation information 324 (not shown). This distributed representation information 324 is a collection of distributed representations of input items in tag format. Each distributed representation is generated by inputting the input items in tag format into the distributed representation model 322. By referring to this distributed representation information 324, it is possible to identify the distributed representation of a target input item.

[0090] The input item identification module 313 identifies the relevant input item in the following procedure. First, the module uses the distributed representation model 322 to convert the template title entered by the user into a distributed representation. Next, the module refers to the distributed representation information 324, calculates the similarity between the distributed representation generated by the conversion and the distributed representations of all input items in tag format, and identifies input items with high similarity.

[0091] The similarity here is an index that indicates how closely two distributed representations (vectors) are oriented in the same direction, such as cosine similarity. Methods for identifying input items include identifying the top n similarities and identifying all input items for which similarity is greater than a threshold.

[0092] The input item identification module 313 sends the tag-format input items identified by the method described above to the document generation server 101. As a result, the template generation module 212 acquires the tag-format input items corresponding to the template title.

[0093] 2-4. How to determine tag-format input items (Part 3) In the above modification (3), the tag-format input items to be used are changed depending on the template title. Instead of this method, the tag-format input items to be used may be changed depending on the entire generated template document. This method also makes it possible to use different tag-format input items for each document template.

[0094] In this case, the template generation module 212 of the document generation server 101 acquires tag-format input items based on the acquired document template and generates another prompt including the tag-format input items. Furthermore, the module includes in the other prompt an instruction for outputting a predetermined input item in the document template as the tag-format input item. The module then inputs the generated other prompt to the generation AI to acquire another document template.

[0095] The document management server 102 further includes an input item identification module 314 (not shown) that receives a request from the document generation server 101 and identifies input items in a tag format whose similarity to the document template satisfies a predetermined condition. The document management server 102 also includes the distributed representation information 324 described in the above modification (3).

[0096] The template generation process according to this modification will be described below with reference to Fig. 18 and Fig. 19. Fig. 18 and Fig. 19 show an example of a template generation sequence according to this modification. In a sequence 1800 shown in these figures, steps common to Fig. 7 are assigned the same reference numerals. A description of these steps will be omitted.

[0097] In step 1801 of this sequence, the template generation module 212 of the document generation server 101 generates a prompt to be sent to the generation AI server 103. At this time, the generated prompt includes the following sentence:

[0098] (1) A statement that instructs the user to generate a document template corresponding to the template title entered by the user. (2) A statement that instructs the document template to be generated in HTML format. (3) Text of the reference template (4) A statement instructing the user to generate a document template using the reference template.

[0099] Unlike step 709 in Figure 7, the prompt generated here does not include "tag-format input items" and "a statement instructing that specified input items be output as tag-format input items in the document template to be generated."

[0100] The above prompt may also include a user request and an example of HTML text to improve the appearance of the document template to be generated. The user request here is a request specified by the user that specifies the appearance (in other words, the design) of the document template to be generated. This request may specify, for example, right-justification, left-justification, centering, font, etc. The HTML text, on the other hand, is an example of HTML text that has the appearance specified by the user.

[0101] The template generation module 212 sends the generated prompt to the generation AI server 103 (step 710).

[0102] When the module acquires the document template from the generation AI server 103 as a result of step 712, the module transmits a search request for the input items to the document management server 102 (step 1802). This search request includes the document template acquired in step 712.

[0103] Upon receiving this search request, the document management server 102 identifies input items in tag format that correspond to the document template (step 1803). Specifically, the input item identification module 314 converts the document template into a distributed representation using the distributed representation model 322. Next, the module refers to the distributed representation information 324, calculates the similarity between the distributed representation generated by the conversion and the distributed representations of all tag-format input items, and identifies input items with high similarity.

[0104] The similarity here is an index that indicates how closely two distributed representations (vectors) are oriented in the same direction, such as cosine similarity. Methods for identifying input items include identifying the top n similarities and identifying all input items for which similarity is greater than a threshold.

[0105] The input item identification module 314 sends the tag-formatted input items identified by the method described above to the document generation server 101 (step 1804). As a result, the template generation module 212 acquires the tag-formatted input items corresponding to the document template.

[0106] Next, the template generation module 212 generates a prompt to be sent to the generation AI server 103 (step 1805). The generated prompt includes the following sentence:

[0107] (1) Text of the document template obtained in step 712 (2) Input items in the tag format obtained in step 1804 (3) A statement instructing that the specified input items in the document template of (1) be output as input items in the tag format of (3).

[0108] The template generation module 212 sends the prompt generated as described above to the generation AI server 103 (step 1806).

[0109] The generation processing module 410 of the generation AI server 103 inputs the prompt sent from the document generation server 101 into the generation AI and acquires the document template (step 1807). Then, the module returns the acquired document template to the document generation server 101 (step 1808). As a result, the template generation module 212 acquires the document template.

[0110] Next, the UI module 210 of the document generation server 101 reflects the acquired document template on the document template creation screen (step 713). The subsequent processing is the same as in FIG.

[0111] In the variant described above, the document management server 102 is made to identify the input items of the tag format to be used, but instead of that server, the generation AI server 103 may be made to identify the input items of the tag format to be used. In this case, the document generation server 101 does not send a search request for the input items to the document management server 102. Therefore, steps 1802 to 1804 are omitted. Instead, in step 1805, the document generation server 101 sends a prompt including the following sentence to the generation AI server 103:

[0112] (1) Text of the document template obtained in step 712 (2) List of candidate input items in tag format (3) A statement that instructs the user to select appropriate input items from the candidate list (2) based on the text of the document template (1) and output the specified input items as the selected input items in the document template (1).

[0113] By sending such a prompt to the generation AI server 103, the generation AI server 103 can be made to specify the tag-format input items to be used. Even in this embodiment, it is possible to use different tag-format input items for each document template.

[0114] 2-5. How to determine tag-format input items (Part 4) 7, tag-format input items that are frequently specified by the user may be included in the prompt, thereby enabling input items that better match the user's usage needs to be embedded in the document template.

[0115] In this case, the document generation server 101 further includes a recording module 219 (not shown) that records the designation history of input items in the tag format designated by the document creator. In addition, in step 709 of Figure 7, the template generation module 212 obtains tag-format input items whose designation frequency is equal to or greater than a predetermined value, or whose designation frequency ranking is equal to or greater than a predetermined value, based on the recorded designation history, and includes them in the prompt.

[0116] The recording module 219 records input items in tag format that are specified when the generation AI is not in use. That is, the module records input items that are specified when the document generation button 1101 shown in Fig. 11 is not selected and the user manually creates a document template. This allows the module to record only input items that the user has specified of their own volition.

[0117] The input items may be recorded without distinguishing between users, or may be recorded with distinguishing between users. When the input items are recorded with distinguishing between users, input items that better match the user's usage needs can be embedded in the document template.

[0118] 2-6. Tags In the above embodiment, the document template is generated in HTML format (see step 709 in FIG. 7). However, the format of the document template is not limited to HTML, and the document template may be generated using other markup languages. When using other markup languages, input items using tags for the language to be used are inserted into the document template.

[0119] 2-7. Input items The tag-format input items inserted into the prompt in the above example (e.g., "Document Creation Date") are merely examples. Other input items (e.g., employee number, employee pension number, etc.) may also be inserted into the prompt.

[0120] 2-8.Functional layout In the above document generation system 100, the document management function and the generation AI function are provided on a server different from the document generation server 101. However, this functional arrangement is merely an example. The functions of each server in this system may be arranged as appropriate depending on the environment in which the document generation service is provided. For example, the document management function and the generation AI function may be provided on the document generation server 101.

[0121] 2-9. How to search for reference templates In the above search processing flow 900, both a full-text search (steps 901 and 902) and a vector search (steps 903 and 904) are performed. However, it is not essential to perform both searches. Only one of the searches may be performed.

[0122] 2-10. Document completion process In the above document creation sequence 1000, the document completion process (see step 1008) is performed in the document generation server 101. This completion process may be performed by the user terminal 104. In this case, the user terminal 104 acquires the document template and employee attribute information from the document generation server 101, and completes the document on the terminal based on this information.

[0123] 2-11. Where to save document templates The above embodiment assumes that the document template generated by the document generation server 101 is used as a reference template. Therefore, in the above embodiment, the document template generated by the document generation server 101 is stored in the document management server 102. However, it is not essential to use the document template generated by the document generation server 101 as a reference template. When the document template generated by the document generation server 101 is not used as a reference template, but a document template stored in advance in the document management server 102 is used as a reference template, the document template generated by the document generation server 101 may be stored in that server.

[0124] 2-12.Other The present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0125] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as the programs, tables, and files that implement each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

[0126] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and do not necessarily show all the control lines and information lines in the product. In reality, it can be assumed that almost all components are interconnected. The above-described embodiments disclose at least the configurations described in the claims. [Explanation of symbols]

[0127] 100...document generation system, 101...document generation server, 102...document management server, 103...generation AI server, 104...user terminal, 1100...document template creation screen, 1101...document generation button, 1102...template name input field, 1103...editing area 1104...editing button group, 1200...document template creation screen, 1201...document template, 1300...document template, 1400...document template, 1500...document template, 1600...document creation screen, 1601...template list, 1602...radio button, 1603...pull-down menu, 1604...check box, 1605...confirmation button, 1700...completed document

Claims

1. a title receiving means for receiving an input of a document title; a prompt generating means for generating a prompt for instructing to generate a document template corresponding to the document title based on the received document title; a template acquisition means for inputting the generated prompt to a generation AI and acquiring a document template corresponding to the document title; An information processing system comprising:

2. The prompt generating means Takes input in the form of tags and includes it in the prompt, Furthermore, the prompt includes an instruction for outputting predetermined input items as input items in the tag format in the document template to be generated.

2. The information processing system according to claim 1, wherein:

3. 3. The information processing system according to claim 2, wherein the predetermined input item is a phrase indicating an input location of the user attribute information.

4. a user accepting means for accepting a designation of a user to be written on the document; attribute acquisition means for acquiring user attribute information of the specified user; a document generating means for completing a document by replacing the tag-format input items in the acquired document template with the acquired user attribute information; The information processing system according to claim 3 , further comprising:

5. 3. The information processing system according to claim 2, wherein the prompt generating means identifies the tag-format input items according to the creator of the document and includes the input items in the prompt.

6. an input item specifying means for specifying input items in an unusable tag format according to the creator of the document; a correction means for converting an input item in the predetermined tag format into an output format that does not use the tag when the input item in the acquired document template corresponds to an input item in the specified unusable tag format; The information processing system according to claim 2 , further comprising:

7. a similar document acquisition unit for acquiring similar documents whose similarity to the document title satisfies a predetermined condition; The prompt generating means includes in the prompt an instruction for generating a document template by referring to the acquired similar documents.

2. The information processing system according to claim 1, wherein:

8. The prompt generating means acquiring tag-format input items based on the acquired document template, and generating another prompt including the tag-format input items; Furthermore, the other prompt includes an instruction for outputting predetermined input items in the document template as input items in the tag format, The template acquisition means inputs the generated different prompt to the generation AI and acquires a different document template.

2. The information processing system according to claim 1, wherein:

9. 3. The information processing system according to claim 2, wherein the prompt generating means identifies the tag-format input items based on the document title and includes the tag-format input items in the prompt.

10. further comprising a recording means for recording a history of input items in a tag format designated by a creator of the document; The prompt generating means Based on the designation history, tag-format input items whose designation frequency is equal to or greater than a predetermined value or whose designation frequency ranking is equal to or greater than a predetermined value are acquired and included in the prompt.

3. The information processing system according to claim 2, wherein:

11. 1. A computer-implemented information processing method, comprising: accepting input of a document title; generating a prompt to instruct the user to generate a document template corresponding to the document title based on the received document title; inputting the generated prompt into a generation AI to obtain a document template corresponding to the document title; An information processing method comprising:

12. On the computer, accepting input of a document title; generating a prompt to instruct the user to generate a document template corresponding to the document title based on the received document title; inputting the generated prompt into a generation AI to obtain a document template corresponding to the document title; A program to execute.

Images