CN-121981148-A - Business system agent processing method, device, equipment and medium

CN121981148ACN 121981148 ACN121981148 ACN 121981148ACN-121981148-A

Abstract

The application discloses a business system agent processing method, a device, equipment and a medium. The method comprises the steps of judging whether at least one preset agent in a current service system needs to be called according to natural language information of a question which is received by a user to the current service system, responding to a judging result that the current service system needs to call at least one preset agent, arranging all preset agents to obtain an agent chain, and triggering response operation to the natural language information according to the agent chain. According to the technical scheme provided by the embodiment of the application, the operation complexity and the learning cost of the service system are reduced, and the efficiency and the reliability of the service processing process are improved.

Inventors

CHEN MAOLIN
TANG XINJIE
JIANG JIQING
Ye Naikan
ZHENG QIN
RAO LEI

Assignees

杭州新中大科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260210

Claims (10)

1. The business system agent processing method is characterized by comprising the following steps: judging whether at least one preset agent in the current service system is required to be called according to natural language information of a question which is received by a user to the current service system; Responding to the judgment result that at least one preset agent needs to be called for the current service system, and arranging each preset agent to obtain an agent chain; And triggering response operation to the natural language information according to the agent chain.
2. The method of claim 1, wherein the determining whether at least one preset agent in the current service system needs to be invoked according to natural language information that a user asks for the current service system, comprises: Carrying out intention reasoning on the natural language information to obtain a reasoning result; Responding to at least one item of behavior action information in the reasoning result, and inquiring in a preset behavior characteristic database; responding to the behavior characteristics in the behavior characteristic database to comprise the behavior action information, and determining that the current service system needs to call at least one preset agent; and generating knowledge solution information corresponding to the natural language information in response to the fact that no behavior action information exists in the reasoning result.
3. The method of claim 2, wherein the arranging each of the preset agents to obtain an agent chain comprises: Responding to the behavior action information to accord with a behavior list of any preset agent, and taking the preset agent as a root node; Predicting follow-up behavior actions according to a root node, and taking preset intelligent agents corresponding to the follow-up behavior actions as leaf nodes of the root node; and determining the agent chain according to the root node and each leaf node.
4. A method according to claim 3, wherein predicting a subsequent behavioral action from the root node comprises: predicting the follow-up behavior action according to the logic sequence relation of each behavior characteristic based on the behavior action information corresponding to the root node; correspondingly, the determining the agent chain according to the root node and each leaf node comprises the following steps: and forming the root node and each leaf node into the agent chain according to the logic sequence relation and the authority priority preset by the user.
5. The method of claim 1, wherein the preset agent is constructed by: scanning a system database of the current service system, and identifying a menu item with an empty link address in a menu table of the system database as a function authority; Determining a function menu and a function module of the system database according to the function authority; For any functional module, if the functional type of the functional module does not belong to the system management type or the system configuration type, a preset agent corresponding to the functional module is constructed.
6. The method of claim 5, wherein the constructing the preset agent corresponding to the functional module comprises: Determining a behavior list of the preset intelligent agent according to each function menu subordinate to the function module; determining return value information corresponding to the function module and an execution link of the action according to the link address of the action list; and constructing the preset intelligent agent according to the behavior list, the return value information and the execution link.
7. The method of claim 6, wherein the method further comprises: And carrying out calling identification on any behavior action of the preset agent, and eliminating the behavior action from the behavior list in response to no logic calculation in the service layer calling code.
8. A business system agent processing device, comprising: the intelligent agent calling judging module is used for judging whether at least one preset intelligent agent in the current service system needs to be called according to natural language information of a question which is received by a user to the current service system; the intelligent agent chain determining module is used for calling at least one preset intelligent agent for the current service system according to the judging result, and arranging the preset intelligent agents to obtain an intelligent agent chain; and the response operation triggering module is used for triggering response operation to the natural language information according to the intelligent agent chain.
9. An electronic device, the electronic device comprising: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the business system agent processing method of any one of claims 1-7.
10. A computer readable storage medium, characterized in that the computer readable storage medium stores computer instructions for causing a processor to implement the business system agent processing method of any one of claims 1-7 when executed.

Description

Business system agent processing method, device, equipment and medium Technical Field The present application relates to the field of artificial intelligence technologies, and in particular, to a method, an apparatus, a device, and a medium for processing an agent in a service system. Background In recent years, as large language model technology matures and prevails, its powerful natural language understanding and generating capabilities are being widely integrated into various types of software products. By means of model fine tuning and the like, the industry can develop intelligent application aiming at specific scenes, and the semantic interaction level of the system is remarkably improved. However, most current business systems still rely on traditional graphical interfaces and fixed operation flows, and users need to complete complex tasks through multi-step clicking, form filling and menu jumping, so that the operation process is heavy and the learning cost is high. How to upgrade the existing system rapidly, and to simplify the interaction and remodel the user operation mode by using the large language model capability has become a key challenge for improving the intelligence and usability of the system. Disclosure of Invention The application provides a method, a device, equipment and a medium for processing an intelligent agent of a service system, so as to improve the efficiency of the service processing process. According to an aspect of the present application, there is provided a service system agent processing method, including: Judging whether at least one preset agent in the current service system is required to be called according to natural language information of a question which is received by a user to the current service system; Responding to the judgment result that the current service system needs to call at least one preset agent, and arranging each preset agent to obtain an agent chain; and triggering response operation to the natural language information according to the agent chain. According to another aspect of the present application, there is provided a service system agent processing apparatus, including: The agent calling judging module is used for judging whether at least one preset agent in the current service system needs to be called according to natural language information of a question which is received by a user to the current service system; the intelligent agent chain determining module is used for calling at least one preset intelligent agent for the current service system according to the judging result, and arranging the preset intelligent agents to obtain an intelligent agent chain; And the response operation triggering module is used for triggering response operation to the natural language information according to the intelligent agent chain. According to another aspect of the present application, there is provided an electronic apparatus including: at least one processor, and A memory communicatively coupled to the at least one processor, wherein, The memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the business system agent processing method of any one of the embodiments of the present application. According to another aspect of the present application, there is provided a computer readable storage medium storing computer instructions for causing a processor to execute a method for processing an agent of a business system according to any embodiment of the present application. According to another aspect of the present application, there is provided a computer program product comprising a computer program which, when executed by a processor, implements a business system agent processing method according to any of the embodiments of the present application. According to the technical scheme of the embodiment of the application, whether at least one preset agent in the current service system needs to be called is judged according to the natural language information of the question which is received by the user to the current service system, at least one preset agent needs to be called for the current service system according to the judging result, each preset agent is arranged to obtain an agent chain, and the response operation to the natural language information is triggered according to the agent chain. For the service system, the scheme of the application innovates an interactive paradigm, a user can drive a complex service processing flow by using natural language, the operation complexity and the learning cost are obviously reduced, and meanwhile, the service system rapidly responds to the user demand by dynamically combining the collaboration of preset intelligent agents, thereby improving the function multiplexing rate and the service flexibility. In the multi-agent cooperative level, the system realizes automatic decomposition and agent execution of complex tasks, ensures acc