A processing method and system for OPC financial reconciliation

Abstract

This application discloses an OPC financial reconciliation processing method and system, relating to the field of automatic review technology for expense reimbursement applications. It solves the problems of low efficiency and error-prone manual review of expense reimbursement applications in existing technologies. The method includes: monitoring expense reimbursement application events; identifying invoices in the attachments of an expense reimbursement application when it is submitted; conducting multi-dimensional review of the invoices; generating an approval conclusion; and marking it in the office workflow approval module. This achieves fully automated approval processing of expense reimbursement applications, eliminating the need for manual operation by financial personnel, significantly improving the efficiency of financial expense reconciliation, completely resolving the backlog of reviews, and improving the overall operational efficiency of corporate finance. Simultaneously, it eliminates financial risks caused by human error and achieves accurate review of invoice information, amount, authenticity, and duplicate reimbursements.

Description

Technical Field

[0001] This application relates to the field of automatic review technology for expense reimbursement applications, and in particular to a method and system for processing OPC financial reconciliation. Background Technology

[0002] An OPC (One Person Company) is a company structure established and solely funded by a single individual shareholder who assumes limited liability. One-person companies are characterized by low barriers to entry, flexible operation, and efficient decision-making, making them a common legal entity form adopted by startups, freelancers, and micro-enterprise owners.

[0003] Traditionally, corporate expense reimbursement and reconciliation processes require employees to submit reimbursement applications and invoice attachments offline or through office software. Finance staff then manually review and reconcile these documents. The specific process involves finance staff manually verifying the accuracy of basic information such as the company name and tax identification number on the invoices, manually comparing the amount entered in the reimbursement application with the invoice amount, manually logging into the national tax system to retrieve invoice information to confirm the authenticity of the invoices, and manually checking for duplicate reimbursements. After review, finance staff manually mark the approval or rejection results in the approval process and manually compile the reimbursement data to generate a reconciliation report and a summary report of employee payables.

[0004] However, existing corporate financial reconciliation methods have the following technical shortcomings:

[0005] 1. Manual review of expense reimbursement applications is extremely inefficient. Finance staff need to spend a lot of time completing tasks such as verifying invoice information, searching the tax system, and checking for duplicate reimbursements. When faced with a large number of expense reimbursement applications from enterprises, backlog of applications can easily occur, affecting the overall progress of finance work.

[0006] 2. Manual operations are prone to errors, such as incorrect amount comparisons, oversights in verifying tax identification numbers, and failure to detect duplicate reimbursements, leading to inaccurate financial reconciliation results and posing financial risks to the company. Manually compiling reimbursement data and generating reports is time-consuming and prone to data discrepancies, making it difficult to guarantee the timeliness and accuracy of report generation. Furthermore, the company's financial personnel costs are high, as a large amount of basic auditing and reconciliation work consumes the financial staff's energy, preventing them from engaging in higher-value financial analysis and other tasks. This shortage of personnel is particularly pronounced under the OPC model. Summary of the Invention

[0007] The purpose of this application is to overcome the problems of low efficiency and error-prone manual review of reimbursement applications in the existing technology, and to provide an OPC financial reconciliation processing method and system.

[0008] Firstly, a method for processing OPC financial reconciliation is provided, including:

[0009] Monitor expense reimbursement application events in the office workflow approval module;

[0010] In response to the submission of a reimbursement application, which includes reimbursement information and invoice attachments, the invoice attachments are input into a preset multimodal model for invoice recognition in order to extract the invoice information;

[0011] The invoice information is subject to multi-dimensional review, which includes basic information review, amount consistency review, invoice authenticity review, and duplicate reimbursement review.

[0012] An approval conclusion is generated based on the results of the multi-dimensional review, and the reimbursement application is marked in the office workflow approval module based on the approval conclusion.

[0013] In some possible implementations, the multi-dimensional audit specifically includes:

[0014] Basic information verification: Compare the extracted invoice header and tax number with the standard header and tax number registered with the enterprise to see if they are consistent;

[0015] Amount consistency verification: Compare whether the reimbursement amount entered by the employee in the office workflow approval module is consistent with the amount recognized on the invoice;

[0016] Invoice authenticity verification: Retrieve the invoice code and invoice number from the invoice attachments in the company's financial database, and confirm whether the invoice is genuine and valid, as well as compare the accuracy of the invoice information;

[0017] Duplicate expense reimbursement review: Retrieve expense invoice data from the company's financial database and verify whether the invoice has been submitted and reimbursed by other employees;

[0018] The enterprise financial database contains basic invoice information, invoice reimbursement information, and pending transfer tasks.

[0019] Some possible implementations also include updating the company's financial database, specifically:

[0020] Retrieve tax bureau invoice files from the specified directory;

[0021] The obtained tax bureau invoice documents are parsed;

[0022] The parsed invoice records are stored in the company's financial database.

[0023] In some possible implementations, the marker includes:

[0024] If all aspects of the approval process for a reimbursement application are approved, the application will be marked as an annotation in the office workflow approval module.

[0025] If any item in the approval conclusion of the reimbursement application fails, the reimbursement application will be marked as rejected in the office workflow approval module, and the reason for the failure will be fed back to the reimbursement application in the office workflow approval module.

[0026] Among some possible implementations is the generation of financial reports, specifically:

[0027] After reviewing a single expense reimbursement application, the approval result data is updated in real time to the company's financial files, and a financial report is generated, which includes:

[0028] The reconciliation report includes the review results of the reimbursement application, invoice information, and details of any issues.

[0029] The employee payable amount summary report includes all approved reimbursement applications by employee to form the total payable amount data.

[0030] Some possible implementations also include data storage, specifically:

[0031] Store the review results and financial reports of expense reimbursement applications in the company's financial file storage;

[0032] Store invoice information and a list of pending transfer tasks in the company's financial database.

[0033] Secondly, an OPC financial reconciliation processing system is provided, including:

[0034] An event-listening agent is used to listen for expense reimbursement application events in the office workflow approval module.

[0035] The reconciliation agent responds to the submission of a reimbursement application, which includes reimbursement information and invoice attachments. The invoice attachments are input into a preset multimodal model for invoice recognition to extract invoice information. The invoice information is then subjected to multi-dimensional review, including basic information review, amount consistency review, invoice authenticity review, and duplicate reimbursement review. An approval conclusion is generated based on the results of the multi-dimensional review, and the reimbursement application is marked in the office workflow approval module based on the approval conclusion.

[0036] Among the possible implementations are:

[0037] The invoice record synchronization intelligent agent is used to obtain tax bureau invoice files in a specified directory, parse the obtained tax bureau invoice files, and store the parsed invoice records in the enterprise's financial database.

[0038] The approval result summary agent is used to obtain the approval results of all reimbursement applications for the day, and output a summary report after statistical analysis and summarization through a preset large language model. The preset large language model performs mathematical calculations in the statistical analysis and summarization process by calling external mathematical calculation tools.

[0039] The pending reconciliation task agent is used to record pending transfer tasks to the enterprise's financial database at the payment node, query transfer tasks according to specified conditions, generate payment orders for the finance department to make payments, and mark pending transfer tasks as completed after payment is completed.

[0040] The Finance Department employee AI agent is used to invoke other AI agents through dialogue.

[0041] Thirdly, a computer-readable storage medium is provided that stores program code for execution by a device, the program code including steps for performing a method as described in any of the implementations of the first aspect above.

[0042] Fourthly, an electronic device is provided, the electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the method as described in any of the implementations of the first aspect above.

[0043] This application has the following beneficial effects:

[0044] 1. This application can monitor and automatically review expense reimbursement applications in the office workflow approval module. The multi-dimensional review of expense reimbursement applications does not require manual operation by financial personnel, which greatly improves the efficiency of financial expense reconciliation, completely solves the problem of backlog of reviews, and improves the overall operational efficiency of corporate finance. At the same time, it can eliminate the financial risks caused by human operation errors. Through model recognition and automatic comparison, it can achieve accurate review of invoice information, amount, authenticity, and duplicate reimbursements. The accuracy of the review results is far higher than that of manual operation.

[0045] 2. This application enables the automated and real-time generation of financial reports. After the review is completed, the data is immediately compiled and a reconciliation report and an accounts payable summary report are generated, ensuring the timeliness of the reports and the accuracy of the data, without the need for manual statistics by financial personnel.

[0046] 3. This application fully automates the monitoring and review of expense reimbursement applications through an intelligent agent, eliminating the need for manual operation. This reduces the company's financial human resource costs, frees up the work energy of financial personnel, and allows them to devote themselves to higher-value work such as financial analysis and cost control, thereby enhancing the work value of the company's finance department. Attached Figure Description

[0047] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an undue limitation of this application.

[0048] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart of the OPC financial reconciliation processing method according to Embodiment 1 of this application;

[0050] Figure 2 This is a simplified flowchart of the OPC financial reconciliation processing method according to Embodiment 1 of this application;

[0051] Figure 3 This is a structural block diagram of the OPC financial reconciliation processing system according to Embodiment 3 of this application;

[0052] Figure 4 This is an example diagram of the summary report generated by the intelligent agent summarizing the approval results in Embodiment 3 of this application;

[0053] Figure 5 This is a schematic diagram of the internal structure of the electronic device according to Embodiment 5 of this application.

[0054] Figure label:

[0055] 100. Event Listening Agent; 200. Reconciliation Agent; 300. Invoice Record Synchronization Agent; 400. Approval Result Summary Agent; 500. Pending Reconciliation Task Agent; 600. Finance Department Employee Agent. Detailed Implementation

[0056] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Example 1

[0059] like Figure 1 and Figure 2 As shown in Embodiment 1 of this application, an OPC financial reconciliation processing method includes:

[0060] S100: Monitor expense reimbursement application events in the office workflow approval module. The office workflow approval module can be other enterprise office modules with workflow approval functions, such as Lark Workflow, WeChat Work approval, or DingTalk approval. Lark Workflow will be used as an example for explanation below.

[0061] Step S100 can be executed by an event listening agent to capture reimbursement application events in the Lark workflow in real time. When a new reimbursement application is detected, the subsequent intelligent review process is automatically triggered. The reimbursement application includes reimbursement information such as the reason for reimbursement and the amount of reimbursement, and the invoice is uploaded as an attachment (the invoice can be in the form of an image or PDF).

[0062] Store tax bureau invoice records into the company's financial database: Finance personnel upload tax bureau invoice files to a designated directory, and the "Invoice Record Synchronization Smart Agent" parses the files and stores the parsed invoice records into the company's financial database.

[0063] S200 In response to the submission of a reimbursement application, which includes reimbursement information and invoice attachments, the invoice attachments are input into a preset multimodal model for invoice recognition in order to extract the invoice information.

[0064] Step S200 can be executed by the reconciliation agent, with the aim of recognizing invoice information. Specifically, the reconciliation agent transmits the invoice attachments from the reimbursement application to a multimodal model. The multimodal model performs fusion recognition of visual and textual information on the invoice. This multimodal model integrates visual and textual encoding capabilities, enabling it to perform layout analysis, text detection and recognition on invoice images (or invoice PDF files), accurately extracting the structured information from the invoice. The extracted invoice information includes, but is not limited to: company header, taxpayer identification number (tax number), invoice amount, invoice code, invoice number, invoice date, and seller name. The multimodal model can be a multimodal model capable of cross-modal image and text recognition, such as Doubao, Qianwen, or ChatGPT-4V.

[0065] S300. Conduct multi-dimensional audits of invoice information, including basic information audit, amount consistency audit, invoice authenticity audit, and duplicate reimbursement audit.

[0066] Step S300 can also be executed by the reconciliation agent. Its purpose is to conduct multi-dimensional verification of invoices. Specifically, the "reconciliation agent," based on the invoice information extracted by the multimodal model, sequentially completes the following four verification operations: ① Basic Information Verification: Checks whether the invoice number has the correct number of digits and whether the amount is a number; compares the identified invoice header and tax number with the standard header and tax number registered with the enterprise; ② Amount Consistency Verification: Compares the reimbursement amount filled in by the employee in the Lark workflow with the amount identified on the invoice; ③ Invoice Authenticity Verification: Retrieves the invoice code and number from the enterprise's financial database to confirm whether the invoice is genuine and valid, and compares the accuracy of the invoice information to ensure the accuracy of the invoice information extracted by the multimodal model in step S200, preventing the multimodal model from creating illusions that lead to inaccurate invoice information extraction; ④ Duplicate Reimbursement Verification: Retrieves the reimbursement invoice data from the enterprise's financial database to verify whether the invoice has been submitted and reimbursed by other employees, in order to determine whether duplicate reimbursement exists.

[0067] The enterprise financial database contains basic invoice information, invoice reimbursement information, and pending transfer tasks. The basic invoice information includes the invoice number, seller and buyer, invoice issuer, remarks, amount, and invoice status. The invoice reimbursement information includes the reimbursement item name, reimbursement item type, reimbursing person, approval number, and reimbursement time. Pending transfer tasks include the amount, transfer account, transferor, reimbursement item name, reimbursement item type, time, and whether the transfer has been completed.

[0068] To verify the accuracy of the invoice information extracted by the multimodal model in step S200, it is necessary to compare and verify the company's invoice information from the State Taxation Administration with the invoice information extracted by the multimodal model (i.e., the invoices uploaded by employees for expense reimbursement):

[0069] First, the company's finance staff will periodically download the company's invoice information (Excel document) from the national tax system. This document contains information such as invoice number, seller and buyer identification numbers and names, invoice date, amount, and issuer. After the finance staff uploads this document, the system program will parse and extract this information and save it to the company's financial database.

[0070] When employees submit expense reports, the uploaded invoices are identified using a multimodal model to extract information such as the invoice number, the seller's and buyer's identification numbers and names, and the amount. The first step is to perform some verification on this information itself, such as: the number of digits in the invoice number, the number of digits in the identification number, whether the buyer's name is our company, and whether the amount is a valid number.

[0071] After the first step of verification is successful, the invoice information extracted by the multimodal model is compared with the invoice information in the company's financial database (i.e., the invoice information in the national tax system). If all invoice information matches, the accuracy verification of the invoice information passes, indicating that the information extracted by the multimodal model is completely accurate. If any information in the invoice information does not match, this may be due to a problem with the multimodal model's recognition of information or the employee uploading an incorrect invoice. In this case, the reimbursement process will be rejected, and the error information (e.g., which invoice and which piece of information is inconsistent) will be recorded in the log and written in the reimbursement process remarks. Finance personnel will periodically review these rejected reimbursements to analyze the reasons, which is equivalent to manual review. At the same time, the employee submitting the reimbursement can also see that their reimbursement has been rejected in the approval process and check whether they uploaded an incorrect invoice based on the remarks in the reimbursement process.

[0072] S400: Generate an approval conclusion based on the results of multi-dimensional review, and mark the reimbursement application in the office process approval module according to the approval conclusion.

[0073] Similarly, step S400 can also be executed by the reconciliation agent to achieve intelligent approval feedback. Specifically, the "reconciliation agent" automatically generates an approval conclusion based on the multi-dimensional review results: if all four reviews in the above multi-dimensional review are passed, the reimbursement application will be automatically marked as "approved" in the Lark workflow; if any one of the four reviews fails, it will be marked as "rejected" and the reason for the failure will be simultaneously fed back to the reimbursement application page of the Lark workflow (e.g., "the invoice header does not match the filing information" or "the invoice has been reimbursed twice") for the employee who initiated the reimbursement application to view.

[0074] In a further embodiment, the system also includes intelligent report generation that can be completed by the approval result summary agent. Specifically, after the approval of a single reimbursement application is completed, the "approval result summary agent" updates the approval result data to the company's financial file storage in real time, and generates two types of financial reports: ① Reconciliation report, which generates a report containing the approval result, invoice information, issue details, and statistical information of a single reimbursement application based on invoice information, reconciliation information, approval results, etc.; ② Employee accounts payable summary report, which calculates the total amount of all approved reimbursement applications for each employee, forming the total accounts payable data. This achieves automated and real-time generation of financial reports, immediately compiling data and generating reconciliation and accounts payable summary reports after the approval is completed, ensuring the timeliness and accuracy of the reports, without requiring manual statistics from financial personnel.

[0075] In this embodiment, through the collaborative work of intelligent agents such as the event listening agent, the account agent, and the approval result summary agent, the entire process from the triggering of the reimbursement application to intelligent review, approval feedback, and report generation is automated. This eliminates the need for manual intervention by financial personnel, significantly improving the efficiency of financial reimbursement and reconciliation, completely resolving the backlog of reviews, and enhancing the overall operational efficiency of the company's financial work. Simultaneously, it eliminates the financial risks caused by human error. Through model recognition and automatic comparison, it achieves accurate review of invoice information, amount, authenticity, and duplicate reimbursements, with the accuracy of the review results far exceeding that of manual operation.

[0076] Meanwhile, the monitoring and approval of expense reimbursement applications are fully automated by intelligent agents, eliminating the need for manual operation. This reduces the company's financial human resource costs, frees up the work energy of finance personnel, and allows them to devote themselves to higher-value work such as financial analysis and cost control, thereby enhancing the work value of the company's finance department.

[0077] Example 2

[0078] Based on Example 1, this example further provides a method for processing OPC financial reconciliation that includes data storage, dividing the data into report type and information type to achieve classified storage of different types of data.

[0079] Specifically, the review results and generated report data of this reimbursement application will be uniformly stored in the company's financial file storage; invoice information and the list of pending transfer tasks will be stored in the company's financial database to support subsequent financial accounting and data traceability. The company's financial database includes: basic invoice information (invoice number, seller and buyer, invoice issuer, remarks, amount, invoice status), invoice reimbursement information (reimbursement project name, reimbursement project type, reimbursing person, approval number, reimbursement time), and pending transfer tasks (amount, transfer account, transferor, reimbursement project name, reimbursement project type, time, and whether the transfer has been completed).

[0080] Example 3

[0081] like Figure 3 As shown in Embodiment 2 of this application, an OPC financial reconciliation processing system includes:

[0082] Event Listener Agent 100 is used to listen for expense reimbursement application events in the office workflow approval module.

[0083] The reconciliation agent 200 responds to the submission of a reimbursement application, which includes reimbursement information and invoice attachments. It inputs the invoice attachments into a preset multimodal model for invoice recognition to extract invoice information and performs multi-dimensional review of the invoice information. The multi-dimensional review includes basic information review, amount consistency review, invoice authenticity review, and duplicate reimbursement review. Based on the results of the multi-dimensional review, it generates an approval conclusion and marks the reimbursement application in the office workflow approval module according to the approval conclusion.

[0084] In a further embodiment, it also includes:

[0085] The Invoice Record Synchronization Smart Agent 300 is used to obtain tax bureau invoice files from a specified directory, parse the obtained tax bureau invoice files, and store the parsed invoice records in the enterprise's financial database.

[0086] The approval result summary agent 400 is used to obtain the approval results of all reimbursement applications for the day, and output a summary report after statistical analysis and summarization through a preset large language model.

[0087] It should be noted that during the process of generating the summary report of approval results through the pre-set large language model, in order to avoid the large language model from experiencing illusions during mathematical calculations, the approval result summary agent 400 utilizes the ability to call external tools. Mathematical calculations involving amounts are all obtained by the approval result summary agent 400 calling external mathematical calculation tools. These external mathematical calculation tools can be self-developed or third-party provided external amount calculation tools. The function of these external mathematical calculation tools is to perform mathematical calculations such as addition, subtraction, multiplication, and division as needed, ensuring 100% accuracy in amount calculations and conforming to financial calculation rules. Therefore, the mathematical calculations in the approval result summary report are not performed by the large language model, but are achieved by calling external mathematical calculation tools, thus avoiding the illusion problem during the mathematical calculation process.

[0088] For example, the generation of the above summary report includes:

[0089] 1. After the reconciliation agent 200 performs a reconciliation, it will record the processing result in the approval result log of the day, which includes detailed approval information: the person making the expense claim, the amount, the reason for the expense claim, the processing result, etc.

[0090] 2. The approval result summary agent 400 is designed to execute in the early morning of the next day and generate a summary report of the approval results of the previous day;

[0091] 3. The approval result summary agent 400 will first read the approval result log of the previous day as the knowledge context of the large language model;

[0092] 4. Then, summarize the processing results, including: (1) calling external mathematical calculation tools to calculate the total amount; (2) classifying the processing results and summarizing the approval of different results through preset prompts; (3) generating a summary report based on the preset template or the template selected by the user. An example of the generated reimbursement approval summary report is as follows: Figure 4 As shown, the approval details and rejection details are read from the approval result log and generated by the big language model through summarization and classification. The amount that needs to be calculated mathematically in the amount summary is determined by the big language model and calculated by calling external mathematical calculation tools to ensure the accuracy of the amount calculation.

[0093] The pending reconciliation task agent 500 is used to record pending transfer tasks to the enterprise's financial database at the payment node, query transfer tasks according to specified conditions, generate payment orders for the finance department to make payments, and mark pending transfer tasks as completed after payment.

[0094] The Finance Department employee intelligent agent 600 is used to call other intelligent agents through dialogue.

[0095] It should be noted that the office workflow approval module, namely Lark Workflow, can be completely replaced by other enterprise office modules with workflow approval functions, such as WeChat Work Approval and DingTalk Approval. The large language model on which the backend intelligent agent relies, in addition to DeepSeek, can also be replaced by other large language models with similar natural language understanding and generation capabilities, such as Qianwen and Doubao. The workflow orchestration tool uses dify workflow orchestration, and the development tool uses Python.

[0096] The intelligent financial reconciliation system provided in this application, by introducing an event-based triggering mechanism, the accurate identification capabilities of a multimodal model, and a multi-agent collaborative processing architecture, completely changes the traditional manual reconciliation model. It significantly improves the efficiency of financial reimbursement reconciliation, shortening the review cycle from hours or even days to minutes or even seconds. Simultaneously, it eliminates the financial risks associated with manual operation, ensuring the accuracy of review results; it achieves automated and real-time report generation; and it significantly reduces corporate financial labor costs, freeing up the workload of financial personnel. Therefore, this invention effectively overcomes the shortcomings of existing technologies and has extremely high practical value and promising industrial application prospects.

[0097] In addition, for other specific implementations of the OPC financial reconciliation processing system in this embodiment, please refer to the specific implementations of the OPC financial reconciliation processing method described above. To avoid redundancy, they will not be repeated here.

[0098] Example 4

[0099] The present application relates to a computer-readable storage medium in embodiment 4, which stores program code for execution by a device, the program code including steps for performing the method as in any implementation of embodiment 1 of the present application;

[0100] The computer-readable storage medium may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM); the computer-readable storage medium may store program code, and when the program stored in the computer-readable storage medium is executed by the processor, the processor is used to perform the steps of the method in any of the implementations of Embodiment 1 of this application.

[0101] Example 5

[0102] like Figure 5 As shown, an electronic device according to Embodiment 5 of this application includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the method in any of the implementations in Embodiment 1 of this application.

[0103] The processor can be a general-purpose central processing unit (CPU), microprocessor, application-specific integrated circuit (ASIC), graphics processing unit (GPU), or one or more integrated circuits, used to execute related programs to implement the method in any of the implementations of Embodiment 1 of this application.

[0104] The processor can also be an integrated circuit electronic device with signal processing capabilities. In implementation, each step of the method in any of the implementations of Embodiment 1 of this application can be completed by the integrated logic circuitry in the processor's hardware or by software instructions.

[0105] The aforementioned processor can also be a general-purpose processor, digital signal processor, application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the functions required by the units included in the data processing apparatus of the embodiments of this application, or executes the methods in any implementation of Embodiment 1 of this application.

[0106] The above are merely preferred embodiments of this application; however, the scope of protection of this application is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in this application, based on the technical solution and its improved concept, should be covered within the scope of protection of this application.

Claims

1. A method for processing OPC financial reconciliation, characterized in that, include: Monitor expense reimbursement application events in the office workflow approval module; In response to the submission of a reimbursement application, which includes reimbursement information and invoice attachments, the invoice attachments are input into a preset multimodal model for invoice recognition in order to extract the invoice information; The invoice information is subject to multi-dimensional review, which includes basic information review, amount consistency review, invoice authenticity review, and duplicate reimbursement review. An approval conclusion is generated based on the results of the multi-dimensional review, and the reimbursement application is marked in the office workflow approval module based on the approval conclusion.

2. The method of claim 1, wherein the method further comprises: The multi-dimensional review specifically includes: Basic information verification: Compare whether the extracted invoice header and tax number are consistent with the standard header and tax number registered with the enterprise; whether the invoice number has the correct number of digits and whether the amount is a number; Amount consistency verification: Compare whether the reimbursement amount entered by the employee in the office workflow approval module is consistent with the amount recognized on the invoice; Invoice authenticity verification: Retrieve the invoice code and invoice number from the invoice attachments in the company's financial database, and confirm whether the invoice is genuine and valid, as well as compare the accuracy of the invoice information; Duplicate expense reimbursement review: Retrieve expense invoice data from the company's financial database and verify whether the invoice has been submitted and reimbursed by other employees; The enterprise financial database contains basic invoice information, invoice reimbursement information, and pending transfer tasks.

3. The method of claim 2, wherein the method further comprises: This also includes data updates to the company's financial database, specifically: Retrieve tax bureau invoice information files from the specified directory; The obtained tax bureau invoice information file is parsed; The parsed invoice records are stored in the company's financial database.

4. The method of claim 1, wherein the method further comprises: The marker includes: If all aspects of the approval process for a reimbursement application are approved, the application will be marked as an annotation in the office workflow approval module. If any item in the approval conclusion of the reimbursement application fails, the reimbursement application will be marked as rejected in the office workflow approval module, and the reason for the failure will be fed back to the reimbursement application in the office workflow approval module.

5. The method of claim 1-4, wherein, This also includes the generation of financial reports, specifically: After reviewing a single expense reimbursement application, the approval result data is updated in real time to the company's financial files, and a financial report is generated, which includes: The reconciliation report includes the review results of the reimbursement application, invoice information, and details of any issues. The employee payable amount summary report includes all approved reimbursement applications by employee to form the total payable amount data.

6. The method of claim 5, wherein the step of generating a report comprises the step of: It also includes data storage, specifically: ​ Store the review results and financial reports of expense reimbursement applications in the company's financial file storage; Store invoice information and a list of pending transfer tasks in the company's financial database.

7. A processing system for OPC financial reconciliation, characterized by, include: An event-listening agent is used to listen for expense reimbursement application events in the office workflow approval module. The reconciliation agent responds to the submission of a reimbursement application, which includes reimbursement information and invoice attachments. The invoice attachments are input into a preset multimodal model for invoice recognition to extract invoice information. The invoice information is then subjected to multi-dimensional review, including basic information review, amount consistency review, invoice authenticity review, and duplicate reimbursement review. An approval conclusion is generated based on the results of the multi-dimensional review, and the reimbursement application is marked in the office workflow approval module based on the approval conclusion.

8. The processing system for OPC financial reconciliation of claim 7, wherein, Also includes: The invoice record synchronization intelligent agent is used to obtain tax bureau invoice information files in a specified directory, parse the obtained tax bureau invoice information files, and store the parsed invoice records in the enterprise's financial database. The approval result summary agent is used to obtain the approval results of all reimbursement applications for the day, and output a summary report after statistical analysis and summarization through a preset large language model. The preset large language model performs mathematical calculations in the statistical analysis and summarization process by calling external mathematical calculation tools. The pending reconciliation task agent is used to record pending transfer tasks to the enterprise's financial database at the payment node, query transfer tasks according to specified conditions, generate payment orders for the finance department to make payments, and mark pending transfer tasks as completed after payment is completed. The Finance Department employee AI agent is used to invoke other AI agents through dialogue.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program code for execution by the device, the program code including steps for performing the method as described in any one of claims 1-6.

10. An electronic device, characterized in that, The electronic device includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the method as described in any one of claims 1-6.

Images