Audit early warning management method and system for enterprise expense bills

Abstract

The invention provides an audit early warning management method and system for an enterprise expense bill, and relates to the technical field of enterprise data management, and the method comprises the steps: receiving a target bill, executing workflow arrangement and audit path matching under bill classification, and determining an audit sequence; triggering a prior node of the auditing platform, scanning bill information, disassembling to determine non-key content and key elements, executing mixed-sequence coding encryption, determining a reconstructed target bill and performing auditing circulation, executing authority authentication at an auditing sequence node, if the authentication is successful, performing reduction display and auditing processing on the reconstructed target bill, and determining an auditing result. And performing early warning management on the target bill. The technical problems of low auditing efficiency and poor data security caused by lack of automatic arrangement of the auditing process and insufficient bill sensitive information protection mechanism in the prior art are solved, and the technical effects of improving the intelligent level of the auditing process, enhancing the data security and realizing accurate early warning management are achieved.

Description

Technical Field

[0001] This application relates to the technical field of enterprise data management, and specifically to an audit warning management method and system for enterprise expense bills. Background Art

[0002] With the continuous expansion of the enterprise business scale and expense types, expense bills have shown the characteristics of diversification, batch quantity, and sensitivity. The traditional manual audit method has been difficult to meet the dual requirements of efficiency and risk control. For this reason, some enterprises have gradually introduced a bill audit and warning management system based on an information platform to improve the audit efficiency and reduce human errors.

[0003] The existing enterprise expense bill audit warning methods usually rely on fixed processes for bill classification, manual identification of key fields, hierarchical approval, and triggering of risk warning rules. Such methods lack flexibility in process design, and the audit path adopts static configuration, making it difficult to adapt to the dynamic audit requirements of different bill types. At the same time, a large amount of manual intervention not only prolongs the processing cycle but also increases the risk of data leakage. In addition, the existing solutions are generally weak in the protection mechanism of bill information and cannot effectively encrypt and protect the sensitive information in the bill, resulting in the bill data being easily leaked and tampered with during the audit process, posing a greater information security risk. Summary of the Invention

[0004] This application provides an audit warning management method and system for enterprise expense bills, which solves the technical problems of low audit efficiency and poor data security in the prior art due to the lack of automated scheduling in the audit process, frequent manual intervention, and insufficient protection mechanism for bill sensitive information, and achieves the technical effects of improving the intelligent level of the audit process, enhancing data security, and realizing precise warning management.

[0005] In view of the above problems, on the one hand, this application provides an audit warning management method for enterprise expense bills. The method includes: deploying routing rules in the bill classification engine of the audit platform, receiving a target bill, performing workflow orchestration and audit path matching under bill classification to determine the audit sequence; with the reception of the target bill, triggering the prior node of the audit platform to scan the bill information, disassembling and determining non-critical content and key elements, performing disordered coding encryption to determine the reconstructed target bill; according to the audit sequence, performing audit transfer on the reconstructed target bill, performing permission authentication at the audit sequence node, if the authentication is successful, performing restored display and audit processing on the reconstructed target bill to determine the audit result; according to the audit result, performing warning management on the target bill.

[0006] Preferably, perform scrambled coding encryption to determine the reconstructed target bill, including: for the non-critical content, perform a first encryption process to determine a first data layer; for the key elements, determine element codes, perform an encryption process to determine a second data layer, where the information position in the bill information is used as the coding standard; concatenate the first data layer and the second data layer to determine the reconstructed target bill.

[0007] Preferably, each key element corresponds to an element code; perform scrambling and a second encryption process on the key elements to determine an element data layer; perform a third encryption process on the element codes to determine a coding data layer; determine the second data layer according to the coding data layer and the element data layer.

[0008] Preferably, the bill classification includes a bill attribute dimension and a bill level dimension; according to the bill attribute dimension and the bill level dimension, perform an encryption execution determination and a scrambling execution determination based on non-critical content and key elements.

[0009] Preferably, perform workflow orchestration under bill classification, including: receiving an external audit instruction, where the external audit instruction is a personalized audit requirement; determining an audit workflow by performing bill classification; adjusting the audit workflow according to the external audit instruction, performing audit orchestration based on a time series to determine an audit workflow.

[0010] Preferably, perform audit path matching to determine an audit sequence, including: matching and determining audit nodes according to the audit workflow; taking the interaction based on the audit nodes as the guide to determine an interaction path, where the optimal interaction thread under multiple interaction threads between nodes is used as the standard; determining the audit sequence according to the audit nodes and the interaction path.

[0011] Preferably, perform permission authentication at the audit sequence nodes. If the authentication is successful, perform restoration display and audit processing on the reconstructed target bill to determine an audit result, including: transferring the reconstructed target bill to the first node of the audit sequence for node audit permission authentication to determine an authentication result; if the authentication result is successful, perform key decryption processing on the first data layer, the element data layer, and the coding data layer, perform scrambling restoration on the element data layer according to the coding data layer to determine non-critical content and key content; perform targeted audit processing on the non-critical content and the key content to determine the audit result of the first node.

[0012] Preferably, after determining the review result of the first node, it includes: for the review sequence, completing the progressive review of nodes, determining the review results of multiple nodes; summarizing the review results of the multiple nodes, determining the comprehensive review result, and integrating the review record chain; using the comprehensive review result and the review record chain as the review result.

[0013] Preferably, for early warning management of the target bill, it includes: traversing the review record chain to determine chain early warning information; determining comprehensive early warning information according to the comprehensive review result; and performing early warning management on the target bill according to the chain early warning information and the comprehensive early warning information.

[0014] On the other hand, the present application also provides an audit and early warning management system for enterprise expense bills. The system includes: an audit path matching module, which is used to deploy routing rules in the bill classification engine of the audit platform, receive the target bill, perform workflow orchestration and audit path matching under bill classification, and determine the audit sequence; an information reconstruction module, which is used to trigger the prior node of the audit platform with the receipt of the target bill, perform bill information scanning, disassemble and determine non-critical content and key elements, and perform disordered coding encryption to determine the reconstructed target bill; an audit processing module, which is used to perform audit transfer on the reconstructed target bill according to the audit sequence, perform permission authentication at the audit sequence node, and if the authentication is successful, perform restored display and audit processing on the reconstructed target bill to determine the audit result; and an early warning management module, which is used to perform early warning management on the target bill according to the audit result.

[0015] One or more technical solutions provided in the present application have at least the following beneficial effects: By deploying routing rules in the bill classification engine of the audit platform, receiving target bills, performing workflow orchestration and audit path matching under bill classification to determine the audit sequence, the automatic classification of bills from different sources and types and the audit path planning are realized, laying a foundation for the subsequent efficient and orderly audit process, and improving the automation degree and efficiency of the audit process. With the receipt of the target bill, the prior node of the audit platform is triggered to scan the bill information, disassemble and determine the non-critical content and key elements, perform disordered coding encryption, and determine the reconstructed target bill. While ensuring the integrity of the bill information, the key elements of the bill are effectively protected, the data security is enhanced, and the risk of information leakage or tampering of the bill in the audit link is prevented, providing a safe and reliable bill data basis for the subsequent audit work. According to the audit sequence, the reconstructed target bill is audited and transferred. Permission authentication is performed at the audit sequence node. If the authentication is successful, the reconstructed target bill is restored and displayed and audited to determine the audit result, realizing the hierarchical permission audit of the bill, ensuring that sensitive information is only viewed and processed at the authorized node, and at the same time ensuring the strict compliance of the audit process. According to the audit result, early warning management is carried out on the target bill, and the audit processing and early warning response are integrated in a closed loop, ensuring that high-risk or abnormal bills enter the early warning process in a timely manner, and improving the scientificity and effectiveness of the enterprise expense bill audit management.

[0016] In summary, this application realizes the automation and intelligence of the enterprise expense bill audit process by introducing a dynamic orchestration mechanism for bill classification and audit path; constructs a strict sensitive information protection system by disassembling and encrypting bill data through prior nodes; combines permission authentication and content restoration mechanisms in the multi-level audit process to effectively control the information access boundary; and finally realizes the linkage closed loop of audit and risk control through the early warning management driven by the audit result. The overall solution significantly improves the processing efficiency and accuracy of expense audit, enhances the data security guarantee ability, and at the same time constructs a rapid-response and multi-scenario adaptable expense early warning management system.

[0017] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the specific embodiments of this application are specifically given below. Brief Description of the Drawings

[0018] Figure 1 It is a schematic flowchart of the audit early warning management method for enterprise expense bills provided by the embodiment of this application.

[0019] Figure 2Schematic flowchart of performing scrambled coding encryption and determining the reconstructed target bill in the audit warning management method for enterprise expense bills provided by the embodiments of the present application.

[0020] Figure 3 Schematic structural diagram of the audit warning management system for enterprise expense bills provided by the embodiments of the present application.

[0021] Explanation of reference numerals: Audit path matching module 10, information reconstruction module 20, audit processing module 30, warning management module 40. Detailed implementation manners

[0022] The embodiments of the present application provide an audit warning management method and system for enterprise expense bills, solving the technical problems in the prior art that due to the lack of automated arrangement in the audit process, frequent manual intervention, and insufficient protection mechanism for bill sensitive information, the audit efficiency is low and the data security is poor, achieving the technical effects of improving the intelligent level of the audit process, enhancing data security, and realizing accurate warning management.

[0023] Embodiment 1, as Figure 1 shown, the embodiments of the present application provide an audit warning management method for enterprise expense bills, and the method includes: Step S100: Deploy routing rules in the bill classification engine of the audit platform, receive the target bill, perform workflow arrangement and audit path matching under bill classification, and determine the audit sequence.

[0024] Specifically, the audit platform is an integrated system for receiving, classifying, auditing, and warning management of enterprise expense bills. The bill classification engine is a functional module in the audit platform that realizes intelligent classification of bills according to attributes, types, etc. The routing rules are a set of rules for defining the audit processes that different types of bills should enter. The audit sequence is the specific execution order of each node in the audit path.

[0025] The routing rules are preset in the bill classification engine of the audit platform. When receiving an enterprise expense bill (target bill), the audit platform first analyzes and classifies the bill through the bill classification engine, identifies the category of the bill (such as travel expenses, procurement expenses, etc.) according to the preset routing rules, and matches the corresponding audit process. Subsequently, according to the classification result, workflow arrangement is performed to automatically generate an audit path including information such as audit nodes, processing roles, and time sequences. Then, through the path matching mechanism, the optimal audit order between each node is identified, and finally a structured audit sequence is output to drive the subsequent audit process.

[0026] In a feasible embodiment provided by this application, the configuration of routing rules is defined by combining attribute matching and logical conditions. For different types of enterprise expense bills, multiple rules are preset in the bill classification engine of the audit platform. Each rule includes conditions such as bill type, amount range, time range, required attachments, etc., as well as the corresponding workflow identifier. For example, for travel expense bills, the matching conditions can be set as the bill type being travel expenses, the amount being between 0 and 20,000 yuan, and it must be accompanied by a travel itinerary and invoice. After successful matching, it enters the specified travel expense audit process. All rules can be dynamically managed and hot-updated through a database or a rule engine to ensure matching accuracy and flexibility.

[0027] Furthermore, when the target bill arrives at the audit platform, the bill classification engine will parse the bill content, quickly match the bill attributes according to the above-mentioned pre-configured routing rules, and determine its category and the corresponding audit process. The matching result is used to call the workflow orchestration engine to generate a specific audit path. Exemplarily, the workflow orchestration adopted can select open-source engines such as Camunda or Activiti, or use a self-developed data structure based on a directed acyclic graph for node orchestration and flow control. The workflow definition includes information such as the roles to be executed by each audit node, the sequence order between nodes, and the completion time limit of each node, ensuring the controllability and traceability of the process execution. Exemplarily, after determining the above-mentioned travel expense audit process, it is deconstructed with a data structure of a directed acyclic graph to determine each audit node and calibrate the element information to be executed by it, that is, to determine multiple node tasks and task cascade relationships.

[0028] Furthermore, after the workflow is generated, the execution order of nodes is further optimized according to the path matching algorithm. The main considerations include approval timeliness, personnel idle time, hierarchical order, and cross-departmental approval costs, etc. Exemplarily, weight calculations are performed on all feasible paths, the path with the lowest total cost is searched from the directed graph, and the optimal node execution sequence is output as the final audit sequence. This sequence is used to drive the subsequent automated audit flow.

[0029] To ensure integration with the enterprise's internal ERP or financial system, the platform provides RESTful API interfaces. The bill data is interacted in JSON or XML format. The interfaces include bill reception, status query, and callback notification after approval completion, etc. It supports OAuth2.0 or JWT authentication and is encrypted and transmitted through HTTPS. At the same time, an IP white list and signature verification mechanism are provided to ensure the security and consistency of data interaction.

[0030] This step realizes the automatic initialization of the enterprise bill review process, ensures the dynamic adaptation of the bill type to the review process, effectively improves the flexibility and intelligence level of the process configuration, and significantly reduces the efficiency bottleneck and error risk brought by manual distribution and configuration.

[0031] Step S200: With the receipt of the target bill, trigger the prior node of the review platform, scan the bill information, disassemble and determine the non-critical content and key elements, perform disordered encoding encryption, and determine the reconstructed target bill.

[0032] Specifically, the prior node is a pre-set preprocessing node in the review process for data preprocessing and content security enhancement. With the receipt of the target bill, the prior node of the review platform is triggered to comprehensively scan and parse all fields in the bill. First, identify which fields belong to non-critical content and which belong to key elements, and process them separately. The non-critical content is simply masked by the first encryption method, while the key elements are first processed by field number encoding, then a higher-strength encryption algorithm is executed, and their order is scrambled (disordered processing) to form a second data layer. Finally, the two data layers are cascaded to construct a fully encrypted and structure-controlled reconstructed target bill. This reconstructed target bill contains the complete bill data body after differential encryption processing and serves as the core data object in the review process. For example, if the bill contains "Travel expenses of 20,000 yuan submitted by Zhang San, invoice number 001XYZ", "Zhang San", "Travel expenses", etc. are encrypted as fuzzy labels as non-critical content, and "20,000 yuan" and "001XYZ" are scrambled and encrypted using AES as key elements to generate encrypted information, which is then cascaded with the non-critical content to finally form the reconstructed target bill.

[0033] In a feasible embodiment provided by the present application, to ensure the executability and security of the bill information preprocessing, the classification of fields is set based on the general sensitive information standard of enterprise expense management and actual business requirements. Among them, key elements usually refer to fields that have a direct impact on the review result and involve sensitive financial data, mainly including amount fields (such as individual amount, total amount, tax amount), bank account numbers, payee and payer account information, invoice numbers, tax numbers, contract numbers, etc., which are directly used for fund flow verification or compliance verification; non-critical content refers to information with a lower correlation with the accounting result or only used for general identity or background description, such as the name of the reimburser, department, application reason, travel destination, etc. This part of the content can be partially masked or marked by means of fuzzification without affecting the correctness of the process.

[0034] Preferably, the field classification rule library can be deployed according to the above classification method, and its optional configuration method is: maintained in the form of a configuration table inside the bill classification engine, and can be dynamically expanded and version-managed in combination with industry compliance requirements or user-defined rules.

[0035] Among them, non-critical content usually adopts simple hashing or masking methods with low computational cost and irreversible results. For example, use SHA-1 or SHA-256 hashing to convert fields such as names into irreversible values, or directly use character masks (such as asterisk masks) to achieve blurred display, meeting the requirements of basic information masking and ensuring efficient processing; key elements are protected by high-strength encryption algorithms that comply with international standards. For example, select AES-256-GCM to perform symmetric encryption on amounts, invoice numbers, bank account numbers, etc. If there is a need for secure cross-system transmission, asymmetric encryption such as RSA-OAEP can be further used to encapsulate some fields. Optionally, the encryption key is uniformly generated and rotated by the platform key management service (KMS) to ensure that the key lifecycle and compliance requirements meet the enterprise security policy.

[0036] Furthermore, the scrambling logic proposed in this application for performing scrambled encoding is based on ensuring that the order of the encrypted key elements cannot be directly restored during transmission. Specifically, a hash value generated using the bill's unique ID or combined with the bill submission timestamp is used as a random seed, and a pseudo-random permutation method (such as Fisher-Yates Shuffle) is called to generate a random permutation index for the keyword field block, and the physical storage order of the encrypted fields is rearranged through this index. The receiving end can perform reverse restoration according to the recorded scrambling index when it has the corresponding key and seed, ensuring data integrity and decryptability. The finally formed reconstructed target bill is the sequential concatenation of the first data layer after hashing or masking of non-critical content and the second data layer after strong encryption and scrambling of key elements, becoming a unified encrypted data object in the subsequent review process, which not only ensures the confidentiality of the data but also facilitates on-demand decryption and verification at each approval node later.

[0037] This step effectively prevents sensitive information from being accessed before authorization by differentiating the bill content, achieving secure isolation during data transmission and storage, and enhancing the information confidentiality and compliance of the bill during the circulation process.

[0038] Step S300: According to the review sequence, conduct review circulation on the reconstructed target bill, perform permission authentication at the review sequence nodes. If the authentication is successful, perform restored display and review processing on the reconstructed target bill to determine the review result.

[0039] Specifically, the reconstructed target bill will sequentially flow to each predetermined audit node during the audit process. At each node, first, permission authentication is performed, that is, by means of user identity verification, role permission verification, etc., to determine whether the current node has the right to access the bill content. If the authentication is successful, the node private key is called to decrypt and restore the encrypted fields related to this node in the reconstructed bill, and it is presented to the auditor in a visual manner. Subsequently, the auditor or the audit platform checks, annotates, or approves the restored bill according to business rules to generate the audit result of this node. The bill then enters the next node until the audit process is completed, obtaining the final audit result, that is, the processing conclusions formed by each audit node on the bill.

[0040] In a feasible embodiment provided by the present application, to ensure the controllable decryption and secure access of the reconstructed target bill during the audit process, for each node in each audit sequence, the platform will pre-generate and allocate a unique key pair for each role or node during the initial deployment phase. The key management is entrusted to the hardware security module (HSM) inside the enterprise or in the cloud to ensure that the private key is not transmitted in plain text over the network. Alternatively, according to actual requirements, a key distribution service can be used to dynamically issue access tokens for the corresponding private key when the node task is triggered.

[0041] Preferably, for scenarios that require higher flexibility, a dynamic session key generation mode can also be adopted, that is, after the user successfully passes the identity verification, the authentication service issues a JWT token carrying a key index or temporary decryption permission based on the user's permissions. The validity period is bound to a single audit task to avoid the leakage risk brought by long-term valid keys.

[0042] Among them, in terms of field-level permission control, the platform defines the specific range of bill fields that each audit node can access by maintaining a role-field mapping table. This mapping table is configured by the system administrator or process designer. For example: the fields that the department manager role can view include amount and reimbursement reason; the finance role can view amount, invoice number, and bank account number; the compliance role can view contract number and payment and receipt accounts. The field access permissions are filtered by the decryption module according to the role mapping table to call the corresponding private key during bill decryption, ensuring that the node can only perform decryption operations on the authorized fields. If a node has no permission for a field, the field remains encrypted or is displayed in a blurred marked manner.

[0043] Further, to address the possible conflict issues in the review results among multiple nodes, the platform is based on the built-in conflict handling and arbitration mechanism: when there is a logical contradiction between the processing conclusion of a subsequent node and that of a previous node (for example, a certain expense is judged compliant by a previous node but rejected by a subsequent node), the preset conflict arbitration rules will be triggered. Common strategies include overriding the results of lower-priority nodes with the opinions of higher-priority nodes according to the priority ranking, or determining the final conclusion by voting among multiple nodes in the same role group according to the majority voting method. When necessary, it also supports automatically activating the superior review node for manual intervention and arbitration. The arbitration process and conflict status will be recorded in real time and appended to the process audit log to ensure the integrity and traceability of the responsibility chain.

[0044] In summary, through the key management, field mapping, and conflict arbitration mechanisms, the process transparency and information security are balanced in the node-by-node decryption, hierarchical display, and review processing of the entire reconstructed target bill, supporting the enterprise's controllable review and risk control of sensitive financial information.

[0045] This step realizes the decentralized display and targeted processing of the bill during the review process, ensuring the minimum exposure of sensitive information, while improving the compliance and efficiency of the review, and effectively preventing unauthorized access and data leakage.

[0046] Step S400: Perform early warning management on the target bill according to the review result.

[0047] Specifically, when the bill completes the processing of all review nodes and generates the final review result, the early warning management module 40 is called to analyze and discriminate the review result, compare whether there is high-risk content in the result (such as rejection, inconsistent opinions of multiple nodes, abnormal amount, etc.), and combine the risk characteristics formed in the historical review record chain to determine whether to trigger an early warning event. If the early warning is triggered, an early warning log will be generated, an early warning notice will be sent, and the bill will be marked in the risk control database for subsequent review and tracking processing. For example, if a bill is rejected by the "department head" but passed by the "financial review", it is judged as a disagreement among nodes. Considering that similar bills have had illegal invoicing in the past, a red early warning is triggered and automatically pushed to the audit department for processing.

[0048] In a feasible embodiment provided by the present application, for the final review result generated after completing the processing of all review nodes, it is comprehensively determined whether to trigger a risk early warning based on multi-dimensional indicators, and different responses are implemented according to the risk level. The early warning classification standard is set to be quantitatively distinguished based on the risk score threshold.

[0049] In the specific implementation process, the risk score is calculated based on factors such as the comprehensive node rejection rate, the consistency of opinions between nodes, the degree of amount fluctuation, and the matching degree with historical similar abnormal events. The specific setting is as follows: If the calculated risk score is greater than or equal to 0.8, it is determined as a high-risk event and marked with a red warning, and the audit or risk control personnel need to intervene immediately for verification; if the risk score is between 0.4 and 0.8, it is determined as a medium-risk event, marked with a yellow warning, and the relevant financial or compliance specialists can conduct a secondary review; if it is lower than 0.4, it is defaulted to pass without triggering a warning, and the bill can enter the normal filing or payment process.

[0050] At the same time, in the calculation process of the risk score, the historical audit record chain will be called as the comparison basis, and a time decay mechanism will be introduced for historical data to ensure greater sensitivity to recent anomalies. In a feasible setting method, the time decay factor is set with a three-month cycle. The weight of the same type of records three months ago is automatically reduced by 50%, and so on, to ensure that the impact of old records on the current risk score is limited and will not cause lagging misjudgment. This weight adjustment continuously optimizes the risk discrimination accuracy by dynamically fine-tuning the feature weights.

[0051] Furthermore, when it is judged that a warning is triggered, structured warning content will be generated. Its specific format includes the warning number, the bill number, the trigger time, the warning level (red or yellow), the main trigger reasons (such as inconsistent node opinions, the percentage of abnormal amount amplitude, suspicious supplier identifiers, etc.), the information of relevant responsible persons, and the subsequent recommended handling measures. Then, the warning information will be immediately written into the warning log, pushed to the message center or email system of the designated audit or risk management personnel, and the relevant bills will be recorded in the risk control database to facilitate subsequent review and trend analysis of repetitive or patterned violations, thus forming a closed-loop bill risk control management system.

[0052] Preferably, the entire warning process is centered on automated processing, and an artificial review and intervention entry is reserved to achieve real-time monitoring and response to financial compliance risks.

[0053] This step integrates the results of the review process with the risk control mechanism in a closed loop, realizes the automatic identification and warning response of review anomalies, significantly enhances the proactive defense ability of enterprise expense control, and improves the scientificity and effectiveness of enterprise expense management.

[0054] Furthermore, as Figure 2 shown, step S200 includes: Step S210: For the non-critical content, perform the first encryption process to determine the first data layer.

[0055] Step S220: For the said key elements, determine element codes, perform encryption processing, and determine the second data layer, where the information position in the bill information is used as the coding standard.

[0056] Step S230: Concatenate the first data layer and the second data layer to determine the reconstructed target bill.

[0057] Specifically, non-critical content refers to information fields with relatively low risks to bill security or privacy, such as applicant name, remarks, purpose, etc. The first encryption processing refers to technical methods used to perform low-intensity, fast encryption or obfuscation processing on non-sensitive fields. The first data layer is a data set formed by all non-keyword fields after the first encryption processing, with an independent and recognizable structure. After identifying non-critical content fields, lightweight encryption algorithms (such as symmetric encryption, masking processing, etc.) are called to encrypt these fields. This processing gives priority to ensuring the reducibility and structural integrity of the fields, while keeping the processing speed and resource consumption low. The processed fields are assembled into the first data layer, retaining the relative structural relationship between the fields, in preparation for subsequent concatenation.

[0058] Key elements include highly sensitive fields such as the amount involved in the bill, invoice number, account information, etc. The element code is a coding identifier set for the key elements according to their positions in the bill information. The second data layer is a data set formed by performing positioning coding and high-intensity encryption on all key elements. For each key element field, record its information position in the original bill structure (such as field row and column numbers or label paths), and use this as its unique coding identifier (element code). For example, if the key element "amount of 20,000 yuan" is in the 3rd row and 2nd column of the bill, then the element code can be set as R3C2. Subsequently, high-security encryption (such as AES-256, RSA, etc.) is performed on this field, and the encryption result is bound to the corresponding code. All processed fields are assembled to form the second data layer.

[0059] After the construction of the first data layer and the second data layer is completed, they are fused according to the original bill field order or logical structure requirements. This process includes steps such as field merging, embedding of structure positioning identifiers, and data layer annotation, ensuring that the reconstructed bill can achieve secure encryption while maintaining field decoupling and permission control. The finally generated reconstructed target bill will have characteristics of being structured, encrypted, and auditable and transferable, ensuring both the integrity and security of the bill information while retaining the structure of the original bill.

[0060] Furthermore, each key element in step S220 corresponds to an element code; the key elements are shuffled and second encryption processing is performed to determine the element data layer; third encryption processing is performed on the element code to determine the coding data layer; according to the coding data layer and the element data layer, the second data layer is determined.

[0061] Specifically, the third encryption process is an additional encryption process performed on the field position code, so that the field structure information cannot be directly parsed. During the bill parsing stage, the structure position information of each key element field identified is recorded, and a unique corresponding element code is generated. This code will serve as an important index for subsequent data binding and tracking.

[0062] In the process of determining the second data layer, the key elements of the determined element codes are first mixed and the order of the key element fields is disrupted (based on random seeds, user policies or rule bases), and then each field value is highly encrypted to form a mixed encrypted field set, namely the element data layer. This data layer does not contain structural position information, and the field order is no longer consistent with the original structure of the bill, further improving security. Next, the aforementioned generated element codes (such as "R3C4" and "R5C2") are asymmetric or hashed (third encryption processing) to generate a set of coded ciphertexts (i.e., the coded data layer), which corresponds to each field in the element data layer. Finally, the generated element data layer and the coded data layer are reassembled into a key-value structure according to the original binding relationship, i.e., the second data layer. This structure retains the field and position relationship, but hides the specific meaning and structure through encryption to achieve dual protection of key fields and structural information, avoid overall risks caused by single-point information exposure, and ensure the encryption integrity and controllable and reducible nature of data during the circulation process.

[0063] Furthermore, the bill classification includes a bill attribute dimension and a bill level dimension; according to the bill attribute dimension and the bill level dimension, encryption execution judgment and mixed order execution judgment based on non-critical content and critical elements are performed.

[0064] Specifically, the bill attribute dimension refers to the function or type attribute of the bill in the business system, such as "procurement", "travel", "project", etc., which is used to determine the field distribution pattern and business processing sensitivity. The bill level dimension refers to the sensitivity level or security level of the bill, such as "normal", "restricted", "confidential", which is automatically determined based on parameters such as amount, source unit, and responsible entity.

[0065] When classifying bills, it is necessary to comprehensively consider the attribute dimension and level dimension of the bills. The bill classification engine first extracts information such as structural tags, keywords, field formats, and business codes from the received bill metadata, and automatically determines the attributes and levels of the bills. For example, if the field tag contains "flight number, accommodation expenses", the attribute is "travel category"; if the amount is greater than 500,000 yuan, or the approval process spans multiple business units, the level is "confidential level".

[0066] Based on the above bill classification determination results, different encryption levels and scrambling processes are selected in step S200. For the encryption execution determination, if the bill belongs to a business type with high sensitivity in terms of attribute dimension or has a high-risk level in terms of level dimension, then a high-strength encryption algorithm is used to encrypt the key elements, and non-critical content is also encrypted. For the scrambling execution determination, if the bill has a high-risk level and the key elements are sensitive, then the scrambling operation is performed, and the scrambling rules and methods are determined according to the specific attributes of the bill. For example, if the bill is "project type - confidential level", then non-keyword fields also need to be strongly encrypted, and all fields need to be fully scrambled; if the bill is "procurement type - restricted level", then non-critical and key elements are encrypted, and partial scrambling (field segmentation and disordering) is performed; if the bill is "travel type - ordinary level", then only the key elements are encrypted, and no scrambling is performed.

[0067] Making encryption execution determination and scrambling execution determination according to the bill attribute dimension and bill level dimension can make the security processing of bills more flexible and accurate, adjust the security policy according to the actual risks and characteristics of the bills, improve the processing efficiency while ensuring the security of bill information, and avoid excessive or insufficient security measures.

[0068] Furthermore, step S100 includes: Step S110: Receive an external audit instruction, where the external audit instruction is a personalized audit requirement.

[0069] Step S120: Determine the audit workflow by classifying the bills.

[0070] Step S130: Adjust the audit workflow according to the external audit instruction, perform audit scheduling based on time series, and determine the audit workflow.

[0071] Specifically, the external audit instruction refers to a personalized audit task request from a third-party platform such as an audit system or an ERP system. Connect to the enterprise OA system or ERP platform, and listen for and receive the external audit instruction sent by the third-party platform through the interface service. This instruction can include audit priority setting, audit role limitation, audit order preference, urgent processing requirements, etc. After the audit platform analyzes these requirements, it uses them as an input template or adjustment parameter for workflow scheduling to construct a personalized audit process. For example, an instruction content is: "For records in project type bills with an amount exceeding 200,000 yuan, a legal review node must be added." After receiving this instruction, the audit platform will automatically insert the "legal node" into the preset workflow under the corresponding bill classification and amount conditions. By receiving and analyzing external personalized audit instructions, it is possible to dynamically respond to business changes and meet differentiated process requirements.

[0072] The audit workflow is a sequence of nodes required to complete the bill audit, including approval roles, node rules, transfer paths, etc. After receiving the target bill, the audit platform uses the bill classification engine to identify and classify the target bill in terms of attribute dimension and level dimension. For example, classifying the "project expenditure type" bill as the "project type". Subsequently, according to the pre-configured standard audit template library of the enterprise, a preset audit workflow is selected for this type of bill, such as "project leader - budget approval - financial audit".

[0073] After selecting the initial workflow, the original workflow structure is dynamically adjusted according to the received external audit instructions. For example, if the audit task needs to be completed within 48 hours, the approval authority can be decentralized or the parallel nodes can be optimized to reduce the time overhead. At the same time, according to information such as the idle time of the auditors and the holiday schedule, the node order is dynamically rearranged, and finally an optimized audit workflow with an optimized time sequence is generated. Introducing the time sequence scheduling mechanism can achieve the optimal control of the process time, improve the overall response speed and execution efficiency on the premise of ensuring the audit accuracy.

[0074] Furthermore, step S100 further includes: Step S140: Determine the audit nodes according to the audit workflow.

[0075] Step S150: Orienting to the interaction based on the audit nodes, determine the interaction path, where the optimal interaction thread under multiple interaction threads between nodes is used as the standard.

[0076] Step S160: Determine the audit sequence according to the audit nodes and the interaction path.

[0077] Specifically, the audit node represents each role unit with independent approval authority in the audit process, such as the financial supervisor, project leader, auditor, etc. According to the determined audit workflow template, analyze the responsibilities, roles and permission requirements of each level of process nodes, then query the permission database and the organizational structure model, match the user accounts or organizations with such responsibilities, and bind them as audit nodes to realize the transformation from the template role to the actual user. If the shift or delegation processing mechanism is supported, the node assignment can also be combined with the available personnel in the current time period. For example, if a certain node in the workflow is "preliminary review by the finance department", then find the valid personnel under the current "financial preliminary reviewer" role in the organizational structure, such as "User A", and bind it to this node.

[0078] The review platform analyzes all available interaction paths between each pair of adjacent review nodes, including local area network communication paths, VPN connection channels, mobile terminal interfaces, instant messaging links, etc. For each interaction path, the following performance parameters are collected: communication latency (determined by the physical length of the path, network topology, and number of hops), thread load (i.e., the number of currently active bill review threads on this path), channel availability (i.e., the idle bandwidth or thread idle degree of the path during the current time period), and node response ability (including dynamic behavior data such as the online status of the reviewer and historical response times). The platform constructs a multi-parameter weighted evaluation model, scores each performance parameter according to the following example weights: communication latency - 30%; thread load - 30%; channel availability - 20%; node response ability - 20%, calculates the weighted total score for each path, denoted as the interaction efficiency score, and then selects the path with the highest score from all available interaction paths as the optimal interaction thread path for bill transfer between nodes. This evaluation mechanism can be continuously learned and optimized in combination with an asynchronous thread scheduler and historical path performance logs. Among them, the weights can be customized and adjusted by users according to the actual execution environment.

[0079] Sort the previously generated review nodes according to the workflow logic and map them in combination with the selected optimal interaction path, so as to generate an audit execution sequence with a communication path identifier. This sequence is used to guide the scheduling execution and communication routing of bills during the actual review process, so as to ensure the stability and efficiency of the entire approval process.

[0080] Furthermore, step S300 includes: Step S310: Transfer the reconstructed target bill to the first node of the review sequence for node audit permission authentication and determine the authentication result.

[0081] Step S320: If the authentication result is successful authentication, perform key decryption processing on the first data layer, element data layer, and coding data layer, and restore the element data layer in a scrambled order according to the coding data layer to determine non-critical content and critical content.

[0082] Step S330: Perform targeted review processing on the non-critical content and the critical content to determine the review result of the first node.

[0083] Specifically, according to the review sequence, the reconstructed target bill is pushed to the first review node, and at the same time, the user credential information of this node (such as digital signature, role identifier, authentication token, etc.) is read, and it is verified whether it has the permission to read the data layers (first data layer, element data layer, coding data layer) in the bill, and an authentication result of successful authentication or failed authentication is generated.

[0084] If the authentication result is successful, use the set of keys bound to the current node's permissions to perform decryption operations on each data layer of the bill. The decryption process includes: decrypting the first data layer with the first key to obtain the original non-critical content; decrypting the element data layer and the encoded data layer with the second key, and reordering the content of the decrypted element data layer according to the position index contained in the decrypted encoded data layer to restore the original critical element content.

[0085] According to the node function configuration, call the corresponding audit logic module to perform compliance and integrity checks on the non-critical content (such as invoice number legality, field filling), and perform targeted verification on the critical content (such as budget matching, authorization limit, expense details comparison); and interactively obtain the feedback information of the node auditors. After the audit is completed, generate the audit result of the first node and return it to the audit platform. The audit result includes statuses such as passed, returned, suspended, etc., and is accompanied by an audit note and a change suggestion.

[0086] Furthermore, after determining the audit result of the first node, it also includes: Step S340: For the audit sequence, complete the progressive transfer audit of the nodes to determine the audit results of multiple nodes.

[0087] Step S350: Summarize the audit results of the multiple nodes to determine the comprehensive audit result and integrate the audit record chain.

[0088] Step S360: Use the comprehensive audit result and the audit record chain as the audit result.

[0089] Specifically, the progressive transfer audit of nodes means that after the audit of the previous node of the bill is completed, it is sequentially transferred to the next node for continuous audit according to the predetermined audit sequence. According to the determined audit sequence, the bill is transferred one by one from the first node to the back. After each node receives the decrypted and restored bill, it performs the permission authentication and targeted audit processing required by this node. Each node only processes the fields within its permission range, and forms the audit sub-result of this node after the audit, which is recorded in the audit log, realizing the hierarchical processing and responsibility sharing of the audit task, improving the modularity and transparency of the audit link, and helping to trace errors and determine responsibility attribution.

[0090] After all nodes have completed the review, summarize the review results submitted by each node: If the review results of all nodes are "passed", generate a comprehensive "passed" result; if any node is marked as "rejected" or "pending", set the comprehensive review result to the corresponding status; at the same time, organize the processing information of each node (including the operator, time, review fields and results) to generate a chained record file, that is, the review record chain, and attach the review record chain to the full-process operation log of the bill, providing a complete traceable and verifiable review history, which helps with post-event auditing, dispute clarification and responsibility confirmation, and at the same time improves the credibility of the review and regulatory compliance. Among them, the review record chain is a sequence of review logs composed of the results of multiple review nodes in time and sequence relationships. During the review process, a blockchain structure or a hash pointer chain structure can be used to solidify the review records to ensure the timestamp and immutability of each operation. For example, each review node generates a review record chain, and uses the hash value of the previous node as the chain head.

[0091] Package the above-mentioned comprehensive review result and the review record chain together into the final "review result" structure, mark a unique review task identifier (such as the review ID), and synchronize it to the bill management system, the warning system and the audit platform for subsequent processing, realizing a unified and structured review output mechanism.

[0092] Furthermore, step S400 includes: Step S410: Traverse the review record chain to determine the chain warning information.

[0093] Step S420: Determine the comprehensive warning information according to the comprehensive review result.

[0094] Step S430: Perform warning management on the target bill according to the chain warning information and the comprehensive warning information.

[0095] Specifically, analyze each node in the review record chain one by one to determine whether there are the following abnormal patterns: the review status is "rejected" or "pending"; the review processing time is abnormal (exceeding the normal processing time limit); the review fields are abnormal (such as amount deviation, missing key fields); the review operator's permissions are abnormal, etc. If any of the above situations exists, mark the node as a warning node, record the abnormal content, form a set of chain warning information, and record the location of the abnormal node, the type of abnormality, the abnormal field and the triggering conditions, etc.

[0096] Execute the following judgment logic on the comprehensive review result: If the review result is "not passed" or contains the result of a "high-risk" node, trigger a comprehensive warning; if there are conflicts in the results of multiple nodes (such as inconsistent field judgments) or the failure rate of key decryption between nodes increases abnormally, mark it as a comprehensive consistency abnormality, and finally form comprehensive warning information covering the overall perspective of the bill review.

[0097] Implement early warning management for target bills by combining chain early warning information and comprehensive early warning information, including sending a detailed early warning report to relevant responsible persons, initiating an internal investigation process, or making special marks on the bills for subsequent tracking. The early warning report will contain key information such as the location of abnormal nodes, abnormal characteristics, comprehensive evaluation results, etc., to ensure that relevant personnel can quickly understand the problem and take measures. Exemplarily, if the chain early warning information shows that there is an abnormality at a specific review node, such as a review failure, and the comprehensive early warning information also indicates that there are problems with the overall review, then the transfer of the bill can be suspended, and relevant review personnel (such as the personnel responsible for reviewing this node and the review personnel of subsequent nodes) can be notified to re-review or correct the bill. If the chain early warning information indicates that there is a situation of review timeout at a certain node, it is judged according to the comprehensive early warning information whether the overall review progress is affected. If so, relevant personnel can be notified to speed up the review or adjust the review process. By implementing early warning management based on chain early warning information and comprehensive early warning information, problems occurring during the review process of target bills can be processed in a timely manner, ensuring the smooth progress of the review process and improving the quality and efficiency of bill review.

[0098] In summary, the method for audit early warning management of enterprise expense bills provided by the embodiments of the present application has the following beneficial effects: By deploying a bill classification engine and routing rules in the audit platform, combining the bill attribute dimension and the level dimension, performing personalized audit workflow orchestration and path matching, and generating the optimal audit sequence; after receiving the bill, first disassemble the bill information, encrypt the non-critical content and key elements respectively, where the key elements form a two-layer encryption structure through element coding and scrambling encryption, and finally concatenate them into a reconstructed bill to achieve the secure isolation and controllable restoration of bill data; subsequently, perform permission authentication and hierarchical decryption on the reconstructed bill according to the audit sequence, perform node-level audits on the non-critical content and key elements and transfer them step by step, and form a comprehensive audit result and a complete audit record chain after completing multi-node audits; finally, through traversing and analyzing the abnormal nodes and comprehensive results of the audit record chain, extract chain early warning information and comprehensive early warning information, and then perform early warning management operations such as bill freezing, marking, notification, and review.

[0099] Overall, the embodiments of the present application significantly improve the intelligent level of audit transfer while ensuring the security of bill data and controllability of audits, and achieve efficient identification and precise early warning management of abnormal bills through the linkage of chain-level and global information, effectively improving the compliance, transparency, and risk prevention and control levels of enterprise expense management.

[0100] Embodiment 2, as Figure 3 shown, based on the same inventive concept as the foregoing Embodiment 1, the embodiments of the present application provide an audit early warning management system for enterprise expense bills, and the system includes: The audit path matching module 10 is used to deploy routing rules in the bill classification engine of the audit platform, receive target bills, perform workflow orchestration and audit path matching under bill classification, and determine the audit sequence.

[0101] The information reconstruction module 20 is used to trigger the prior nodes of the audit platform with the receipt of the target bill, perform bill information scanning, disassemble and determine non-critical content and key elements, perform disordered coding encryption, and determine the reconstructed target bill.

[0102] The audit processing module 30 is used to perform audit transfer on the reconstructed target bill according to the audit sequence, perform permission authentication at the audit sequence nodes. If the authentication is successful, perform restoration display and audit processing on the reconstructed target bill to determine the audit result.

[0103] The early warning management module 40 is used to perform early warning management on the target bill according to the audit result.

[0104] Furthermore, the information reconstruction module 20 of the embodiment of the present application is further used to perform the following steps: Perform first encryption processing on the non-critical content to determine the first data layer; for the key elements, determine element codes, perform encryption processing to determine the second data layer, where the information position in the bill information is used as the coding standard; cascade the first data layer and the second data layer to determine the reconstructed target bill.

[0105] Furthermore, the information reconstruction module 20 of the embodiment of the present application is further used to perform the following steps: Perform disordering and second encryption processing on the key elements to determine the element data layer; perform third encryption processing on the element codes to determine the coding data layer; determine the second data layer according to the coding data layer and the element data layer.

[0106] Furthermore, the bill classification includes a bill attribute dimension and a bill level dimension; the information reconstruction module 20 performs encryption execution determination and disorder execution determination based on non-critical content and key elements according to the bill attribute dimension and the bill level dimension.

[0107] Furthermore, the audit path matching module 10 of the embodiment of the present application is further used to perform the following steps: Receive an external audit instruction, where the external audit instruction is a personalized audit requirement; determine the audit workflow by performing bill classification; adjust the audit workflow according to the external audit instruction, and perform audit orchestration based on the time sequence to determine the audit workflow.

[0108] Further, the audit path matching module 10 of the embodiment of the present application is further configured to perform the following steps: According to the audit workflow, match and determine the audit nodes; taking the interaction based on the audit nodes as the orientation, determine the interaction path, where the optimal interaction thread under multiple interaction threads between nodes is used as the standard; according to the audit nodes and the interaction path, determine the audit sequence.

[0109] Further, the audit processing module 30 of the embodiment of the present application is further configured to perform the following steps: Transfer the reconstructed target bill to the first node of the audit sequence for node audit permission authentication to determine the authentication result; if the authentication result is successful authentication, perform key decryption processing on the first data layer, the element data layer, and the coding data layer, and perform scrambling reduction on the element data layer according to the coding data layer to determine the non-critical content and the critical content; perform targeted audit processing on the non-critical content and the critical content to determine the audit result of the first node.

[0110] Further, the audit processing module 30 of the embodiment of the present application is further configured to perform the following steps: For the audit sequence, complete the progressive transfer audit of nodes to determine the audit results of multiple nodes; summarize the audit results of multiple nodes to determine the comprehensive audit result and integrate the audit record chain; use the comprehensive audit result and the audit record chain as the audit result.

[0111] Further, the early warning management module 40 of the embodiment of the present application is further configured to perform the following steps: Traverse the audit record chain to determine the chain early warning information; determine the comprehensive early warning information according to the comprehensive audit result; perform early warning management on the target bill according to the chain early warning information and the comprehensive early warning information.

[0112] Through the foregoing detailed description of the audit early warning management method for enterprise expense bills in this specification, those skilled in the art can clearly know the audit early warning management system for enterprise expense bills in this embodiment. For the system disclosed in Embodiment 2, since it corresponds to the method disclosed in Embodiment 1, it has corresponding functional modules and beneficial effects. For the relevant parts, refer to the description in the method part.

[0113] The foregoing description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for auditing and warning management of enterprise expense bills, characterized in that, The method includes: Deploying a routing rule in the bill classification engine of the audit platform, receiving a target bill, performing workflow orchestration and audit path matching under bill classification, and determining an audit sequence; With the receipt of the target bill, triggering the prior node of the audit platform, performing bill information scanning, disassembling and determining non-critical content and key elements, performing disordered coding encryption, and determining a reconstructed target bill; According to the audit sequence, performing audit transfer on the reconstructed target bill, performing permission authentication at the audit sequence node. If the authentication is successful, performing restored display and audit processing on the reconstructed target bill, and determining an audit result; According to the audit result, performing early warning management on the target bill.

2. The audit warning management method for enterprise expense bills according to claim 1, wherein Performing disordered coding encryption and determining a reconstructed target bill, including: Performing a first encryption process on the non-critical content to determine a first data layer; For the key elements, determining element codes, performing an encryption process, and determining a second data layer, where the information position in the bill information is used as the coding standard; Cascading the first data layer and the second data layer to determine the reconstructed target bill.

3. The audit warning management method for enterprise expense bills according to claim 2, wherein Each key element corresponds to an element code; Performing disordering and a second encryption process on the key elements to determine an element data layer; Performing a third encryption process on the element codes to determine a coding data layer; According to the coding data layer and the element data layer, determining the second data layer.

4. The audit warning management method for enterprise expense bills according to claim 3, characterized in that The bill classification includes a bill attribute dimension and a bill level dimension; According to the bill attribute dimension and the bill level dimension, performing an encryption execution determination and a disorder execution determination based on non-critical content and key elements.

5. The audit warning management method for enterprise expense bills according to claim 1, characterized in that Performing workflow orchestration under bill classification, including: Receiving an external audit instruction, where the external audit instruction is a personalized audit requirement; Determining an audit workflow by performing bill classification; According to the external audit instruction, adjusting the audit workflow, performing audit orchestration based on a time series, and determining an audit workflow.

6. The audit warning management method for enterprise expense bills according to claim 5, characterized in that, Performing audit path matching and determining an audit sequence, including: Matching and determining audit nodes according to the audit workflow; Taking the interaction based on the audit nodes as the guide, determining an interaction path, where the optimal interaction thread under multiple interaction threads between nodes is used as the standard; According to the audit nodes and the interaction path, determining the audit sequence.

7. The audit warning management method for enterprise expense bills according to claim 3, characterized in that, Performing permission authentication at the audit sequence node. If the authentication is successful, performing restored display and audit processing on the reconstructed target bill, and determining an audit result, including: Transferring the reconstructed target bill to the first node of the audit sequence for node audit permission authentication to determine an authentication result; If the authentication result is authentication success, performing key decryption processing on the first data layer, the element data layer, and the coding data layer, and performing disorder restoration on the element data layer according to the coding data layer to determine non-critical content and key content; Performing targeted audit processing on the non-critical content and the key content to determine the audit result of the first node.

8. The audit warning management method for enterprise expense bills according to claim 7, wherein After determining the audit result of the first node, including: For the audit sequence, completing progressive transfer audits of nodes to determine the audit results of multiple nodes; Summarize the audit results of the multiple nodes, determine the comprehensive audit result, and integrate the audit record chain; Use the comprehensive audit result and the audit record chain as the audit result.

9. The audit warning management method for enterprise expense bills according to claim 8, wherein Perform early warning management on the target bill, including: Traverse the audit record chain to determine chain early warning information; Determine comprehensive early warning information based on the comprehensive audit result; Perform early warning management on the target bill according to the chain early warning information and the comprehensive early warning information.

10. An audit warning management system for enterprise expense bills, characterized in that, The system is used to execute the audit and early warning management method for enterprise expense bills described in any one of claims 1-9, including: An audit path matching module, configured to deploy routing rules in the bill classification engine of the audit platform, receive a target bill, perform workflow orchestration and audit path matching under bill classification, and determine the audit sequence; An information reconstruction module, configured to trigger the prior nodes of the audit platform with the receipt of the target bill, perform bill information scanning, disassemble and determine non-critical content and key elements, and perform disordered coding encryption to determine the reconstructed target bill; An audit processing module, configured to perform audit transfer on the reconstructed target bill according to the audit sequence, perform permission authentication at the audit sequence nodes, and if the authentication is successful, perform restored display and audit processing on the reconstructed target bill to determine the audit result; An early warning management module, configured to perform early warning management on the target bill according to the audit result.

Images