Base-to-hot-line work order full-process intelligent circulation system

Abstract

The invention provides a hotline-oriented work order full-process intelligent circulation system, and belongs to the crossing field of artificial intelligence and digital government affair technologies. According to the method, the whole process of a hotline work order is disassembled into sub-tasks which are cooperatively completed by six special intelligent agents, namely work order acceptance and semantic understanding, classification and routing, dispatching and scheduling, disposal execution, result checking and feedback, and citizen evaluation and data analysis. The intelligent agents communicate and cooperate based on an MCP protocol, a tool is called in a unified mode through an MCP server, a knowledge base is inquired, context is managed, and accurate understanding, efficient dispatching, closed-loop disposal and intelligent analysis of citizen appeals are achieved. The system effectively solves the problems of dependence on manpower, low efficiency, unsmooth circulation, difficulty in tracing and the like in traditional hotline work order processing, and realizes automation, intellectualization and collaboration of the whole process of work order processing.

Description

Technical Field

[0001] This invention relates to the intersection of artificial intelligence and digital government technology, and in particular to a work order intelligent workflow system based on hotline. Background Technology

[0002] Government service hotlines serve as an important bridge between the government and citizens, but their traditional processing procedures rely heavily on manual labor, resulting in the following pain points:

[0003] During the call processing stage: Manual recording by operators is prone to errors, especially when dealing with dialects or complex descriptions.

[0004] In the classification and assignment process, the manual determination of which department should be responsible for a matter requires a high level of expertise and is prone to errors, leading to departmental buck-passing. This falls short of the "precise assignment" requirement emphasized in the Anhui Province regulations.

[0005] In the handling and feedback process: the progress is not transparent and there is a lack of automated supervision. The latest opinion from the General Office of the State Council clearly requires strengthening the follow-up visits and closed-loop management of the handling of complaints.

[0006] Analysis and evaluation phase: The value of massive amounts of data has not been fully explored, and it has failed to effectively support "proactive handling before complaints are filed" and proactive governance.

[0007] In recent years, although some studies have attempted to use single AI models for automatic work order classification, they have struggled to handle complex, multi-step closed-loop processes. Multi-agent systems (MAS) offer a new approach to collaborative complex tasks, but traditional inter-agent communication is inefficient and lacks a unified context management standard. Model Context Protocol (MCP), as an emerging protocol, standardizes the interaction between agents and tools / data sources, providing an ideal framework for multi-agents to share context and collaboratively invoke resources. However, directly applying MCP to the 12345 hotline scenario requires addressing challenges in security (such as context isolation and authorization authentication), integration with existing government systems, and meeting government data security and compliance requirements. Currently, no mature solution adequately addresses these issues. Summary of the Invention

[0008] To address the above technical problems, this invention provides a basic-to-hotline intelligent workflow system for the entire process of work orders.

[0009] The technical solution of this invention is:

[0010] A fully intelligent workflow system for hotline-based work orders includes:

[0011] The work order processing and semantic understanding intelligent agent receives original requests from citizens through various channels and performs semantic analysis.

[0012] The work order classification and routing intelligent agent determines the department, priority, and processing time limit of the work order. The work order dispatch and scheduling intelligent agent dispatches the work order to the corresponding processing department or the next level intelligent agent.

[0013] Work order processing and execution intelligent agents, representing the processing department to automatically process or assist in processing work orders.

[0014] The intelligent agent for verifying and providing feedback on the handling results automatically verifies the completeness and rationality of the handling results and generates feedback reports to notify citizens.

[0015] The citizen evaluation and data analysis intelligent agent collects citizen evaluations, conducts satisfaction analysis, and mines the value of work order data.

[0016] Each agent collaborates based on the MCP protocol, and all agents register and communicate through the MCP server. Each agent assumes specific responsibilities and accesses shared tools, knowledge bases, and contexts through the MCP server.

[0017] in,

[0018] Tools: Provide a unified API call entry point for intelligent agents, including speech recognition (ASR), natural language processing (NLP) tools, database query tools, SMS gateway interfaces, and business system API call tools;

[0019] Knowledge base: Stores structured knowledge, such as policy and regulation knowledge base, departmental power and responsibility list knowledge base linked with dynamically updated work order catalog, historical work order handling case library, and standard handling procedure (SOP) knowledge base;

[0020] Context: Records the status changes of each work order throughout the entire process and the operation history of each intelligent agent.

[0021] Furthermore,

[0022] It also integrates the A2A protocol, which is used for point-to-point high-frequency, low-latency information interaction between the work order dispatch and scheduling intelligent agent, the work order processing and execution intelligent agent, and the processing result verification and feedback intelligent agent.

[0023] Furthermore,

[0024] The system operates as follows:

[0025] Receive original requests from citizens through various channels;

[0026] The original request is parsed by a work order acceptance and semantic understanding intelligent agent to generate structured work order data;

[0027] By using a single-classification and routing intelligent agent, combined with a pre-built knowledge base, the department to which a work order belongs and its priority are determined.

[0028] Work orders are automatically dispatched to the target processing system through a work order dispatch and scheduling intelligent agent.

[0029] The work order processing intelligent agent assists or automates the work order processing and records the processing process.

[0030] The intelligent agent for verifying and feeding back the handling results will be used to verify the compliance of the handling results and automatically feed back the results to the citizens.

[0031] Through citizen evaluation and data analysis intelligence, citizen satisfaction is collected and analyzed, and the data from the entire process is mined and analyzed to generate analysis reports.

[0032] Furthermore,

[0033] The MCP protocol is used to provide unified tool invocation, knowledge base query and context management services for all intelligent agents, so as to realize information sharing and collaborative cooperation among intelligent agents.

[0034] The work order classification and routing intelligent agent makes work order dispatch decisions by querying the departmental authority and responsibility list knowledge base and historical work order case database accessed based on the MCP protocol and using a machine learning model; wherein, the departmental authority and responsibility list knowledge base is dynamically linked to the government authority and responsibility list.

[0035] The intelligent agent for verifying and responding to the handling results uses natural language processing technology to compare the work order requests with the handling responses and automatically determines whether the handling meets the standards. For work orders that are not satisfactory during the follow-up visit or fail the verification, a reassignment or escalation supervision process is initiated.

[0036] The beneficial effects of this invention are

[0037] (1) Full-process intelligence and efficiency improvement: Through the division of labor and cooperation of multiple intelligent agents, the full-process automation from acceptance to evaluation is realized, which greatly reduces manual operation and improves work efficiency.

[0038] (2) More precise decision-making and fewer errors: The rich knowledge base and tools integrated based on the MCP protocol make the decision-making of each link (such as classification, dispatching and handling) more accurate and reduce errors and shirking responsibility.

[0039] (3) System coordination, flexibility and scalability: The MCP protocol, as an "operating system", enables intelligent agents to be plug-and-play, and it is easy to expand new functional modules or connect new external tools (such as integrating the A2A protocol for high-frequency interaction between specific intelligent agents). The system architecture is flexible and highly adaptable.

[0040] (4) Closed-loop management and traceability: Based on the shared context, the entire life cycle of the work order is clearly visible, which facilitates supervision, management and traceability, and truly realizes the closed-loop management of "every item is implemented and every matter is responded to", which meets the closed-loop management requirements of government services.

[0041] (5) Data-driven optimization and proactive governance: The final analytical agent transforms work order data into valuable insights to help the government optimize services and make scientific decisions, promote the transformation from "responding to complaints immediately" to "addressing issues before complaints arise", and improve social governance capabilities. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the workflow of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0044] This invention comprises a series of dedicated intelligent agents built upon the MCP protocol, all of which register and communicate through an MCP server. Each agent assumes specific responsibilities and accesses shared tools, knowledge bases, and context through the MCP server. The MCP server manages the following resources:

[0045] Tools: Provide a unified API call entry point for intelligent agents, such as speech recognition (ASR), natural language processing (NLP) tools, database query tools, SMS gateway interfaces, business system API call tools, etc.

[0046] Knowledge base: Stores structured knowledge, such as policy and regulation knowledge base, departmental authority and responsibility list knowledge base (linked with dynamically updated work order catalog), historical work order handling case library, and standard handling procedure (SOP) knowledge base.

[0047] Context: Record the status changes of each work order throughout the entire process and the operation history of each intelligent agent to ensure unambiguous collaboration and achieve process traceability.

[0048] The core intelligent agents, their responsibilities, and their technical implementation are as follows:

[0049] (1) Work order acceptance and semantic understanding intelligent agent:

[0050] Responsibilities: Receive original requests from citizens through various channels such as hotlines, WeChat, and websites, and perform semantic analysis.

[0051] Technical implementation: The system calls speech recognition (ASR) and natural language understanding (NLU) tools through the MCP protocol, and performs entity recognition (such as names, locations, and events), sentiment analysis, and intent classification to generate structured document information.

[0052] (2) Work order classification and routing agent:

[0053] Responsibilities: Determine the department to which a work order belongs, its priority, and the processing time limit.

[0054] Technical implementation: Access the knowledge base of policies, regulations and departmental responsibilities through the MCP protocol, combine it with historical work order data, and use machine learning models (such as text classification and graph neural networks) for accurate classification.

[0055] (3) Work order dispatch and scheduling agent:

[0056] Responsibilities: Dispatch work orders to the corresponding processing department or the next level of intelligent agent.

[0057] Technical Implementation: Work orders are automatically pushed to the target department's business system by calling the workflow engine API via the MCP protocol. For work orders with unclear responsibilities, a collaborative session can be initiated to invite human agents or other intelligent agents to participate in the decision-making process.

[0058] (4) Work order processing and execution intelligent agent:

[0059] Responsibilities: To automatically process or assist in processing work orders on behalf of the relevant department.

[0060] Technical Implementation: Access to the business system APIs of various departments is achieved through the MCP protocol. For common issues (such as information inquiries and simple complaints), responses or solutions can be automatically generated; for complex issues, staff are provided with suggested templates and knowledge support, and the handling process is tracked and recorded.

[0061] (5) Intelligent agent for verifying and providing feedback on processing results:

[0062] Responsibilities: Automatically verify the completeness and rationality of the handling results, and generate feedback reports to notify citizens.

[0063] Technical Implementation: The system uses the MCP protocol to call NLP tools, performing a matching analysis between the handling results and the work order requests to detect whether the response was irrelevant or the problem was not fully resolved. Once verified, feedback is automatically sent to the citizen via SMS, mobile app, or other means.

[0064] (6) Citizen Evaluation and Data Analysis Intelligent Agent:

[0065] Responsibilities: Collect citizen feedback, conduct satisfaction analysis, and explore the value of work order data.

[0066] Technical Implementation: The system integrates with the evaluation system via the MCP protocol to analyze citizens' sentiments. Simultaneously, it aggregates and analyzes work order data across the entire process to generate hotspot issue maps, departmental performance reports, and public opinion trend reports, providing support for decision-making.

[0067] The following functions are achieved through this invention.

[0068] (1) Achieve full-process automation: Achieve full-process automation and intelligence from work order acceptance to evaluation feedback, break down the entire chain of work order processing tasks, and have multiple dedicated intelligent agents execute them automatically, reducing human intervention.

[0069] (2) Improve accuracy and efficiency: Improve the accuracy and efficiency of work order classification and dispatch through professional intelligent agents and knowledge bases.

[0070] (3) Achieve closed-loop management and traceability: Based on the shared context of the MCP protocol, achieve closed-loop management and traceability of the entire life cycle of work orders, ensuring that every link from acceptance to evaluation of work orders is traceable and supervised, forming a business closed loop of "every item is implemented and every matter is responded to".

[0071] (4) Enhance the system's intelligent decision-making capabilities: Through intelligent analysis of work order data throughout the entire process, provide data support for hotspot early warning, departmental performance evaluation, and policy optimization, and assist the government in "proactive action before complaints" and scientific decision-making.

[0072] The above description is merely a preferred embodiment of the present invention and is used only to illustrate the technical solution of the present invention, and is 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 are included within the scope of protection of the present invention.

Claims

1. A work order intelligent workflow system based on a hotline, characterized in that, include: The work order acceptance and semantic understanding intelligent agent receives original requests from citizens from various channels and performs semantic analysis; The work order classification and routing intelligent agent determines the department to which the work order belongs, its priority, and its processing time limit; The work order dispatch and scheduling intelligent agent dispatches work orders to the corresponding handling department or the next level intelligent agent; The work order processing and execution intelligent agent automatically or assistedly processes work orders on behalf of the processing department. The intelligent agent for verifying and providing feedback on the handling results automatically verifies the completeness and rationality of the handling results and generates feedback reports to notify citizens. The citizen evaluation and data analysis intelligent agent collects citizen evaluations, conducts satisfaction analysis, and mines the value of work order data. Each agent collaborates based on the MCP protocol, and all agents register and communicate through the MCP server. Each agent assumes specific responsibilities and accesses shared tools, knowledge bases, and contexts through the MCP server.

2. The system according to claim 1, characterized in that, The tools provide a unified API call entry point for intelligent agents, including speech recognition (ASR), natural language processing (NLP) tools, database query tools, SMS gateway interfaces, and business system API call tools.

3. The system according to claim 1, characterized in that, The knowledge base stores structured knowledge, including a policy and regulation knowledge base, a departmental power and responsibility list knowledge base linked with a dynamically updated work order directory, a historical work order handling case library, and a standard handling procedure (SOP) knowledge base.

4. The system according to claim 1, characterized in that, The context records the status changes of each work order throughout the entire process and the operation history of each intelligent agent.

5. The system according to claim 1, characterized in that, It also integrates the A2A protocol, which is used for point-to-point high-frequency, low-latency information interaction between the work order dispatch and scheduling intelligent agent, the work order processing and execution intelligent agent, and the processing result verification and feedback intelligent agent.

6. The system according to any one of claims 1-5, characterized in that, The system operates as follows: Receive original requests from citizens through various channels; The original request is parsed by a work order acceptance and semantic understanding intelligent agent to generate structured work order data; By using a single-classification and routing intelligent agent, combined with a pre-built knowledge base, the department to which a work order belongs and its priority are determined. Work orders are automatically dispatched to the target processing system through a work order dispatch and scheduling intelligent agent. The work order processing intelligent agent assists or automates the work order processing and records the processing process. The intelligent agent for verifying and feeding back the handling results will be used to verify the compliance of the handling results and automatically feed back the results to the citizens. Through citizen evaluation and data analysis intelligence, citizen satisfaction is collected and analyzed, and the data from the entire process is mined and analyzed to generate analysis reports.

7. The system according to claim 6, characterized in that, The MCP protocol is used to provide unified tool invocation, knowledge base query and context management services for all intelligent agents, so as to realize information sharing and collaborative cooperation among intelligent agents.

8. The system according to claim 6, characterized in that, The work order classification and routing intelligent agent makes work order dispatch decisions by querying the departmental authority and responsibility list knowledge base and historical work order case database accessed based on the MCP protocol and using a machine learning model; wherein, the departmental authority and responsibility list knowledge base is dynamically linked to the government authority and responsibility list.

9. The system according to claim 6, characterized in that, The intelligent agent for verifying and responding to the handling results uses natural language processing technology to compare the work order requests with the handling responses and automatically determines whether the handling meets the standards. For work orders that are not satisfactory during the follow-up visit or fail the verification, a reassignment or escalation supervision process is initiated.

Images