Annuity payment business processing method and device based on large model and MCP protocol

Abstract

The application provides an annuity payment business processing method and device based on a large model and an MCP protocol, relates to the technical field of artificial intelligence, and comprises the following steps: receiving an annuity payment business request, wherein the annuity payment business request comprises payment application materials; obtaining current effective interface protocol metadata from an MCP server dynamically through the MCP protocol; inputting the payment application materials and the interface protocol metadata into a business processing intelligent agent based on a large model, performing multi-modal analysis and extraction on the payment application materials through the business processing intelligent agent to obtain first structured business data; determining a target interface protocol and field requirements matched with a current business stage according to the interface protocol metadata; performing data assembly on the first structured business data according to the target interface protocol and the field requirements, obtaining second structured business data packets conforming to the interface specification of an account management system, and sending the second structured business data packets to the corresponding account management system to complete the annuity payment business interaction.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and more particularly to a method and apparatus for processing annuity payment transactions based on a large model and the MCP protocol. Background Technology

[0002] In the current operational model of pension (especially enterprise annuity) contribution business, the trustee institution needs to initiate the contribution business with the account management institution (such as a bank). (Reference) Figure 1 This is a schematic diagram illustrating the system relationships in existing payment service scenarios. For example... Figure 1 As shown, the specific process involves the entrusted institution's staff collecting payment application materials from the client (usually including the payer's identity information, payment institution information, payment amount, etc., provided in Excel, Word, or image formats), entering them into the entrusted institution's entrusted system 100, and then the entrusted system generating a data file in a specific format, which interacts multiple times with the accounting management institution's accounting management system 200 to complete a series of processes such as business acceptance review and payment receipt entry.

[0003] However, the current model has significant technical flaws: different accounting systems (currently, the entrusted system may need to interface with up to 19 different bank accounting systems) have varying requirements for the business data formats, field requirements, and validation rules submitted by the entrusted system. Due to protocol heterogeneity, the entrusted system needs to develop and maintain a separate set of interface adaptation code for each accounting system, resulting in extremely high development and maintenance costs. Changes in business rules and regulatory requirements lead to frequent protocol changes, causing frequent upgrades and updates to the accounting system's interface protocols (including fields, validation logic, etc.). Each change requires the entrusted system to invest significant manpower in code modification, integration testing, and system upgrades, resulting in cumbersome processes, delayed responses, and severely impacting business efficiency. Existing solutions are typically based on hard-coding, embedding business processes and interface protocols in the source code. The system cannot automatically understand and process unstructured payment application materials (such as document images and documents), still requiring manual identification, input, and validation, resulting in low automation, error-proneness, and inefficiency.

[0004] In summary, there is an urgent need for an annuity payment processing solution that can overcome the above-mentioned shortcomings, automatically adapt to different accounting system interfaces, dynamically respond to protocol changes, and intelligently handle multimodal business materials. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention proposes a method and apparatus for processing annuity payment transactions based on a large model and the MCP protocol.

[0006] In a first aspect of this invention, a method for processing annuity contribution transactions based on a large model and the MCP protocol is proposed. This method is executed by a trustee system in a pension contribution transaction processing scenario, and includes: Receive an annuity payment request, wherein the annuity payment request includes payment application materials; Dynamically obtain the currently active interface protocol metadata from the MCP server registered in the accounting system via the MCP protocol; The payment application materials and the interface protocol metadata are input into the business processing intelligent agent based on the large model. The business processing intelligent agent performs multimodal parsing on the payment application materials to extract the first structured business data. Based on the interface protocol metadata, determine the target interface protocol and field requirements that match the current business stage; According to the target interface protocol and field requirements, the first structured business data is assembled to obtain a second structured business data packet that conforms to the accounting system interface specification, and the second structured business data packet is sent to the corresponding accounting system to complete the annuity payment business interaction.

[0007] In a second aspect of this invention, a method for processing annuity contribution transactions based on a large model and the MCP protocol is proposed. This method is executed by an account management system in a pension contribution transaction processing scenario, and includes: Register and configure the MCP server, and provide the currently effective interface protocol metadata to the entrusted system through the MCP server; The system receives a second structured business data packet sent by the entrusted system; wherein the second structured business data packet is generated by the entrusted system after performing multimodal parsing and data assembly on the payment application materials according to the interface protocol metadata, and conforms to the interface specification of the accounting management system. Perform business processing based on the second structured business data packet and generate business processing results; The business processing result is returned to the entrusted system.

[0008] In a third aspect of this invention, an annuity contribution processing device based on a large model and the MCP protocol is proposed. This device is deployed in a trustee system for pension contribution processing scenarios, and includes: The business request receiving module is used to receive annuity payment business requests, wherein the annuity payment business request includes payment application materials; The interface protocol acquisition module is used to dynamically obtain the currently effective interface protocol metadata from the MCP server registered in the accounting system via the MCP protocol. The data parsing module is used to input the payment application materials and the interface protocol metadata into the business processing intelligent agent based on the large model, and to extract the first structured business data by performing multimodal parsing on the payment application materials through the business processing intelligent agent; The protocol processing module is used to determine the target interface protocol and field requirements that match the current business stage based on the interface protocol metadata. The data assembly and interaction module is used to assemble the first structured business data according to the target interface protocol and field requirements to obtain a second structured business data packet that conforms to the accounting system interface specification, and send the second structured business data packet to the corresponding accounting system to complete the annuity payment business interaction.

[0009] In a fourth aspect of this invention, an annuity contribution processing device based on a large model and the MCP protocol is proposed. This device is deployed in an account management system for pension contribution processing scenarios, and includes: The configuration module is used to register and configure the MCP server, and provide the currently effective interface protocol metadata to the entrusted system through the MCP server; The data receiving module is used to receive the second structured business data packet sent by the entrusted system; wherein the second structured business data packet is generated by the entrusted system after performing multimodal parsing and data assembly on the payment application materials according to the interface protocol metadata, and conforms to the interface specification of the accounting management system; The business processing module is used to perform business processing based on the second structured business data packet and generate business processing results; The result return module is used to return the business processing result to the entrusted system.

[0010] In a fifth aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement an annuity payment processing method based on a large model and the MCP protocol.

[0011] In a sixth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements an annuity payment business processing method based on a large model and the MCP protocol.

[0012] In a seventh aspect of the present invention, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements an annuity payment business processing method based on a large model and the MCP protocol.

[0013] The annuity payment processing method and apparatus proposed in this invention, based on a large model and the MCP protocol, dynamically acquires interface protocol metadata through the MCP protocol, enabling the entrusted system to perceive changes in the accounting system interface in real time. This allows for automatic adaptation without manual intervention to modify the code, significantly reducing system maintenance costs and upgrade cycles. Simultaneously, this invention introduces a large model as a business processing intelligent agent, leveraging its powerful multimodal understanding capabilities to process non-standardized payment application materials, improving the accuracy and generalization ability of data extraction, and effectively enhancing the processing efficiency and automation level of annuity payment transactions. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the system relationships in the existing payment business scenario.

[0016] Figure 2 This is a schematic diagram of the annuity payment business processing method based on a large model and MCP protocol according to an embodiment of the present invention.

[0017] Figure 3 This is a schematic flowchart of multimodal analysis according to an embodiment of the present invention.

[0018] Figure 4 This is a schematic diagram of the data assembly process according to an embodiment of the present invention.

[0019] Figure 5 This is a schematic diagram of the compatibility verification and event stream storage process according to an embodiment of the present invention.

[0020] Figure 6 This is a schematic diagram of the annuity payment business processing method based on a large model and MCP protocol according to another embodiment of the present invention.

[0021] Figure 7 This is a schematic diagram of the architecture of an annuity payment processing device based on a large model and MCP protocol according to an embodiment of the present invention.

[0022] Figure 8 This is a schematic diagram of the annuity payment processing device architecture based on a large model and the MCP protocol, according to another embodiment of the present invention.

[0023] Figure 9 This is a schematic diagram of a computer device structure according to an embodiment of the present invention.

[0024] Figure 10This is a schematic diagram of the annuity payment process using existing technology.

[0025] Figure 11 This is a schematic diagram of the architecture of annuity payment business processing according to a specific embodiment of the present invention. Detailed Implementation

[0026] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are given merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make this disclosure more thorough and complete, and to fully convey the scope of this disclosure to those skilled in the art.

[0027] Those skilled in the art will recognize that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, this disclosure can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0028] According to an embodiment of the present invention, a method and apparatus for processing annuity payment business based on a large model and the MCP protocol are proposed, which relates to the field of artificial intelligence technology.

[0029] The principles and spirit of the present invention will be explained in detail below with reference to several representative embodiments.

[0030] Figure 2 This is a schematic diagram of a method for processing annuity payment transactions based on a large model and the MCP protocol, according to an embodiment of the present invention. Figure 2 As shown, this method is executed by a trustee system based on the pension payment business processing scenario, and includes: S201, Receive an annuity payment request, wherein the annuity payment request includes payment application materials; S202, dynamically obtain the currently effective interface protocol metadata from the MCP server registered in the accounting system through the MCP protocol; S203, input the payment application materials and the interface protocol metadata into the business processing intelligent agent based on the large model, and use the business processing intelligent agent to perform multimodal parsing on the payment application materials to extract the first structured business data; S204, Determine the target interface protocol and field requirements that match the current business stage based on the interface protocol metadata; S205, according to the target interface protocol and field requirements, the first structured business data is assembled to obtain a second structured business data packet that conforms to the accounting system interface specification, and the second structured business data packet is sent to the corresponding accounting system to complete the annuity payment business interaction.

[0031] The annuity payment processing method proposed in this invention, based on a large model and the MCP protocol, fundamentally solves the system rigidity problem caused by protocol heterogeneity and changes in the traditional model by dynamically acquiring interface protocol metadata and using a business processing intelligent agent based on the large model. In application scenarios, the entrusted system does not need to pre-configure and hard-code interface protocols for each accounting system. Instead, it can pull the latest protocol specifications in real time and, combined with the intelligent parsing and assembly capabilities of the large model, automatically transform unstructured business materials into structured data packets that conform to specific interface specifications. The overall solution can achieve a qualitative leap in business processing from fixed coding and manual adaptation to dynamic perception and intelligent adaptation, significantly improving system flexibility, maintainability, and processing efficiency.

[0032] In one embodiment, for S201, an annuity payment service request is received, wherein the annuity payment service request includes payment application materials.

[0033] Specifically, the payment application materials include at least image materials and payment report documents.

[0034] In practical applications, enhanced multimodal OCR technology can be used to support high-precision recognition of common document images such as ID cards, household registration books, and business licenses. It can also automatically process payment reports in Word and Excel formats based on a document semantic parsing engine, accurately extract key fields such as "payer's ID number", "company code", and "amount due", and automatically map them to the internal unified data model.

[0035] In one embodiment, for S202, the currently active interface protocol metadata is dynamically obtained from the MCP server registered in the accounting system via the MCP protocol.

[0036] Specifically, a query request is initiated to the resource list interface of the MCP server to obtain the interface protocol metadata sent by the MCP server; wherein, the interface protocol metadata includes at least field requirements, verification rules and version metadata.

[0037] The MCP protocol defines a standardized and programmable protocol acquisition mechanism for obtaining metadata. This allows queries to be sent to a specified resource list interface, retrieving metadata containing details such as field requirements and validation rules in structured data (e.g., JSON). This makes the exchange of protocol information machine-readable and parsable. This step provides accurate and reliable input for subsequent automatic validation and data assembly by the business processing agent, serving as a key technological guarantee for realizing contract-driven automated processes.

[0038] Before dynamically obtaining interface protocol metadata from the MCP server of the accounting system via the MCP protocol, the following is also included: The entrusted system accesses the MCP server through an authentication mechanism, wherein the MCP server only allows authorized entrusted systems to access it; When the accounting system does not provide an MCP service address, the entrusted system builds its own MCP service address and modifies the interface protocol metadata based on the communication results provided by the accounting system administrators.

[0039] The authentication and authorization mechanisms enhance the security and deployment flexibility of the solution, ensuring that only legitimate entrusted systems can access sensitive interface protocol metadata, preventing unauthorized access, and safeguarding business boundaries. Secondly, a backup solution (self-built and offline configuration) is provided when the accounting system does not offer an MCP server, enhancing the compatibility and feasibility of this invention in practical applications. This eliminates the need for the accounting system to provide standardized services, broadening the overall solution's applicability.

[0040] In one embodiment, for S203, the payment application materials and the interface protocol metadata are input to a business processing agent based on a large model, and the business processing agent performs multimodal parsing on the payment application materials to extract the first structured business data.

[0041] For details, please refer to Figure 3 The specific process of multimodal analysis is as follows: S301, the image material is processed by optical character recognition, and the payment report document is processed by a document semantic parsing engine to extract the payer's identity information, unit information and payment amount information; S302, the extracted information is mapped to a unified data model to generate the first structured business data.

[0042] This invention employs a multimodal parsing approach, combining optical character recognition (OCR) technology with a document semantic parsing engine. It can cover various unstructured materials commonly used in pension payment transactions (e.g., image documents and forms), achieving automated and high-precision information extraction across all material types. The extracted information is mapped to a unified data model, establishing a standardized and consistent data foundation for subsequent data processing. This avoids processing errors caused by inconsistent data formats, improving the robustness and accuracy of the entire process.

[0043] The business processing intelligent system based on the large model embeds a state machine, which drives the annuity payment business processing process; wherein, the state machine includes at least an information entry and inspection stage, a business acceptance stage, and a business completion stage.

[0044] Information entry and verification phase: During the information entry check phase, the business processing intelligent agent performs a completeness and consistency check on the payment application materials according to the verification rules. Business acceptance stage: After the verification is passed, the process proceeds to the business acceptance stage. The first structured business data is assembled according to the target interface protocol and field requirements to obtain a second structured business data packet that conforms to the accounting system interface specification. The second structured business data packet is then sent to the corresponding accounting system. Business completion stage: Upon receiving a successful accounting confirmation from the accounting system, the business status is advanced to the business completion stage, and all process data and execution logs are stored in the database.

[0045] This invention introduces a state machine to drive business processes, modularizing and state-based the complex annuity payment process, making the agent's operational logic clear, orderly, and fault-tolerant. By dividing the process into clearly defined stages (checking, accepting, and completing), and executing specific tasks at each stage (such as verification, assembly and sending, and persistence), refined management of business logic is achieved. The state machine mechanism ensures that business processes can only proceed according to predetermined order and conditions, preventing logical chaos and facilitating state rollback or retries in the event of anomalies, thus improving system reliability and automation.

[0046] In one embodiment, for S204, the target interface protocol and field requirements that match the current business stage are determined based on the interface protocol metadata.

[0047] To ensure seamless integration between the entrusted system and the accounting system, the latest interface field definitions, validation rules, and version metadata are dynamically retrieved in real time from the accounting system's MCP server using the standard MCP (Managed Communication Protocol). Unlike static configuration, this dynamic retrieval method features "contract adaptation": when the accounting system upgrades its interfaces (e.g., adding an "electronic receipt number" field to the 0209 interface), the intelligent agent can automatically synchronize the new specifications and update the internal prompt word template and data dictionary mapping before the next business trigger, avoiding business interruptions caused by interface changes.

[0048] In one embodiment, for S205, the first structured business data is assembled according to the target interface protocol and field requirements to obtain a second structured business data packet that conforms to the accounting system interface specification.

[0049] refer to Figure 4 The specific process for data assembly is as follows: S401, Based on the business knowledge graph, perform semantic alignment and mapping between the fields and field requirements in the first structured business data; S402, perform calculations or format conversions on the data in the first structured business data according to predefined conversion rules; S403, assemble the processed fields and data to obtain the second structured business data packet, and perform desensitization or encryption processing on the sensitive fields in the second structured business data packet.

[0050] The detailed data assembly process mainly includes semantic alignment based on business knowledge graphs to resolve issues of field naming and semantic differences between different systems, achieving intelligent field mapping. Secondly, data calculation and format conversion are performed using predefined transformation rules to ensure that the source data meets the business logic requirements of the target interface. Finally, sensitive fields are anonymized or encrypted, embedding data security and privacy protection mechanisms while achieving automated interaction, making the solution meet the high security requirements of financial businesses.

[0051] Furthermore, to ensure stable business interaction processing, compatibility checks can be performed before sending, as per [reference needed]. Figure 5 The specific process is as follows: S501, based on the version metadata in the interface protocol metadata, verify the compatibility between the field version of the second structured business data packet and the version of the target interface protocol; S502, after the compatibility verification is passed, the second structured business data is sent to the corresponding accounting system to complete the annuity payment business interaction; wherein, the payment application materials, the first structured business data, the interface interaction logs and the business response results generated in the entire process of this business are converted into a structured event stream with timestamps and operation traceability identifiers and stored.

[0052] This invention verifies the compatibility between the data packet version and the interface protocol version, enabling the detection of potential protocol mismatch issues before interaction, avoiding interface call failures due to version inconsistencies, and improving the success rate of interaction and system stability.

[0053] The entire process data is transformed into a structured event stream with timestamps and traceability identifiers for storage, realizing complete data archiving. More importantly, it provides a high-quality, highly structured data foundation for business auditing, problem investigation, process backtracking, and subsequent data analysis and model optimization, realizing the observability and traceability of the processing process.

[0054] In another embodiment, reference Figure 6 This is a schematic diagram of a pension payment processing method based on a large model and the MCP protocol, according to another embodiment of the present invention. Figure 6 As shown, this method is implemented in an accounting system based on a pension contribution processing scenario, and includes: S601, Register and configure the MCP server, and provide the currently effective interface protocol metadata to the entrusted system through the MCP server; S602, receive the second structured business data packet sent by the entrusted system; wherein, the second structured business data packet is generated by the entrusted system after performing multimodal parsing and data assembly on the payment application materials according to the interface protocol metadata, and conforms to the interface specification of the account management system; S603, Perform business processing based on the second structured business data packet and generate business processing results; S604, return the business processing result to the entrusted system.

[0055] It should be noted that although the operation of the method of the present invention has been described in a specific order in the above embodiments and figures, this does not require or imply that the operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0056] After introducing the method of exemplary embodiments of the present invention, the following references are made. Figures 7 to 8An annuity payment processing device based on a large model and the MCP protocol, according to an exemplary embodiment of the present invention, is described.

[0057] The implementation of the annuity payment processing device based on the large model and MCP protocol can refer to the implementation of the above method, and the repeated parts will not be described again. The term "module" or "unit" used below can be a combination of software and / or hardware to achieve a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0058] Based on the same inventive concept, this invention also proposes an annuity payment processing device based on a large model and the MCP protocol, such as... Figure 7 As shown, the device is deployed in a trustee system for pension payment processing scenarios, and includes: The business request receiving module 710 is used to receive annuity payment business requests, wherein the annuity payment business request includes payment application materials; The interface protocol acquisition module 720 is used to dynamically obtain the currently effective interface protocol metadata from the MCP server registered in the accounting system via the MCP protocol. The data parsing module 730 is used to input the payment application materials and the interface protocol metadata into the business processing intelligent agent based on the large model, and to extract the first structured business data by performing multimodal parsing on the payment application materials through the business processing intelligent agent; Protocol processing module 740 is used to determine the target interface protocol and field requirements that match the current business stage based on the interface protocol metadata; The data assembly and interaction module 750 is used to assemble the first structured business data according to the target interface protocol and field requirements to obtain a second structured business data packet that conforms to the accounting system interface specification, and send the second structured business data packet to the corresponding accounting system to complete the annuity payment business interaction.

[0059] refer to Figure 8 This is a schematic diagram of the annuity payment processing device architecture based on a large model and MCP protocol, according to another embodiment of the present invention. Figure 8 As shown, the device is deployed in the accounting system for pension payment processing scenarios, and includes: Configuration module 810 is used to register and configure the MCP server, and to provide the currently effective interface protocol metadata to the entrusted system through the MCP server; The data receiving module 820 is used to receive the second structured business data packet sent by the entrusted system; wherein the second structured business data packet is generated by the entrusted system after performing multimodal parsing and data assembly on the payment application materials according to the interface protocol metadata, and conforms to the interface specification of the accounting management system; The business processing module 830 is used to perform business processing based on the second structured business data packet and generate business processing results; The result return module 840 is used to return the business processing result to the entrusted system.

[0060] It should be noted that although several modules of the annuity payment processing device based on the large model and MCP protocol are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules described above can be embodied in one module. Conversely, the features and functions of one module described above can be further divided and embodied by multiple modules.

[0061] Based on the aforementioned inventive concept, such as Figure 9 As shown, the present invention also proposes a computer device 900, including a memory 910, a processor 920, and a computer program 930 stored in the memory 910 and executable on the processor 920. When the processor 920 executes the computer program 930, it implements the aforementioned annuity payment processing method based on the large model and MCP protocol.

[0062] Based on the aforementioned inventive concept, this invention proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned annuity payment processing method based on the large model and MCP protocol.

[0063] Based on the aforementioned inventive concept, this invention proposes a computer program product, which includes a computer program that, when executed by a processor, implements an annuity payment processing method based on a large model and the MCP protocol.

[0064] The following describes the annuity payment processing method and apparatus based on the large model and MCP protocol of the present invention with reference to a specific embodiment.

[0065] refer to Figure 10 This is a schematic diagram of the annuity payment process using existing technology. For example... Figure 10 As shown, targeted code development is carried out based on business processes and interface protocols, embedding the business processes and interface protocols into the source code. The main business interaction process between the entrusted system and the accounting system is as follows, which includes multiple interface protocols: The entrusted system pushes business acceptance and review data (0201, 0202 interfaces) to the accounting system; The accounting system sends the acceptance and review results (0203, 0204 interfaces) back to the entrusted system. The entrusted system pushes payment receipt data (0207 interface) to the accounting system; The accounting system sends payment receipt data (0208, 0209 interfaces) back to the entrusted system. All of the above interface protocol fields are hard-coded. If the accounting system modifies the field requirements, the entrusted system must modify the relevant code accordingly, complete the interface test with the accounting system, and then upgrade and go live.

[0066] Based on this, this application improves the entrusted system, the accounting system, and the corresponding processing procedures.

[0067] For the accounting system, registration and configuration MCP services are provided.

[0068] The accounting system must register and configure the MCP service on an MCP service management platform (e.g., MSE Nacos). The specific steps are as follows: Log in to the MSE Registration and Configuration Center console.

[0069] Select region and instance.

[0070] Configure according to the service type, selecting standard MCP service, HTTP-converted MCP service, or dynamically registered MCP service.

[0071] Configure the service IP, port, and access path.

[0072] The MCP server needs to expose resources and list interfaces, listing the field requirements of all interface protocols and the text description of the returned results.

[0073] You need to configure the service path and SSE address, and ensure that the URL is publicly accessible.

[0074] If the accounting system does not provide a service address, the entrusted system can build its own MCP service to meet business needs and modify the required fields of the agreement based on the results of offline communication with the accounting system administrators.

[0075] For the entrusted system, a business processing intelligent agent based on a large-scale artificial intelligence model is built. This agent deeply integrates multimodal perception, structured knowledge reasoning, and workflow engine capabilities to construct an end-to-end intelligent business closed loop. The core architecture consists of four units and a dynamically state-driven business processing intelligent agent, referencing... Figure 11 The specific architecture is as follows: The payment application material processing unit is used for multimodal semantic understanding and structured parsing.

[0076] Employing enhanced multimodal OCR technology, it not only supports high-precision recognition of common document images such as ID cards, household registration books, and business licenses, but also integrates a document semantic parsing engine that can automatically process payment reports in Word and Excel formats, accurately extract key fields such as "payer's ID number," "company code," and "amount due," and automatically map them to the internal unified data model.

[0077] The query interface unit is used to obtain the dynamic contract driven by the MCP protocol.

[0078] To ensure seamless integration with the accounting system, the latest interface field definitions, validation rules, and version metadata are dynamically retrieved in real time from the accounting system's MCP server using the standard MCP (Managed Communication Protocol). Unlike static configuration, the dynamic retrieval method has "contract adaptive" capabilities: when the accounting system upgrades its interfaces (for example, adding the "electronic receipt number" field to the 0209 interface), the agent can automatically synchronize the new specifications and update the internal prompt word template and data dictionary mapping relationship before the next business trigger, avoiding business interruptions caused by interface changes.

[0079] Interface encapsulation unit, used to assemble intelligent data guided by knowledge graph.

[0080] Data assembly is achieved using the system's built-in payment business data dictionary knowledge base. This knowledge base is organized in the form of a lightweight knowledge graph, including field semantics, source paths, transformation rules, and compliance constraints. In addition, the system supports dynamic desensitization and encryption encapsulation of sensitive fields (such as ID card numbers) to ensure secure transmission.

[0081] Storage unit, used for event-based structured and persistent storage.

[0082] During the business completion phase, the system transforms all raw materials, intermediate parsing results, interface interaction logs, and final accounting vouchers generated throughout the entire process into a structured event stream with timestamps and operation traceability. Data is stored in the entrusted system database using a three-level key-value pair: "Business ID + Stage + Operation Type," and a dual-channel storage mechanism of full-text indexing and vectorized embedding is established—supporting both traditional SQL queries and providing underlying support for subsequent intelligent customer service or audit analysis based on RAG (Retrieval Enhanced Generation).

[0083] The entrusted system's business processing intelligent agent is equipped with a state machine-driven autonomous decision-making engine. Using a large-scale artificial intelligence model as its core inference engine, the agent orchestrates AI workflows through an AI platform and embeds a three-stage state machine to achieve autonomous business progression and self-healing from anomalies.

[0084] Information entry and verification phase: After the operator uploads the materials, the business processing agent calls the "Payment Application Material Processing Unit" and the "Query Interface Unit" in parallel to dynamically generate a compound prompt containing a "List of Required Fields," a "Format Specification Example," and "Historical Similar Cases." The large model performs dual checks for completeness and consistency based on this. If there are any missing or conflicting information, the system will prompt for supplementary submission.

[0085] During the business acceptance phase: After the materials are complete, the business processing agent, based on the account management interface contract, schedules the "interface encapsulation unit" to generate data packets conforming to multiple interface specifications from 0201 to 0209, and initiates asynchronous calls according to business priority. If an interface temporarily fails due to network latency or service anomalies, the system activates the "interface circuit breaker and retry strategy": temporarily storing the current state and automatically resuming transmission after service recovery, avoiding overall rollback and ensuring high throughput and high availability.

[0086] Business completion stage: After receiving the successful accounting confirmation from the accounting system, the business processing agent immediately triggers the "storage unit" to complete the data write-to-disk and simultaneously pushes the business completion notification to the operator's workbench.

[0087] The entrusted system automatically records the processing time, call chain, and model decision logs for this business, which are used to continuously optimize the prompt word engineering and state transition strategy.

[0088] This invention, by deeply coupling the generalization and understanding capabilities of large models with vertical domain skills, not only achieves "unmanned" processing of payment services, but also constructs an interpretable, traceable, and evolvable intelligent service foundation, laying a solid foundation for future expansion.

[0089] Regarding security and access control, the entrusted system accesses the accounting system's field configurations through the MCP authentication mechanism (e.g., OAuth2). The accounting system's MCP server only allows authorized entrusted systems to access field resources.

[0090] Taking a payment transaction initiated through the entrusted system interface as an example, the operator uploads application materials including ID card, business license, and payment report (Word or Excel format). At this time, the business process processing agent is immediately activated and enters the "information entry check stage" according to the embedded three-stage state machine.

[0091] The business processing agent invokes the "Payment Application Material Processing Unit" and the "Query Interface Unit" in parallel. The Payment Application Material Processing Unit utilizes an enhanced multimodal OCR and document semantic parsing engine to accurately extract key fields such as "payer's ID number," "company code," and "amount due" from images and documents, mapping them to a unified data model. The "Query Interface Unit" connects in real-time to the MCP server of the accounting system via the standard MCP protocol (using JSON-RPC as the transmission medium) to retrieve metadata for currently effective interface contracts, including field definitions, validation rules, and version information for interfaces 0201 to 0209. For example, fields such as transaction sequence number, company name, payment amount, and due date are shown in Table 1.

[0092] Table 1 Field Names and Descriptions

[0093] Based on the above dual inputs, the business processing intelligent agent dynamically generates compound prompts, which are then checked for completeness and consistency by an AI big data model. For example, it verifies whether the ID card is valid, whether the payment amount meets the prescribed limit, and whether the enterprise code matches the organization code. If anything is missing (e.g., a verification form without the company seal), the entrusted system not only prompts for resubmission but also intelligently recommends the most likely missing document type based on historical cases.

[0094] Once all materials are complete and verified, the business status automatically advances to the "business acceptance stage." At this point, the business processing intelligent agent schedules the "interface encapsulation unit," which performs semantic alignment and intelligent assembly based on the built-in payment business knowledge graph.

[0095] For example, if the accounting interface requires "contribution base" while the original materials only provide "total monthly salary," the entrusted system will automatically complete the logical conversion based on preset salary rules (such as rounding after deducting tax-exempt items). It can also accurately map semantically equivalent fields such as "document number" and "ID card number" that have different names. Simultaneously, sensitive fields (such as ID card number and social security number) undergo dynamic desensitization and encryption before encapsulation to ensure compliant and secure transmission.

[0096] After encapsulation, the business processing agent asynchronously calls the corresponding interface of the accounting system (e.g., interface 0203, 0209, etc.) according to business priority. After the accounting system returns a successful entry confirmation (usually in XML or TXT format, including transaction serial number, electronic receipt number, etc.), the business processing agent enters the "business completion stage".

[0097] First, the interface processing module calls the "query interface unit" again to obtain the field descriptions of the returned message. It then performs semantic comparison and verification with the actual response content. After accurately extracting and verifying the key business identifier fields, the business processing agent immediately triggers the "storage unit" to record the entire process data, including original materials, intermediate parsing results, interface interaction logs, accounting vouchers, and model decisions. Simultaneously, it transforms this data into a structured event stream with timestamps and operation traceability, persisting it to the database using a three-level key-value pair: "Business ID + Stage + Operation Type." Finally, the entrusted system automatically archives the call chain and decision logs to the business processing agent's local memory for continuous optimization of prompt word engineering and state transition strategies, enabling the business processing agent's self-evolution.

[0098] The annuity payment processing method and apparatus proposed in this invention, based on a large model and the MCP protocol, dynamically acquires interface protocol metadata through the MCP protocol, enabling the entrusted system to perceive changes in the accounting system interface in real time. This allows for automatic adaptation without manual intervention to modify the code, significantly reducing system maintenance costs and upgrade cycles. Simultaneously, this invention introduces a large model as a business processing intelligent agent, leveraging its powerful multimodal understanding capabilities to process non-standardized payment application materials, improving the accuracy and generalization ability of data extraction, and effectively enhancing the processing efficiency and automation level of annuity payment transactions.

[0099] The acquisition, storage, use, and processing of data in this application all comply with relevant laws and regulations.

[0100] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0102] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0103] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0104] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for processing annuity payment transactions based on a large model and the MCP protocol, characterized in that, This method is based on the entrusted system execution in the pension contribution business processing scenario, and includes: Receive an annuity payment request, wherein the annuity payment request includes payment application materials; Dynamically obtain the currently active interface protocol metadata from the MCP server registered in the accounting system via the MCP protocol; The payment application materials and the interface protocol metadata are input into the business processing intelligent agent based on the large model. The business processing intelligent agent performs multimodal parsing on the payment application materials to extract the first structured business data. Based on the interface protocol metadata, determine the target interface protocol and field requirements that match the current business stage; According to the target interface protocol and field requirements, the first structured business data is assembled to obtain a second structured business data packet that conforms to the accounting system interface specification, and the second structured business data packet is sent to the corresponding accounting system to complete the annuity payment business interaction.

2. The annuity payment processing method based on the large model and MCP protocol according to claim 1, characterized in that, The payment application materials shall at least include image materials and payment report documents; The business processing intelligent agent performs multimodal parsing on the payment application materials to extract the first structured business data, including: The image material is processed by optical character recognition, and the payment report document is processed by a document semantic parsing engine to extract the payer's identity information, unit information and payment amount information. The extracted information is mapped to a unified data model to generate the first structured business data.

3. The annuity payment processing method based on the large model and MCP protocol according to claim 1, characterized in that, The system dynamically obtains the currently active interface protocol metadata from the MCP server registered with the accounting system via the MCP protocol, including: A query request is sent to the resource list interface of the MCP server to obtain the interface protocol metadata sent by the MCP server; wherein, the interface protocol metadata includes at least field requirements, verification rules and version metadata.

4. The annuity payment processing method based on the large model and MCP protocol according to claim 1, characterized in that, The business processing intelligent system based on the large model embeds a state machine, which drives the annuity payment business processing process; wherein, the state machine includes at least an information entry and inspection stage, a business acceptance stage, and a business completion stage. During the information entry check phase, the business processing intelligent agent performs a completeness and consistency check on the payment application materials according to the verification rules. After the verification is passed, the process proceeds to the business acceptance stage. The first structured business data is assembled according to the target interface protocol and field requirements to obtain a second structured business data packet that conforms to the accounting system interface specification. The second structured business data packet is then sent to the corresponding accounting system. Upon receiving a successful accounting confirmation from the accounting system, the business status is advanced to the business completion stage, and all process data and execution logs are stored in the database.

5. The annuity payment processing method based on the large model and MCP protocol according to claim 1, characterized in that, The first structured business data is assembled according to the target interface protocol and field requirements to obtain a second structured business data package that conforms to the accounting system interface specification, including: Based on the business knowledge graph, semantic alignment and mapping are performed on the fields and field requirements in the first structured business data; According to predefined conversion rules, the data in the first structured business data is calculated or converted in format; The processed fields and data are assembled to obtain a second structured business data packet, and sensitive fields in the second structured business data packet are desensitized or encrypted.

6. The annuity payment processing method based on the large model and MCP protocol according to claim 1, characterized in that, The method also includes: Based on the version metadata in the interface protocol metadata, verify the compatibility between the field version of the second structured business data packet and the version of the target interface protocol; After the compatibility verification is passed, the second structured business data is sent to the corresponding accounting system to complete the annuity payment business interaction. The payment application materials, the first structured business data, the interface interaction logs and the business response results generated in the entire process of this business are converted into a structured event stream with timestamps and operation traceability identifiers and stored.

7. The annuity payment processing method based on the large model and MCP protocol according to claim 1, characterized in that, Before dynamically obtaining interface protocol metadata from the MCP server of the accounting system via the MCP protocol, the following is also included: The entrusted system accesses the MCP server through an authentication mechanism, wherein the MCP server only allows authorized entrusted systems to access it; When the accounting system does not provide an MCP service address, the entrusted system builds its own MCP service address and modifies the interface protocol metadata based on the communication results provided by the accounting system administrators.

8. A method for processing annuity payment transactions based on a large model and the MCP protocol, characterized in that, This method is implemented based on the accounting system in the pension contribution processing scenario, and includes: Register and configure the MCP server, and provide the currently effective interface protocol metadata to the entrusted system through the MCP server; The system receives a second structured business data packet sent by the entrusted system; wherein the second structured business data packet is generated by the entrusted system after performing multimodal parsing and data assembly on the payment application materials according to the interface protocol metadata, and conforms to the interface specification of the accounting management system. Perform business processing based on the second structured business data packet and generate business processing results; The business processing result is returned to the entrusted system.

9. A device for processing annuity payment transactions based on a large model and the MCP protocol, characterized in that, The device is deployed in the entrusted system for pension payment processing scenarios, including: The business request receiving module is used to receive annuity payment business requests, wherein the annuity payment business request includes payment application materials; The interface protocol acquisition module is used to dynamically obtain the currently effective interface protocol metadata from the MCP server registered in the accounting system via the MCP protocol. The data parsing module is used to input the payment application materials and the interface protocol metadata into the business processing intelligent agent based on the large model, and to extract the first structured business data by performing multimodal parsing on the payment application materials through the business processing intelligent agent; The protocol processing module is used to determine the target interface protocol and field requirements that match the current business stage based on the interface protocol metadata. The data assembly and interaction module is used to assemble the first structured business data according to the target interface protocol and field requirements to obtain a second structured business data packet that conforms to the accounting system interface specification, and send the second structured business data packet to the corresponding accounting system to complete the annuity payment business interaction.

10. A device for processing annuity payment transactions based on a large model and the MCP protocol, characterized in that, This device is deployed in the accounting system for pension contribution processing scenarios, including: The configuration module is used to register and configure the MCP server, and provide the currently effective interface protocol metadata to the entrusted system through the MCP server; The data receiving module is used to receive the second structured business data packet sent by the entrusted system; wherein the second structured business data packet is generated by the entrusted system after performing multimodal parsing and data assembly on the payment application materials according to the interface protocol metadata, and conforms to the interface specification of the accounting management system; The business processing module is used to perform business processing based on the second structured business data packet and generate business processing results; The result return module is used to return the business processing result to the entrusted system.

Images