CN-122022826-A - Collaborative system and method for agent and human expert based on achieving common goal

CN122022826ACN 122022826 ACN122022826 ACN 122022826ACN-122022826-A

Abstract

The invention provides an agent and human expert cooperative system and method based on achieving a common target, which belong to the technical field of data processing, and concretely comprise the steps of determining processing delay data of an expert system in a number interval of each interactive processing object by switching processing data of the agent to the expert system, determining an active switching control interval based on the processing delay data, determining real-time interactive data of the agent and the interactive processing object in the active control interval, determining an updating strategy of switching control strategy of active switching to the expert system based on the real-time interactive data and the active switching control interval of the agent to the expert system, and determining a management and control method of active switching control processing in other number intervals according to adjustment data of switching control strategy of service management of the active switching to the expert system, thereby improving the operation reliability of the agent.

Inventors

LI BIN
ZHAO ZHONGZHOU
HAN JUN
DI WENHUA
WANG ZUOBIN

Assignees

智海星河(浙江)科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. An agent and human expert cooperation method based on achieving a common goal is characterized by comprising the following steps: utilizing interaction processing objects of the intelligent agent in different time periods to determine that the intelligent agent can be actively switched to an expert system, and entering the next step; Processing data are switched from an intelligent agent to an expert system, processing delay data of the expert system in the number interval of each interactive processing object are determined, and an active switching control interval is determined based on the processing delay data; And in the active control interval, determining real-time interaction data of the agent and the interaction processing object, determining an update strategy of the switching control strategy of the active switching to the expert system based on the real-time interaction data and the active switching control interval of the agent to the expert system, and determining a management and control method of the active switching control processing in other quantity intervals according to the adjustment data of the switching control strategy of the active switching to the expert system for service management.
2. The co-ordination method of claim 1, wherein the interactive data of the agent comprises interactive process objects and interactive text data of the interactive process objects in different time periods.
3. The method of claim 1, wherein determining that the agent can actively switch to an expert system comprises: determining the number of the interactive processing objects in different time periods based on the interactive processing object data in different time periods; Determining a busy period of the interactive processing in different days according to the number of the interactive processing objects in different periods; Based on the distribution data of the interactive processing busy period in different days, it is determined whether the agent can actively switch to an expert system.
4. The co-agent and human expert based method according to claim 1, wherein the determining of the active switching control interval within the number interval is: Based on switching processing data from an agent to an expert system in a number interval of each interactive processing object, determining a period in which the processing delay of the expert system in the number interval does not meet the requirement, and taking the period as a processing delay period; according to the distribution data of the processing delay time periods in different quantity intervals, determining the time length duty ratio of the processing delay time periods in the different quantity intervals in all the interactive processing time periods, and taking the time length duty ratio as the delay duty ratio; An active handoff control interval within a different number interval is determined based on delay duty cycles within the number interval.
5. The co-agent and human expert based co-process for achieving the common goal according to claim 4, wherein the time delay period is a time period duty cycle of processing time delay periods in the number interval in all interactive processing periods in the number interval.
6. The collaborative method for an agent and a human expert based on achieving a common goal according to claim 1, wherein the method for determining an update strategy of a handover control strategy of an active handover to expert system is: based on the real-time interaction data, determining the interaction times between the intelligent agent and the interaction processing object; Determining an interaction processing object with the interaction times meeting a switching control strategy based on the interaction times, and taking the interaction processing object as a switching processing object; And determining an updating strategy of the switching control strategy of the active switching control section to the expert system according to the switching processing object data and the active switching control section data.
7. The collaborative method for an agent and a human expert based on achieving a common goal according to claim 6, wherein the interaction processing object whose interaction times satisfy a switching control policy is an interaction processing object whose interaction times is greater than a threshold of interaction times corresponding to the switching control policy.
8. The collaborative method for an agent and a human expert based on achieving a common goal according to claim 6, wherein when not in an active switching control interval, the agent is required to switch to an expert system for processing interactive data of an interactive processing object when the agent cannot meet the interactive processing requirement of the interactive processing object.
9. The method according to claim 6, wherein determining an update policy of a handover control policy of an active handover control section to an expert system according to the handover processing object data and the active handover control section data, comprises: and acquiring the data of the switching processing objects in the latest preset time, and determining that the adjustment processing of the switching control strategy is needed if the number of the switching processing objects in the latest preset time does not meet the requirement.
10. An agent and human expert coordination system for achieving a common goal, which adopts the agent and human expert coordination method for achieving a common goal according to any one of claims 1-9, and is characterized by comprising the following steps: the switching identification module, expert system switching module, access processing module; the switching identification module is responsible for carrying out the process by utilizing a central intelligent agent when the switching identification module is not in the number interval of the interactive processing objects of the active switching control processing, and the front-end intelligent agent is identifying the problem type of the user problem of the processing; the expert system switching module is responsible for automatically switching to an expert system when determining that the suspension processing of the front-end intelligent agent is required based on the recognition result of the problem type; The access processing module is responsible for solving the problem of the user and automatically switching the front-end agent for processing after the solution is completed.

Description

Collaborative system and method for agent and human expert based on achieving common goal Technical Field The invention belongs to the technical field of data processing, and particularly relates to an intelligent agent and human expert cooperation system and method based on achieving a common target. Background Along with the large-scale popularization of the intelligent agent, the intelligent agent is applied to a customer service system in a large scale, the processing pressure of the artificial customer service is greatly improved, and a similar technical scheme is provided in the invention patent application CN202511710824.2, a customer service processing method and system of a multi-intelligent agent architecture, but the technical scheme has the following defects: The prior art scheme of the intelligent agent in the operation process usually reminds the manual intervention treatment when the intelligent agent cannot be treated, but when the identification of the intelligent agent has deviation, the user experience is often not affected little at this time, so how to actively switch to the control scheme of the expert system in the interaction quantity interval of different interaction treatment objects according to the switching data between the intelligent agent and the expert system is determined, thereby improving the reliability of the identification treatment when the intelligent agent has abnormal identification, and further ensuring the reliability of the interaction treatment of the intelligent agent and the user on the basis of improving the user experience to be the technical problem to be solved urgently. Specifically, the application provides an agent and human expert collaborative system and method based on achieving a common goal. Disclosure of Invention In order to achieve the purpose of the invention, the invention adopts the following technical scheme: specifically, the application provides an agent and human expert cooperation method based on achieving a common goal, which specifically comprises the following steps: s1, utilizing interaction processing objects of an agent in different time periods, and entering a next step when determining that the agent can be actively switched to an expert system; s2, processing data are switched from an agent to an expert system, processing delay data of the expert system in the number interval of each interactive processing object are determined, and an active switching control interval is determined based on the processing delay data; S3, in the active control interval, real-time interaction data of the agent and the interaction processing object are determined, based on the real-time interaction data and the active switching control interval from the agent to the expert system, the update strategy of the switching control strategy of the active switching to the expert system is determined, and the management and control method of the active switching control processing in other quantity intervals is determined according to the adjustment data of the switching control strategy of the active switching to the expert system for service management. The invention has the beneficial effects that: The method comprises the steps of determining an active switching control interval in a quantity interval based on processing delay data, and actively switching the intelligent agent to an expert system, namely determining the quantity of manual quantity intervals according to the angle of the influence degree of the overall processing delay due to the fact that the intelligent agent is in identification deviation at present, so that the technical problems of poor identification reliability and efficiency of identification deviation of the intelligent agent caused by the original passive switching scheme are avoided. Based on real-time interaction data and an active switching control interval from an agent to an expert system, the update strategy of the switching control strategy of the active switching to the expert system is determined, the active switching strategy of the agent is dynamically adjusted by monitoring the interaction data of the agent and a user and introducing the processing result of the expert system (manual) as feedback, so that the self problems (such as knowledge blind areas and understanding deviation) of the agent are identified, the service reliability of the agent and the accuracy of problem identification are finally improved, an optimization closed loop of 'monitoring-identification-adjustment-learning' is formed, and the intelligent system is enabled to evolve from a static and preset rule to a dynamic agent capable of being learned from manual intervention. Further, the interactive data of the agent comprises interactive processing objects in different time periods and interactive text data of the interactive processing objects. Further, determining that the agent can actively switch to an expert system specifically includes: d