Input support method and input support program

Abstract

To reduce the workload involved in entering information when applying for travel expenses. [Solution] The input support device receives text data tx1 entered in response to the chatbot CT's message, "Please enter the details of your business trip." The input support device analyzes the received text data tx1 based on a language model and obtains the analysis results. The input support device refers to the obtained analysis results and extracts information corresponding to each input item in the drafting screen S1. Based on the extracted results, the input support device determines whether information is missing for any of the multiple input items in the drafting screen S1. If information is missing for any of the input items, the input support device outputs chatbot CT's message m9. Message m9 is a message stating, "Some of the business trip details are missing. Please add the missing information or re-enter it."

Description

Technical Field

[0001] The present invention relates to an input support method and an input support program.

Background Art

[0002] Conventionally, there is a business expense application system for employees of companies, government offices, etc. to apply for business expenses. Business expenses are expenses incurred by business trips, and include, for example, transportation expenses, accommodation expenses, etc. In the business expense application system, for example, input of various information such as the place of duty, duties, business trip date, means of transportation, etc. is required.

[0003] As the prior art, for example, a predetermined word sequence is cut out from an input character string, each cut-out word is searched in a word dictionary, and a corresponding keyword group is extracted from a dictionary column in which a keyword of the word dictionary or its analogical word is obtained by comparison and match, and is displayed all at once for a plurality of item correspondences of a display means.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the prior art, there is a problem that it takes time and effort to input information when applying for business expenses. For example, when a user manually inputs one by one at a plurality of input locations on the screen of the business expense application system, the work is performed while determining what should be input at each location, which increases the work load and may lead to input omission.

[0006] In one aspect, an object of the present invention is to reduce the work load required for information input when applying for business expenses.

Means for Solving the Problems

[0007] In one embodiment, an input support method is provided to assist in inputting information related to the application for travel expenses, which receives text data entered in response to a first input request prompting the user to input the details of the travel in natural language, extracts information corresponding to each of the multiple input items on the screen related to the application for travel expenses from the received text data, determines whether information is missing for any of the multiple input items when inputting the extracted information into the corresponding input items on the screen, and if it is determined that information is missing for any of the input items, outputs a second input request prompting the user to input information for that input item. [Effects of the Invention]

[0008] According to one aspect of the present invention, the workload involved in entering information when applying for travel expenses can be reduced. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 is an explanatory diagram showing one embodiment of the input support method according to the embodiment. [Figure 2] Figure 2 is an explanatory diagram showing an example of the system configuration of the travel expense application system 200. [Figure 3] Figure 3 is a block diagram showing an example of the hardware configuration of the input support device 201. [Figure 4] Figure 4 is a block diagram showing an example of the functional configuration of the input support device 201. [Figure 5] Figure 5 is an explanatory diagram (part 1) showing an example of screen transitions on the business trip expense application screen. [Figure 6] Figure 6 is an explanatory diagram (part 2) showing an example of screen transitions on the travel expense application screen. [Figure 7] Figure 7 is an explanatory diagram (part 3) showing an example of screen transitions on the business trip expense application screen. [Figure 8] Figure 8 is an explanatory diagram (part 4) showing an example of screen transitions on the travel expense application screen. [Figure 9] FIG. 9 is an explanatory diagram (Part 5) showing an example of screen transition of a business trip expense application screen. [Figure 10] FIG. 10 is an explanatory diagram (Part 6) showing an example of screen transition of a business trip expense application screen. [Figure 11] FIG. 11 is an explanatory diagram (Part 7) showing an example of screen transition of a business trip expense application screen. [Figure 12] FIG. 12 is an explanatory diagram (Part 8) showing an example of screen transition of a business trip expense application screen. [Figure 13] FIG. 13 is an explanatory diagram showing a specific example of an analysis result. [Figure 14] FIG. 14 is a flowchart showing an example of an input support processing procedure of the input support device 201. [Figure 15] FIG. 15 is a flowchart (Part 1) showing an example of a specific processing procedure of normal mode processing. [Figure 16] FIG. 16 is a flowchart (Part 2) showing an example of a specific processing procedure of normal mode processing. [Figure 17] FIG. 17 is a flowchart showing an example of a specific processing procedure of GPT mode processing.

MODE FOR CARRYING OUT THE INVENTION

[0010] Hereinafter, embodiments of an input support method and an input support program according to the present invention will be described in detail with reference to the drawings.

[0011] (Embodiment) FIG. 1 is an explanatory diagram showing an example of an input support method according to an embodiment. In FIG. 1, the information processing device 101 is a computer that supports information input related to the application for business trip expenses. A business trip means being temporarily dispatched to another region or place for work at a company, office, etc.

[0012] Business trip expenses refer to the expenses incurred due to business trips. The application of business trip expenses is an act for employees or others to apply to the company or the like for reimbursement of the expenses they have paid on behalf of the company during the business trip, or an act for them to apply to the company or the like in advance for payment of the expenses to be incurred during the business trip. In the case of the duties of civil servants, business trips are sometimes called "travels", and business trip expenses are sometimes called "travel expenses".

[0013] The application of business trip expenses is carried out, for example, in a business trip expense application system. A business trip expense application system is a computer system for settling and applying business trip expenses (travel expenses). In a conventional business trip expense application system, when applying for business trip expenses, the user manually enters information one by one for a plurality of input locations (input items) prepared on the screen of the business trip expense application system.

[0014] Therefore, in a conventional business trip expense application system, the user has to perform the work while judging one by one which part of the screen to enter what according to the content of the business trip, and it takes time and effort to input information when applying for business trip expenses.

[0015] Here, when applying for business trip expenses, for example, by using a system that applies LLM (Large Language Model) technology, it is possible to input business trip information without being restricted by form, extract the information required for the application from the input information, and input it into the corresponding location on the screen.

[0016] However, it is assumed that a situation may occur where the information required for the application of business trip expenses is insufficient due to the user inputting business trip information without being restricted by form. For example, if an input omission is found after applying for business trip expenses, the application will be sent back and the user will have to do it again, increasing the user's workload.

[0017] Therefore, in this embodiment, when applying for travel expenses, a method for inputting travel details in natural language and prompting the user to input any missing information in natural language will be described. Here, an example of processing by the information processing device 101 (corresponding to the processing in (1) to (5) below) will be described. Here, an example of supporting information input by user 102 for applying for travel expenses will be used.

[0018] (1) The information processing device 101 outputs an input request 110 (first input request) prompting the user to input the details of the business trip in natural language. Here, the details of the business trip are the details of the business trip for which the business trip expense claim is made. Natural language is a natural form of writing, and may be, for example, a conversational style of writing or a bulleted list.

[0019] Specifically, for example, the information processing device 101 outputs an input request 110 as a message to the chatbot 103. The chatbot 103 is a computer program that engages in conversation with a user (for example, user 102) regarding the application for travel expenses. The input request 110 is a message from the chatbot 103 such as, for example, "Please enter the details of your trip," "Please enter the details of your trip in natural language," or "Please enter the details of your trip in the way you think is appropriate."

[0020] (2) The information processing device 101 receives the text data 120 that was input in response to the outputted input request 110 (first input request). The text data 120 is a natural language expression of the business trip details as intended by the user 102. The text data 120 is entered as text input or voice input, for example, as the user 102's response to the chatbot 103.

[0021] (3) The information processing device 101 extracts information corresponding to each of the multiple input items in the screen 105 from the received text data 120. Specifically, for example, the information processing device 101 extracts information corresponding to each input item by referring to the analysis result 130 obtained by analyzing the text data 120 based on the language model 104.

[0022] Here, the language model 104 is a machine learning model in the field of natural language processing, and corresponds to, for example, a generative AI (document generation AI) for generating text. The language model 104 is, for example, an LLM (Large Language Model). An LLM is a language model that is constructed by performing deep learning training on a large amount of text data. For example, the information processing device 101 can obtain the analysis result 130 by giving the language model 104 a prompt (instruction) to analyze the text data 120 and output the desired information in the desired format.

[0023] Screen 105 is the operation screen for applying for travel expenses. Screen 105 includes several pre-configured input fields. Each of these input fields is an item for which information must be entered when applying for travel expenses. Examples of input fields include place of business, purpose of business, date of business, departure point, departure time, and destination.

[0024] The information processing device 101 may, for example, store candidate information to be entered into an input item in the screen 105 in association with that input item, and extract information corresponding to the candidate information from the text data 120, thereby extracting information corresponding to the candidate information. The information corresponding to the candidate information may be, for example, the same information as the candidate information, or information that has the same meaning but a different word form than the candidate information.

[0025] In the example shown in Figure 1, screen 105 is assumed to contain input items 105a, 105b, and 105c. In this case, the information processing device 101 refers to the analysis result 130 and extracts information corresponding to each input item 105a to 105c. The analysis result 130 includes the information analysis results for each input item 105a to 105c. The information analysis results represent the analysis results for each input item 105a to 105c obtained by analyzing the text data 120 based on the language model 104.

[0026] For example, the information analysis result for input item 105a shows the information about input item 105a extracted from the text data 120. However, there are cases where the text data 120 does not contain information about input item 105a. In this case, the information analysis result for input item 105a will indicate that there is no information corresponding to input item 105a.

[0027] Here, the information analysis results for each input item 105a and 105b represent the information extracted from the text data 120 for each input item 105a and 105b. Furthermore, the information analysis result for input item 105c represents the absence of information corresponding to input item 105c. In this case, information corresponding to input items 105a and 105b is extracted, while information corresponding to input item 105c is not.

[0028] (4) When the information processing device 101 inputs the extracted information into the corresponding input field on the screen 105, it determines whether or not information is missing for any of the input fields 105a to 105c. In this case, the information corresponding to input field 105c has not been extracted. In this case, it is determined that information for input field 105c is missing.

[0029] (5) If the information processing device 101 determines that there is insufficient information for input item 105c, it outputs an input request 140 (a second input request). The input request 140 prompts the user to input information about input item 105c. Specifically, for example, the information processing device 101 outputs the input request 140 as a message to the chatbot 103.

[0030] For example, if input field 105c is set to "Departure location," then input request 140 is a message from chatbot 103 such as, "There is insufficient information about the departure location. Please enter additional information."

[0031] Thus, the information processing device 101 can reduce the workload involved in entering information when applying for travel expenses. For example, the information processing device 101 can reduce the effort and time required for information entry by allowing users to input details of their travel trips in freely written text (natural language) when applying for travel expenses.

[0032] Furthermore, if the information processing device 101 is missing information for any input items in the entered travel details, it can prevent input omissions by prompting the user to enter the information for those missing items. In the example in Figure 1, the travel details entered by user 102 are missing the "departure location" information. In this case, the information processing device 101 can prevent input omissions by prompting user 102 to enter the "departure location" information, thereby alerting the user that the entered travel details did not include the "departure location" information.

[0033] (Example of system configuration for the travel expense application system 200) Next, we will describe an example of the system configuration of the travel expense application system 200, which includes the information processing device 101 shown in Figure 1. Here, we will explain using the example of applying the information processing device 101 shown in Figure 1 to the input support device 201 within the travel expense application system 200.

[0034] Figure 2 is an explanatory diagram showing an example of the system configuration of the travel expense application system 200. In Figure 2, the travel expense application system 200 includes an input support device 201, a client device 202, and an analysis server 203. In the travel expense application system 200, the input support device 201, the client device 202, and the analysis server 203 are connected via a wired or wireless network 210. The network 210 is, for example, the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network).

[0035] Here, the travel expense application system 200 is a computer system for company or government employee to apply for travel expenses. The input support device 201 assists in inputting information related to the travel expense application. The input support device 201 is, for example, a server.

[0036] Client device 202 is a computer used by users of the travel expense application system 200. Users are, for example, employees of companies or government offices. Client device 202 can be, for example, a PC (Personal Computer) or a tablet PC.

[0037] The analysis server 203 is a computer that provides services that make language model M available. Language model M is an LLM (Large-Scale Language Model), such as GPT (Generative Pre-trained Transformer).

[0038] In the travel expense application system 200, the client device 202 can obtain analysis results from the analysis server 203 by, for example, accessing the analysis server 203 and providing the language model M with prompts (commands) for analyzing text data.

[0039] In this example, the input support device 201 and the client device 202 are provided as separate components, but this is not the only option. For example, the input support device 201 may be implemented by the client device 202. Furthermore, the input support device 201 may also have the functionality of the analysis server 203. The travel expense application system 200 may include multiple client devices 202.

[0040] (Example hardware configuration of input support device 201) Next, we will describe an example of the hardware configuration of the input support device 201.

[0041] Figure 3 is a block diagram showing an example of the hardware configuration of the input support device 201. In Figure 3, the input support device 201 includes a CPU (Central Processing Unit) 301, memory 302, disk drive 303, disk 304, communication interface 305, portable recording medium interface 306, and portable recording medium 307. Each component is connected by a bus 300.

[0042] Here, the CPU 301 is responsible for the overall control of the input assistance device 201. The CPU 301 may have multiple cores. The memory 302 includes, for example, ROM (Read Only Memory) and RAM (Random Access Memory). The program stored in the memory 302 is loaded into the CPU 301, causing the CPU 301 to execute the coded process.

[0043] The disk drive 303 controls the reading and writing of data to the disk 304 according to the control of the CPU 301. The disk 304 stores the data written under the control of the disk drive 303. The disk 304 is, for example, a magnetic disk, an optical disk, etc.

[0044] The communication interface 305 is connected to the network 210 via a communication line, and through the network 210, it is connected to an external computer (for example, the client device 202 and analysis server 203 shown in Figure 2). The communication interface 305 manages the interface between the network 210 and the inside of the device, and controls the input and output of data from the external computer. The communication interface 305 is, for example, a modem or a LAN adapter.

[0045] The portable recording medium interface 306 controls the reading and writing of data to the portable recording medium 307 according to the control of the CPU 301. The portable recording medium 307 stores the data written under the control of the portable recording medium interface 306. The portable recording medium 307 is, for example, a CD (Compact Disc)-ROM, a DVD (Digital Versatile Disk), or a USB (Universal Serial Bus) memory.

[0046] The input support device 201 may also have, in addition to the components described above, an input device, a display, etc. Furthermore, the input support device 201 does not necessarily have to include, for example, the portable recording medium I / F 306 and the portable recording medium 307 among the components described above. The client device 202 and analysis server 203 shown in Figure 2 can also be implemented with the same hardware configuration as the input support device 201. However, the client device 202 may have, in addition to the components described above, an input device, a display, a speaker, a microphone, etc.

[0047] (Example of functional configuration of input support device 201) Next, we will describe an example of the functional configuration of the input support device 201.

[0048] Figure 4 is a block diagram showing an example of the functional configuration of the input support device 201. In Figure 4, the input support device 201 includes a conversation control unit 401, a reception unit 402, an extraction unit 403, a determination unit 404, and an input unit 405. The conversation control unit 401 to the input unit 405 function as the control unit 400, and specifically, this function is realized by having the CPU 301 execute a program stored in a storage device such as the memory 302, disk 304, or portable recording medium 307 shown in Figure 3, or by using the communication I / F 305. The processing results of each functional unit are stored in a storage device such as the memory 302 or disk 304.

[0049] The conversation control unit 401 controls the conversation regarding the application for travel expenses. The conversation regarding the application for travel expenses takes place between the user and the chatbot (for example, the chatbot CT shown in Figure 5 below). The user is the person applying for the travel expenses. The user may also be the person applying for the travel expenses on behalf of someone else.

[0050] For example, a user can use the client device 202 shown in Figure 2 to have a conversation with the chatbot CT regarding the application for travel expenses. For instance, by logging into the travel expense application system 200 from the client device 202 and accessing the input support device 201, a conversation regarding the application for travel expenses can be initiated.

[0051] Conversations regarding the application for travel expenses take place, for example, on a conversation screen S2 included in the travel expense application screen AS, as shown in Figure 5 below, which is displayed on the client device 202. However, conversations regarding travel expense applications (e.g., the travel expense application screen AS) may also take place using a display not shown on the input support device 201.

[0052] The conversation control unit 401 outputs a first input request prompting the user to input details of the business trip in natural language. The first input request, for example, asks the user to input details of the business trip in the way they would like. Specifically, for example, in a conversation regarding the application for business trip expenses, the conversation control unit 401 outputs the first input request as a message (statement) from the chatbot CT. In this case, the conversation control unit 401 may output the first input request along with an input box that allows the user to input details of the business trip in natural language.

[0053] More specifically, for example, the conversation control unit 401 sends a first input request to the client device 202, causing the client device 202 to display the first input request as a chatbot CT message. Alternatively, the conversation control unit 401 may display the first input request as a chatbot CT message on a display (not shown) of its own device.

[0054] A concrete example of the first input request will be described later using Figure 6.

[0055] The reception unit 402 receives the text data entered in response to the first input request. The text data is entered as the user's message in response to the first input request, for example, in a conversation regarding the application for travel expenses. The text data expresses the details of the travel in natural language as the user intends.

[0056] The details of the business trip are those for which travel expenses are claimed, such as when, where, and for what purpose the trip will take place. Some users are able to provide all the information required for travel expense claims, while others are not.

[0057] Specifically, for example, the reception unit 402 accepts the input text data by receiving the text data entered in response to the first input request from the client device 202. Alternatively, the reception unit 402 may also accept the input text data through user operation input using an input device (not shown) of its own device.

[0058] The text data is, for example, text data entered using an input device such as a keyboard or mouse (not shown) of the client device 202. Alternatively, the text data may be, for example, audio data entered using a microphone (not shown) of the client device 202. The text data is converted into text data, for example, by speech recognition processing.

[0059] Specific examples of text data will be discussed later using Figure 7. The input text data will be output as user messages (statements) in a conversation, for example, regarding a travel expense claim.

[0060] The extraction unit 403 extracts information corresponding to each of the multiple input fields on the screen related to the travel expense application from the received text data. Specifically, for example, the extraction unit 403 obtains the analysis results by analyzing the received text data based on the language model M. Then, the extraction unit 403 refers to the obtained analysis results and extracts information corresponding to each input field. The language model M is, for example, an LLM such as GPT.

[0061] The screen for applying for travel expenses is, for example, the drafting screen S1 included in the travel expense application screen AS, as shown in Figure 5 below. Each of the multiple input fields is an item for which information must be entered when applying for travel expenses. The input fields include, for example, the place of business, the purpose of business, and the date of the business trip.

[0062] Multiple input fields may include input fields related to the itinerary. An itinerary is the route from one place (departure point) to another place (destination). Input fields related to the itinerary may include, for example, the departure point, departure time, mode of transport, and destination. In addition, multiple input fields may include input fields related to accommodation. Input fields related to accommodation may include, for example, the place of stay and the cost of accommodation.

[0063] More specifically, for example, the extraction unit 403 identifies the target information from the received text data and creates a prompt (instruction statement) to output it in the desired format. The target information is, for example, the information for each input item required when applying for travel expenses. The desired format is, for example, a format that the computer (input support device 201) can recognize for the information for each input item.

[0064] The prompts indicate, for example, what information should be considered as belonging to which input field. The prompts should specify, as adequately as possible, the descriptions of each element (input field) to be extracted from the document data (natural language), examples of each element, and the procedures for extraction.

[0065] For example, the prompts may be tuned to determine "the 21st of this month" as "November 21st" if the current month is November, or to determine "go by train" as "public transport." The input support device 201 may also pass pre-set transportation options and brief descriptions to the language model M for each mode of transport (e.g., train) extracted from the text data. In this case, the language model M will determine the option that is most appropriate for that mode of transport, for example, based on the training data.

[0066] Regarding "business," the language model M extracts content that it estimates to be the purpose from the text data by prompting it to extract the purpose of the business trip. In this case, it may be possible to tune the system so that it only extracts content that is explicitly stated as the purpose, such as "for the purpose of ○○," while providing information on what kind of content is appropriate as the purpose and avoiding forcing the extraction of content that merely "looks like a purpose." For example, if it is questionable whether it is the purpose, the prompt may be tuned to ask the input support device 201 for clarification without extracting the content.

[0067] Furthermore, tuning can be done to absorb variations in wording, such as classifying "SDGs meeting" as a task or "SDGs conference" as a task. Also, for relative dates such as "next Monday," the prompt can be tuned to use today's date as a reference point by passing today's date, obtained from another program, to the language model M.

[0068] Furthermore, for information that is difficult to extract using only the language model M, the input support device 201 may extract only the components of that information and process them using another program. Alternatively, the system may be tuned to first determine whether the document data is for claiming travel expenses, and if not, to return that fact as part of the analysis results.

[0069] Next, the extraction unit 403 sends the created prompt to the analysis server 203 shown in Figure 2. The analysis server 203 provides the received prompt to the language model M, generates an analysis result by analyzing the text data, and sends the generated analysis result to the input support device 201. By receiving the analysis result from the analysis server 203, the extraction unit 403 can obtain the analysis result obtained by analyzing the received text data based on the language model M.

[0070] The analysis results include, for example, information analysis results identified by analyzing the text data for each input item included in multiple input items. The information analysis results show, for example, information about the input items extracted from the text data using a language model M. However, there may be input items for which no information was extracted from the text data. In this case, the information analysis results indicate that there is no information corresponding to that input item.

[0071] The language model M may be provided by the input support device 201. In this case, the extraction unit 403 may obtain the analysis result of analyzing the received text data by providing the created prompt to the language model M of its own device.

[0072] Specific examples of the analysis results will be discussed later using Figure 13.

[0073] The extraction unit 403 then refers to the acquired analysis results and extracts information corresponding to each input item in the drafting screen S1. The input items to be extracted include, for example, the place of business, the purpose of business, the date of the business trip, and input items related to the itinerary (departure point, departure time, means of transportation, destination).

[0074] Specifically, for example, the extraction unit 403 extracts information corresponding to the input items "place of business," "business," and "date of business trip" from the acquired analysis results. Furthermore, the extraction unit 403 extracts information corresponding to the input items "departure point," "departure time," "means of transportation," and "destination" for each itinerary from the acquired analysis results.

[0075] However, there are cases where information corresponding to an input item is not extracted. For example, if the information analysis results for an input item included in the analysis results indicate that there is no information corresponding to that input item, then that information will not be extracted.

[0076] Furthermore, the extraction unit 403 may determine whether the business trip dates are two or more consecutive days based on the information in the input fields related to the extracted business trip dates. For example, if the information corresponding to the extracted input field "business trip date" is "November 24th", the extraction unit 403 will determine that the business trip dates are not two or more consecutive days.

[0077] Furthermore, if the information corresponding to the extracted input item "Business Trip Date" is "November 24th" and "November 25th", the extraction unit 403 determines that the business trip dates are two or more consecutive days. Also, if the information corresponding to the extracted input item "Business Trip Date" is "November 24th to November 26th", the extraction unit 403 determines that the business trip dates are two or more consecutive days.

[0078] The extraction unit 403 then extracts information on input items related to accommodation from the received text data if the business trip dates are two or more consecutive days. Specifically, for example, if the business trip dates are two or more consecutive days, the extraction unit 403 extracts information corresponding to the input items "location of accommodation" and "accommodation cost" from the acquired analysis results. On the other hand, if the business trip dates are not two or more consecutive days, the extraction unit 403 does not extract information on input items related to accommodation.

[0079] Furthermore, the extraction unit 403 may refer to the acquired analysis results and extract information on input items related to accounts. Input items related to accounts include, for example, account names and code numbers. Specifically, for example, the extraction unit 403 may refer to the analysis results and extract the code number for general accounts, the code number for accounting management expenses, the code number for ordinary travel expenses, etc.

[0080] Based on the extracted results, the determination unit 404 determines whether or not information is missing for any of the multiple input fields on the screen related to the application for travel expenses. Specifically, for example, the determination unit 404 determines whether or not the corresponding information has been extracted for each of the input fields "Place of business," "Business," and "Date of business trip."

[0081] Furthermore, the determination unit 404 determines, for each iteration, whether the corresponding information has been extracted for each of the input items "Departure point," "Departure time," "Means of transport," and "Destination." If the corresponding information has not been extracted for any of the input items, the determination unit 404 determines that there is insufficient information for that input item.

[0082] Furthermore, the determination unit 404 may determine that there is insufficient information for the input items related to accommodation if the business trip days are two or more consecutive days and information for the input items related to accommodation has not been extracted. Specifically, for example, the determination unit 404 determines whether the corresponding information for the input items "location of accommodation" and "accommodation expenses" has been extracted for each business trip day other than the last day of a business trip of two or more consecutive days.

[0083] The determination unit 404 then determines that information is missing for any input item if corresponding information is not extracted for any of the input items. Furthermore, if corresponding information is extracted for all input items to be extracted, the determination unit 404 determines that information is not missing for any of the input items.

[0084] The input unit 405 inputs the extracted information into the corresponding input fields on the screen. Information that identifies the input location on the screen where the information corresponding to each input field is entered is stored in advance, for example. Specifically, for example, the input unit 405 inputs the information corresponding to the extracted input field "Business Trip Date" into the input box corresponding to the input field "Business Trip Date" on the drafting screen S1 (see Figure 10, for example).

[0085] Furthermore, if the conversation control unit 401 determines that information is missing for any input item, it outputs a second input request prompting the user to input information for that input item. The second input request, for example, asks the user to input additional information that was missing from the travel details entered by the user.

[0086] Specifically, for example, in a conversation regarding the application for travel expenses, the conversation control unit 401 outputs a second input request as a message from the chatbot CT. For example, suppose it is determined that there is insufficient information for the input item "Place of business". In this case, the second input request would be a message from the chatbot CT such as, "There is insufficient information about the place of business. Please enter the place of business."

[0087] More specifically, for example, the conversation control unit 401 sends a second input request to the client device 202, causing the client device 202 to display the second input request as a message from the chatbot CT. Alternatively, the conversation control unit 401 may display the second input request as a message from the chatbot CT on a display (not shown) of its own device.

[0088] Furthermore, the conversation control unit 401 may output the extraction results for each of the multiple input items along with the second input request. The extraction results for the input items indicate information corresponding to the extracted input items, or that no information corresponding to the input items was extracted.

[0089] In the following explanation, information from any of the input fields that are determined to be missing may be referred to as "information from the first input field."

[0090] Furthermore, the reception unit 402 receives the information of the first input item entered in response to the second input request. The information of the first input item is entered, for example, as a user message in response to the second input request in a conversation regarding the application for travel expenses. The information of the first input item may be entered as text or as voice input.

[0091] Specifically, for example, the reception unit 402 receives information about the first input item from the client device 202 in response to a second input request. Alternatively, the reception unit 402 may also receive information about the first input item through user input using an input device (not shown) of its own device.

[0092] A concrete example of the information in the first input field will be described later using Figure 9.

[0093] When the input unit 405 receives information for the first input item, it inputs that information into the first input item on the screen. For example, if the first input item is "Place of work", the input unit 405 inputs the received information for the first input item "Place of work" into the input box corresponding to the first input item "Place of work" on the drafting screen S1.

[0094] However, the information for the first input item may also be directly entered into the input box corresponding to the first input item "Place of Work" on the drafting screen S1 by user input in response to a second input request.

[0095] Furthermore, if the first input item is "business trip date," and the system receives information for the first input item "business trip date" in response to the second input request, the extraction unit 403 may, for example, determine whether the business trip dates are two or more consecutive days based on the received information for the first input item "business trip date." If the business trip dates are two or more consecutive days, the extraction unit 403 may then extract information for input items related to accommodation facilities from the received text data.

[0096] The second input request may, for example, ask the user to enter additional information that was missing from the business trip details they entered, or to re-enter the business trip details. For example, if the first input item is "Place of business," the second input request may be a chatbot CT message such as, "There is insufficient information about the place of business. Please enter the place of business or re-enter the business trip details."

[0097] In this case, when text data is input in response to the second input request, the reception unit 402 receives the re-entered text data in response to the second input request. Then, the processing of the received text data is performed by the respective functional units described above (extraction unit 403, determination unit 404, input unit 405, conversation control unit 401).

[0098] The processing of the input unit 405 may be performed either before or after the processing of the determination unit 404, or it may be performed in parallel with the processing of the determination unit 404. For example, if the determination unit 404 determines that there is insufficient information for the first input item, and then the input unit 405 receives information for the first input item in response to a second input request, it may perform the input of information for all input items, including the first input item, into the screen all at once.

[0099] Furthermore, reception unit 402 may accept the selection of either the first mode or the second mode when applying for travel expenses. Here, the first mode is a mode in which travel details are entered in natural language (GPT mode). The second mode is a mode in which information is entered in a question-and-answer format (normal mode).

[0100] Furthermore, the conversation control unit 401 may output a first input request when the first mode is selected. On the other hand, when the second mode is selected, the conversation control unit 401 may output a third input request prompting the input of information about each input item included in the multiple input items.

[0101] Specifically, for example, when the second mode is selected, the conversation control unit 401 outputs a third input request as a chatbot CT message for each input item in a conversation regarding the application for travel expenses. The reception unit 402 may also receive information on the input items entered in response to the third input request.

[0102] The input unit 405 may then input the information of the received input item into the corresponding input item on the screen. This allows the input support device 201 to sequentially prompt the user to input information about the input items while switching between them. The user can sequentially input information for each input item in a question-and-answer format.

[0103] The functional units of the input support device 201 (conversation control unit 401 to input unit 405) may be implemented by multiple computers within the travel expense application system 200 (for example, the input support device 201 and the client device 202). In this case, communication between functional units of different computers is performed, for example, by sending and receiving data between functional units via the network 210.

[0104] (Example of screen transitions when submitting a travel expense claim) Next, we will explain an example of screen transitions when submitting a travel expense application in the travel expense application system 200. Here, we assume that the user (applicant) accesses the input support device 201 from the client device 202 to submit a travel expense application. In this case, the screen for submitting the travel expense application is displayed on the client device 202. User operations are performed, for example, by inputting data using an input device (not shown) on the client device 202, or by voice input using a microphone (not shown).

[0105] Figures 5 to 12 are explanatory diagrams showing examples of screen transitions in the travel expense application screen. In Figure 5, the travel expense application screen AS is an example of an operation screen displayed when applying for travel expenses. The travel expense application screen AS may be output to the client device 202 by, for example, the input support device 201.

[0106] The business trip expense application screen AS includes a drafting screen S1 and a conversation screen S2. Drafting screen S1 is an example of an operation screen for applying for business trip expenses. Drafting screen S1 includes multiple input fields for which information must be entered when applying for business trip expenses. Specifically, drafting screen S1 includes basic information d1, payment information d2, traveler information d3, route information d4 (see Figure 10), route details d5 (see Figure 10), accommodation expense information d6 (see Figure 11), and payee information d7 (see Figure 11).

[0107] Basic Information d1 shows basic information about the business trip (travel). Basic Information d1 includes, for example, input boxes b1 for the input field "Period of Business (corresponding to the business trip dates)", input box b2 for the input field "Place of Business", and input box b3 for the input field "Purpose of Business". Payment Information d2 shows information about the payment of business trip expenses (travel expenses). Traveler Information d3 shows information about the business traveler (traveler).

[0108] Route information d4 shows information about the route to the destination. Route information d4 includes, for example, the input box b4 for the input item "Travel date (corresponding to the business trip date)", the checkbox b5 for the input item "Means of transportation" in Itinerary 2, the input box b6 for the input item "Departure point", and the input box b7 for the input item "Destination" (see Figure 10).

[0109] Route details d5 displays detailed information about the route to the destination. Route details d5 includes, for example, an input box b8 for the input field "Place of business" (see Figure 10). Route details d5 also includes, for example, display fields f1~f4 that display the route search results for each leg of the journey (see Figure 11).

[0110] Accommodation expense information d6 shows information about accommodation expenses. Accommodation expense information d6 includes, for example, input boxes b9 for "Travel date (corresponding to business trip date)", input box b10 for "Location of accommodation", and input box b11 for "Accommodation expense" (see Figure 11). Payee information d7 shows information about the payee of business trip expenses.

[0111] Conversation screen S2 is an example of an operation screen for conducting a conversation regarding the application for travel expenses. Conversation screen S2 displays messages from the chatbot CT and messages from the user. In Figure 5, conversation screen S2 displays messages m1~m3,m5 from the chatbot CT and message m4 from the user.

[0112] Here, we will explain using the example of a user designated as a "test traveler" who initiates a business trip (travel). In this case, if the user selects button 501, the user's message m4, "Yes," will be entered in response to chatbot CT's message m3. If the user selects button 502, they can change the traveler (business traveler).

[0113] Furthermore, the chatbot CT's message m5 prompts the user to select a method for drafting a proposal. "New Proposal" corresponds to the second mode (normal mode), where information is entered in a question-and-answer format. "Copy" corresponds to the copy mode, where information entered in a previous application is copied for the current application. "GPT Mode" corresponds to the first mode, where travel details are entered in natural language.

[0114] This section explains the case where GPT mode is selected. In this case, the user selects button 503 to select GPT mode. The user can also select normal mode by selecting button 504. Furthermore, the user can select copy mode by selecting button 505.

[0115] In Figure 6, in the conversation screen S2, the user's message m6, "GPT mode," is input in response to chatbot CT's message m5. In this case, the conversation control unit 401 outputs chatbot CT's message m7. Message m7 is chatbot CT's message, "Please enter details of your business trip," and includes an input box 601 and buttons 602 and 603.

[0116] Input box 601 is a text box where the details of the business trip can be entered in natural language. Button 602 is an operation button to confirm the input of the business trip details. Button 603 is an operation button to restart from the selection of the proposal method. Message m7 corresponds to the first input request prompting the user to enter the details of the business trip in natural language.

[0117] Here, we assume that the user enters text data tx1 (details of the business trip) into input box 601 and selects button 602.

[0118] In Figure 7, the user's message m8 is entered in the conversation screen S2. Message m8 represents the text data tx1 (business trip details) entered in the input box 601 shown in Figure 6. In this case, the reception unit 402 receives the text data tx1 entered in response to the chatbot CT's message m7. Then, the extraction unit 403 obtains the analysis result by analyzing the received text data tx1 based on the language model M.

[0119] Here, using Figure 13, we will explain a specific example of the analysis results obtained by analyzing the text data tx1 based on the language model M.

[0120] Figure 13 is an explanatory diagram showing a specific example of the analysis results. In Figure 13, analysis result 1300 shows the information analysis results for each input item identified by analyzing the text data tx1 (see Figure 6). The input items are the place of business, business, business date, accommodation, accommodation expenses, and the departure point, departure time, mode of transportation, and destination for each leg of the trip.

[0121] For example, the data analysis result for the input field "Place of business" indicates "Kinki Bureau of Economy, Trade and Industry." The data analysis result for the input field "Date of business trip" indicates "November 24, 2024 (Sunday)" and "November 25, 2024 (Monday)." The data analysis result for the input field "Departure point" in Itinerary 1 indicates "Home." The data analysis result for the input field "Means of transportation" in Itinerary 3 indicates "-." "-(Null)" indicates that there is no information corresponding to the input field.

[0122] The extraction unit 403 refers to the acquired analysis results 1300 and extracts information corresponding to each input item in the drafting screen S1. Specifically, for example, the extraction unit 403 extracts the information "Kinki Bureau of Economy, Trade and Industry" corresponding to the input item "Place of work" from the analysis results 1300. The extraction unit 403 also extracts the information "Attended the SDGs Policy Review Meeting" corresponding to the input item "Duties" from the analysis results 1300.

[0123] Furthermore, the extraction unit 403 extracts the information "November 24, 2024 (Sun)" and "November 25, 2024 (Mon)" from the analysis result 1300, which correspond to the input item "Business Trip Date". In addition, the extraction unit 403 extracts the information "Home", "9:00", "Public Transportation", and "Kinki Bureau of Economy, Trade and Industry" from the analysis result 1300, which correspond to each input item "Departure Point", "Departure Time", "Means of Transportation", and "Destination".

[0124] Furthermore, from the analysis results 1300, for itinerary 2 on "November 24, 2024 (Sunday)", information corresponding to each input item "Departure point", "Departure time", "Means of transportation", and "Destination" is extracted: "Kinki Bureau of Economy, Trade and Industry", "17:30", "Official vehicle / Rental vehicle", and "Dormy Inn XX".

[0125] Furthermore, the extraction unit 403 extracts information corresponding to each input item "Departure point," "Departure time," and "Destination" for itinerary 3 on "November 25, 2024 (Monday)" from the analysis result 1300: "Dormy Inn XX," "8:30," and "Kinki Bureau of Economy, Trade and Industry." However, the information corresponding to the input item "Means of Transportation" for itinerary 3 is not extracted because the information analysis result of analysis result 1300 is "-".

[0126] Furthermore, the extraction unit 403 extracts information corresponding to each input item "Departure point" and "Destination," namely "Kinki Bureau of Economy, Trade and Industry" and "Home," for itinerary 4 on "November 25, 2024 (Monday)," from the analysis result 1300. However, information corresponding to each input item "Time" and "Means of Transportation" for itinerary 4 is not extracted because the information analysis result of the analysis result 1300 is "-".

[0127] Furthermore, the extraction unit 403 determines that the business trip dates are two or more consecutive days based on the information "November 24, 2024 (Sun)" and "November 25, 2024 (Mon)" corresponding to the extracted input item "Business Trip Date". In this case, the extraction unit 403 extracts the information "Dormy Inn XX" and "9,000 yen" corresponding to the input items "Accommodation Location" and "Accommodation Cost" respectively from the analysis result 1300 for the business trip date "November 24, 2024 (Sun)".

[0128] Next, the determination unit 404 determines, based on the extracted results, whether or not information is missing for any of the multiple input items in the drafting screen S1. In this case, the information corresponding to the input item "means of transportation" for Itinerary 3 on the business trip date "November 25, 2024 (Monday)" and the information corresponding to the input items "time" and "means of transportation" for Itinerary 4 on the business trip date "November 25, 2024 (Monday)" have not been extracted.

[0129] Therefore, the determination unit 404 determines that the input item "means of transportation" for Itinerary 3 on the business trip date "November 25, 2024 (Monday)" and the input items "time" and "means of transportation" for Itinerary 4 on the business trip date "November 25, 2024 (Monday)" are missing. These input items correspond to the "first input items".

[0130] In this case, as shown in Figures 7 and 8, the conversation control unit 401 outputs the extracted result information 701 to 706 along with the chatbot CT's message m9. Message m9 is the chatbot CT's message, "Some of the business trip details are missing. Please add the missing information or re-enter it." Extracted result information 701 shows the extracted results for the input items "Place of business" and "Business."

[0131] Extraction result information 702 shows the extraction results for the input items "Departure point", "Time (Departure time)", "Means of transportation", and "Destination" for Itinerary 1 on the business trip date "November 24, 2024 (Sun)". Extraction result information 703 shows the extraction results for the input items "Departure point", "Time (Departure time)", "Means of transportation", and "Destination" for Itinerary 2 on the business trip date "November 24, 2024 (Sun)". Extraction result information 704 shows the extraction results for the input items "Accommodation" and "Accommodation cost" for the business trip date "November 24, 2024 (Sun)".

[0132] Extraction result information 705 shows the extraction results for the input items "Departure point", "Time (Departure time)", "Means of transportation", and "Destination" for Itinerary 3 on the business trip date "November 25, 2024 (Monday)". Extraction result information 706 shows the extraction results for the input items "Departure point", "Time (Departure time)", "Means of transportation", and "Destination" for Itinerary 4 on the business trip date "November 25, 2024 (Monday)".

[0133] According to message m9, the user can understand that there is missing information in the travel details they entered, and that they need to either add the missing information or re-enter the travel details. Furthermore, according to the extracted results information 701-706, the user can check the information of each input item extracted from the travel details they entered.

[0134] Furthermore, according to the extracted result information 705, the user can see that the input item "Means of Transportation" for Itinerary 3 of the business trip date "November 25, 2024 (Monday)" is missing. Furthermore, according to the extracted result information 706, the user can see that the input items "Departure Time" and "Means of Transportation" for Itinerary 4 of the business trip date "November 25, 2024 (Monday)" are missing.

[0135] Here, input fields with missing information (means of transportation for itinerary 3, departure time and means of transportation for itinerary 4) are displayed in a different way than other input fields (for example, by changing the font color or boldness). This allows the user to intuitively identify which input fields are missing information.

[0136] The user can add information for the "Means of Transportation" input field in Itinerary 3 by selecting input box 801. The user can also add information for the "Departure Time" input field in Itinerary 4 by selecting input box 802. Furthermore, the user can add information for the "Means of Transportation" input field in Itinerary 4 by selecting input box 803.

[0137] After entering information in input boxes 801-803, the user can select button 804 to add information for each input item (means of transportation for itinerary 3, departure time and means of transportation for itinerary 4). Alternatively, the user can select button 805 to restart the input process from the beginning of the business trip details.

[0138] Here, we assume that, as shown in Figure 9, the information "Public Transportation" is entered into input box 801, the information "18:00" is entered into input box 802, the information "Public Transportation" is entered into input box 803, and button 804 is selected. When information is entered into each of the input boxes 801 to 803, check marks 901 to 903 will be displayed accordingly.

[0139] When button 804 is selected, the input unit 405 inputs the information for each input item into the corresponding input item on the drafting screen S1. The information for each input item is either information extracted from the analysis result 1300 or information entered into each input box 801 to 803.

[0140] As shown in Figure 10, while information is being entered into the corresponding input fields in the proposal screen S1, for example, the conversation screen S2 displays information as if it were being entered in a question-and-answer format. For example, in response to the chatbot CT's message m10, the checkbox b5 for "Means of Transportation" in the input field for Step 2 in the proposal screen S1 is automatically checked as if the user had responded with message m11.

[0141] Furthermore, in response to chatbot CT's message m12, "Dormy Inn ○○" is automatically entered into the "Destination" input box b7 in the "Destination" input field for "Plan 2" on the proposal screen S1, as if the user had responded with message m13.

[0142] Similarly, in the example in Figure 10, "November 24, 2006 (Sunday)" is automatically entered into input box b4 of the input item "Travel Date (corresponding to business trip date)" in the drafting screen S1 (however, the Western calendar notation is converted to the Japanese calendar notation). Also, "Kinki Bureau of Economy, Trade and Industry" is automatically entered into input box b6 of the input item "Departure Point" in Itinerary 2 of the drafting screen S1. Also, "Kinki Bureau of Economy, Trade and Industry" is automatically entered into input box b8 of the input item "Place of Business" in the drafting screen S1.

[0143] Although not shown in the diagram, the input box b1 (see Figure 5) for the "Period of Business (corresponding to the business trip dates)" in the drafting screen S1 will be automatically populated with "November 24th (Sun) - November 25th (Mon), 2024". Additionally, the input box b2 (see Figure 5) for the "Place of Business" in the drafting screen S1 will be automatically populated with "Kinki Bureau of Economy, Trade and Industry". Furthermore, the input box b3 (see Figure 5) for the "Purpose of Business" in the drafting screen S1 will be automatically populated with "SDGs Policy Review Meeting".

[0144] Additionally, the input box b9 (see Figure 11) for the "Travel Date (corresponding to business trip date)" field in the drafting screen S1 will be automatically populated with "November 24, 2024 (Sunday)". Furthermore, the input box b10 (see Figure 11) for the "Accommodation Location" field in the drafting screen S1 will be automatically populated with "Dormy Inn ○○". Finally, the input box b11 (see Figure 11) for the "Accommodation Cost" field in the drafting screen S1 will be automatically populated with "9,000".

[0145] Once information has been entered into the input fields on the proposal screen S1, the conversation control unit 401 outputs the travel information created based on the input content as chatbot CT message m14 on the conversation screen S2. The travel information includes the traveler, travel date, destination, purpose, route search results for each of the four itineraries, accommodation costs, and travel expenses.

[0146] Furthermore, the route search results for each of the steps 1 to 4 are displayed in the respective display fields f1 to f4 within the drafting screen S1. The input support device 201 performs route searches for each step based, for example, on the departure point, departure time, mode of transport, and destination of each step. Route searches may be performed using, for example, an existing route search website.

[0147] The user can refer to chatbot CT's message m14 to check if there are any problems with the travel information. If there are no problems with the travel information, the user can select button 1201 to end the conversation with chatbot CT and proceed with the expense claim process. If there are problems with the travel information, the user can select button 1202 to restart the process from entering the travel details.

[0148] In this way, users can input their travel details as they wish, and the various information required when applying for travel expenses will be automatically entered into the corresponding input fields on the drafting screen S1, thus reducing the time and effort required for information entry. Furthermore, if any information is missing, the chatbot CT will inform the user, preventing input errors.

[0149] (Input support processing procedure for input support device 201) Next, the input support processing procedure of the input support device 201 will be explained using Figure 14.

[0150] Figure 14 is a flowchart showing an example of the input support processing procedure of the input support device 201. In the flowchart of Figure 14, first, the input support device 201 determines whether or not a change operation has been performed by the traveler (business traveler) (step S1401).

[0151] If a change operation is performed for the traveler (step S1401: Yes), the input support device 201 searches for the traveler (step S1402) and proceeds to step S1403. The traveler search is performed, for example, based on the user's input, using name, ID, department name, etc.

[0152] On the other hand, if no change operation is performed by the traveler (step S1401: No), the input support device 201 determines whether or not GPT mode is selected (step S1403). If GPT mode is not selected (step S1403: No), the input support device 201 performs normal mode processing (step S1404) and proceeds to step S1406.

[0153] The specific processing procedure for normal mode will be described later using Figures 15 and 16. Here, we will explain the case where GPT mode is not selected (step S1403: No) and the system switches to normal mode, but it is also possible to switch to copy mode.

[0154] On the other hand, if GPT mode is selected (step S1403: Yes), the input support device 201 performs GPT mode processing (step S1405). The specific processing procedure for GPT mode will be described later with reference to Figure 17.

[0155] Next, the input support device 201 outputs travel information based on the input content (step S1406). The travel information is information for the user to confirm the input content, and corresponds, for example, to message m14 of the chatbot CT shown in Figures 11 and 12.

[0156] The input support device 201 then determines whether or not the user has confirmed the input content (step S1407). Specifically, for example, the input support device 201 may determine that the input content has been confirmed if button 1201 shown in Figure 12 is selected. Alternatively, the input support device 201 may determine that the input content has not been confirmed if button 1202 shown in Figure 12 is selected.

[0157] If input content confirmation is not performed (step S1407: No), the input support device 201 returns to step S1401. On the other hand, if input content confirmation is performed (step S1407: Yes), the input support device 201 terminates the conversation with the chatbot CT on the conversation screen S2 of the travel expense application screen AS, outputs only the drafting screen S1 (step S1408), and ends the series of processes according to this flowchart.

[0158] As a result, the input support device 201 can reduce the workload associated with entering information when applying for travel expenses (business trip expenses).

[0159] Next, using Figures 15 and 16, we will explain the specific processing procedure for the normal mode processing of step S1404 shown in Figure 14.

[0160] Figures 15 and 16 are flowcharts illustrating an example of a specific processing procedure in normal mode. In the flowchart of Figure 15, first, the input support device 201 outputs a message from the chatbot CT prompting the user to input information about the input item "Place of work / task" (step S1501).

[0161] Next, the input support device 201 determines whether or not information for the input item "Place of work / task" has been entered (step S1502). Here, the input support device 201 waits for information for the input item "Place of work / task" to be entered (step S1502: No).

[0162] Then, if information for the input field "Place of work / task" is entered (Step S1502: Yes), the input support device 201 outputs a message from the chatbot CT prompting the user to enter information for the input field "Subject" (Step S1503).

[0163] Next, the input support device 201 determines whether or not information for the input item "Subject" has been entered (step S1504). Here, the input support device 201 waits for information for the input item "Subject" to be entered (step S1504: No).

[0164] Then, if information for the input field "Subject" is entered (Step S1504: Yes), the input support device 201 outputs a message from the chatbot CT prompting the user to enter information for the input field "Travel Date" (Step S1505).

[0165] Next, the input support device 201 determines whether or not information for the input item "travel date" has been entered (step S1506). Here, the input support device 201 waits for information for the input item "travel date" to be entered (step S1506: No).

[0166] Then, if information for the input field "Travel Date" is entered (Step S1506: Yes), the input support device 201 outputs a message from the chatbot CT prompting the user to enter information for the input field "Means of Transportation" (Step S1507).

[0167] Next, the input support device 201 determines whether or not information for the input item "means of transportation" has been entered (step S1508). Here, the input support device 201 waits for information for the input item "means of transportation" to be entered (step S1508: No). If information for the input item "means of transportation" has been entered (step S1508: Yes), the input support device 201 proceeds to step S1601 shown in Figure 16.

[0168] In the flowchart of Figure 16, first, the input support device 201 outputs a message from the chatbot CT prompting the user to enter information about the input item "Departure point / Destination" (step S1601). Next, the input support device 201 determines whether or not information about the input item "Departure point / Destination" has been entered (step S1602).

[0169] Here, the input support device 201 waits for the input item "Departure / Destination" information to be entered (Step S1602: No). If the input item "Departure / Destination" information is entered (Step S1602: Yes), the input support device 201 outputs a message from the chatbot CT prompting the user to enter information about the input item "Departure Time" (Step S1603).

[0170] Next, the input support device 201 determines whether or not information for the input item "departure time" has been entered (step S1604). Here, the input support device 201 waits for information for the input item "departure time" to be entered (step S1604: No). If information for the input item "departure time" has been entered (step S1604: Yes), the input support device 201 performs a route search based on the entered departure point, destination, and departure time (step S1605) and outputs the route search results (step S1606).

[0171] Next, the input support device 201 determines whether or not to create the next step in response to the user's input (step S1607). If the next step is to be created (step S1607: Yes), the input support device 201 returns to step S1505 shown in Figure 15.

[0172] On the other hand, if the next itinerary is not created (step S1607: No), the input support device 201 determines whether or not it is a day trip (step S1608). Specifically, for example, if the entered travel dates are two or more consecutive days, the input support device 201 determines that it is not a day trip. Also, if the entered travel dates are not two or more consecutive days, the input support device 201 determines that it is a day trip.

[0173] Here, if it is a day trip (step S1608: Yes), the input support device 201 returns to the step that called normal mode processing. On the other hand, if it is not a day trip (step S1608: No), the input support device 201 outputs a message from the chatbot CT prompting the user to enter information about the input item "accommodation information" (step S1609). Accommodation information includes, for example, the location of the accommodation and the cost of the accommodation.

[0174] Next, the input support device 201 determines whether or not information for the input item "accommodation information" has been entered (step S1610). Here, the input support device 201 waits for information for the input item "accommodation information" to be entered (step S1610: No). If information for the input item "accommodation information" has been entered (step S1610: Yes), the input support device 201 returns to the step that called normal mode processing.

[0175] This allows the input support device 201 to sequentially input information for each input item in a question-and-answer format.

[0176] Next, using Figure 17, we will explain the specific processing procedure for the GPT mode processing in step S1405 shown in Figure 14.

[0177] Figure 17 is a flowchart showing an example of a specific processing procedure for GPT mode processing. In the flowchart of Figure 17, first, the input support device 201 outputs a chatbot CT message (first message) prompting the user to input details of the business trip in natural language (step S1701).

[0178] Next, the input support device 201 determines whether or not it has received the input text data in response to the output message (step S1702). Here, the input support device 201 waits for the text data to be received (step S1702: No). Then, if the input support device 201 has received the text data (step S1702: Yes), it obtains the analysis result by analyzing the received text data based on the language model M (step S1703).

[0179] Next, the input support device 201 refers to the acquired analysis results and extracts information corresponding to each input item in the drafting screen S1 (step S1704). Specifically, for example, the input support device 201 extracts information from the acquired analysis results that corresponds to input items such as "place of business," "purpose of business," "travel dates (business trip dates)," and "departure point," "departure time," "means of transportation," and "destination" for each itinerary. In this case, if the travel dates are two or more consecutive days based on the information of the input item "travel dates," the input support device 201 refers to the analysis results and extracts information on input items related to accommodation (location, accommodation cost, etc.).

[0180] The input support device 201 then determines, based on the extracted results, whether or not information is missing for any of the multiple input items in the drafting screen S1 (step S1705). Specifically, for example, the input support device 201 determines whether or not corresponding information has been extracted for each input item such as "Place of business," "Business," "Travel date (business trip date)," and for each itinerary, "Departure point," "Departure time," "Means of transportation," and "Destination." The input support device 201 then determines that information is missing for any of the input items if corresponding information has not been extracted for any of the input items. In addition, the input support device 201 determines that information is missing for the input items related to accommodation if the business trip dates are two or more consecutive days and information related to accommodation has not been extracted.

[0181] If information is missing for the first input item (one or more input items) (step S1705: Yes), the input support device 201 outputs a chatbot CT message (second message) prompting the user to correct the travel details (step S1706). The message prompting the user to correct the travel details may, for example, prompt the user to add information to the first input item or to re-enter the travel details.

[0182] The input support device 201 then determines whether additional information has been entered for the first input item, or whether the details of the business trip have been re-entered (step S1707). If the details of the business trip have been re-entered (step S1707: re-entry), the input support device 201 returns to step S1703. On the other hand, if additional information has been entered for the first input item (step S1707: additional entry), the input support device 201 proceeds to step S1710.

[0183] Furthermore, if no information is missing for any of the input items in step S1705 (step S1705: No), the input support device 201 outputs a request to confirm the acquired analysis results as a message from the chatbot CT on the conversation screen S2 (step S1708).

[0184] Then, the input support device 201 confirms the analysis results in response to the user's input and determines whether or not the input of the business trip details has been finalized (step S1709). If the input of the business trip details has not been finalized (step S1709: No), the process returns to step S1702.

[0185] On the other hand, once the input of the business trip details is confirmed (step S1709: Yes), the input support device 201 inputs the extracted information or the additionally entered information into the corresponding input field in the drafting screen S1 (step S1710), and returns to the step that called the GPT mode processing.

[0186] As a result, the input support device 201 can extract information for each input item required when applying for travel expenses (business trip expenses) from the travel details entered in natural language, and automatically input it into the corresponding input items on the drafting screen S1. In addition, if there is any missing information in the travel details entered in natural language, the input support device 201 can prevent input omissions by prompting the user to add that information.

[0187] As described above, the input support device 201 according to the embodiment receives text data entered in response to a first input request prompting the user to input details of a business trip in natural language, and can extract information corresponding to each of the multiple input items in the screen for applying for business trip expenses (for example, the drafting screen S1) from the received text data. Specifically, for example, the input support device 201 extracts information corresponding to each input item by referring to the analysis results obtained by analyzing the received text data based on a language model M. The language model M is, for example, LLM. Then, according to the input support device 201, when inputting the extracted information into the corresponding input item on the screen, it determines whether information is missing for any of the multiple input items, and if it determines that information is missing for any of the input items, it can output a second input request prompting the user to input information for that input item.

[0188] As a result, the input support device 201 can reduce the workload associated with entering information when applying for travel expenses. For example, the input support device 201 can reduce the effort and time required for information entry by allowing users to enter travel details in freely written text (natural language) when applying for travel expenses. In addition, if there is any missing information in the entered travel details, the input support device 201 can prevent input omissions by prompting the user to enter the information for those missing items.

[0189] Furthermore, the input support device 201 allows the extracted information to be entered into the corresponding input fields on the screen.

[0190] As a result, the input support device 201 can automatically input information extracted from the travel details entered by the user in natural language into the corresponding input fields on the screen, reducing the effort and time required for information input compared to when the user manually enters each piece of information one by one.

[0191] Furthermore, according to the input support device 201, if it receives information for any of the input items in response to the outputted second input request, it can input that information into any of the input items on the screen.

[0192] This allows the input support device 201 to input any missing information from the travel details entered by the user in natural language into the corresponding input fields on the screen.

[0193] Furthermore, the input support device 201 determines whether the business trip dates are two or more consecutive days based on the extracted information of the input items related to the business trip dates. If the business trip dates are two or more consecutive days, it can extract information of the input items related to accommodation facilities from the received text data. However, multiple input items include input items related to business trip dates and input items related to accommodation facilities. The input support device 201 can determine that there is insufficient information regarding the input items related to accommodation facilities if the business trip dates are two or more consecutive days and information of the input items related to accommodation facilities is not extracted.

[0194] As a result, the input support device 201 can determine whether information regarding accommodations is insufficient in natural language input only when the business trip dates are two or more consecutive days, and if it is insufficient, it can prompt the user to input information regarding accommodations. For example, if the business trip dates are not two or more consecutive days, the input support device 201 does not determine whether information regarding accommodations is insufficient, thus reducing the number of excessive messages prompting information input.

[0195] Furthermore, the input support device 201 can output the first input request as a message to the chatbot CT, which is engaged in a conversation regarding the application for travel expenses.

[0196] This allows the input support device 201 to input details of business trips in a conversational format with the chatbot CT, making it easier for users unfamiliar with application procedures to input information related to business trip expense claims.

[0197] Furthermore, according to the input support device 201, when applying for travel expenses, the user can select either a first mode (GPT mode) or a second mode (normal mode). According to the input support device 201, if the first mode is selected, the first input request can be output. According to the input support device 201, if the second mode is selected, a third request can be output for each input item included in the multiple input items, prompting the user to input information about that input item. If the information for the input item is received in response to the outputted third input request, that information can be entered into the corresponding input item on the screen.

[0198] This allows the input support device 201 to select between GPT mode, where the details of the business trip are entered as text, and normal mode, where information for each input item is entered in a question-and-answer format, enabling the user to input information for the business trip expense application in a mode that is easy for them to use.

[0199] Furthermore, according to the input support device 201, an LLM (Large-Scale Language Model) can be used as the language model M.

[0200] As a result, the input support device 201 can accurately obtain the desired information in the desired format from the travel details entered in natural language.

[0201] Based on these findings, the input support method and input support program according to the embodiment can assist users in inputting information into the travel expense application system 200, thereby reducing the burden on employees of companies and government offices who apply for travel expenses and improving convenience.

[0202] The input support method described in this embodiment can be implemented by executing a pre-prepared program on a computer such as a personal computer or workstation. This input support program is recorded on a computer-readable recording medium such as a hard disk, flexible disk, CD-ROM, DVD, or USB memory, and is executed when read from the recording medium by the computer. This input support program may also be distributed via a network such as the Internet.

[0203] Furthermore, the information processing device 101 (input support device 201) described in this embodiment can also be implemented using application-specific ICs such as standard cells and structured ASICs (Application Specific Integrated Circuits), or PLDs (Programmable Logic Devices) such as FPGAs.

[0204] With regard to the embodiments described above, the following additional information is disclosed.

[0205] (Note 1) An input support method that assists in entering information related to the application for travel expenses, The system accepts the text data entered in response to the first input request, which prompts the user to enter details of their business trip in natural language. From the received text data, information corresponding to each of the multiple input fields on the screen for the application for the aforementioned travel expenses is extracted. When inputting the extracted information into the corresponding input fields on the screen, it is determined whether or not information is missing in any of the input fields. If it is determined that information is missing for any of the aforementioned input fields, a second input request prompting the user to enter information for that input field is output. An input assistance method characterized by having a computer perform the processing.

[0206] (Note 2) The extracted information is entered into the corresponding input field on the screen. The input support method according to Appendix 1, characterized in that the computer performs the processing.

[0207] (Note 3) When information of any of the above input items is received in response to the outputted second input request, the information is entered into any of the above input items on the screen. The input support method according to Appendix 1 or 2, characterized in that the computer performs the processing.

[0208] (Note 4) The aforementioned multiple input fields include input fields related to the business trip date and input fields related to accommodation, The extraction process described above is: Based on the information in the input fields related to the extracted business trip dates, it is determined whether the business trip dates are two or more consecutive days. If the aforementioned business trip days are two or more consecutive days, information regarding the input items related to the accommodation will be extracted from the aforementioned text data. The process for making the determination is as follows: If the aforementioned business trip dates are two or more consecutive days, and information regarding the accommodation facility is not extracted, it is determined that there is insufficient information regarding the accommodation facility. An input support method as described in any one of the appendices 1 to 3, characterized by the above.

[0209] (Note 5) The input support method according to any one of Notes 1 to 4, characterized in that the first input request is output as a message to a chatbot that engages in a conversation regarding the application for travel expenses.

[0210] (Note 6) When applying for the aforementioned travel expenses, the applicant may select either the first mode or the second mode. When the first mode is selected, the first input request is output. When the second mode is selected, a third request prompting the input of information about each of the multiple input items is output, and when the information for the input item is received in response to the outputted third input request is received, that information is entered into the corresponding input item on the screen. An input support method according to any one of the appendices 1 to 5, characterized in that the computer performs the processing.

[0211] (Note 7) The input support method according to any one of Notes 1 to 6, characterized in that the language model is an LLM (Large Language Model).

[0212] (Note 8) An input support method that assists in entering information related to the application for travel expenses, The system accepts the text data entered in response to the first input request, which prompts the user to enter details of their business trip in natural language. From the received text data, information corresponding to each of the multiple input fields on the screen for the application for the aforementioned travel expenses is extracted. When inputting the extracted information into the corresponding input fields on the screen, it is determined whether or not information is missing in any of the input fields. If it is determined that information is missing for any of the aforementioned input fields, a second input request prompting the user to enter information for that input field is output. An input assistance program characterized by having a computer perform the processing. [Explanation of symbols]

[0213] 101 Information Processing Device 102 users 103, CT Chatbot 104,M language model 105a, 105b, 105c Input items 105 screens 110,140 input requests 120,tx1 Text data 130,1300 analysis results 200 Business Trip Expense Application System 201 Input support device 202 Client Devices 203 Analysis Server 210 Network 300 bus 301 CPU 302 memory 303 Disk Drive 304 disks 305 Communication I / F 306 Portable recording medium interface 307 Portable recording media 400 Control Unit 401 Conversation Control Unit 402 Reception Department 403 Extraction part 404 Judgment section 405 Input section Buttons 501, 502, 503, 504, 505, 602, 603, 804, 805, 1201, 1202 601,801,802,803,b1,b2,b3,b4,b6,b7,b8,b9,b10,b11 Input Box 701,702,703,704,705,706 Extraction result information 901, 902, 903 Checkmark b5 Checkbox f1,f3,f3,f4 display field

Claims

1. An input support method that assists in entering information for business trip expense applications, The system accepts the text data entered in response to the first input request, which prompts the user to enter details of their business trip in natural language. From the received text data, information corresponding to each of the multiple input fields on the screen for the application for the aforementioned travel expenses is extracted. When inputting the extracted information into the corresponding input fields on the screen, it is determined whether or not information is missing in any of the input fields. If it is determined that information for any of the aforementioned input items is missing, a second input request prompting the user to enter information for that input item is output. An input assistance method characterized by having a computer perform the processing.

2. The extracted information is entered into the corresponding input field on the screen. The input support method according to claim 1, characterized in that the computer performs the processing.

3. When information for any of the input items entered in response to the outputted second input request is received, that information is entered into any of the input items on the screen. The input support method according to claim 1, characterized in that the computer performs the processing.

4. The aforementioned multiple input fields include input fields related to the business trip date and input fields related to accommodation, The extraction process described above is: Based on the information in the input fields related to the extracted business trip dates, it is determined whether the business trip dates are two or more consecutive days. If the aforementioned business trip days are two or more consecutive days, information regarding the input items related to the accommodation will be extracted from the aforementioned text data. The process for making the determination is as follows: If the aforementioned business trip dates are two or more consecutive days, and information regarding the accommodation facility is not extracted, it is determined that there is insufficient information regarding the accommodation facility. The input support method according to feature 1.

5. The input support method according to any one of claims 1 to 4, characterized in that the language model is an LLM (Large Language Model).

6. An input support method that assists in entering information for business trip expense applications, The system accepts the text data entered in response to the first input request, which prompts the user to enter details of their business trip in natural language. From the received text data, information corresponding to each of the multiple input fields on the screen for the application for the aforementioned travel expenses is extracted. When inputting the extracted information into the corresponding input fields on the screen, it is determined whether or not information is missing in any of the input fields. If it is determined that information for any of the aforementioned input items is missing, a second input request prompting the user to enter information for that input item is output. An input assistance program characterized by having a computer perform the processing.

Images