Filling method and system based on artificial intelligence

Abstract

The invention provides a filling method and system based on artificial intelligence, and the method comprises the steps: obtaining the basic information of a current user, obtaining the authority of the user according to the basic information, and building the man-machine interaction based on a large language model; application information of a user is collected through human-computer interaction, application form pre-filling information is generated and submitted, and application form filling is completed; and querying the effective application form under the name of the user, collecting reimbursement information related to the effective application form by utilizing man-machine interaction, generating reimbursement pre-filling information, and submitting the reimbursement pre-filling information to finish reimbursement form filling. By utilizing a large language model and AIGC capability, entity information is extracted from dialogues, dialogue intention is judged, and contents of receipts needing to be submitted are identified and generated, so that automatic input of the details of the receipts is realized, the workload of manual input is reduced, the automation level of a reimbursement process is improved, and the reimbursement efficiency is improved. Through deep fusion of artificial intelligence and the reimbursement process, the reimbursement process is intelligent and automatic, non-perceptual reimbursement experience is realized, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent form filling technology, and in particular to an artificial intelligence-based form filling method and system. Background Art

[0002] In the wave of digital transformation, enterprises are increasingly in need of improving internal management efficiency and optimizing resource allocation. As key links in the daily operations of enterprises, the optimization and intelligent upgrading of expense control and reimbursement management processes have become the focus of enterprises. Traditional reimbursement systems are mostly designed based on fixed business processes, with complex page layouts and scattered functions, which means that users often need to shuttle between multiple pages when filling out reimbursement documents, which not only increases the difficulty of operation but also reduces work efficiency. For users who are not familiar with the system, finding the correct information filling entry is often a difficult task. Due to the lack of intuitiveness of the page layout, users may need to browse and try repeatedly to find the content they need to fill in, which not only wastes time but also reduces the user experience.

[0003] With the rapid development of artificial intelligence (AI) technology, especially the increasing maturity of deep learning and natural language processing technologies, it has become possible to transform the field of expense control and reimbursement into an intelligent one. As an important application form of AI technology, the private large model platform integrates internal data resources of the enterprise to build customized and efficient AI models. Therefore, exploring the application of expense control and reimbursement scenarios based on artificial intelligence has become an important research direction. Summary of the invention

[0004] In order to solve the technical problems that traditional reimbursement systems have numerous and complex pages, it is difficult to understand the operation process and it is easy to cause omissions and errors in filling in forms, the present invention proposes an artificial intelligence-based form filling method and system, which implements form filling operations based on artificial intelligence, reduces omissions and improves form filling efficiency.

[0005] The present invention provides the following specific solutions:

[0006] An artificial intelligence-based form filling system, comprising: a basic module, an application form filling module, an application form submission module, a reimbursement form filling module and a reimbursement form submission module;

[0007] The basic module includes an information permission submodule for obtaining basic information of the current user and obtaining user permissions based on the basic information; the basic information includes employee role, company, accounting subject, department, cost center, position, and rank; the user permissions include form filling permissions and data permissions;

[0008] The application form filling module includes: an application information collection submodule for inquiring the user's application information through AI dialogue; an application AI front-end submodule for receiving the user's application information, an application AI model docking platform for interacting with the application AI front-end submodule and the application large model submodule, an application large model submodule for generating AI dialogue inquiries for the user based on the large language model, extracting entity information based on the application information answered by the user, determining the dialogue intention, identifying and generating a pre-filled application form; the application information includes: the user's business trip start date, business trip end date, departure city, arrival city, and business trip reason;

[0009] The application form submission module is used to call the reimbursement system interface and submit the pre-filled application form generated by the application large model submodule to the reimbursement system;

[0010] The reimbursement form filling module includes: a reimbursement information collection submodule for inquiring the user's reimbursement information through AI dialogue; a reimbursement AI front-end submodule for receiving the user's reimbursement information, a reimbursement AI model docking platform for interacting with the reimbursement AI front-end submodule and the reimbursement large model submodule, a reimbursement large model submodule for generating AI dialogue inquiries for users based on a large language model, extracting reimbursement entity information based on the reimbursement information answered by the user, determining the dialogue intention, and identifying and generating a pre-filled reimbursement form; the reimbursement entity information extraction content includes: associating application forms, confirming summary information, collecting invoice information, confirming expense reimbursement information, and subsidy information;

[0011] The reimbursement form submission module is used to call the reimbursement system interface and submit the pre-filled reimbursement form generated by the reimbursement large model submodule to the reimbursement system.

[0012] Preferably, the reimbursement form filling module also includes: based on the basic information and permissions, a valid application information query sub-module is used to query the valid application form under the user name and the reimbursement information associated with the application form, the valid application form is an application form for completing actual expense expenditures, and the reimbursement information includes unified settlement information, invoice information and subsidy information associated with the valid reimbursement application form.

[0013] Preferably, the step of extracting entity information in the application form filling module is: extracting entity information based on the user application information of the application information collection submodule, and automatically entering travel expense details in the application form.

[0014] Preferably, the system also includes a policy standard query module, which conducts a group reimbursement policy AI dialogue with the user through the user interface, and sequentially calls the reimbursement AI front-end submodule, the reimbursement AI model docking platform, the reimbursement large model submodule and the group reimbursement policy standard database, generates policy question and answer answers in the large model submodule based on the group's reimbursement policy standard database, and then displays the reimbursement policy consulted by the user and the information to be filled in for reimbursement on the user interface through the reimbursement AI model docking platform and the reimbursement AI front-end submodule.

[0015] Preferably, the reimbursement information collection submodule includes: a unified settlement unit, an invoice collection unit, an expense information collection unit and a subsidy collection unit;

[0016] The unified settlement unit queries the consolidated air ticket order and hotel order booked according to the valid application form, and collects the consolidated reimbursement information of the air ticket order and hotel order;

[0017] The invoice collection unit selects the invoice related to the application form, uses OCR recognition technology to identify and collect the selected invoice information, and obtains invoice reimbursement information; the invoice reimbursement information includes invoice type, cabin seat, date information, itinerary description, and amount;

[0018] The expense information collection unit supplements the actual expense information that cannot be fully collected from the invoice collection unit; the actual expense information includes expense items, transportation expenses and accommodation expenses, the transportation expenses include transportation tools, departure date, and itinerary description, and the accommodation expenses include check-in location, start date, end date, and number of days of stay;

[0019] The subsidy collection unit collects subsidy reimbursement information of the user's transportation subsidy and meal subsidy, and the subsidy reimbursement information includes location, subsidy start date, subsidy end date, and subsidy collection days.

[0020] Preferably, the invoice information unit also includes: collecting the selected invoice information and bringing out the reimbursement amount by default; when the invoice type is a paper air ticket itinerary or an electronic invoice-air transport electronic ticket itinerary, the reimbursement amount includes three detailed amounts of ticket price, fuel, and civil aviation development fund.

[0021] Preferably, in the invoice information unit, the default values ​​of the expense items and means of transport are set according to the invoice type as follows:

[0022] When the invoice type is a paper air ticket itinerary or an electronic invoice - an air transport electronic ticket itinerary, the expense item defaults to domestic travel expenses - airplane, and the means of transportation defaults to airplane;

[0023] When the invoice type is train ticket or electronic invoice - railway electronic ticket, the expense item defaults to domestic travel expenses - train, and the means of transportation defaults to high-speed rail / motor vehicle;

[0024] When the invoice type is a taxi ticket, the expense item defaults to in-transit transportation fee, and the transportation tool defaults to taxi / online car-hailing service;

[0025] When the invoice type is not among the above types, the expense reimbursement collection unit collects and confirms the corresponding expense items from the user.

[0026] Preferably, the means of transportation is set by default according to the expense item, further comprising: when the expense item is domestic travel transportation fee-ship ticket, the means of transportation is defaulted to ship; when the expense item is domestic travel transportation fee-long-distance bus, the means of transportation is defaulted to "long-distance bus".

[0027] A form filling method based on artificial intelligence, comprising:

[0028] S1. Obtain basic information: Obtain the basic information of the current user, obtain the user's permissions based on the basic information, and establish human-computer interaction based on a large language model; the basic information includes employee role, company, accounting entity, department, cost center, position, and rank; the permissions include form filling permissions and data permissions;

[0029] S2. Application form filling: Based on the basic information and permissions, using the human-computer interaction, collect the user's application information, generate pre-filled information for the application form based on the application information, the user checks and submits the pre-filled information for the application form, and completes the application form filling;

[0030] S3. Filling in reimbursement forms: Based on the basic information and permissions, query the valid application forms under the name of the user, and based on the valid application forms, use the human-computer interaction to collect the user's reimbursement information related to the valid application forms, generate reimbursement pre-filled information based on the reimbursement information, and the user checks and submits the pre-filled reimbursement information to complete the filling in of the reimbursement form; the valid application form is an application form that completes actual expense expenditures, and the reimbursement information includes the unified settlement information, invoice information and subsidy information associated with the valid application form.

[0031] Preferably, the method further comprises:

[0032] S4. Policy and Standards Query: Based on the group’s policy and standards, the user’s questions about relevant policy and standards are answered through human-computer interaction with a large language model.

[0033] Preferably, the specific steps of S2 are:

[0034] S21. Application form information collection: Through the interaction of the large language model, the user's application information is inquired, and the application information includes the business trip start date, business trip end date, departure city, arrival city, and business trip reason;

[0035] S22. Generate pre-filled information for the application form: Based on the collected application information, inquire the user whether it is correct, and after confirming the information, generate pre-filled information for the application form;

[0036] S23. Submit the application form: display the pre-filled information of the application form, submit the pre-filled information of the application form to the reimbursement system, receive the returned result, and complete the application form filling.

[0037] Preferably, the specific steps of S3 are:

[0038] S31. Associate the application form and collect reimbursement information: query the user's valid application form, select the application form that needs to be reimbursed through the human-computer interaction, and collect the unified settlement information, invoice information and subsidy information associated with the application form that needs to be reimbursed;

[0039] S32. Generate pre-filled information for reimbursement form: Send the collected reimbursement information to the user for confirmation, and generate pre-filled information for reimbursement form after confirmation;

[0040] S33. Submit the reimbursement form: display the pre-filled information of the reimbursement form, submit the pre-filled information of the reimbursement form to the reimbursement system, receive the returned result, and complete the filling of the reimbursement form.

[0041] Beneficial effects:

[0042] The present invention provides a form filling method and system based on artificial intelligence, which obtains basic information: obtains the basic information of the current user, obtains the user's authority according to the basic information, and establishes a human-computer interaction based on a large language model; uses human-computer interaction to collect the user's application information, and generates pre-filled information for the application form, and the user checks and submits the pre-filled information for the application form to complete the application form filling; queries the valid application form under the user's name, collects the reimbursement information related to the valid application form based on the valid application form by using the human-computer interaction, generates pre-filled reimbursement information according to the reimbursement information, and the user checks and submits it to complete the reimbursement form filling. Using the large language model and AIGC capabilities, entity information is extracted from the dialogue, the dialogue intention is determined, and the content of the required submitted documents is identified and generated, so as to realize automatic entry of document details, reduce the workload of manual input, and thus improve the automation level of the reimbursement process. User experience optimization: by deeply integrating intelligent technology with the reimbursement process, an automated and imperceptible reimbursement experience is realized, and the reimbursement efficiency and employee satisfaction are improved. By integrating multi-party information and reducing manual selection through intelligent calculation, automatic filling is realized, making the reimbursement process more intelligent and automated, and improving work efficiency.

[0043] In particular, during the reimbursement information collection process, OCR (optical character recognition) technology is used to automatically identify invoices and image attachments, extract structured data, and ensure that key information in the reimbursement documents can be accurately extracted.

[0044] In particular, by leveraging large language models and AIGC capabilities, it provides instant and intelligent financial reimbursement system inquiry functions that can answer employees' questions about reimbursement policies. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A framework diagram of an artificial intelligence-based form filling system provided by the present invention.

[0046] Figure 2 A flowchart of an artificial intelligence-based form filling method provided by the present invention. DETAILED DESCRIPTION

[0047] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0048] Figure 1 This is a framework diagram of an artificial intelligence-based form filling system provided by the present invention. Figure 1 As shown, an artificial intelligence-based form filling system includes: a basic module, an application form filling module, an application form submission module, a reimbursement form filling module and a reimbursement form submission module;

[0049] The basic module includes an information permission submodule for obtaining basic information of the current user and obtaining user permissions based on the basic information; the basic information includes employee role, company, accounting subject, department, cost center, position, and rank; the user permissions include form filling permissions and data permissions;

[0050] The application form filling module includes: an application information collection submodule for inquiring the user's application information through AI dialogue; an application AI front-end submodule for receiving the user's application information, an application AI model docking platform for interacting with the application AI front-end submodule and the application large model submodule, an application large model submodule for generating AI dialogue inquiries for the user based on the large language model, extracting entity information based on the application information answered by the user, determining the dialogue intention, identifying and generating a pre-filled application form; the application information includes: the user's business trip start date, business trip end date, transportation, departure city, arrival city, and business trip reason;

[0051] The application form submission module is used to call the reimbursement system interface and submit the pre-filled application form generated by the application large model submodule to the reimbursement system;

[0052] The reimbursement form filling module includes: a reimbursement information collection submodule for inquiring the user's reimbursement information through AI dialogue; a reimbursement AI front-end submodule for receiving the user's reimbursement information, a reimbursement AI model docking platform for interacting with the reimbursement AI front-end submodule and the reimbursement large model submodule, a reimbursement large model submodule for generating AI dialogue inquiries to the user based on the large language model, extracting reimbursement entity information based on the reimbursement information answered by the user, determining the dialogue intention, and identifying and generating a pre-filled reimbursement form; the reimbursement entity information extraction content includes: associating application forms, confirming summary information, collecting invoice information, confirming expense reimbursement information, and subsidy information;

[0053] The reimbursement form submission module is used to call the reimbursement system interface and submit the pre-filled reimbursement form generated by the reimbursement large model submodule to the reimbursement system.

[0054] The system provided in this application collects application information and reimbursement information by means of dialogue with user AI, automatically generates application form pre-filled information and reimbursement pre-filled information based on the collected application information and reimbursement information, and submits them to the reimbursement system to complete the filling of application form and reimbursement information. By using large language models and AIGC capabilities, entity information is extracted from the dialogue, the dialogue intent is determined, and the content of the required documents is identified and generated, so as to realize automatic entry of document details and reduce the workload of manual input, thereby improving the automation level of the reimbursement process.

[0055] In some embodiments, the application AI front-end submodule and the reimbursement AI front-end submodule are integrated into the AI ​​front-end submodule, the application AI model docking platform and the reimbursement AI model docking platform are integrated into the AI ​​model docking platform, and the application large model submodule and the reimbursement large model submodule are integrated into the large model submodule. In the actual data processing process, the application form filling module or the reimbursement form filling module calls the AI ​​front-end submodule, the AI ​​model docking platform or the large model submodule to implement the above functions.

[0056] Specifically, the reimbursement form filling module also includes: based on the basic information and permissions, a valid application information query sub-module is used to query the valid application forms under the user name and the reimbursement information associated with the application form. The valid application form is an application form that has completed the approval and has not been closed by other reimbursement references. The reimbursement information includes the unified settlement information, invoice information and subsidy information associated with the valid reimbursement application form.

[0057] Specifically, the step of extracting entity information in the application form filling module is: based on the user application information of the application information collection submodule, extracting entity information, and automatically entering travel expense details in the application form.

[0058] Specifically, the system also includes a policy standard query module, which conducts a group reimbursement policy AI dialogue with the user through the user interface, and sequentially calls the reimbursement AI front-end submodule, the reimbursement AI model docking platform, the reimbursement large model submodule and the group reimbursement policy standard database, generates policy question and answer answers in the large model submodule based on the group's reimbursement policy standard database, and then displays the reimbursement policy consulted by the user and the information to be filled in for reimbursement on the user interface through the reimbursement AI model docking platform and the reimbursement AI front-end submodule.

[0059] Specifically, the reimbursement information collection submodule includes: a unified settlement unit, an invoice collection unit, a cost information collection unit and a subsidy collection unit. The unified settlement unit queries the consolidated air ticket order and hotel order booked according to the valid application form, and collects the consolidated reimbursement information of the air ticket order and hotel order; the invoice collection unit selects the invoice related to the application form, uses OCR recognition technology to identify and collect the selected invoice information, and obtains the invoice reimbursement information, wherein the invoice reimbursement information includes the invoice type, cabin seat, date information, itinerary description, and amount; the cost information collection unit supplements the actual cost information that cannot be fully collected from the invoice collection unit; the actual cost information includes cost items, transportation fees and accommodation fees, the transportation fees include transportation tools, departure date, itinerary description, and the accommodation fees include check-in location, start date, end date, and number of days of stay; the subsidy collection unit collects the subsidy reimbursement information of the user's transportation subsidy and food subsidy, and the subsidy reimbursement information includes location, subsidy start date, subsidy end date, and subsidy collection days.

[0060] Specifically, the invoice information unit also includes: collecting the selected invoice information and bringing out the reimbursement amount by default; when the invoice type is a paper air ticket itinerary or an electronic invoice - an air transport electronic ticket itinerary, the reimbursement amount includes three detailed amounts of ticket price, fuel, and civil aviation development fund.

[0061] Specifically, in the invoice information unit, the default values ​​of the expense items and means of transportation are set according to the invoice type: when the invoice type is a paper air ticket itinerary or an electronic invoice-air transport electronic ticket itinerary, the expense item defaults to domestic travel transportation expenses-airplane, and the means of transportation defaults to airplane; when the invoice type is a train ticket or an electronic invoice-railway electronic ticket, the expense item defaults to domestic travel transportation expenses-train, and the means of transportation defaults to high-speed rail / motor vehicle; when the invoice type is a taxi ticket, the expense item defaults to in-transit transportation expenses, and the means of transportation defaults to taxi / online car-hailing; when the invoice type is not among the above types, the corresponding expense items are collected and confirmed from the user through the reimbursement expense collection unit.

[0062] Specifically, the means of transportation is set by default according to the expense item, and also includes: when the expense item is domestic travel transportation fee-ship ticket, the means of transportation is defaulted to ship; when the expense item is domestic travel transportation fee-long-distance bus, the means of transportation is defaulted to "long-distance bus".

[0063] Figure 2 1 is a flow chart of a form filling method based on artificial intelligence provided by the present invention. Figure 2 As shown, a form filling method based on artificial intelligence includes:

[0064] S1. Obtain basic information: obtain the basic information of the current user, obtain the user's permissions based on the basic information, and establish human-computer interaction based on a large language model; the basic information includes employee role, company, accounting entity, department, cost center, position, and rank; the permissions include form filling permissions and data permissions.

[0065] S2. Application form filling: Based on the basic information and permissions, using the human-computer interaction, collect the user's application information, generate pre-filled information for the application form based on the application information, and the user checks and submits the pre-filled information for the application form to complete the application form filling.

[0066] In some embodiments, the above S2 specifically comprises the following steps:

[0067] S21. Application form information collection: Through the interaction of the large language model, the user's application information is inquired, and the application information includes the business trip start date, business trip end date, departure city, arrival city, and business trip reason;

[0068] S22. Generate pre-filled information for the application form: Based on the collected application information, inquire the user whether it is correct, and after confirming the information, generate pre-filled information for the application form;

[0069] S23. Submit the application form: display the pre-filled information of the application form, submit the pre-filled information of the application form to the reimbursement system, receive the returned result, and complete the application form filling.

[0070] S3. Filling in reimbursement forms: Based on the basic information and permissions, query the valid application forms under the name of the user, and based on the valid application forms, use the human-computer interaction to collect the user's reimbursement information related to the valid application forms, generate reimbursement pre-filled information based on the reimbursement information, and the user checks and submits the pre-filled reimbursement information to complete the filling in of the reimbursement form; the valid application form is an application form that completes actual expense expenditures, and the reimbursement information includes the unified settlement information, invoice information and subsidy information associated with the valid application form.

[0071] In some embodiments, the above S3 specifically comprises the following steps:

[0072] S31. Associate the application form and collect reimbursement information: query the user's valid application form, select the application form that needs to be reimbursed through the human-computer interaction, and collect the unified settlement information, invoice information and subsidy information associated with the application form that needs to be reimbursed;

[0073] S32. Generate pre-filled information for reimbursement form: Send the collected reimbursement information to the user for confirmation, and generate pre-filled information for reimbursement form after confirmation;

[0074] S33. Submit the reimbursement form: display the pre-filled information of the reimbursement form, submit the pre-filled information of the reimbursement form to the reimbursement system, receive the returned result, and complete the filling of the reimbursement form.

[0075] In some embodiments, the method further comprises:

[0076] S4. Policy and Standards Query: Based on the group’s policy and standards, the user’s questions about relevant policy and standards are answered through human-computer interaction with a large language model.

[0077] It should be noted that the above-described specific implementations can enable those skilled in the art to more fully understand the invention, but do not limit the invention in any way. Therefore, although this specification has described the invention in detail with reference to the drawings and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents. In short, all technical solutions and improvements that do not deviate from the spirit and scope of the invention should be included in the protection scope of the patent for the invention.

Claims

1. A form filling system based on artificial intelligence, characterized in that: include: Basic module, application form filling module, application form submission module, reimbursement form filling module and reimbursement form submission module; The basic module includes an information permission submodule for obtaining basic information of the current user and obtaining user permissions based on the basic information; the basic information includes employee role, company, accounting subject, department, cost center, position, and rank; the user permissions include form filling permissions and data permissions; The application form filling module includes: an application information collection submodule for inquiring the user's application information through AI dialogue; an application AI front-end submodule for receiving the user's application information, an application AI model docking platform for interacting with the application AI front-end submodule and the application large model submodule, an application large model submodule for generating AI dialogue inquiries for the user based on the large language model, extracting entity information based on the application information answered by the user, determining the dialogue intention, identifying and generating a pre-filled application form; the application information includes: the user's business trip start date, business trip end date, transportation, departure city, arrival city, and business trip reason; The application form submission module is used to call the reimbursement system interface and submit the pre-filled application form generated by the application large model submodule to the reimbursement system; The reimbursement form filling module includes: a reimbursement information collection submodule for inquiring the user's reimbursement information through AI dialogue; a reimbursement AI front-end submodule for receiving the user's reimbursement information, a reimbursement AI model docking platform for interacting with the reimbursement AI front-end submodule and the reimbursement large model submodule, a reimbursement large model submodule for generating AI dialogue inquiries to the user based on the large language model, extracting reimbursement entity information based on the reimbursement information answered by the user, determining the dialogue intention, and identifying and generating a pre-filled reimbursement form; the reimbursement entity information extraction content includes: associating application forms, confirming summary information, collecting invoice information, confirming expense reimbursement information, and subsidy information; The reimbursement form submission module is used to call the reimbursement system interface and submit the pre-filled reimbursement form generated by the reimbursement large model submodule to the reimbursement system.

2. The form filling system based on artificial intelligence according to claim 1, characterized in that: The reimbursement form filling module also includes: based on the basic information and permissions, a valid application information query sub-module is used to query the valid application forms under the user name and the reimbursement information associated with the application form. The valid application form is an application form that has completed the approval and has not been closed by other reimbursement references. The reimbursement information includes the unified settlement information, invoice information and subsidy information associated with the valid reimbursement application form.

3. The form filling system based on artificial intelligence according to claim 1, characterized in that: The step of extracting entity information in the application form filling module is: extracting entity information based on the user application information in the application information collection submodule, and automatically entering travel expense details in the application form.

4. The form filling system based on artificial intelligence according to claim 1, characterized in that: The system also includes a policy standard query module, which conducts a group reimbursement policy AI dialogue with the user through a user interface, and sequentially calls the reimbursement AI front-end submodule, the reimbursement AI model docking platform, the reimbursement large model submodule and the group reimbursement policy standard database, generates policy question and answer answers in the large model submodule based on the group's reimbursement policy standard database, and then displays the reimbursement policy consulted by the user and the information to be filled in for reimbursement on the user interface through the reimbursement AI model docking platform and the reimbursement AI front-end submodule.

5. The form filling system based on artificial intelligence according to claim 1, characterized in that: The reimbursement information collection submodule includes: a unified settlement unit, an invoice collection unit, an expense information collection unit and a subsidy collection unit; The unified settlement unit queries the consolidated air ticket order and hotel order booked according to the valid application form, and collects the consolidated reimbursement information of the air ticket order and hotel order; The invoice collection unit selects the invoice related to the application form, uses OCR recognition technology to identify and collect the selected invoice information, and obtains invoice reimbursement information; the invoice reimbursement information includes invoice type, cabin seat, date information, itinerary description, and amount; The expense information collection unit supplements the actual expense information that cannot be fully collected from the invoice collection unit; the actual expense information includes expense items, transportation expenses and accommodation expenses, the transportation expenses include transportation tools, departure date, and itinerary description, and the accommodation expenses include check-in location, start date, end date, and number of days of stay; The subsidy collection unit collects subsidy reimbursement information of the user's transportation subsidy and meal subsidy, and the subsidy reimbursement information includes location, subsidy start date, subsidy end date, and subsidy collection days.

6. The form filling system based on artificial intelligence as claimed in claim 5, characterized in that: The invoice information unit further includes: The reimbursement amount is included by default when collecting the selected invoice information; when the invoice type is a paper air ticket itinerary or an electronic invoice - an air transport electronic ticket itinerary, the reimbursement amount includes the three detailed amounts of fare, fuel, and civil aviation development fund.

7. The form filling system based on artificial intelligence as claimed in claim 5, characterized in that: In the invoice information unit, the default values ​​of the expense items and transportation means are set according to the invoice type: When the invoice type is a paper air ticket itinerary or an electronic invoice - an air transport electronic ticket itinerary, the expense item defaults to domestic travel expenses - airplane, and the means of transportation defaults to airplane; When the invoice type is train ticket or electronic invoice - railway electronic ticket, the expense item defaults to domestic travel expenses - train, and the means of transportation defaults to high-speed rail / motor vehicle; When the invoice type is a taxi ticket, the expense item defaults to in-transit transportation fee, and the transportation tool defaults to taxi / online car-hailing service; When the invoice type is not among the above types, the expense reimbursement collection unit collects and confirms the corresponding expense items from the user.

8. The form filling system based on artificial intelligence as claimed in claim 7, characterized in that: The means of transportation is set by default according to the expense item, and also includes: when the expense item is domestic travel transportation fee-ship ticket, the means of transportation is set by default as ship; when the expense item is domestic travel transportation fee-long-distance bus, the means of transportation is set by default as "long-distance bus".

9. An artificial intelligence-based form filling method applied to the system of any one of claims 1 to 8, characterized in that: include: S1. Obtain basic information: Obtain the basic information of the current user, obtain the user's permissions based on the basic information, and establish human-computer interaction based on a large language model; the basic information includes employee role, company, accounting entity, department, cost center, position, and rank; the permissions include form filling permissions and data permissions; S2. Application form filling: Based on the basic information and permissions, using the human-computer interaction, collect the user's application information, generate pre-filled information for the application form based on the application information, the user checks and submits the pre-filled information for the application form, and completes the application form filling; S3. Filling in the reimbursement form: Based on the basic information and permissions, query the valid application form under the name of the user, collect the reimbursement information of the user related to the valid application form by using the human-computer interaction based on the valid application form, generate reimbursement pre-filled information according to the reimbursement information, and the user checks and submits the reimbursement pre-filled information to complete the filling in the reimbursement form; the valid application form is an application form that completes the actual expense expenditure, and the reimbursement information includes the unified settlement information, invoice information and subsidy information associated with the valid application form; S4. Policy and Standards Query: Based on the group’s policy and standards, the user’s questions about relevant policy and standards are answered through human-computer interaction with a large language model.

10. The form filling method based on artificial intelligence according to claim 9, characterized in that: The specific steps of S2 are: S21. Application form information collection: Through the interaction of the large language model, the user's application information is inquired, and the application information includes the business trip start date, business trip end date, departure city, arrival city, and business trip reason; S22. Generate pre-filled information for the application form: Based on the collected application information, inquire the user whether it is correct, and after confirming the information, generate pre-filled information for the application form; S23. Submit the application form: display the pre-filled information of the application form, submit the pre-filled information of the application form to the reimbursement system, receive the returned result, and complete the application form filling.

11. The form filling method based on artificial intelligence according to claim 9, characterized in that: The specific steps of S3 are: S31. Associate the application form and collect reimbursement information: query the user's valid application form, select the application form that needs to be reimbursed through the human-computer interaction, and collect the unified settlement information, invoice information and subsidy information associated with the application form that needs to be reimbursed; S32. Generate pre-filled information for reimbursement form: send the collected reimbursement information to the user for confirmation, and generate pre-filled information for reimbursement form after confirmation; S33. Submit the reimbursement form: display the pre-filled information of the reimbursement form, submit the pre-filled information of the reimbursement form to the reimbursement system, receive the returned result, and complete the filling of the reimbursement form.

Images