A multi-agent collaborative enterprise employee welfare system and its implementation method

Abstract

This invention relates to the field of enterprise management technology, specifically to a multi-agent collaborative enterprise employee welfare system and its implementation method. The system constructs a four-layer architecture: a data layer, an agent layer, an application layer, and an interaction layer. It is configured with a real-time price comparison agent, an employee welfare adaptation agent, an intelligent e-commerce platform operation and maintenance agent, and a multi-scenario collaborative agent. At the application layer, it implements intelligent applications for various core scenarios. Each agent achieves data interoperability and collaborative operation through standardized data interfaces, completing an automated closed loop from product pricing, personalized recommendations, order processing to cross-scenario resource scheduling. It can replace manual operations, significantly improving welfare management efficiency and employee experience, while achieving full-process compliance supervision. It can be widely applied to employee welfare management scenarios in various enterprises and administrative institutions.

Description

Technical Field

[0001] This invention relates to the field of enterprise management technology, specifically to a multi-agent collaborative enterprise employee welfare system and its implementation method. Background Technology

[0002] As a carrier for the distribution, consumption and management of employee welfare, the welfare mall of enterprises currently has the following defects: (1) It only focuses on the single scenario of welfare product redemption and cannot cover other related scenarios. Cross-scenario resources cannot be integrated and cannot meet the diversified and full-scenario welfare needs of employees. (2) In the existing technology, AI agents are only applied to individual functional modules of the welfare mall (such as simple product recommendations) and are independent of each other without a collaborative operation mechanism. (3) The pricing, delisting and other operations of the welfare mall are all done manually, with a high proportion of repetitive work and low operational efficiency. (4) The product / service recommendations of the existing welfare mall are generalized and homogeneous, with extremely low recommendation accuracy, which easily leads to the waste of welfare funds. (5) The data of each functional module and each extended scenario of the existing welfare mall are independent of each other and cannot achieve data interoperability and linkage analysis. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide a multi-agent collaborative enterprise employee welfare system and implementation method, so as to solve the problems of existing welfare malls, such as single scenario, inefficient manual operation and maintenance, lack of personalized services, and cross-scenario data incompatibility.

[0004] According to a first aspect of the present invention, a multi-agent collaborative enterprise employee welfare system is provided, comprising:

[0005] The data layer is used to collect and store relevant data from the entire scenario using pre-opened API interfaces; The intelligent agent layer is configured with a real-time price comparison intelligent agent, an employee benefits adaptation intelligent agent, an e-commerce intelligent operation and maintenance intelligent agent, and a multi-scenario collaborative intelligent agent. Each intelligent agent interacts with data through a standardized data interface. Each intelligent agent has an independent built-in AI decision-making and execution module, which is used to obtain data from the data layer and complete the corresponding functions to generate decision results. The real-time price comparison agent is used to calculate the real-time price difference between the prices of each product in the welfare mall and the lowest prices of each product on external shopping platforms, based on the product-related data in the data layer. The employee benefits adaptation intelligent agent is used to build a personalized profile based on the employee's personal information in the data layer, and push products or services from the benefits mall to the employee based on the personalized profile. The intelligent operation and maintenance agent of the mall is used to delist and relist products in the welfare mall, assign logistics tracking numbers to orders, and respond to user inquiries. The multi-scenario collaborative intelligent agent is used to associate account information corresponding to each employee's identity; integrate operation and maintenance data of each core scenario in the application layer to build a full-scenario operation database; and intelligently schedule service resources of each core scenario. The application layer is used to build intelligent applications for various core scenarios based on the decision results of the intelligent agent layer. Each core scenario has an independent functional unit. Through the collaborative linkage of intelligent agents in multiple scenarios, cross-scenario data linkage, resource scheduling and business connection can be achieved. The interaction layer connects to various types of operation ports for user authentication and human-computer interaction.

[0006] Preferably, the real-time price comparison agent in the agent layer includes: The multi-source data acquisition module is used to integrate the data layer and collect product price data from various shopping platforms in real time through pre-opened API interfaces; The product normalization matching module is used to normalize and classify the product price data that belong to the same category of products; The price cleaning and comparison module is used to clean the normalized product price data and then calculate the real-time price difference between the lowest price of each product and the price of each product in the welfare mall. The price fluctuation warning module is used to automatically trigger dynamic price adjustment instructions based on the real-time price difference and the preset tiered price difference threshold; The dynamic price adjustment module is used to automatically adjust the price of the product in the welfare mall based on the lowest price of the product after receiving the dynamic price adjustment instruction; The front-end price comparison display module is used to display the real-time prices of each product in the welfare mall, as well as the average price of each product on the shopping platform.

[0007] Preferably, the employee benefits adaptation agent includes: The multi-dimensional data acquisition module is used to integrate the multi-dimensional information of each employee in the data layer, collect the shopping behavior data of each employee in the welfare mall, and build a data profile for each employee. The user profile building module is used to generate exclusive consumption tags and health tags for each employee based on their data profile; and to build a personalized profile for each employee based on their consumption tags and health tags. The welfare allowance calculation module is used to read the welfare allowance information of each employee in real time, filter the set of redeemable items for that employee from the welfare mall, and generate the optimal use plan for the welfare allowance based on the employee's personalized profile. The personalized recommendation engine module is used to use collaborative filtering and profile matching algorithms to filter and display recommendation results from the set of redeemable goods based on the employee's personalized profile. The health service linkage module is used to push medical and health services from the benefits mall to employees based on their health tags.

[0008] Preferably, the intelligent operation and maintenance agent for the e-commerce platform includes: The automated product management module is used to identify and remove overpriced products from the welfare mall based on the real-time price difference of each product, and to identify and remove products with a push rate below the threshold from the welfare mall based on the push data of the employee welfare adaptation intelligent agent. The intelligent order processing module is used to receive order and user inquiry data from various operation ports, assign logistics tracking numbers to the orders, and generate response information corresponding to the user inquiry data. The intelligent inventory monitoring module is used to monitor inventory levels in real time. When the inventory level falls below the warning threshold, it automatically pushes replenishment reminders to the enterprise management system and suppliers. The data statistics and analysis module is used to generate multi-dimensional operation reports based on the full-scenario operation data of the welfare mall in the data layer; The system linkage maintenance module is used to monitor the operation status of the multi-agent collaborative enterprise employee welfare system in real time. When an anomaly occurs, it automatically triggers an early warning message and pushes it to the technical maintenance terminal.

[0009] Preferably, the multi-scenario collaborative intelligent agent includes: The identity and account linkage module is used to identify an employee's multiple identities based on the employee's personal information in the data layer and associate the account information corresponding to each identity. The cross-scenario data integration module is used to integrate operation and maintenance data from various core scenarios in the application layer, so that the employee benefits matching agent can update the personalized profile of employees based on the corresponding operation and maintenance data; it is also used to build a full-scenario operation database based on the integrated operation and maintenance data. The intelligent resource scheduling module is used to intelligently schedule service resources for various core scenarios based on user needs, geographical distance, and professional suitability. The compliance and supervision module is used to monitor the data in the full-scenario operation database, automatically identify abnormal data, and push abnormal warnings to the enterprise management end and the supervision end.

[0010] Preferably, the data layer is also used to encrypt and desensitize sensitive data in the collected full-scene related data.

[0011] The preferred application layer includes multiple core scenarios, such as: corporate welfare procurement scenario, employee welfare consumption scenario, state-owned enterprise meal card integration scenario, medical and health service scenario, college student entrepreneurship incubation scenario, community aging service scenario, offline health station operation scenario, and full-process data supervision scenario.

[0012] According to a second aspect of the present invention, a method for implementing enterprise employee benefits through multi-agent collaboration is provided, applied to the multi-agent collaborative enterprise employee benefits system described in any one of the above claims, comprising: Connect the real-time price comparison intelligent agent, the employee benefits adaptation intelligent agent, the mall intelligent operation and maintenance intelligent agent, and the multi-scenario collaborative intelligent agent with the welfare mall basic system, debug the data interface, and set the operating parameters and compliance thresholds of each intelligent agent. Collect and store relevant data from all scenarios through pre-opened API interfaces; Based on the relevant data of the entire scenario, the four intelligent agents independently complete their corresponding functions, generate decision results, and interact with each other through a standardized data interface. Intelligent applications for various core scenarios are built based on the decision results of each intelligent agent. Each core scenario is set to run independently. Cross-scenario data linkage, resource scheduling and business connection are achieved through multi-scenario collaborative linkage of intelligent agents. Real-time recording and multi-dimensional analysis of operation and maintenance data for each core scenario; automatic labeling of abnormal data and push notifications; continuous optimization of algorithm parameters for each intelligent agent based on operation and maintenance data.

[0013] Preferably, after collecting relevant data across the entire scenario through pre-opened API interfaces, the process also includes: Preprocess the collected data from all scenarios, including cleaning, deduplication, and desensitization. The preprocessed, all-scenario-related data is encrypted and stored in the corresponding database.

[0014] Preferably, the method further includes: The interaction layer receives service feedback data from various operation ports and synchronizes the service feedback data to the intelligent agent layer.

[0015] The technical solution provided by this invention may include the following beneficial effects: It is understood that the technical solution presented in this invention constructs a four-layer architecture: a data layer, an intelligent agent layer, an application layer, and an interaction layer. It configures a real-time price comparison intelligent agent, an employee benefits adaptation intelligent agent, an intelligent mall operation and maintenance intelligent agent, and a multi-scenario collaborative intelligent agent. At the application layer, it realizes intelligent applications for various core scenarios. Each intelligent agent achieves data interoperability and collaborative operation through standardized data interfaces, completing an automated closed loop from product pricing, personalized recommendations, order processing to cross-scenario resource scheduling. It can replace manual operation, significantly improve the efficiency of welfare management and employee experience, while achieving full-process compliant supervision, and can be widely applied to employee welfare management scenarios in various enterprises and administrative institutions.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0018] Figure 1 This is a schematic block diagram illustrating a multi-agent collaborative enterprise employee welfare system according to an exemplary embodiment; Figure 2 This is a schematic diagram illustrating the steps of a multi-agent collaborative enterprise employee welfare implementation method according to an exemplary embodiment. Detailed Implementation

[0019] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0020] In one embodiment, Figure 1 This is a schematic block diagram illustrating a multi-agent collaborative enterprise employee benefits system according to an exemplary embodiment. See also: Figure 1 A multi-agent collaborative enterprise employee benefits system is provided, comprising: The data layer is used to collect and store relevant data from the entire scenario using pre-opened API interfaces.

[0021] In practice, the data layer provides fundamental data support for the system. Located at the bottom of the four-layer architecture, it serves as the comprehensive data source for all intelligent agents to conduct decision-making and execution operations. It is responsible for the collection, cleaning, storage, encryption, and synchronization of relevant data across all scenarios. Its core collected data covers six major categories: basic employee data (including job level, employee ID, welfare benefits, meal card information, facial recognition information, etc.), commodity supply chain data (including product costs, qualification documents, inventory quantities, supply cycles, etc.), cross-platform price data (including mainstream e-commerce platforms, offline supermarket guidance prices, and quotations from compliant procurement platforms such as Guocaiyun), employee consumption and health data (including historical consumption records, purchase categories, consumption frequency, medical examination reports, chronic disease records, and health station testing data, etc.), multi-scenario operational data (including entrepreneurial service records, community service needs, health station operational data, and order processing data, etc.), and regulatory data (including data related to fiscal and state-owned asset supervision requirements, audit standards, and compliance verification data, etc.).

[0022] The data layer collects all data only through the official open API interface, and the use of illegal crawlers or other unauthorized methods is strictly prohibited.

[0023] In a preferred embodiment, the data layer is further used to encrypt and de-identify sensitive data in the collected data related to the entire scenario. For example, for sensitive data such as employee identity information and health data, encryption and de-identification processing is performed. When storing, the data storage architecture meets the requirements of Level 3 Information Security Protection, wherein medical and health-related data is encrypted and stored for a period of not less than 3 years; the system is generally compatible with domestic database systems, comprehensively ensuring the security and compliance of data throughout the entire process of collection, storage, and transmission.

[0024] The intelligent agent layer is configured with a real-time price comparison intelligent agent, an employee benefits adaptation intelligent agent, an e-commerce intelligent operation and maintenance intelligent agent, and a multi-scenario collaborative intelligent agent. Each intelligent agent interacts with data through a standardized data interface. Each intelligent agent has an independent built-in AI decision-making and execution module, which is used to obtain data from the data layer and complete the corresponding functions to generate decision results.

[0025] In practice, the intelligent agent layer is the core decision-making and execution layer of the system, located at the second layer of the four-layer architecture. It connects to the data layer above and interfaces with the application layer below. This layer has four types of intelligent agents; each intelligent agent has an independent built-in AI algorithm model and decision-making execution module, and can independently complete its corresponding function based on standardized data output from the data layer. At the same time, the four intelligent agents achieve real-time data exchange, result synchronization, and collaborative operation through standardized data interfaces. The entire layer accepts unified scheduling and parameter optimization from intelligent agents that coordinate across multiple scenarios, ensuring that the decision outputs and execution actions of each intelligent agent meet the operational needs of all scenarios and compliance regulatory requirements.

[0026] The real-time price comparison agent is used to calculate the real-time price difference between the prices of each product in the welfare mall and the lowest prices of each product on external shopping platforms, based on the product-related data in the data layer.

[0027] The employee benefits adaptation intelligent agent is used to build personalized profiles based on employee personal information in the data layer, and push products or services from the benefits mall to employees based on the personalized profiles.

[0028] The intelligent operation and maintenance agent of the mall is used to delist and relist products in the welfare mall, assign logistics tracking numbers to orders, and respond to user inquiries.

[0029] The multi-scenario collaborative intelligent agent is used to associate account information corresponding to each employee's identity; integrate operation and maintenance data of each core scenario in the application layer to build a full-scenario operation database; and intelligently schedule service resources of each core scenario.

[0030] The application layer is used to build intelligent applications for various core scenarios based on the decision results of the intelligent agent layer. Each core scenario has an independent functional unit. Through the collaborative linkage of intelligent agents in multiple scenarios, cross-scenario data linkage, resource scheduling and business connection can be achieved.

[0031] In practice, the application layer is the third layer of the four-layer architecture, connecting the intelligent agent layer and the interaction layer. Based on the decision-making results, data support, and resource scheduling instructions of the intelligent agent layer, it implements intelligent applications across all enterprise welfare scenarios and builds a business closed loop. This layer corely covers multiple application scenarios, with each scenario module having an independent functional unit. At the same time, through multi-scenario collaborative intelligent agents, it achieves cross-scenario data linkage, intelligent resource scheduling, and seamless business integration.

[0032] In a preferred embodiment, the application layer constructs multiple core scenarios, including: corporate welfare procurement scenario, employee welfare consumption scenario, state-owned enterprise meal card integration scenario, medical and health service scenario, university student entrepreneurship incubation scenario, community aging service scenario, offline health station operation scenario, and full-process data supervision scenario. Each scenario can be modularly expanded according to actual operational needs, and the expanded modules can be seamlessly integrated into the existing architecture to achieve data interoperability and collaborative operation with the original scenarios.

[0033] The interaction layer connects to various types of operation ports for user authentication and human-computer interaction.

[0034] In practice, the interaction layer serves as the entry point for multi-terminal operations and service feedback within the system. Located at the top of the four-layer architecture, it acts as the interaction carrier between the system and various users. In practical applications, this layer adapts to various operation ports, including enterprise management, employee user, startup service, health service, and regulatory ports. Each port is configured with dedicated functional modules and user interfaces based on user type, while also supporting unified identity authentication methods such as facial recognition, employee ID password, and ID card authentication. Operational data from each port is synchronized to the intelligent agent layer in real time.

[0035] In a preferred embodiment, service feedback data from various operation ports is received through the interaction layer, and this service feedback data is synchronized to the intelligent agent layer. The service feedback data provides a basis for system algorithm optimization and service upgrades.

[0036] Preferably, the employee user interface is optimized specifically for the elderly, with simplified operation processes, and supports functions such as voice input, operation on behalf of children, appointment on behalf of children, and order placement on behalf of children, fully adapting to the operation and usage needs of different users.

[0037] The real-time price comparison AI is primarily used for price control of welfare goods, addressing the technical issues of opaque pricing, procurement premiums, and low efficiency of manual price comparison in existing welfare malls. The AI's modules operate systematically according to a process of data collection, product matching, price comparison, anomaly alerts, dynamic price adjustments, and front-end display. Data between modules is synchronized in real time, forming an automated closed loop for price control.

[0038] In a preferred embodiment, the real-time price comparison agent in the agent layer includes: The multi-source data acquisition module is used to integrate the data layer and collect product price data from various shopping platforms in real time through pre-opened API interfaces. In practical applications, this module utilizes data from the data layer, which continuously collects product prices, promotional information, and coupon deduction amounts from mainstream e-commerce platforms, offline supermarkets, and compliant procurement platforms such as Guocaiyun 24 / 7 through officially open API interfaces. It also simultaneously collects supplier quotations and cost data from the welfare mall to build a comprehensive price database.

[0039] The product normalization matching module is used to normalize and classify product price data belonging to the same category. In practical applications, this module extracts core feature values ​​such as product barcodes, model parameters, brand specifications, and product materials, matches cross-platform product data, normalizes and categorizes products with the same style / parameters, and eliminates interfering data with different specifications, models, and configurations, ensuring the consistency and accuracy of price comparisons.

[0040] The price cleaning and comparison module is used to clean the normalized product price data and then calculate the real-time price difference between the lowest price of each product and the price of each product in the welfare mall. In practical applications, this module performs deduplication, outlier removal, and invalid data removal on the normalized price data, calculates the average price and lowest price of products on external platforms, and compares the cleaned lowest price of external products with the selling price and cost data of products within the mall in real time, automatically calculating the price difference ratio and amount.

[0041] The price fluctuation warning module is used to automatically trigger dynamic price adjustment commands based on the real-time price difference and preset tiered price difference thresholds. In practical applications, this module presets tiered price difference thresholds: when the price of a product in the online store is more than 5% higher than the lowest price across the entire network, a dynamic price adjustment command is automatically triggered; when the price is 3%-5% higher than the lowest price across the entire network, a price fluctuation reminder is automatically pushed to the enterprise management terminal for manual confirmation of whether to adjust the price; for special medical and health products, a dedicated price difference threshold is set according to the compliant guidance price to ensure that the price complies with industry regulatory requirements.

[0042] The dynamic pricing module, upon receiving the dynamic pricing instruction, automatically modifies the price of the product in the employee welfare mall based on the lowest price available. In practical application, after receiving the pricing instruction, this module automatically calls the mall's product management API to modify the selling price, lowering it to the lowest price across the entire network + 2%, and simultaneously updating the product cost ledger and selling price ledger to ensure data traceability. Separate pricing rules are set for employee welfare products to ensure that the selling price of welfare products is 100% lower than the standard price of mainstream self-operated platforms.

[0043] The front-end price comparison module displays the real-time prices of various products within the welfare mall, as well as the average price of each product across shopping platforms. In practical applications, this module synchronizes cross-platform price comparison results, data sources, comparison times, and price differences to the product details page in real time, intuitively displaying the mall's selling price, the lowest external price, and the average external price. It also presents information such as product qualifications, after-sales guarantees, and other details, achieving price transparency.

[0044] The employee benefits matching intelligent agent addresses the technical issues of homogenized recommendations, wasted benefit quotas, and a disconnect between health services and consumer demand in existing benefit malls. The agent's modules operate systematically according to a process of data collection, profile building, quota calculation, precise recommendation, and health-related linkage. Simultaneously, it receives user feedback data to continuously optimize the recommendation algorithm, forming a personalized service closed loop.

[0045] In a preferred embodiment, the employee benefits adaptation agent includes: The multi-dimensional data acquisition module integrates multi-dimensional information about each employee in the data layer, collects shopping behavior data of each employee in the employee benefits mall, and builds a data profile for each employee. In practical applications, this module integrates multi-dimensional information such as employee basic data, consumption data, health data, meal card account data, and welfare account data in the data layer, and synchronously collects employee browsing, favorites, and add-to-cart behavior data in the mall to build a comprehensive data profile for each employee.

[0046] The user profile building module generates unique consumption and health tags for each employee based on their data profile. Based on these tags, a personalized profile is constructed. In practical application, this module uses AI clustering algorithms and tagging technology to analyze multi-dimensional employee data, generating unique consumption and health tags. Consumption tags cover dimensions such as purchase category preferences, purchase frequency, price sensitivity, and brand preferences, while health tags cover dimensions such as physical examination results, chronic disease types, health risks, dietary restrictions, and medical service needs. A complete personalized profile of the employee is built based on these dual tags, and the profile is dynamically updated in real time according to the employee's behavioral data.

[0047] The welfare allowance calculation module is used to read each employee's welfare allowance information in real time, filter the set of redeemable goods for that employee from the welfare mall, and generate the optimal use plan for the welfare allowance based on the employee's personalized profile. In practical applications, this module reads the employee's available welfare balance, meal card balance, welfare points, and other asset information in real time, and intelligently determines the range of goods / services that the employee can redeem based on the mall's product / service prices; at the same time, it generates the optimal use plan for the welfare allowance based on the employee's consumption preferences and health needs, supporting various forms such as full redemption, mixed payment, and combination redemption.

[0048] The personalized recommendation engine module utilizes collaborative filtering and profile matching algorithms to filter and display recommendations from the set of redeemable goods based on employees' personalized profiles. In practical applications, this module, based on collaborative filtering and profile matching algorithms combined with a real-time price comparison agent to filter high-value products, pushes personalized product / service lists to employees, tailored to each individual. Recommendations are prioritized based on employee spending tags, health tags, and welfare allowances, and are displayed through multiple channels including the mall homepage, personal center, and push notifications.

[0049] The health service linkage module is used to push medical and health services from the employee benefits mall to employees based on their health tags. In practical applications, this module deeply integrates employee health tags with medical and health services and offline health kiosks, pushing appropriate health testing services, chronic disease management products, dietary advice, and other content based on the employee's health status. When an employee's health data is abnormal, it automatically pushes health risk warnings and access to medical consultation services, achieving a deep integration of consumer services and health services.

[0050] The intelligent operation and maintenance agent for the e-commerce platform addresses the technical problems of inefficiency, error-proneness, and lack of coordination among various operational stages in existing welfare e-commerce platforms. Each module of this agent operates independently yet shares data, capable of replacing manual labor in most repetitive tasks of daily e-commerce operations, while simultaneously enabling real-time statistics of operational data and real-time monitoring of system status.

[0051] In a preferred embodiment, the intelligent operation and maintenance agent for the e-commerce platform includes: The automated product management module identifies and removes overpriced items from the employee benefits mall based on real-time price differences, and also removes items with a push rate below a threshold based on push data from the employee benefits matching agent. In practical application, this module links price data from the real-time price comparison agent with user demand data from the employee benefits matching agent to automatically remove items with excessive premiums, no price advantage, or low user demand. Simultaneously, it automatically adds best-selling, high-demand, and cost-effective items; and it synchronizes supplier supply data with mall sales data in real time to ensure that product categories and inventory match operational needs.

[0052] The intelligent order processing module receives order and user inquiry data from various operation ports, assigns tracking numbers to the orders, and generates corresponding response information for the user inquiries. In practical applications, this module automatically receives orders for goods purchases, service reservations, etc., from various operation ports, automatically assigns tracking numbers and synchronizes logistics information; intelligently answers common inquiries such as logistics tracking, return and exchange rules, and the use of welfare points; and automatically transfers complex questions to human customer service; it also updates order status in real time to ensure efficient and accurate order processing.

[0053] The intelligent inventory monitoring module is used to monitor inventory levels in real time. When inventory levels fall below a warning threshold, it automatically pushes replenishment reminders to the enterprise management system and suppliers. In practical applications, this module monitors the inventory levels of online store products and health station supplies in real time, sets inventory warning thresholds, and automatically pushes replenishment reminders to the enterprise management system and suppliers when inventory falls below the threshold. It also intelligently calculates replenishment quantities and cycles based on supply chain data to ensure sufficient inventory without excessive backlog.

[0054] The data statistics and analysis module is used to generate multi-dimensional operation reports based on the full-scenario operation data of the welfare mall in the data layer. In practical applications, this module automatically collects full-scenario operation data of the mall and generates multi-dimensional operation reports on a daily, weekly, monthly, and quarterly basis. These reports cover product sales, price comparison data, employee redemption preferences, entrepreneurial service performance, health station operation, user satisfaction, and other report types, providing objective data support for corporate welfare policy adjustments, supplier selection, and entrepreneurial support strategy optimization.

[0055] The system linkage maintenance module is used to monitor the operational status of the multi-agent collaborative enterprise employee benefits system in real time. When an anomaly occurs, it automatically triggers an early warning and pushes it to the technical maintenance end. In practical applications, this module monitors the operational status of each module in the data layer, agent layer, application layer, and interaction layer in real time, as well as official API interface calls and data transmission status. When system failures, interface anomalies, or data transmission interruptions occur, it automatically triggers an early warning and pushes it to the technical maintenance end, while simultaneously initiating an automatic repair program to ensure the stable and continuous operation of the system.

[0056] Preferably, the intelligent operation and maintenance system of the e-commerce platform also includes an automatic commission calculation module, which has preset commission calculation rules, including general rules set by product category, distribution level, and performance tier, as well as preferential rules for service-oriented entrepreneurs; AI automatically matches data such as the entrepreneur's business type, sales performance, service frequency, and customer satisfaction to calculate commissions and additional subsidies; after a user places an order, the entrepreneur's ID is immediately linked, the entrepreneur's account income is updated in real time, a commission calculation ledger is generated, and audit export is supported.

[0057] The multi-scenario collaborative intelligent agent addresses the technical challenges of existing welfare malls, such as single-scenario operation, data silos across scenarios, and inefficient resource scheduling. As the core scheduling unit coordinating cross-scenario operations across the entire system, this intelligent agent enables unified monitoring and compliance verification of data across all scenarios.

[0058] In a preferred embodiment, the multi-scenario collaborative intelligent agent includes: The identity and account linkage module is used to identify an employee's multiple identities based on their personal information in the data layer and associate the corresponding account information for each identity. In practical applications, this module supports intelligent recognition of multiple employee identities, automatically identifying employees who are also corporate employees, university students, or community residents, and automatically associating each identity with corresponding welfare accounts, meal card accounts, entrepreneurial accounts, and distribution accounts, achieving integrated management of identity and account information; it also supports cross-scenario consumption, cross-account settlement, and commission calculation for employees, ensuring the consistency and synchronization of identity and account data.

[0059] The cross-scenario data integration module integrates operational data from various core scenarios within the application layer. This allows the employee benefits matching agent to update individual employee profiles based on the corresponding operational data. It also constructs a full-scenario operational database based on the integrated operational data. In practical applications, this module integrates operational data, employee behavior data, and service record data from various application layer scenarios, synchronizing cross-scenario data to the employee benefits matching agent to continuously iterate and improve individual user profiles. Based on the integrated cross-scenario data, a full-scenario operational database is built, providing comprehensive data support for intelligent resource scheduling and compliance supervision.

[0060] The intelligent resource scheduling module is used to intelligently schedule service resources for various core scenarios based on user needs, geographical distance, and professional suitability. In practical applications, this module uses a three-dimensional algorithm based on user need type, geographical distance, and professional suitability to intelligently schedule various service resources within the platform, such as university student entrepreneurship resources, health kiosk service resources, community merchant resources, and medical and health professional resources, to match users with nearby and suitable service supply; it builds a "15-minute employee / resident service circle" to enable convenient access to welfare consumption, health services, and community services, improving service efficiency and user experience.

[0061] Compliance supervision module, which is used to monitor the data in the full-scenario operation database, automatically identify abnormal data, and push abnormal warnings to the enterprise management terminal and the supervision terminal. In practical applications, this module realizes unified collection, classified storage and multi-dimensional linkage analysis of full-scenario data, and real-time records all-domain information such as employee consumption data, overall enterprise welfare consumption data, entrepreneurship service data, operation data of health huts, and community service data; automatically marks abnormal data through AI (such as employees purchasing the same type of goods frequently in a single month, abnormal commission calculation for entrepreneurial teams, abnormal consumption of health hut supplies inventory, etc.), and pushes abnormal warnings to the enterprise management terminal and the supervision terminal in real time; supports exporting Excel audit reports classified by multiple dimensions, and can be directly connected to the financial and state-owned assets supervision systems to meet the requirements of audit supervision, financial state-owned assets compliance and information creation security.

[0062] The above four agents do not operate independently, but form a collaborative operation system with the multi-scenario collaborative linkage agent as the core dispatching center and the other three agents as the execution units. The four agents achieve real-time data interconnection, decision result synchronization, and two-way instruction transmission through standardized data interfaces, forming a full-scenario and full-process intelligent closed loop of "data support - decision execution - dispatching and overall planning - optimization and feedback".

[0063] According to the second aspect of the present invention, refer to Figure 2 , a method for realizing enterprise employee welfare through multi-agent collaboration is provided, which is applied to the multi-agent collaborative enterprise employee welfare system described in any one of the above, including: Step S11: Connect the real-time price comparison agent, employee welfare adaptation agent, mall intelligent operation and maintenance agent, and multi-scenario collaborative linkage agent to the welfare mall basic system, debug the data interface, and set the operation parameters and compliance thresholds of each agent.

[0064] This step is system deployment and data initialization. First, complete the technical connection of the four agents with the enterprise welfare mall basic system, debug and connect the data interfaces between each module to ensure normal data flow; synchronously complete the basic training of the AI algorithm model to make it adapt to the actual scenario requirements of enterprise welfare management. Then deploy the software and hardware architecture, covering devices such as application servers, data storage servers, and firewalls, and be compatible with the operating system and database. Then import all-dimensional basic data such as enterprise employee basic data, commodity supply chain data, and official API interface permissions of each platform; finally, configure the operation parameters and compliance thresholds of each agent according to the enterprise operation requirements and compliance supervision requirements, including price difference warning thresholds, commission calculation rules, welfare distribution standards, data retention periods, abnormal data determination criteria, etc., to complete the overall system initialization and ensure the normal startup and stable operation of each module.

[0065] Step S12: Collect and store full-scenario related data through the pre-opened API interface.

[0066] After the system initialization is complete, the data layer continuously collects comprehensive data from all dimensions, including basic data of enterprise employees, commodity supply chain data, cross-platform price data, employee consumption and health data, and multi-scenario operation data, through the official open API interfaces of various platforms.

[0067] Preferably, the system can also preprocess the collected data from all scenarios, including cleaning, deduplication, and anonymization. Sensitive data such as employee identity information and health data in the preprocessed data are then encrypted and anonymized before being stored in the corresponding database. A data synchronization mechanism is established to ensure that data in each database is updated in real time, providing comprehensive, accurate, and secure basic data support for decision-making and execution at the intelligent agent level.

[0068] Step S13: The four intelligent agents independently complete their corresponding functions and generate decision results based on the relevant data of the entire scenario, while interacting with each other through a standardized data interface.

[0069] Based on standardized data output from the data layer, the intelligent agent layer activates four intelligent agents to achieve collaborative decision-making and execution: the real-time price comparison agent completes cross-platform price collection, product normalization matching, price cleaning and comparison, price anomaly warning, and dynamic price adjustment, generating a list of high-value products; the employee benefits adaptation agent completes multi-dimensional employee data integration, personalized profile construction, benefit amount calculation, and personalized recommendations; the e-commerce intelligent operation and maintenance agent completes automated product management, intelligent order processing, intelligent inventory monitoring, automatic commission calculation, and operational data statistical analysis; the multi-scenario collaborative linkage agent, as the core scheduling module, coordinates the operation of the other three intelligent agents, completes automatic association of multiple employee identities and accounts, cross-scenario data integration, intelligent resource scheduling, and preliminary compliance supervision of full-scenario data. The four intelligent agents achieve real-time data exchange and result synchronization through standardized data interfaces, forming a decision-making and execution closed loop according to the established collaborative logic, and synchronizing decision results, execution instructions, and resource scheduling information to the application layer.

[0070] Step S14: Based on the decision results of each intelligent agent, construct intelligent applications for multiple core scenarios. Each core scenario is set to run independently. Through multi-scenario collaborative linkage of intelligent agents, cross-scenario data linkage, resource scheduling and business connection are realized.

[0071] Based on the decision results of the intelligent agent layer, the application layer constructs and initiates the intelligent operation of core scenarios. Each scenario module independently completes its own business process, while achieving cross-scenario data linkage, intelligent resource scheduling, and business closure through multi-scenario collaborative intelligent agents.

[0072] Step S15: Record and analyze the operation and maintenance data of each core scenario in real time and in multiple dimensions, automatically label abnormal data and push early warnings; continuously optimize the algorithm parameters of each intelligent agent based on the operation and maintenance data.

[0073] The following is a specific application case 1, taking the closed loop of employee welfare consumption and health kiosk services as an example: After employees log in to the employee user terminal through unified identity authentication, the multi-scenario collaborative intelligent agent automatically verifies the employee's identity and binds their welfare account and health record. Employees can view their welfare balance and the health kiosk appointment portal in real time within the terminal; after employees make an online appointment for chronic disease testing services at the health kiosk, the multi-scenario collaborative intelligent agent intelligently schedules the appointment queue and pushes appointment reminders and navigation information to employees; after employees complete the testing at the store, the test data is automatically synchronized to the online health record through the IoT interface, and the employee welfare matching intelligent agent generates a health report and intervention suggestions based on the test data; based on the health report, employees use welfare points to redeem and purchase AI-recommended chronic disease management products in the mall, and can choose local express delivery provided by e-commerce college student entrepreneurs, or self-pickup at the health kiosk; the welfare verification data, health service data, consumption data, and college student delivery service data generated in this process are all synchronized to the multi-scenario collaborative intelligent agent in real time and included in the unified supervision of the entire scenario.

[0074] The following is a specific application case 2, taking a closed loop of college student entrepreneurship and community service as an example: College students register on the entrepreneurship service platform through a school-specific invitation code / link. AI automatically verifies their school information and student status. The multi-scenario collaborative intelligent agent recommends suitable entrepreneurial directions for college students based on their professional suitability, market demand, and personal interest tags, and automatically activates their entrepreneurial and distribution permissions. Elderly residents in the community register through the senior citizen mode on the employee user platform, entering their personal needs such as care for those living alone, fresh vegetable delivery, and chronic disease follow-up services. The multi-scenario collaborative intelligent agent combines the elderly's needs with the college students' entrepreneurial directions, matching them with nearby college student entrepreneurs based on geographical distance and professional suitability, and pushing the elderly's address and needs information to the entrepreneurs. College students deliver corresponding goods or provide related services from the platform's supply chain according to their needs, and the elderly complete the payment through their welfare account / meal card. After the elderly submit service evaluations within the platform, the mall's intelligent operation and maintenance agent calculates additional entrepreneurial subsidies for the college students based on the evaluation results, service frequency, and delivery time, while simultaneously synchronizing order data and commissions to the college students' entrepreneurial accounts in real time. The startup order data, consumption data, and service evaluation data generated by this process are all recorded in real time by a multi-scenario collaborative intelligent agent and incorporated into the overall data statistics and supervision.

[0075] The following is a specific application case 3, taking the closed-loop linkage between medical students and health stations as an example: After medical students complete their entrepreneurial registration, a multi-scenario collaborative intelligent agent recommends entrepreneurial directions for them based on their professional attributes, such as health station health testing assistance and chronic disease follow-up services, and activates corresponding entrepreneurial permissions. After the health station receives an employee's physical examination appointment order, the multi-scenario collaborative intelligent agent intelligently dispatches the medical student to participate in offline services; the student assists health station staff in completing health testing operations such as blood pressure and blood sugar, and the test data is automatically synchronized to the employee's health record. The employee welfare adaptation intelligent agent generates personalized health suggestions for the employee based on the test data. According to the chronic disease follow-up plan pushed by AI, the student regularly provides online chronic disease consultation services to the employee. The system records the service duration and service effect in real time, forming a complete service process data archive. The platform calculates commissions and subsidies based on the student's service frequency, employee satisfaction, and service effectiveness, and synchronizes them to the student's entrepreneurial account in real time; the data dashboard displays the student's service performance data in real time, providing data support for the optimization and adjustment of their entrepreneurial endeavors.

[0076] The following uses the intelligent management of employee welfare across all scenarios in state-owned enterprises as an example to illustrate the specific implementation of the present invention. This embodiment is only used to explain the present invention: First, the hardware and software are deployed: a fully domestically produced hardware and software architecture is adopted, including 3 domestically produced high-performance application servers, 2 domestically produced data storage servers with master-slave backup architecture (to ensure no data loss), and 1 domestically produced distributed data acquisition server. A domestically produced next-generation firewall, load balancing equipment, and data encryption gateway are also deployed to ensure network security and data transmission stability. On the software side, a domestically produced operating system and database system are used, with the development language based on a Python / Java hybrid architecture, and the algorithm framework relying on a domestically produced artificial intelligence framework. The entire system uses HTTPS encryption for data transmission, and sensitive data is stored in encrypted form, meeting the overall requirements of Level 3 Information Security Protection.

[0077] Through official open API interfaces, we have completed technical integration with the internal human resources system, meal card management system, and fiscal and state-owned asset supervision system of state-owned enterprises, enabling real-time synchronization of employee data, meal card data, and supervision data; we have completed integration with the price comparison API interfaces of mainstream e-commerce platforms, enabling cross-platform price data collection; we have completed integration with the compliance API interfaces of medical and health platforms, enabling the collection of medical product qualification data and health service data; and we have completed integration with the official API interface of the local university student registration management system, enabling automatic verification of the identity and student registration information of college student entrepreneurs.

[0078] Personnel and resource integration: Connect with long-term supply chain merchants of state-owned enterprises, covering categories such as welfare materials, daily necessities, and medical and health products, and complete the entry of commodity supply chain data; connect with offline health kiosks, community experience centers, and surrounding merchants around state-owned enterprises to integrate offline service resources; connect with entrepreneurial teams of college students majoring in economics and management, e-commerce, medical school, social work, design, etc. in local universities to provide sufficient human resources and resource support for the full-scenario implementation.

[0079] In the specific implementation, the system is first initialized: basic data such as the employee's job level, employee number, welfare amount, meal card information, and facial recognition information are imported, as well as supply chain commodity data such as product cost, qualification documents, inventory quantity, and supply cycle are imported; system operating parameters are configured according to the welfare management requirements and compliance regulations of state-owned enterprises, and the price difference threshold is set to automatically adjust the price when the selling price of welfare goods is lower than the self-operated standard price of the shopping platform, and the selling price of goods in the mall is higher than the lowest price on the entire network by 5%, while the 3%-5% range is confirmed manually; the commission calculation rules are set to the commission rate for service-oriented entrepreneurial students is 3%-5% higher than that for pure commodity distribution, and additional subsidies are issued according to the frequency of service and customer satisfaction; the data retention period is set to retain health data for no less than 3 years, and all operational data is retained until the end of the audit cycle, completing the overall system initialization.

[0080] Employee identity and account binding: Employees register and log in to the employee user terminal by using their employee ID and facial recognition. The identity and account linkage sub-module of the multi-scenario collaborative intelligent agent automatically verifies the employee's identity information and links their welfare account and meal card account. Employees can view asset information such as welfare balance, meal card balance, and welfare points in real time within the terminal. After completing unified identity authentication, they can perform cross-scenario operations through this identity on various terminals.

[0081] Full-scenario data collection: After the system starts, the data layer collects employee data such as employee consumption records, physical examination reports, and meal card consumption data 24 / 7 through various official API interfaces; supply chain data such as commodity costs, inventory, and qualifications; price data such as cross-platform commodity prices and promotional information; operational data such as health station equipment testing and material inventory; and entrepreneurial data such as college student entrepreneurship services and order delivery, to build a full-dimensional, real-time updated basic database.

[0082] Multi-agent collaborative operation: The four intelligent agents start collaborative operation based on standardized data at the data layer. The real-time price comparison agent completes cross-platform price collection and dynamic price adjustment to ensure that the price of goods is compliant and competitive. The employee welfare matching agent builds exclusive consumption and health tags for each employee to achieve personalized recommendations of goods and services. The e-commerce intelligent operation and maintenance agent completes product listing and delisting, order processing, and inventory monitoring, and automatically calculates commissions and subsidies for college student entrepreneurial teams without manual intervention. The multi-scenario collaborative linkage agent coordinates the operation of the three intelligent agents, completes cross-scenario resource scheduling, matches employees with nearby health kiosks and community services, matches college student entrepreneurs with corresponding community needs, and achieves real-time monitoring of data across all scenarios.

[0083] Full-scenario business implementation: Employees can complete all-scenario operations such as welfare product consumption, meal card consumption, health station service reservation, community service ordering, and health consultation through the employee user terminal. The senior employee exclusive elder mode simplifies the operation process and supports voice input and operation by children. College student entrepreneurs can complete operations such as entrepreneurial information management, commission inquiry, order receipt, and service feedback through the entrepreneurial service terminal. The system displays their entrepreneurial performance data in real time. Health service personnel can complete operations such as health station reservation management, employee health record management, health report generation, and chronic disease follow-up through the health service terminal. Enterprise managers can complete operations such as welfare policy setting, product management, data supervision, and abnormal data handling through the enterprise management terminal. Regulatory departments can complete operations such as full-scenario data viewing, audit report export, and compliance supervision through the regulatory terminal. Data from all terminals is synchronized in real time, forming a closed loop of full-scenario business.

[0084] System Optimization and Iteration: Based on full-scenario operational data, user service feedback data from various terminals, and abnormal data processing results, the system continuously and automatically optimizes the algorithm parameters and operating parameters of each intelligent agent. For example, it adjusts the recommendation algorithm based on employee consumption preferences, adjusts the price difference threshold based on price fluctuations, and optimizes the resource scheduling algorithm based on the effectiveness of entrepreneurial services. Enterprise managers can update the system operating parameters in real time on the enterprise management terminal according to the adjustment of state-owned enterprise welfare policies and changes in fiscal and state-owned asset supervision requirements, ensuring that the system always meets the requirements of enterprise operation and compliance supervision.

[0085] After being applied to the intelligent management of employee welfare across all scenarios in state-owned enterprises, the implementation examples achieved significant technical results and social value, as detailed below: Significantly improved operational efficiency: Relying on the collaborative operation of four intelligent agents, AI replaces more than 80% of manual operations. Repetitive tasks such as supply chain pricing, commission calculation, order processing, and inventory management are fully automated without human intervention. The operational efficiency of corporate welfare management is improved by more than 60%, and the rate of human error is significantly reduced.

[0086] Significantly improved employee experience: Personalized profiles are built based on multi-dimensional employee data, enabling precise services tailored to each individual, and the utilization rate of welfare allowances has increased by more than 70%; employees can purchase high-quality welfare products without manually comparing prices, and can enjoy health kiosks and community services nearby, increasing employee welfare satisfaction from 70 points to over 95 points.

[0087] Price transparency and cost control: Relying on the real-time price comparison and dynamic price adjustment mechanism across the entire network, the prices of goods are transparent and traceable throughout the process. Employees can intuitively view the cross-platform price comparison results on the product details page; corporate welfare procurement costs are reduced by 15%-20%, achieving precise control over welfare management costs.

[0088] Full-process compliance supervision implemented: The intelligent agent with multi-scenario collaboration enables real-time recording, multi-dimensional analysis and compliance supervision of welfare data throughout the entire process. All data is traceable, and audit reports can be directly connected to the fiscal and state-owned assets supervision system to meet audit and supervision requirements. No data leakage, illegal operation or price premium issues occurred during the implementation period.

[0089] The social value is fully demonstrated: by empowering college student entrepreneurship incubation with AI, it matches college students from various majors with suitable entrepreneurial directions and resource support, effectively increasing college students' entrepreneurial income; by integrating community services and health station resources, it accurately meets the health service needs of the aging population in the community, builds a sustainable development model of "public welfare and commerce", and drives the coordinated development of local consumption and industry.

[0090] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0091] It should be noted that in the description of this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means at least two.

[0092] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0093] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0094] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it includes one or a combination of the steps of the method embodiments.

[0095] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0096] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0097] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0098] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A multi-agent collaborative enterprise employee welfare system, characterized in that, include: The data layer is used to collect and store relevant data from the entire scenario using pre-opened API interfaces; The intelligent agent layer is configured with a real-time price comparison intelligent agent, an employee benefits adaptation intelligent agent, an e-commerce intelligent operation and maintenance intelligent agent, and a multi-scenario collaborative intelligent agent. Each intelligent agent interacts with data through a standardized data interface. Each intelligent agent has an independent built-in AI decision-making and execution module, which is used to obtain data from the data layer and complete the corresponding functions to generate decision results. The real-time price comparison agent is used to calculate the real-time price difference between the prices of each product in the welfare mall and the lowest prices of each product on external shopping platforms, based on the product-related data in the data layer. The employee benefits adaptation intelligent agent is used to build a personalized profile based on the employee's personal information in the data layer, and push products or services from the benefits mall to the employee based on the personalized profile. The intelligent operation and maintenance agent of the mall is used to delist and relist products in the welfare mall, assign logistics tracking numbers to orders, and respond to user inquiries. The multi-scenario collaborative intelligent agent is used to associate the account information corresponding to each employee's identity. Integrate operation and maintenance data from various core application scenarios to build a full-scenario operation database; intelligently schedule service resources for each core scenario; The application layer is used to build intelligent applications for various core scenarios based on the decision results of the intelligent agent layer. Each core scenario has an independent functional unit. Through the collaborative linkage of intelligent agents in multiple scenarios, cross-scenario data linkage, resource scheduling and business connection are realized. The interaction layer connects to various types of operation ports for user authentication and human-computer interaction.

2. The system according to claim 1, characterized in that, The real-time price comparison agents in the agent layer include: The multi-source data acquisition module is used to integrate the data layer and collect product price data from various shopping platforms in real time through pre-opened API interfaces; The product normalization matching module is used to normalize and classify the product price data that belong to the same category of products; The price cleaning and comparison module is used to clean the normalized product price data and then calculate the real-time price difference between the lowest price of each product and the price of each product in the welfare mall. The price fluctuation warning module is used to automatically trigger dynamic price adjustment instructions based on the real-time price difference and the preset tiered price difference threshold; The dynamic price adjustment module is used to automatically adjust the price of the product in the welfare mall based on the lowest price of the product after receiving the dynamic price adjustment instruction; The front-end price comparison display module is used to display the real-time prices of each product in the welfare mall, as well as the average price of each product on the shopping platform.

3. The system according to claim 1, characterized in that, The employee benefits adaptation agent includes: The multi-dimensional data acquisition module is used to integrate the multi-dimensional information of each employee in the data layer, collect the shopping behavior data of each employee in the welfare mall, and build a data profile for each employee. The user profile building module is used to generate exclusive consumption tags and health tags for each employee based on their data profile; and to build a personalized profile for each employee based on their consumption tags and health tags. The welfare allowance calculation module is used to read the welfare allowance information of each employee in real time, filter the set of redeemable items for that employee from the welfare mall, and generate the optimal use plan for the welfare allowance based on the employee's personalized profile. The personalized recommendation engine module is used to use collaborative filtering and profile matching algorithms to filter and display recommendation results from the set of redeemable goods based on the employee's personalized profile. The health service linkage module is used to push medical and health services from the benefits mall to employees based on their health tags.

4. The system according to claim 1, characterized in that, The intelligent operation and maintenance agent for the e-commerce platform includes: The automated product management module is used to identify and remove overpriced products from the welfare mall based on the real-time price difference of each product, and to identify and remove products with a push rate below the threshold from the welfare mall based on the push data of the employee welfare adaptation intelligent agent. The intelligent order processing module is used to receive order and user inquiry data from various operation ports, assign logistics tracking numbers to the orders, and generate response information corresponding to the user inquiry data. The intelligent inventory monitoring module is used to monitor inventory levels in real time. When the inventory level falls below the warning threshold, it automatically pushes replenishment reminders to the enterprise management system and suppliers. The data statistics and analysis module is used to generate multi-dimensional operation reports based on the full-scenario operation data of the welfare mall in the data layer; The system linkage maintenance module is used to monitor the operation status of the multi-agent collaborative enterprise employee welfare system in real time. When an anomaly occurs, it automatically triggers an early warning message and pushes it to the technical maintenance terminal.

5. The system according to claim 1, characterized in that, The multi-scenario collaborative intelligent agent includes: The identity and account linkage module is used to identify an employee's multiple identities based on the employee's personal information in the data layer and associate the account information corresponding to each identity. The cross-scenario data integration module is used to integrate operation and maintenance data from various core scenarios in the application layer, so that the employee benefits matching agent can update the personalized profile of employees based on the corresponding operation and maintenance data; it is also used to build a full-scenario operation database based on the integrated operation and maintenance data. The intelligent resource scheduling module is used to intelligently schedule service resources for various core scenarios based on user needs, geographical distance, and professional suitability. The compliance and supervision module is used to monitor the data in the full-scenario operation database, automatically identify abnormal data, and push abnormal warnings to the enterprise management end and the supervision end.

6. The system according to claim 1, characterized in that, The data layer is also used to encrypt and desensitize sensitive data in the collected data related to the entire scene.

7. The system according to claim 1, characterized in that, The application layer is built with multiple core scenarios, including: corporate welfare procurement scenario, employee welfare consumption scenario, state-owned enterprise meal card integration scenario, medical and health service scenario, college student entrepreneurship incubation scenario, community aging service scenario, offline health station operation scenario, and full-process data supervision scenario.

8. A method for implementing enterprise employee benefits through multi-agent collaboration, characterized in that, The multi-agent collaborative enterprise employee benefits system as described in any one of claims 1 to 7 includes: Connect the real-time price comparison intelligent agent, the employee benefits adaptation intelligent agent, the mall intelligent operation and maintenance intelligent agent, and the multi-scenario collaborative intelligent agent with the welfare mall basic system, debug the data interface, and set the operating parameters and compliance thresholds of each intelligent agent. Collect and store relevant data from all scenarios through pre-opened API interfaces; Based on the relevant data of the entire scenario, the four intelligent agents independently complete their corresponding functions, generate decision results, and interact with each other through a standardized data interface. Intelligent applications for various core scenarios are built based on the decision results of each intelligent agent. Each core scenario is set to run independently. Cross-scenario data linkage, resource scheduling and business connection are achieved through multi-scenario collaborative linkage of intelligent agents. Real-time recording and multi-dimensional analysis of operation and maintenance data for each core scenario; automatic labeling of abnormal data and push notifications; continuous optimization of algorithm parameters for each intelligent agent based on operation and maintenance data.

9. The method according to claim 8, characterized in that, After collecting relevant data across the entire scenario through pre-opened API interfaces, it also includes: Preprocess the collected data from all scenarios, including cleaning, deduplication, and desensitization. The preprocessed, all-scenario-related data is encrypted and stored in the corresponding database.

10. The method according to claim 8, characterized in that, Also includes: The interaction layer receives service feedback data from various operation ports and synchronizes the service feedback data to the intelligent agent layer.

Images