A method and system for automatic verification of compliance of a ghost settlement based on four-in-one

Abstract

The application discloses a kind of based on four flow integration's spirit worker settlement compliance automatic checking method and system, belong to spirit worker service and financial compliance verification field, including acquisition contract flow, fund flow, invoice flow, business flow data, after encryption transmission, data cleaning, main body and time standardization processing, carry out multidimensional cross-checking and cross-flow comprehensive checking.System to abnormal data weighted score realizes risk classification, and matches corresponding disposal strategy, simultaneously using hash digest generates tamper-proof electronic evidence chain, finally feedback result to spirit worker end and operation end.System uses layered microservice architecture, and realizes intelligent processing in combination with SpringAI.The four flow integration's spirit worker settlement compliance automatic checking method and system based on it, solve the existing scheme data fragmentation, format disorder, rule solidification, artificial checking inefficiency, risk control lag, traceability difficult and other problems, realize spirit worker settlement compliance verification automation, intelligentization and whole-process traceability, effectively reduce compliance risk.

Description

Technical Field

[0001] This invention relates to the field of flexible employment services and financial compliance verification technology, and in particular to an automatic verification method and system for flexible employment settlement compliance based on the integration of four flows. Background Technology

[0002] With the rapid development of the flexible economy, compliance requirements for flexible settlements are becoming increasingly stringent. The industry generally requires consistency verification of the contract flow, fund flow, invoice flow, and business flow (hereinafter referred to as "four flows") to mitigate tax and legal compliance risks. However, existing flexible settlement compliance verification solutions have several technical shortcomings, as detailed below: First, the four flows of data are fragmented, lacking a unified technical path and implementation solution. Contract flow, fund flow, invoice flow, and business flow data are stored in different business systems, making it impossible to achieve cross-flow linkage verification. This results in a single verification dimension and makes it easy to have compliance vulnerabilities. Second, the data sources are scattered and the formats are inconsistent. There is a lack of standardized preprocessing technical paths and specific technical solutions. The data output by different systems differ in text format, date format, numerical precision, and encoding rules. There is a lot of dirty data and abnormal data, which seriously affects the accuracy of the verification results. Third, the verification rules are fixed and there is no flexible and configurable technical path. The verification rules of the existing system are mostly hard-coded, which cannot adapt to the dynamic needs of different industries and business scenarios. Rule iteration requires restarting the system, resulting in long iteration cycles and poor adaptability. Fourth, manual verification is inefficient and highly subjective, lacking automated verification solutions. Traditional verification methods rely on manual comparison of four streams of data, which is not only time-consuming and labor-intensive, but also prone to omissions due to human error. Furthermore, it cannot form a complete chain of verification evidence, making subsequent traceability difficult. Fifth, risk identification is lagging behind, and there is a lack of technical pathways for intelligent classification and handling. Existing solutions can only identify simple verification anomalies, but cannot classify and determine abnormal risks, and there are no clear handling strategies, resulting in compliance risks not being managed in a timely manner.

[0003] To address the aforementioned technical pain points, there is an urgent need for a flexible settlement compliance verification system that possesses a complete technical path and specific technical solutions, can achieve automatic verification of the integration of four flows, adapts to multiple scenarios, and is traceable, and can combine existing mature technical frameworks to solve the shortcomings of existing solutions. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of fragmented data across four streams, inability to perform cross-stream linkage verification, inconsistent data formats, excessive dirty data, fixed verification rules, inability to dynamically adapt to multiple scenarios, low efficiency of manual verification, susceptibility to errors, and lack of a complete chain of evidence.

[0005] To achieve the above objectives, the present invention employs the following technology: an automatic verification method for the compliance of flexible employment settlement based on the integration of four flows, comprising the following steps: S1. Data Acquisition: Conduct four-flow data acquisition, connect with various business systems to obtain raw data of contract flow, fund flow, invoice flow, and business flow, transmit data through HTTPS encrypted protocol, temporarily store data using Redis, and perform non-empty and format preliminary checks to filter invalid data; S2. Data Preprocessing: Perform data preprocessing, sequentially completing data cleaning, subject information standardization, and time standardization to obtain standardized four-stream data; S3, Four-Flow Linkage Verification: Conduct consistency verification of the four flows, namely contract-business flow verification, fund-business flow verification, invoice-business flow verification, and cross-flow comprehensive verification, and mark various verification anomalies; S4. Intelligent Risk Assessment: Conduct risk assessment and graded handling, collect abnormal data and quantify scores, classify risk levels and match corresponding handling strategies, and complete the closed-loop handling of abnormalities. S5. Electronic Evidence Chain Management: Generates and manages electronic evidence chains, collects data throughout the entire process, uses hash digest algorithms to prevent data tampering, archives and stores evidence chains according to settlement tasks, and supports querying and exporting. S6. Result Push: Feedback the verification results, risk level and handling plan to the flexible worker terminal and the operation management terminal.

[0006] As a further description of the above technical solution: the data preprocessing in step S2 specifically includes: Data cleaning: Use text processing capabilities to remove spaces and special characters from text, and standardize capitalization; use time processing components to convert all time data to a uniform format and time zone; standardize numerical precision and units, and remove invalid values; use regular expressions to standardize various encoding formats; use hash comparison to remove duplicate data and mark abnormal data; Standardization of entity information: Extract flexible worker and employment entity information from the four streams of data, verify the legality of document codes through regular expressions, achieve fuzzy matching of entity names with the help of semantic comparison models, and store the unified entity identifier; Time standardization: Extract timestamps from each stage, complete time zone conversion and format unification, and associate and mark time data according to settlement batches.

[0007] As a further description of the above technical solution: Step S3, the four-stream linkage verification, specifically includes: Contract-Business Flow Validation: Retrieve two types of standardized data, align core fields, and compare them according to preset rules and configurable deviation thresholds. If the threshold is exceeded or there is a time logic anomaly, an anomaly is marked. Funds-Business Flow Validation: Standardize fund amount data and perform validation from three dimensions: amount, time, and entity. If any dimension fails to meet the standard, it is marked as abnormal. Invoice-Business Flow Verification: Key information is extracted and standardized by parsing invoice images using OCR technology. The reasonableness of invoice items, amounts, and dates is verified. At the same time, the system is connected to the tax system to verify the authenticity of invoices. Any problems are marked as abnormal. Cross-flow integrated verification: Construct a four-flow data association model to verify the consistency of the subject and amount, and analyze the continuity of the timeline. If data contradictions or timeline chaos occur, an integrated anomaly is marked. The default deviation threshold for amount is 5%, and the default deviation threshold for time is 24 hours.

[0008] As a further description of the above technical solution: Step S4, intelligent risk determination, specifically includes: Collect all verification anomaly data, generate anomaly vectors, and calculate risk scores by weighting anomaly type (40%), anomaly severity (30%), and anomaly quantity (30%). Based on the score, the risk level is divided into four levels: a score ≥ 80 is high risk, 50 ≤ score < 80 is medium risk, 20 ≤ score < 50 is low risk, and a score < 20 is normal. Matching corresponding handling strategies: High-risk cases will have their settlements frozen and require manual review; medium-risk cases will require manual review; low-risk cases will receive automatic reminders and a deadline for rectification; and normal cases will have their settlements automatically approved. Record information throughout the entire process of handling anomalies, update the anomaly status, and form a closed-loop management system; the system is configured with an automatic retry mechanism, which retryes 3 times by default, and transfers to manual handling after a retry failure.

[0009] As a further description of the above technical solution: Step S5, electronic evidence chain management, specifically includes: collecting data from the entire process of data collection, preprocessing, verification, risk assessment, and anomaly handling, and standardizing it; converting unstructured files to their formats; generating a unique hash digest for all evidence data and storing it independently; performing secondary hash digest calculation to complete data anti-tampering verification; and archiving and storing the complete evidence chain according to settlement tasks, providing query and export interfaces to the outside world.

[0010] An automatic compliance verification system for flexible settlement based on the integration of four flows is used to implement the method described in any of the above. The system adopts a layered architecture, which includes, from top to bottom, a user access layer, an API gateway layer, a microservice business layer, a data storage layer, and an AI capability layer. The user access layer includes the flexible mini-program developed by Uniapp and the operation and management web client developed by Vue3, which realizes multi-terminal interaction, page display, function operation and permission control. The API gateway layer is built on Spring Cloud Alibaba Gateway and integrates Spring Security to implement identity verification, routing forwarding, traffic limiting, and exception interception. The microservice business layer is built on Spring Boot and consists of seven core microservices: data collection service, data preprocessing service, four-flow verification service, rule platform service, risk assessment service, evidence chain management service, and operation management service. Nacos is used for service registration, discovery, and configuration management, Feign is used for communication between microservices, and a distributed transaction mechanism is configured to ensure data consistency. The data storage layer uses MySQL to store structured data, MongoDB to store unstructured data, and Redis to cache hot and temporary data; sensitive data is stored with de-identification, and full and incremental backup mechanisms are used to ensure data security. The AI ​​capability layer is built on SpringAI and SpringAIAlibaba, deploying text standardization models, semantic comparison models, OCR parsing models, and risk scoring models to complete intelligent data processing, comparison and verification, and risk quantification.

[0011] As a further description of the above technical solution: the rule platform service supports visual rule configuration, rule review, rule publishing, dynamic loading and version management. The configured rules and deviation thresholds are dynamically pushed to each microservice through Nacos and take effect without restarting the system.

[0012] As a further description of the above technical solution: the system is also equipped with an exception handling module and a security module; The exception handling module sets up an exception interception mechanism in each key link of the system, records exception information and pushes reminders through the message queue, and completes the closed-loop management of exceptions in conjunction with automatic retry and manual handling. The security module uses HTTPS to encrypt data transmission and employs a de-identification algorithm to protect sensitive data storage. Based on Spring Security, it classifies three roles—flexible workers, operations personnel, and administrators—and assigns them permissions. It records operation logs throughout the process and uses hash digests and digital signature technologies to prevent core data from being tampered with.

[0013] As a further description of the above technical solution: the data preprocessing service integrates text processing components, time processing components, regular expression verification tools, and semantic comparison models to complete the entire process of data cleaning, subject information standardization, and time standardization.

[0014] As a further description of the above technical solution: the evidence chain management service integrates a hash digest algorithm to complete evidence data digest generation, anti-tampering verification, classification and archiving, and provides evidence chain query and export interfaces to the outside world.

[0015] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. By adopting a unified data collection and preprocessing technology path and solution, the standardization and linkage verification of four-stream data have been achieved. The verification coverage is comprehensive, effectively avoiding compliance loopholes and solving the problems of fragmented four-stream data and inability to perform cross-stream linkage verification. 2. By using SpringAI-driven data cleaning and standardization technology, data quality was improved, the impact of dirty data on verification results was reduced, the data standardization rate was significantly improved, verification efficiency was greatly improved, and the problems of inconsistent data formats and a large amount of dirty data were solved. 3. Through the dynamic configuration technology path of the rule platform, the verification rules are configurable, updatable, and version-manageable. Rule iteration can be completed without restarting the system, adapting to the needs of different industries and business scenarios. The rule iteration cycle is greatly shortened, solving the problem of fixed verification rules and inability to dynamically adapt to multiple scenarios. 4. Through automated verification and evidence chain management technology, the verification process is fully automated, greatly reducing manual intervention. At the same time, a complete electronic evidence chain is generated to ensure the traceability of the settlement process, reduce labor costs and traceability difficulty, and solve the problems of low efficiency, error-proneness and lack of complete evidence chain in manual verification. 5. Through SpringAI-driven intelligent risk scoring and graded handling technology, the system has achieved accurate identification and graded handling of abnormal risks, significantly shortened the risk identification response time, effectively improved compliance risk management capabilities, reduced compliance risks of the flexible work platform, and solved the problems of delayed risk identification and inability to handle risks in a graded manner. 6. Through the collaborative application of multiple technology frameworks, the system achieves high scalability, high availability, and high security, supports the business expansion of the flexible platform, and ensures data security and stable system operation. The system response speed is significantly improved, and the failure rate is greatly reduced. Attached Figure Description

[0016] Figure 1 A flowchart of a method according to an embodiment of the present invention is shown; Figure 2 A system block diagram according to an embodiment of the present invention is shown. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Reference Figures 1-2 This embodiment provides an automatic compliance verification system for flexible settlement based on the integration of four flows. The system adopts a layered architecture, which includes, from top to bottom, a user access layer, an API gateway layer, a microservice business layer, a data storage layer, and an AI capability layer. The system is deployed and built through the above five-layer architecture.

[0019] Deploy the backend infrastructure: Seven core microservices are built based on Spring Boot: data acquisition service, data preprocessing service, four-flow verification service, rule platform service, risk assessment service, evidence chain management service, and operation management service. All microservices are registered, discovered, and centrally configured using Spring Cloud Alibaba's Nacos component, with unified configuration of interface parameters, database parameters, and AI model call parameters. Communication rules between microservices are configured through the Feign component, and a distributed transaction mechanism is enabled to ensure data consistency during multi-service collaborative operations.

[0020] Deploy the API gateway layer: Build a unified API gateway using Spring Cloud Alibaba Gateway, integrate Spring Security components, and configure identity verification, request routing, traffic limiting, and exception interception strategies to serve as the unified entry point for all frontend requests.

[0021] Deploy the data storage layer: Deploy three types of databases: MySQL, MongoDB, and Redis. MySQL is used to store four-stream structured data, rule data, risk results, etc.; MongoDB is used to store unstructured data such as invoice images and vouchers; Redis is used to cache hot data and temporary data; Configure sensitive data anonymization rules, and enable a data backup mechanism of full backup + incremental backup.

[0022] Deploy the AI ​​capability layer: Based on SpringAI and SpringAIAlibaba, complete the environment setup, deploy text standardization model, semantic comparison model, OCR parsing model, and risk scoring model, and complete model debugging and interface integration.

[0023] Deploy the front-end application: Use Uniapp to develop the flexible worker mini-program to realize functions such as task submission, settlement viewing, exception reception, and rectification operations; use Vue3 to develop the operation and management web client, integrate the ElementPlus component library, and realize functions such as rule configuration, risk review, evidence chain query and export, and personnel permission management; rely on route guards and Spring Security to divide the roles into three categories: flexible workers, operation personnel, and administrators, and assign corresponding operation permissions.

[0024] Function Configuration: Operations personnel log in to the web interface to access the rules platform, where they can visually configure various verification rules, configurable deviation thresholds (default deviation of 5% for amount, default deviation of 24 hours for time), risk scoring standards, handling strategies, and automatic retry counts (default 3 times). After the rules are approved, they are dynamically published through Nacos and take effect without restarting the system. The rules platform also enables version management, supporting rule traceability and rollback.

[0025] An automatic compliance verification method for flexible employment settlement based on the integration of four flows (information, capital, logistics, and capital flow) is proposed. This method uses the system deployed above and includes the following steps: S1, Data Acquisition After the flexible worker completes the labor task, the four-stream data collection process is initiated: Linggong submits raw business process data through a mini-program, including task name, task period, main information, task amount, task voucher, etc.

[0026] The system's data acquisition service uses the Feign component to connect with the contract management system, financial system, and invoice system, simultaneously collecting raw data from the contract flow, fund flow, and invoice flow.

[0027] All cross-system and front-end / back-end interactive data is transmitted using the HTTPS encrypted protocol, and the collected raw data is temporarily stored in Redis.

[0028] The data acquisition service performs non-empty checks and basic format checks on the temporarily stored data, filtering out invalid data such as null values ​​and severely formatted errors, thus completing the data acquisition process.

[0029] S2, Data Preprocessing The data preprocessing service retrieves the collected raw data from Redis and sequentially performs data cleaning, subject information standardization, and time standardization. Data cleaning utilizes AI text processing capabilities to remove spaces and special characters from text content and standardize capitalization; it calls time processing components to convert various time data into a unified format and complete time zone alignment; it standardizes the precision and units of measurement for numerical values ​​such as amounts and removes invalid values; it standardizes various encoding formats such as contract numbers and invoice numbers through regular expressions; and it uses hash comparison algorithms to remove duplicate data and marks abnormal data.

[0030] The standardization of entity information involves extracting the entity information of freelancers and employers from the four streams of data, verifying the legality of the document and unified social credit code format through regular expressions, and calling the semantic comparison model to perform fuzzy matching on entity names to identify information that has inconsistent names but belongs to the same entity. After unifying the entity identifier, the information is temporarily stored.

[0031] Time standardization extracts timestamps from all four stages of the process, unifying time zones and time formats; all time data are associated and marked according to settlement batches to ensure that the time dimension of settlement data in the same batch is comparable.

[0032] After all processing is complete, the standardized four-stream data will be categorized and stored in MySQL and MongoDB.

[0033] S3, Four-stream linkage verification The four-stream verification service retrieves standardized four-stream data, performs four types of verification in sequence, and marks abnormal data: The contract-business flow verification retrieves standardized data from the contract flow and business flow, aligning core fields such as flexible worker information, task information, amount, and period; it calls the verification rules configured in the rule platform and compares them with the deviation threshold. If the data deviation exceeds the threshold, or if there are time logic anomalies such as the contract signing time being later than the task start time, the anomaly is immediately marked.

[0034] The funds-business flow verification standardizes the amount of receipts and payments in the funds flow and conducts verification from three dimensions: amount, time, and entity. It verifies the payment amount against the task amount, verifies whether the payment time is later than the task completion time, and verifies whether the receiving and paying entities match the flexible workers and the employing entities. If any dimension fails to meet the verification standard, it is marked as abnormal.

[0035] The invoice-business flow verification process calls an OCR parsing model to recognize invoice images, extracts key information such as invoice amount, invoice issuer, invoice time, invoice item, and invoice number, and standardizes it; it verifies the reasonableness of invoice item and task type, invoice amount and task amount, and invoice time and task completion time; at the same time, it connects to the tax system interface to verify the authenticity of invoices, and marks abnormalities if there are fake invoices or information mismatches.

[0036] Cross-flow comprehensive verification constructs a four-flow data association model, integrating contract flow, fund flow, invoice flow, and business flow data; it verifies the consistency of the main information and total amount of the four flows, and analyzes the entire process timeline to determine whether it conforms to the normal logic of "contract signing, business execution, invoice issuance, and fund payment"; if data contradictions or timeline confusion occur, they are marked as cross-flow comprehensive anomalies. All verification results and anomaly data are uniformly pushed to the rule platform and risk assessment service.

[0037] S4, Intelligent Risk Assessment The risk assessment service collects all verification anomaly data and performs risk quantification scoring, level classification, strategy matching, and closed-loop management. All abnormal information is integrated to generate an anomaly vector, and a risk score is calculated by weighting the anomaly type (40%), severity (30%), and number (30%).

[0038] The risk level is divided into four levels based on the score, and corresponding handling strategies are applied: a score of ≥80 is considered high risk, and settlement will be frozen and manual review will be implemented; a score of 50 ≤ score < 80 is considered medium risk, and manual review will be implemented; a score of 20 ≤ score < 50 is considered low risk, and automatic reminders and time-limited rectification will be implemented; a score < 20 is considered normal, and the settlement process will be automatically approved.

[0039] The system sets up an anomaly interception mechanism in key stages such as data collection, preprocessing, and verification, and configures a default automatic retry mechanism of 3 times. For retryable anomalies, it automatically retryes, and if the retry fails, it is transferred to manual handling.

[0040] The entire process of handling anomalies is recorded, along with the results. The anomaly status is updated in real time, forming a closed-loop control system of "anomaly identification - notification - handling - feedback". At the same time, anomaly alerts are pushed to the corresponding operators through a message queue.

[0041] S5, Electronic Evidence Chain Management The evidence chain management service collects data from the entire settlement and verification process, completing the construction, tamper-proofing, and archiving of electronic evidence chains. It summarizes and standardizes data from the entire process of data collection, data preprocessing, four-stream verification, risk assessment, and anomaly handling; and performs general format conversion on unstructured files such as invoice images and vouchers.

[0042] A unique hash digest is generated for all evidence data and stored independently. When verifying data integrity later, the hash digest of the current data is recalculated and compared with the original digest to achieve data tamper-proof verification.

[0043] All evidence data is categorized and archived according to the settlement task dimension, and the complete electronic evidence chain is uniformly stored in MongoDB.

[0044] The system provides external query and export interfaces, allowing operators and regulators to input settlement task numbers to query and export the corresponding electronic evidence chains.

[0045] S6, Results Push The system will push the final verification results, risk level, corresponding handling plan, cause of anomaly, and rectification requirements to the flexible worker's mini-program and the operations management web interface, respectively. Flexible workers can view their own settlement progress and anomaly information and complete rectification through the mini-program; operations personnel can view the verification status and risk distribution of all settlement tasks through the web interface, and carry out manual review and handling to ensure the compliant progress of the flexible worker settlement process.

[0046] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An automatic compliance verification method for flexible employment settlement based on the integration of four flows, characterized in that, Includes the following steps: S1. Data Acquisition: Conduct four-flow data acquisition, connect with various business systems to obtain raw data of contract flow, fund flow, invoice flow, and business flow, transmit data through HTTPS encrypted protocol, temporarily store data using Redis, and perform non-empty and format preliminary checks to filter invalid data; S2. Data Preprocessing: Perform data preprocessing, sequentially completing data cleaning, subject information standardization, and time standardization to obtain standardized four-stream data; S3, Four-Flow Linkage Verification: Conduct consistency verification of the four flows, namely contract-business flow verification, fund-business flow verification, invoice-business flow verification, and cross-flow comprehensive verification, and mark various verification anomalies; S4. Intelligent Risk Assessment: Conduct risk assessment and graded handling, collect abnormal data and quantify scores, classify risk levels and match corresponding handling strategies, and complete the closed-loop handling of abnormalities. S5. Electronic Evidence Chain Management: Generates and manages electronic evidence chains, collects data throughout the entire process, uses hash digest algorithms to prevent data tampering, archives and stores evidence chains according to settlement tasks, and supports querying and exporting. S6. Result Push: Feedback the verification results, risk level and handling plan to the flexible worker terminal and the operation management terminal.

2. The automatic compliance verification method for flexible employment settlement based on the integration of four flows as described in claim 1, characterized in that, The data preprocessing step S2 specifically includes: data cleaning: using text processing capabilities to remove spaces and special characters from text, and unifying capitalization; using a time processing component to convert all time data into a unified format and unify the time zone; unifying numerical precision and units, and removing invalid values; standardizing various encoding formats using regular expressions; using hash comparison to remove duplicate data and marking abnormal data; Standardization of entity information: Extract flexible worker and employment entity information from the four streams of data, verify the legality of document codes through regular expressions, achieve fuzzy matching of entity names with the help of semantic comparison models, and store the unified entity identifier; Time standardization: Extract timestamps from each stage, complete time zone conversion and format unification, and associate and mark time data according to settlement batches.

3. The automatic compliance verification method for flexible employment settlement based on the integration of four flows as described in claim 1, characterized in that, The step S3 four-flow linkage verification specifically includes: Contract-business flow verification: retrieve two types of standardized data, align core fields, and compare them according to preset rules and configurable deviation thresholds. If the threshold is exceeded or there is a time logic abnormality, an abnormality is marked. Funds-Business Flow Validation: Standardize fund amount data and perform validation from three dimensions: amount, time, and entity. If any dimension fails to meet the standard, it is marked as abnormal. Invoice-Business Flow Verification: Key information is extracted and standardized by parsing invoice images using OCR technology. The reasonableness of invoice items, amounts, and dates is verified. At the same time, the system is connected to the tax system to verify the authenticity of invoices. Any problems are marked as abnormal. Cross-flow integrated verification: Construct a four-flow data association model to verify the consistency of the subject and amount, and analyze the continuity of the timeline. If data contradictions or timeline chaos occur, an integrated anomaly is marked. The default deviation threshold for amount is 5%, and the default deviation threshold for time is 24 hours.

4. The automatic compliance verification method for flexible employment settlement based on the integration of four flows as described in claim 1, characterized in that, The intelligent risk assessment step S4 specifically includes: Collect all verification anomaly data, generate anomaly vectors, and calculate risk scores by weighting anomaly type (40%), anomaly severity (30%), and anomaly quantity (30%). Based on the score, the risk level is divided into four levels: a score ≥ 80 is high risk, 50 ≤ score < 80 is medium risk, 20 ≤ score < 50 is low risk, and a score < 20 is normal. Matching corresponding handling strategies: High-risk cases will have their settlements frozen and require manual review; medium-risk cases will require manual review; low-risk cases will receive automatic reminders and a deadline for rectification; and normal cases will have their settlements automatically approved. Record information throughout the entire process of handling anomalies, update the anomaly status, and form a closed-loop management system; the system is configured with an automatic retry mechanism, which retryes 3 times by default, and transfers to manual handling after a retry failure.

5. The automatic compliance verification method for flexible employment settlement based on the integration of four flows as described in claim 1, characterized in that, The S5 step of electronic evidence chain management specifically includes: collecting and standardizing data from the entire process of data collection, preprocessing, verification, risk assessment, and anomaly handling; converting unstructured files to the correct format; generating a unique hash digest for all evidence data and storing it independently; performing secondary hash digest calculation to complete data tamper-proof verification; and archiving and storing the complete evidence chain according to settlement tasks, providing query and export interfaces to external parties.

6. An automatic compliance verification system for flexible employment settlement based on the integration of four flows, used to implement the method described in any one of claims 1-5, characterized in that, The system adopts a layered architecture, which includes, from top to bottom, the user access layer, API gateway layer, microservice business layer, data storage layer, and AI capability layer. The user access layer includes the flexible mini-program developed by Uniapp and the operation and management web client developed by Vue3, which realizes multi-terminal interaction, page display, function operation and permission control. The API gateway layer is built on Spring Cloud Alibaba Gateway and integrates Spring Security to implement identity verification, routing forwarding, traffic limiting, and exception interception. The microservice business layer is built on Spring Boot and consists of seven core microservices: data collection service, data preprocessing service, four-flow verification service, rule platform service, risk assessment service, evidence chain management service, and operation management service. Nacos is used for service registration, discovery, and configuration management, Feign is used for communication between microservices, and a distributed transaction mechanism is configured to ensure data consistency. The data storage layer uses MySQL to store structured data, MongoDB to store unstructured data, and Redis to cache hot and temporary data; sensitive data is stored with de-identification, and full and incremental backup mechanisms are used to ensure data security. The AI ​​capability layer is built on SpringAI and SpringAIAlibaba, deploying text standardization models, semantic comparison models, OCR parsing models, and risk scoring models to complete intelligent data processing, comparison and verification, and risk quantification.

7. The automatic compliance verification system for flexible employment settlement based on the integration of four flows as described in claim 6, characterized in that, The rule platform service supports visual rule configuration, rule review, rule publishing, dynamic loading, and version management. The configured rules and deviation thresholds are dynamically pushed to each microservice through Nacos and take effect without restarting the system.

8. The automatic compliance verification system for flexible employment settlement based on the integration of four flows as described in claim 6, characterized in that, The system is also equipped with an anomaly handling module and a security module; The exception handling module sets up an exception interception mechanism in each key link of the system, records exception information and pushes reminders through the message queue, and completes the closed-loop management of exceptions in conjunction with automatic retry and manual handling. The security module uses HTTPS to encrypt data transmission and employs a de-identification algorithm to protect sensitive data storage. Based on Spring Security, it classifies three roles—flexible workers, operations personnel, and administrators—and assigns them permissions. It records operation logs throughout the process and uses hash digests and digital signature technologies to prevent core data from being tampered with.

9. The automatic compliance verification system for flexible employment settlement based on the integration of four flows as described in claim 6, characterized in that, The data preprocessing service integrates text processing components, time processing components, regular expression validation tools, and semantic comparison models to complete the entire process of data cleaning, subject information standardization, and time standardization.

10. The automatic compliance verification system for flexible employment settlement based on the integration of four flows as described in claim 6, characterized in that, The evidence chain management service integrates a hash digest algorithm to generate evidence data digests, perform anti-tampering verification, classify and archive evidence, and provide external interfaces for evidence chain querying and exporting.

Images