Voucher rechecking method, device, equipment and product

Abstract

The invention discloses a voucher rechecking method, device, equipment and product, and relates to the technical field of accounting, the method is applied to an enterprise terminal, and the method comprises the steps: receiving business data, carrying out the scene classification of the business data, and obtaining a business scene; performing accounting subject mapping on the business data based on the business scene to obtain a subject mapping result, and generating an accounting document according to the subject mapping result; and determining the business complexity of the accounting voucher according to the business scene, and rechecking the accounting voucher based on the business complexity to obtain a rechecking result. Therefore, the problems of low efficiency and high risk caused by poor business rule adaptability of an existing accounting document processing scheme and rechecking depending on manpower are solved, and the document rechecking efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of accounting technology, and in particular to a method, apparatus, equipment and product for voucher review. Background Technology

[0002] Currently, the mainstream accounting voucher processing solutions on the market are based on automated generation by software programs and are widely used in enterprise ERP systems and financial software. By pre-customizing accounting voucher templates and establishing fixed accounting subject mapping rules, they automatically convert specific business document data such as purchase orders, sales invoices, and bank statements generated by the business system into vouchers that meet the basic accounting standards for review. This replaces manual entry of each item and achieves data linkage between business and finance and improves the efficiency of basic processing.

[0003] However, this solution has significant drawbacks. First, it has poor adaptability to business rules. When companies face situations such as new cross-border business, changes in cost accounting methods, or updates to tax policy-related items, the preset templates and mapping logic are completely incompatible. Companies need to invest a lot of resources and require the technical and financial teams to work together to modify the program, which usually takes several weeks, severely slowing down the timeliness of financial processing. Second, the review process is inefficient and risky. Traditional review relies heavily on manual verification of each document. Especially during peak periods such as the end of the month and the end of the year when the volume of documents surges, the efficiency bottleneck becomes particularly prominent. Moreover, repetitive manual work is prone to fatigue errors, and the differences in experience among personnel lead to inconsistent review standards, further increasing the risk of financial operations and subsequent audit costs for companies, becoming a key obstacle to improving the quality and efficiency of financial processes.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a method, apparatus, equipment and product for voucher review, which aims to solve the technical problems of poor business rule adaptability of existing accounting voucher processing solutions and low efficiency and high risk caused by reliance on manual review.

[0006] To achieve the above objectives, this application proposes a voucher verification method, which is applied at the enterprise level. The voucher verification method includes: Receive business data, classify the business data into scenarios, and obtain business scenarios; Based on the business scenario, the business data is mapped to accounting subjects to obtain the subject mapping results, and accounting vouchers are generated based on the subject mapping results. The business complexity of the accounting voucher is determined based on the business scenario, and the accounting voucher is reviewed based on the business complexity to obtain the review result.

[0007] In one embodiment, the steps of receiving service data, classifying the service data into scenarios, and obtaining service scenarios include: The business data is analyzed to obtain the data source, and the business data is classified into structured data, semi-structured data, and unstructured data based on the data source. The structured data is cleaned to obtain cleaned structured data. The semi-structured data is scanned to obtain character scanning results. The unstructured data is labeled with features using a natural language processing model to obtain labeling results. A set of structured business elements is generated based on the cleaned structured data, character scanning results, and annotation results; The business data is classified into scenarios by using the structured business element set to obtain business scenarios.

[0008] In one embodiment, the step of classifying the business data into scenarios using the structured business element set to obtain business scenarios includes: The business attributes are obtained by performing basic attribute judgments on the business data using the structured business element set. The business data is classified by transaction nature based on the structured business element set to obtain the nature classification result; Additional feature recognition is performed on the business data using the structured business element set to obtain feature recognition results; Based on the business attributes, property classification results, and feature recognition results, a multi-dimensional analysis is performed using a business scenario recognition model to obtain the business scenario.

[0009] In one embodiment, the step of mapping accounting subjects to the business data based on the business scenario to obtain the subject mapping result includes: If the business scenario is a simplified scenario, the enterprise's basic accounting rule base is called, and the business data is mapped to accounting subjects through the basic accounting rule base to obtain the subject mapping result; If the business scenario is complex, the business data and the business scenario are combined for contextual understanding to obtain the business behavior. Based on the business behavior, the core knowledge base of the enterprise is called through the retrieval enhancement generation framework to perform accounting subject mapping on the business behavior to obtain the subject mapping result.

[0010] In one embodiment, the step of reviewing the accounting voucher based on the business complexity to obtain the review result includes: When the business complexity is simple, the accounting voucher is queried through the review rule engine to obtain the review rule code; The review rules are extracted using the review rule code, and the accounting vouchers are reviewed using the review rules to obtain the review results.

[0011] In one embodiment, the step of reviewing the accounting voucher based on the business complexity to obtain the review result further includes: When the business complexity is complex, a large language model is used to perform semantic understanding and compliance checks on the accounting vouchers and business data to obtain compliance check results. The business logic rationality is checked by using the associated information corresponding to the accounting voucher and the business scenario to obtain the logic rationality check result; Based on the historical abnormal voucher records of the enterprise, abnormal signals are identified in the accounting vouchers to obtain abnormal identification results; The results of the compliance check, the logical rationality check, and the anomaly identification are reviewed from multiple dimensions to obtain the review results.

[0012] In one embodiment, after the steps of determining the business complexity of the accounting voucher based on the business scenario, reviewing the accounting voucher based on the business complexity, and obtaining the review result, the method further includes: Record the verification items, pass items, and exception items corresponding to the verification results; A review report is generated based on the verification items, pass items, and exception items. The accounting subject basic rule base is updated based on the review report, and the update result is obtained.

[0013] Furthermore, to achieve the above objectives, this application also proposes a voucher verification device, wherein the voucher verification is applied at the enterprise level, and the voucher verification device includes: The classification module is used to receive business data, classify the business data into scenarios, and obtain business scenarios. The generation module is used to map the business data to accounting subjects based on the business scenario, obtain the subject mapping result, and generate accounting vouchers based on the subject mapping result; The review module is used to determine the business complexity of the accounting voucher based on the business scenario, review the accounting voucher based on the business complexity, and obtain the review result.

[0014] In addition, to achieve the above objectives, this application also proposes a voucher verification device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the voucher verification method as described above.

[0015] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the credential verification method described above.

[0016] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the credential verification method described above.

[0017] One or more technical solutions proposed in this application have at least the following technical effects: This application proposes a voucher review method, apparatus, device, and product. The method, applied to an enterprise, receives business data, classifies the data into scenarios to obtain business scenarios, maps the business data to accounting subjects based on these scenarios, generates accounting vouchers based on the mapping results, determines the business complexity of the accounting vouchers based on the business scenarios, and reviews the accounting vouchers based on the business complexity to obtain a review result. Thus, after receiving business data, the method analyzes the data to obtain the corresponding business scenarios, maps the business data to accounting subjects based on these scenarios, generates accounting vouchers based on the mapping results, determines the business complexity of the accounting vouchers based on the business scenarios, and reviews the accounting vouchers based on the business complexity to obtain a voucher review result. This solves the problems of poor business rule adaptability in existing accounting voucher processing solutions and low efficiency and high risk due to reliance on manual review, thereby improving the efficiency of voucher review. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments of the present application and, together with the specification, serve to explain the principles of the present application.

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the first embodiment of the document verification method for this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the document verification method for this application. Figure 3 A simplified flowchart illustrating the document verification method provided in Embodiment 2 of this application; Figure 4 This is a schematic diagram of the module structure of the certificate verification device according to an embodiment of this application; Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the credential verification method in this application embodiment.

[0021] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0022] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0023] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0024] The main solution of this application embodiment is as follows: Parse the business data to obtain the data source; classify the business data by type based on the data source to obtain structured data, semi-structured data, and unstructured data; clean the structured data to obtain cleaned structured data; scan the semi-structured data to obtain character scanning results; annotate the unstructured data using a natural language processing model to obtain annotation results; generate a structured business element set based on the cleaned structured data, character scanning results, and annotation results; classify the business data by scenario using the structured business element set to obtain business scenarios; determine the basic attributes of the business data using the structured business element set to obtain business attributes; classify the business data by transaction nature based on the structured business element set to obtain nature classification results; identify additional features of the business data using the structured business element set to obtain feature recognition results; and perform multi-dimensional analysis using a business scenario recognition model based on the business attributes, nature classification results, and feature recognition results to obtain business scenarios. If the business scenario is simple, the enterprise's basic accounting rule base is invoked, and the business data is mapped to accounting subjects using this base to obtain a mapping result. If the business scenario is complex, the business data and scenario are combined for contextual understanding to obtain business behavior. Based on this behavior, the enterprise's core knowledge base is invoked through a retrieval-enhanced generation framework to map the business behavior to accounting subjects, resulting in a mapping result. In the case of simple business complexity, the accounting voucher is queried using a review rule engine to obtain review rule code. Review rules are extracted from this code, and the accounting voucher is reviewed using these rules to obtain a review result. Under the condition of complex business complexity, a large language model is used to perform semantic understanding and compliance checks on the accounting vouchers and business data to obtain compliance check results; the business logic rationality check is performed based on the associated information corresponding to the accounting vouchers and the business scenario to obtain logical rationality check results; anomaly signal identification is performed on the accounting vouchers based on the historical abnormal voucher records of the enterprise to obtain anomaly identification results; a multi-dimensional review is performed based on the compliance check results, logical rationality check results, and anomaly identification results to obtain review results. The verification items, pass items, and anomaly items corresponding to the review results are recorded; a review report is generated based on the verification items, pass items, and anomaly items; the basic rule base of accounting subjects is updated based on the review report to obtain update results. This solves the problems of poor business rule adaptability in existing accounting voucher processing solutions and low efficiency and high risk caused by reliance on manual review, realizing voucher review and improving the efficiency of voucher review.Based on the present invention, addressing the problem that traditional accounting voucher review relies heavily on manual verification of each voucher and item, resulting in low efficiency and difficulty in handling massive amounts of documents, a voucher review and identification method was designed. The effectiveness of the voucher review method of the present invention was verified when reviewing vouchers, and the efficiency of voucher review was significantly improved by the method of the present invention.

[0025] In this embodiment, for ease of description, the following description uses the voucher verification device as the execution subject.

[0026] Due to the limitations of existing automated accounting voucher processing solutions, their business adaptability and review efficiency still need improvement. One issue is poor adaptability to business rules. When enterprises face new cross-border business, changes in cost accounting methods, or updates to tax policy-related accounts, the preset templates and mapping logic are completely incompatible. Enterprises need to invest significant costs and require collaborative modifications between the technical and financial teams, a process that typically takes several weeks, severely slowing down financial processing efficiency. This leads to a decline in the business adaptability of accounting voucher processing. Another issue is the low efficiency and high risk in the review process. Traditional review heavily relies on manual verification of each voucher, especially during peak periods at the end of the month and year when the volume of documents surges, making efficiency bottlenecks particularly prominent. Furthermore, repetitive manual labor is prone to fatigue errors, and the workload of personnel is also a concern. Inconsistent verification standards lead to different review criteria, further increasing the risks of corporate financial operations and subsequent audit costs. This affects the overall efficiency of accounting voucher processing. Furthermore, there are differences in corporate business scenarios. Efficient accounting voucher processing needs to adapt to the business models and financial processes of different companies. However, different companies have significant differences in business scope, accounting standards, and policy adaptation. Some companies are also involved in special industry regulations, which also affects the universality of accounting voucher processing solutions. Therefore, in the current corporate financial processes, there are also challenges in improving the quality and efficiency of finance. Because different companies have different business scenarios, accounting requirements, and policy-related needs, if accounting voucher processing solutions are not specifically customized, optimized, and flexibly adapted, their efficiency and security in practical applications will also decline.

[0027] This application provides a solution whereby, on the enterprise side, business data is received, categorized into scenarios to obtain business scenarios; accounting subjects are mapped to the business data based on the business scenarios to obtain mapping results, and accounting vouchers are generated based on the mapping results; the business complexity of the accounting vouchers is determined based on the business scenarios, and the accounting vouchers are reviewed based on the business complexity to obtain review results. Thus, after receiving business data, the solution analyzes the business data to obtain the corresponding business scenarios, then maps the business data to accounting subjects based on the business scenarios, generates accounting vouchers based on the mapping results, and finally determines the business complexity of the accounting vouchers based on the business scenarios and reviews the accounting vouchers based on the business complexity to obtain voucher review results. This solves the problems of poor business rule adaptability and low efficiency and high risk caused by manual review in existing accounting voucher processing solutions, improving the efficiency of voucher review and providing users with a better service.

[0028] Based on this, the embodiments of this application provide a method for credential verification, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the document verification method of this application.

[0029] In this embodiment, the voucher verification method is applied to the enterprise side, and the voucher verification method includes S01~S03: Step S01: Receive business data, classify the business data into scenarios, and obtain business scenarios; Before describing the solution in this embodiment, it should be clear that the current mainstream accounting voucher processing solution is based on automated generation by software programs. It is widely used in enterprise ERP systems and financial software. By pre-setting voucher templates and fixed accounting subject mapping rules, business documents such as purchase orders and sales invoices are automatically converted into standardized vouchers, replacing manual entry to improve business-finance linkage and processing efficiency. However, this solution has obvious defects. First, it has poor adaptability to business rules. When enterprises face new cross-border business or changes in accounting standards, the pre-set templates and logic cannot be adapted. The technical and financial teams need to work together to modify the program, which is costly and time-consuming (usually several weeks), slowing down financial efficiency. Second, the review process is inefficient and risky. It relies on manual verification of each voucher. The efficiency bottleneck is prominent during peak closing periods. Moreover, repetitive manual work is prone to errors, and differences in experience lead to inconsistent review standards, increasing financial operation risks and audit costs.

[0030] To address the aforementioned issues, this embodiment receives business data sent by users or customers on the enterprise side. Then, it categorizes the business data into scenarios to obtain the corresponding business scenarios. The criteria for classifying business scenarios include, but are not limited to, business type (e.g., procurement, sales, expense reimbursement, asset disposal), transaction nature (e.g., domestic transactions, cross-border transactions, related-party transactions), accounting entity (e.g., parent company, subsidiary, business unit), and whether special accounting treatment is involved (e.g., financial instruments, leasing business, debt restructuring). For example, when business data includes information such as import customs declarations, foreign exchange settlement documents, and cross-border service contracts, the system can classify it as a "cross-border procurement business scenario."

[0031] Step S02: Based on the business scenario, perform accounting subject mapping on the business data to obtain the subject mapping result, and generate accounting vouchers according to the subject mapping result; Once the current business scenario is determined, accounting subjects can be mapped to the business data based on the business scenario to obtain the corresponding subject mapping results. Then, corresponding accounting vouchers can be generated based on the subject mapping results. The subject mapping rules are implemented through a preset scenario-subject association model. This model integrates enterprise accounting standards, industry accounting systems, and enterprise internal accounting norms. For example, in the "cross-border procurement business scenario", the purchase price, customs duties, and consumption tax of imported goods, which can be directly attributed to inventory costs, will be automatically mapped to the "raw materials" or "inventory goods" accounts.

[0032] Step S03: Determine the business complexity of the accounting voucher based on the business scenario, and review the accounting voucher based on the business complexity to obtain the review result.

[0033] Finally, the generated accounting vouchers need to be reviewed. Therefore, the solution in this embodiment needs to determine the business complexity corresponding to the accounting vouchers based on the business scenario, and then review the accounting vouchers based on the business complexity to obtain the corresponding review results. The classification of business complexity can be comprehensively judged by combining multiple dimensions such as the number of transaction elements involved in the business scenario, special accounting treatment requirements, and regulatory compliance risk level. For example, for the "domestic standard procurement business scenario" which only involves a single domestic supplier, regular goods procurement and no special taxes and fees, its business complexity can be judged as "low". However, for the "cross-border related party hedging business scenario" which includes cross-border multi-party transaction entities, involves forward foreign exchange contract hedging, and also has the need for related party pricing fairness assessment, its business complexity is judged as "high".

[0034] Specifically, step S01 above, which involves receiving business data and classifying the business data into scenarios to obtain the business scenarios, includes the following steps: Step S011: Parse the business data to obtain the data source, and classify the business data by type based on the data source to obtain structured data, semi-structured data, and unstructured data; Step S012: Clean the structured data to obtain cleaned structured data; perform character scanning on the semi-structured data to obtain character scanning results; and use a natural language processing model to annotate the unstructured data to obtain annotation results. Step S013: Generate a structured business element set based on the cleaned structured data, character scanning results, and annotation results; Step S014: Classify the business data into scenarios using the structured business element set to obtain business scenarios.

[0035] This embodiment adopts a customized processing strategy based on the characteristics of different types of data to achieve efficient information extraction and standardization: For structured data, targeted data cleaning is carried out on structured data such as ERP business documents and bank statements. By removing redundant fields, correcting outliers, and filling in missing information, cleaned structured data with uniform format and accurate data is obtained, ensuring the reliability of the basic data.

[0036] For semi-structured data, we perform character-by-character scanning on semi-structured data such as expense report "explanation of purpose" and email payment instructions, focusing on capturing key information such as amount, date, and keywords of purpose, and forming character scanning results containing core business identifiers, thus overcoming the processing difficulties of semi-structured data with inconsistent formats.

[0037] For unstructured data processing, a fusion of information extraction and structuring technologies is adopted. For unstructured data such as invoices (PDF / images), contracts (PDF / Word), and receipts (images), OCR technology is first used to convert images and scanned documents into editable text. Then, natural language processing models and document understanding models are combined to annotate elements, accurately extracting key business elements such as payer, payee, amount, and business type. Finally, annotated results containing complete business information are obtained, realizing in-depth analysis and value mining of unstructured data.

[0038] Subsequently, the cleaned structured data, the character scanning results of semi-structured data, and the annotation results of unstructured data are fused from multiple sources. According to the business logic association rules, various elements are integrated, deduplicated, and standardized to generate a structured business element set covering core dimensions such as business subject, amount, time, business type, and related document number. This element set breaks down the information barriers between different types of data, realizes the transformation of business data from scattered to centralized and from disordered to ordered, and provides a unified and efficient data foundation for subsequent scenario classification.

[0039] Finally, based on the structured business element set, the system classifies the original business data into business scenarios through a preset scenario classification rule base and intelligent matching algorithm. During the classification process, the system automatically identifies key features in the element set (such as the combination of "purchase order + warehouse receipt" corresponding to the purchase scenario, and the combination of "reimbursement reason + invoice" corresponding to the expense reimbursement scenario, etc.), and finally outputs accurate business scenarios to provide scenario-based support for subsequent targeted voucher processing, process adaptation and other links, thereby improving the intelligence and accuracy of the overall business processing.

[0040] More specifically, step S014 above, which involves classifying the business data into scenarios using the structured business element set to obtain the business scenarios, includes: Step S0141: Perform basic attribute judgment on the business data through the structured business element set to obtain business attributes; Step S0142: Based on the structured business element set, classify the business data according to transaction nature to obtain the nature classification result; Step S0143: Perform additional feature recognition on the business data using the structured business element set to obtain feature recognition results; Step S0144: Based on the business attributes, property classification results, and feature recognition results, a multi-dimensional analysis is performed using a business scenario recognition model to obtain the business scenario.

[0041] First, key dimensions such as the business entity (e.g., payer and payee names and types), transaction time (date of occurrence, settlement cycle), core amount (total transaction amount, taxes, and detailed amounts), and related document numbers (order number, invoice number) are extracted from the structured business elements. Through rule validation and format standardization, basic feature tags for the business data are determined. These basic attributes are prerequisites for a deeper understanding of the business substance. For example, attributes such as "the payee is the first supplier" and "the transaction time is within the current month's procurement cycle" provide initial basis for determining the business type.

[0042] Subsequently, based on information such as business type identifiers, fund flows, and related accounts in the structured business elements, the business data is precisely classified according to transaction nature, resulting in a classification result. The classification dimensions cover the core economic attributes of the transaction, such as revenue (product sales, service revenue, etc.), expenditure (purchase payments, expense reimbursements, asset purchases, etc.), accounts receivable / payable (accounts receivable collection, accounts payable payment, etc.), and equity (investment, dividends, etc.). This embodiment achieves preliminary classification of core business behaviors by constructing a keyword library of transaction natures and logical matching rules. For example, when the elements show a combination of "purchase order + warehouse receipt + supplier", it can be initially classified as "purchase expenditure nature", narrowing the scope for subsequent in-depth analysis of the business substance.

[0043] Next, a feature extraction algorithm is used to perform a deep scan of the structured business elements to obtain feature recognition results. In this embodiment, additional features include specific transaction attributes, such as whether it involves cross-border transactions (including foreign currency settlement, international suppliers / customers), whether it is a prepayment / prepayment, whether it is associated with individual employees (e.g., employee name, department identifier), and whether it involves a specific project (project number, project name). These features are key to distinguishing similar business scenarios. For example, even with the same expenditure nature, the difference in additional features between "associated with employees + travel expense invoices" and "associated with suppliers + equipment contracts" directly points to different specific scenarios.

[0044] Finally, the business scenario identification model is used to conduct a multi-dimensional comprehensive analysis of the business attributes, property classification results, and feature recognition results output from the preceding stages. This model integrates the deep semantic understanding capabilities of large models and can comprehensively analyze all input information (including structured fields, text content extracted from unstructured data, and the contextual logic of business occurrence), thereby gaining a deep understanding of the economic substance of business occurrence.

[0045] The specific steps described above can be as follows: When the model obtains information such as "payer is the company's administrative department + payee is the hotel + amount is 5,000 yuan + additional features include 'employee travel approval form'", it will accurately determine that the payment is essentially "reimbursement of employee travel expenses". If the information combination is "payer is the purchasing department + payee is the second equipment manufacturer + amount is 100,000 yuan + additional features include 'prepayment clause'", it will be judged as "prepayment for equipment". At the same time, because the model has dynamic learning and innovative classification capabilities, when encountering new business combinations that are not preset, it can intelligently judge and generate new reasonable classifications based on economic substance logic, breaking the limitations of traditional fixed templates, realizing flexible adaptation and accurate identification of business scenarios, and finally outputting accurate business scenario results.

[0046] Further, step S02 above, which involves mapping the business data to accounting subjects based on the business scenario to obtain the subject mapping result, includes: Step S021: If the business scenario is a simplified scenario, the accounting subject basic rule library of the enterprise is called, and the business data is mapped to accounting subjects through the accounting subject basic rule library to obtain the subject mapping result; Step S022: If the business scenario is a complex scenario, the business data and the business scenario are combined for contextual understanding to obtain the business behavior. Based on the business behavior, the core knowledge base of the enterprise is called through the retrieval enhancement generation framework to perform accounting subject mapping on the business behavior to obtain the subject mapping result.

[0047] When the business scenario is determined to be a simplified scenario, a rule-enhanced processing approach is adopted. The system directly calls the enterprise's basic accounting rule library to complete the account mapping and obtain the account mapping result. In this embodiment, the simplified scenario usually refers to routine business with clear business logic, fixed transaction patterns, and clear element identification, such as standard purchase payments and routine product sales receipts. The basic accounting rule library has built-in enterprise-specific basic mapping rules, which are driven by the "accounting voucher template" rule engine. For example, there are clear correspondences such as "payment to suppliers -> accounts payable / bank deposits" and "sales receipts -> main business revenue / bank deposits". During processing, the system accurately matches the core elements in the business data (such as transaction type, fund flow, and related party type) with the rule library. It can quickly output the account mapping result without complex semantic understanding. This rule-based processing can greatly improve the account mapping efficiency of routine business and reduce resource consumption.

[0048] When a business scenario is determined to be complex, an intelligent processing flow led by a large model is initiated. This flow uses a full-chain logic of "combined analysis - contextual understanding - RAG-assisted decision-making" to obtain the subject mapping results. Complex scenarios encompass ambiguous descriptions of business (such as non-standard descriptions of reimbursement reasons), multi-dimensional complex business (such as cross-border mixed procurement including freight and miscellaneous cost allocation), and new business types (such as newly added digital asset transactions by the enterprise). A deep assessment based on the actual business situation is required. The specific process consists of three steps: First, combine business data and business scenarios to understand the context. Integrate data from multiple dimensions, such as structured business elements, unstructured text information extraction, and business background, to accurately extract complete business behaviors, such as "purchasing server software for R&D, including 3 years of service fees, with payment in two installments."

[0049] Secondly, based on the extracted business behaviors, the Search Enhancement Generation (RAG) framework is launched to retrieve the enterprise's core knowledge base in real time. This knowledge base includes customized content such as accounting policy manuals, detailed instructions on the use of accounts, and historical similar business processing cases.

[0050] Finally, the big model combines an understanding of the business substance, the company's specific accounting chart, and the results of knowledge base retrieval to make a comprehensive judgment to achieve accurate mapping. For example, for "purchasing server software for R&D", it will combine the company's R&D expense capitalization policy to determine whether to record it as "intangible assets" or "R&D expenses - capitalized expenditures / expenses", ensuring the accuracy and compliance of the account mapping in complex scenarios.

[0051] The innovation of this scenario-based strategy lies in the fact that it not only ensures the efficiency of simple business through rule-based processing, but also breaks through the limitations of traditional fixed templates in adapting to complex and new businesses by leveraging the intelligent combination of "large model + RAG", thus achieving the dual goals of "efficiency and accuracy" in accounting subject mapping.

[0052] This embodiment, through the above-described scheme, specifically receives business data, classifies the business data into scenarios to obtain business scenarios; maps the business data to accounting subjects based on the business scenarios to obtain subject mapping results, and generates accounting vouchers based on the subject mapping results; determines the business complexity of the accounting vouchers based on the business scenarios, and reviews the accounting vouchers based on the business complexity to obtain review results. Thus, after receiving business data, it analyzes the business data to obtain the corresponding business scenarios, then maps the business data to accounting subjects based on the business scenarios, generates accounting vouchers based on the mapping results, and finally determines the business complexity of the accounting vouchers based on the business scenarios, and reviews the accounting vouchers based on the business complexity to obtain voucher review results. This solves the problems of poor business rule adaptability and low efficiency and high risk caused by manual review in existing accounting voucher processing schemes, thereby improving the efficiency of voucher review.

[0053] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 In step S03, the voucher review method further includes steps S031-S032, whereby the voucher is reviewed based on the business complexity to obtain the review result. Step S031: When the business complexity is simple, the accounting voucher is queried through the review rule engine to obtain the review rule code; Step S032: Extract the review rules through the review rule code, and review the accounting voucher through the review rules to obtain the review result.

[0054] When the business complexity is determined to be simple, the system prioritizes the activation of the review rule engine for basic verification preprocessing. The core of this step is to initiate rule queries on accounting vouchers through the review rule engine and ultimately obtain the corresponding review rule code. The review rule engine integrates key rule transformation logic, which pre-converts the basic rules in traditional reviews (such as debit and credit balance, compliance of account usage, correctness of amount calculation, date logic, and completeness of voucher elements) into standardized code rules and stores them in the rule base. These code rules cover all the basic verification dimensions that frequently occur in simple business scenarios. By identifying the business type and simplicity / complexity tag corresponding to the accounting voucher, the system accurately matches and extracts the appropriate review rule code, providing a clear execution basis for subsequent automated reviews and avoiding the tedious process of manually checking each rule.

[0055] After obtaining the review rule code, the system automatically extracts the corresponding specific review rules from the code, and then performs full-element automated review of accounting vouchers based on these rules, and finally generates review results. For example, the extracted "debit and credit balance" rule code will drive the system to verify whether the total debit amount in the voucher is equal to the total credit amount, and the "account usage compliance" rule code will match the enterprise's accounting chart of accounts to check whether the accounts used in the voucher are within the prescribed range and meet the business type adaptation requirements.

[0056] The aforementioned rule engine, with its high-speed execution capability of coded rules, achieves a second-level response for simple business voucher review, improving the efficiency and accuracy of basic review, while also releasing the computing power and capabilities of large models, allowing the entire review system to achieve a balance between efficiency and the ability to handle complex scenarios.

[0057] Specifically, step S03 above, which involves reviewing the accounting voucher based on the business complexity to obtain the review result, further includes: Step S033: When the business complexity is complex, semantic understanding and compliance checks are performed on the accounting vouchers and business data using a large language model to obtain compliance check results. Step S034: Perform a business logic rationality check using the associated information corresponding to the accounting voucher and the business scenario to obtain the logic rationality check result; Step S035: Based on the historical abnormal voucher records of the enterprise, perform abnormal signal identification on the accounting voucher to obtain the abnormal identification result; Step S036: Perform a multi-dimensional review based on the compliance check results, logical rationality check results, and anomaly identification results to obtain the review results.

[0058] First, integrate the related information corresponding to accounting vouchers (such as purchase orders, receipts, contracts, travel plans, historical transaction data, etc.) with business scenarios to determine whether the transaction behavior conforms to normal business logic and generate a logical reasonableness check result. The specific implementation process is as follows: The system enables intelligent judgment across multiple typical scenarios. For example, if a supplier makes five purchases of less than 2,000 yuan each within a week, the model will identify the abnormal logic of "multiple small purchases in a short period of time." If the purchase price is 30% higher than the supplier's historical average price without any explanation of the price change, it will be marked as "purchase price deviates significantly from historical price." When employees submit travel expense reimbursements, if the reimbursement location is "Beijing" but the business trip plan is registered as "Shanghai," it will trigger a logic warning of "reimbursement location does not match business trip plan." Payment dates earlier than the "payment within 30 days after receipt" deadline stipulated in the contract or the actual receipt date will also be judged as illogical, achieving deep linkage and verification between business data and scenario information.

[0059] Subsequently, through the pattern learning and feature extraction capabilities of the large model, abnormal signals are identified in the current accounting vouchers to obtain abnormal identification results. In this embodiment, the large model automatically learns the core features in historical abnormal data to form a dynamically updated abnormal pattern library, covering high-frequency error or fraud features such as "specific suppliers (such as suppliers that have previously issued false invoices), specific accounts (such as the "Other Payables" detail account that frequently has errors), specific amount ranges (such as multiple reimbursements of the integer amount of "9999 yuan"), and specific time points (such as concentrated payments on the last day of the month)". When the current voucher matches an abnormal pattern in the library, such as paying a large "service fee" to a supplier that has previously been involved in false invoices, the model will trigger an abnormal signal in real time, realizing the transformation from "post-event tracing" to "in-event early warning" and improving the foresight of risk identification.

[0060] Finally, by establishing a weighted evaluation system, the system performs multi-dimensional integrated analysis of compliance inspection results, logical rationality inspection results, and anomaly identification results, ultimately outputting a comprehensive review result. The system assigns corresponding weights to different inspection items based on their risk levels (e.g., "anomaly pattern recognition" matching historical fraud features has the highest weight, while "vague semantic description" has a relatively lower weight). It also sets warning thresholds based on the company's risk preferences. If a single result triggers a high-risk warning (e.g., matching historical fraud patterns), it is directly marked as "review failed" and prompted for key verification. If multiple results trigger medium- or low-risk warnings (e.g., price deviation + supplier business discrepancy), it is marked as "requires manual review" and the risk points are summarized. If all results are normal, it is judged as "review passed." This integrated model avoids the one-sidedness of single-dimensional inspections and achieves a dual improvement in the accuracy and risk coverage of review results in complex business scenarios.

[0061] More specifically, after step S03 above, which involves determining the business complexity of the accounting voucher based on the business scenario, reviewing the accounting voucher based on the business complexity, and obtaining the review result, the method further includes: Step S04: Record the verification items, pass items, and exception items corresponding to the verification results; Step S05: Generate a review report using the verification items, pass items, and exception items; Step S06: Update the basic rule base of accounting subjects based on the review report to obtain the update result.

[0062] The specific content recorded in this embodiment includes: (1) Detailed list of verification items, covering all verification dimensions called in this review (such as loan balance, account compliance, logical rationality, abnormal pattern matching, etc.), and marking the execution rule number and priority of each verification item; (2) Through the item list, record in detail the specific content of the verification, including the business elements corresponding to the verified item (such as "the account mapping corresponding to purchase order number PO2025001 is compliant"), the verification timestamp and the judgment basis for automatic verification; (3) In-depth information on anomalies: In addition to recording the anomaly content, supplementary information includes the anomaly risk level (high / medium / low), anomaly-related data (such as the business document number that triggered the anomaly, the amount and subject involved), the preliminary cause of the anomaly (such as "the rule does not cover new business scenarios" or "data entry error"), and the manual annotation entry point.

[0063] Based on the multi-dimensional data recorded above, a review report that is both visually appealing and practical is generated. The report design incorporates three innovative functions: First, multi-dimensional visualization, using pie charts to show the "percentage of passed / abnormal items," bar charts to show the "distribution of abnormal items at each risk level," and heatmaps to annotate "high-frequency abnormal verification items," making the review results intuitive and easy to understand. Second, intelligent attribution analysis of abnormal items, automatically linking historical similar cases and business scenario information for medium- and high-risk abnormal items, providing attribution suggestions such as "missing rules," "data anomalies," and "special business logic," along with relevant evidence (such as "this abnormality is related to the rule not being updated in the cross-border business scenario added in March 2025"). Third, interactive traceability function, each data node in the report supports clicking to jump to view the corresponding original business data, the original text of the verification rules, and the review process log, making it convenient for finance personnel to quickly locate the root cause of the problem. In addition, the report also supports custom filtering (such as filtering by business scenario, time range, and abnormality type) to meet the analysis needs of users at different levels.

[0064] Finally, through data analysis of the review report, the intelligent update of the accounting subject basic rule base is driven, forming a closed loop of "review-feedback-optimization". The update logic in this embodiment includes three paths: First, rule supplementation. For frequently occurring "medium-to-high risk anomalies caused by rule omissions" in the report (such as "no corresponding rule for the subject mapping of new digital asset procurement"), the rule addition application is automatically triggered, a rule draft is generated and pushed to the finance and technology teams for review, and it is included in the rule base after approval. Second, rule optimization. For verification items in the report with "low pass rate but high business rationality" (such as "the rule for the collection of expenses for specific R&D projects"), it is analyzed whether there are problems with overly strict rules or vague descriptions, and suggestions for adjusting rule parameters are proposed (such as relaxing the reasonable threshold for "the number of R&D expense attachments"). Third, rule priority adjustment. According to the frequency of abnormal triggering and risk impact of each verification item in the report, the execution priority of verification items in the rule base is dynamically adjusted (such as raising the priority of "abnormal pattern matching" from "medium" to "high") to ensure that high-risk verification items are executed first.

[0065] This embodiment provides a dynamic update mechanism based on actual review data, enabling the rule base to continuously adapt to changes in enterprise business, avoiding the lag of traditional "manual periodic updates", and improving the long-term adaptability of the entire review system.

[0066] This embodiment, through the above-described scheme, specifically by using a review rule engine to query the accounting vouchers under the condition of simple business complexity, obtains review rule codes; extracts review rules from the review rule codes, and reviews the accounting vouchers using the review rules to obtain review results. Thus, after receiving business data, it analyzes the business data to obtain the corresponding business scenario, then maps the business data to accounting subjects based on the business scenario, generates accounting vouchers based on the mapping results, and finally determines the business complexity corresponding to the accounting vouchers based on the business scenario, and reviews the accounting vouchers based on the business complexity to obtain voucher review results. This solves the problems of poor business rule adaptability and low efficiency and high risk caused by reliance on manual review in existing accounting voucher processing schemes, improving the efficiency of voucher review and providing users with better services.

[0067] For example, to help understand the implementation flow of the voucher verification method obtained by combining this embodiment with the above embodiment one, please refer to... Figure 3 , Figure 3 A simplified flowchart of a voucher verification method is provided, specifically: First, business elements are extracted in two ways. The first is to process structured data such as purchase orders and bank statements, and directly extract business elements (such as transaction amount and supplier information). The second is to extract business elements (such as price and tax information on invoices) from unstructured / semi-structured data such as invoice PDFs and scanned receipts through NLP and OCR technologies.

[0068] Next, the process moves to mapping business elements to voucher accounts. If it is a simple or existing business (such as regular purchase payments), the basic rules of the "Accounting Voucher Template" are directly called to complete the mapping. If a complex scenario such as a vague description or a new business is encountered, taking "purchasing server software for R&D" as an example, the large model will use RAG to search the enterprise accounting policy manual in real time to determine whether the business should be included in "R&D expenses - capitalized expenditures" and update this processing logic to the basic rule base.

[0069] Then, based on the above mapping rules, the corresponding accounting vouchers are automatically generated, binding the business elements with the matching accounting subjects (e.g., the debit side of "R&D expenses - capitalized expenditures" corresponds to the credit side of "bank deposits").

[0070] Finally, the voucher is automatically verified by the review rules engine to complete basic checks such as debit and credit balance and compliance of account usage.

[0071] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the certificate verification method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0072] This application also provides a credential verification device; please refer to [reference needed]. Figure 4 The voucher verification device is applied at the enterprise level, and the voucher verification device includes: Classification module 10 is used to receive business data, classify the business data into scenarios, and obtain business scenarios; The generation module 20 is used to map the business data to accounting subjects based on the business scenario, obtain the subject mapping result, and generate accounting vouchers based on the subject mapping result. The review module 30 is used to determine the business complexity of the accounting voucher based on the business scenario, review the accounting voucher based on the business complexity, and obtain the review result.

[0073] The voucher verification device provided in this application, employing the voucher verification method in the above embodiments, can solve the technical problems of poor business rule adaptability in existing accounting voucher processing schemes and low efficiency and high risk caused by reliance on manual verification. Compared with the prior art, the beneficial effects of the voucher verification device provided in this application are the same as those of the voucher verification method provided in the above embodiments, and other technical features in the voucher verification device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0074] This application provides a credential verification device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the credential verification method in Embodiment 1 above.

[0075] The following is for reference. Figure 5 The diagram illustrates a structural schematic suitable for implementing the credential verification device in the embodiments of this application. The credential verification device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The credential verification device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0076] like Figure 5As shown, the credential verification device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the credential verification device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the credential verification device to communicate wirelessly or wiredly with other devices to exchange data. While the figures show credential verification devices with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0077] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0078] The voucher verification device provided in this application, employing the voucher verification method described in the above embodiments, can solve the technical problems of poor business rule adaptability in existing accounting voucher processing schemes and low efficiency and high risk caused by reliance on manual verification. Compared with the prior art, the beneficial effects of the voucher verification device provided in this application are the same as those of the voucher verification method provided in the above embodiments, and other technical features of this voucher verification device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0079] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0080] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0081] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the credential verification method in the above embodiments.

[0082] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0083] The aforementioned computer-readable storage medium may be included in the credential verification device; or it may exist independently and not be assembled into the credential verification device.

[0084] The aforementioned computer-readable storage medium carries one or more programs. When the aforementioned one or more programs are executed by the voucher review device, the voucher review device: receives business data, classifies the business data into scenarios to obtain business scenarios; maps the business data to accounting subjects based on the business scenarios to obtain subject mapping results, and generates accounting vouchers based on the subject mapping results; determines the business complexity of the accounting vouchers based on the business scenarios, and reviews the accounting vouchers based on the business complexity to obtain review results.

[0085] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0086] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0087] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0088] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., computer programs) for executing the above-described voucher review method. This solves the technical problems of poor business rule adaptability in existing accounting voucher processing schemes and low efficiency and high risk caused by reliance on manual review. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the voucher review method provided in the above embodiments, and will not be repeated here.

[0089] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the credential verification method described above.

[0090] The computer program product provided in this application can solve the technical problems of poor adaptability to business rules in existing accounting voucher processing schemes and low efficiency and high risk caused by reliance on manual review. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the voucher review method provided in the above embodiments, and will not be repeated here.

[0091] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for verifying vouchers, characterized in that, The voucher review method is applied at the enterprise level, and the voucher review method includes: Receive business data, classify the business data into scenarios, and obtain business scenarios; Based on the business scenario, the business data is mapped to accounting subjects to obtain the subject mapping results, and accounting vouchers are generated based on the subject mapping results. The business complexity of the accounting voucher is determined based on the business scenario, and the accounting voucher is reviewed based on the business complexity to obtain the review result.

2. The voucher verification method as described in claim 1, characterized in that, The steps of receiving service data and classifying the service data into scenarios to obtain service scenarios include: The business data is analyzed to obtain the data source, and the business data is classified into structured data, semi-structured data, and unstructured data based on the data source. The structured data is cleaned to obtain cleaned structured data. The semi-structured data is scanned to obtain character scanning results. The unstructured data is labeled with features using a natural language processing model to obtain labeling results. A set of structured business elements is generated based on the cleaned structured data, character scanning results, and annotation results; The business data is classified into scenarios by using the structured business element set to obtain business scenarios.

3. The voucher verification method as described in claim 2, characterized in that, The step of classifying the business data into scenarios using the structured business element set to obtain business scenarios includes: The business attributes are obtained by performing basic attribute judgments on the business data using the structured business element set. The business data is classified by transaction nature based on the structured business element set to obtain the nature classification result; Additional feature recognition is performed on the business data using the structured business element set to obtain feature recognition results; Based on the business attributes, property classification results, and feature recognition results, a multi-dimensional analysis is performed using a business scenario recognition model to obtain the business scenario.

4. The voucher verification method as described in claim 1, characterized in that, The step of mapping accounting subjects to the business data based on the business scenario to obtain the subject mapping result includes: If the business scenario is a simplified scenario, the enterprise's basic accounting rule base is called, and the business data is mapped to accounting subjects through the basic accounting rule base to obtain the subject mapping result; If the business scenario is complex, the business data and the business scenario are combined for contextual understanding to obtain the business behavior. Based on the business behavior, the core knowledge base of the enterprise is called through the retrieval enhancement generation framework to perform accounting subject mapping on the business behavior to obtain the subject mapping result.

5. The voucher verification method as described in claim 1, characterized in that, The step of reviewing the accounting voucher based on the business complexity to obtain the review result includes: When the business complexity is simple, the accounting voucher is queried through the review rule engine to obtain the review rule code; The review rules are extracted using the review rule code, and the accounting vouchers are reviewed using the review rules to obtain the review results.

6. The voucher verification method as described in claim 5, characterized in that, The step of reviewing the accounting voucher based on the business complexity to obtain the review result further includes: When the business complexity is complex, a large language model is used to perform semantic understanding and compliance checks on the accounting vouchers and business data to obtain compliance check results. The business logic rationality is checked by using the associated information corresponding to the accounting voucher and the business scenario to obtain the logic rationality check result; Based on the historical abnormal voucher records of the enterprise, abnormal signals are identified in the accounting vouchers to obtain abnormal identification results; The results of the compliance check, the logical rationality check, and the anomaly identification are reviewed from multiple dimensions to obtain the review results.

7. The voucher verification method as described in claim 4, characterized in that, After the steps of determining the business complexity of the accounting voucher based on the business scenario, reviewing the accounting voucher based on the business complexity, and obtaining the review result, the method further includes: Record the verification items, pass items, and exception items corresponding to the verification results; A review report is generated based on the verification items, pass items, and exception items. The accounting subject basic rule base is updated based on the review report, and the update result is obtained.

8. A voucher verification device, characterized in that, The voucher verification device is used at the enterprise level, and the voucher verification device includes: The classification module is used to receive business data, classify the business data into scenarios, and obtain business scenarios. The generation module is used to map the business data to accounting subjects based on the business scenario, obtain the subject mapping result, and generate accounting vouchers based on the subject mapping result; The review module is used to determine the business complexity of the accounting voucher based on the business scenario, review the accounting voucher based on the business complexity, and obtain the review result.

9. A voucher verification device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the credential verification method as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the credential verification method as described in any one of claims 1 to 7.

Images