CN-122022822-A - Intelligent processing method, system, equipment and storage medium for customer service

CN122022822ACN 122022822 ACN122022822 ACN 122022822ACN-122022822-A

Abstract

The application relates to a customer service intelligent processing method, a system, equipment and a storage medium, which belong to the technical field of intelligent customer service, wherein the method comprises the steps of receiving a problem processing request of a user, wherein the problem processing request at least comprises problem contents, analyzing the problem contents, identifying the problem types to which the problem contents belong, matching corresponding question-answering services for the problem contents based on a plurality of preset question-answering services and the identified problem types, and routing the problem contents to the question-answering services obtained by matching for processing, wherein the question-answering services are configured to respectively process the problem contents with different complexity, acquiring answer contents generated by the question-answering services obtained by matching, and feeding the answer contents back to the user so as to enable the user to know. The application has the effect of realizing targeted treatment of the problems with different complexity, ensuring the experience of the user and reducing the frequency of manual intervention.

Inventors

SUN QI
TAN RENFENG
XU HUICHAO
YANG JIANJIAN
ZHANG SHICHENG

Assignees

苏州创旅天下信息技术有限公司

Dates

Publication Date: 20260512
Application Date: 20251211

Claims (10)

1. An intelligent processing method for customer service is characterized by comprising the following steps: Receiving a problem processing request of a user, wherein the problem processing request at least comprises problem content; Analyzing the problem content and identifying the problem type to which the problem content belongs; Based on a plurality of preset question-answering services and the identified question types, matching corresponding question-answering services for the question contents, and routing the question contents to the question-answering services obtained by matching for processing, wherein the question-answering services are configured to process the question contents with different complexity respectively; and obtaining reply content generated by the question and answer service obtained by matching, and feeding back the reply content to the user so as to enable the user to know.
2. The intelligent processing method for customer service according to claim 1, wherein the matching the question content with the corresponding question-answering service based on the preset question-answering services and the identified question types, and routing the question content to the matched question-answering service for processing, includes: If the identified question type is a single question type, matching a corresponding question-answer service for the question content, packaging the corresponding question content into a message, and processing the message through a message queue routing to the question-answer service obtained by matching to obtain answer content; if the identified question type is a plurality of question types, the question content is disassembled into a plurality of subtasks, corresponding question and answer services are respectively matched for each subtask, each subtask is distributed to the matched question and answer services for processing, and processing results returned by the question and answer services matched corresponding to each subtask are integrated to generate answer content.
3. The intelligent processing method for customer service according to claim 2, wherein said assigning each of said subtasks to its matching question-answer service for processing comprises: analyzing whether processing time sequence association exists among all the subtasks, wherein the processing time sequence association refers to a relation that a processing result of a front subtask needs to be used as input and is transmitted to a question-answer service corresponding to a rear subtask for processing; And generating a routing link for guiding the execution sequence of each subtask based on the processing time sequence association and the question-answer service matched with each subtask, and distributing each subtask to the corresponding matched question-answer service for processing based on the execution sequence indicated by the routing link, wherein when the subtask with the processing time sequence association is distributed to the corresponding question-answer service for processing, the processing result of the front subtask corresponding to the rear subtask is provided as input information to the question-answer service of the rear subtask.
4. A customer service intelligence processing method according to claim 3, wherein said assigning each sub-task to a corresponding matched question-answer service for processing based on the execution order indicated by said routing link, previously comprises: Analyzing and determining uncontrollable properties of each subtask, wherein the uncontrollable properties are used for representing the degree of dependence of a processing result of the corresponding subtask on user feedback; Marking the subtasks with uncontrollable properties as high risk properties as user confirmation checkpoints in the routing links; the assigning each subtask to a corresponding matched question-answer service for processing based on the execution sequence indicated by the routing link comprises the following steps: And distributing each subtask to a corresponding matched question-answer service for processing based on the execution sequence indicated by the routing link, feeding back a processing result corresponding to the subtask of the user confirmation check point to the user when the subtask marked as the user confirmation check point is executed each time, and waiting for receiving a confirmation instruction of the user, and then continuing to execute the subsequent subtask in the routing link.
5. The intelligent processing method for customer service according to claim 4, wherein the assigning each sub-task to a corresponding matched question-answer service for processing based on the execution sequence indicated by the routing link further comprises: Pre-generating a branch mapping relation for sub-tasks marked as user confirmation checkpoints, wherein the branch mapping relation comprises a plurality of user selections and implicit tasks corresponding to each user selection, and the user selections are prediction results of confirmation instructions of the users; the method distributes each subtask to the corresponding matched question-answer service for processing based on the execution sequence indicated by the routing link, and further comprises the following steps: When the subtasks marked as user confirmation checkpoints are executed each time, all the user selections which are predicted and obtained by the subtasks corresponding to the user confirmation checkpoints are synchronously provided for the user; and determining a corresponding implicit task from the branch mapping relation according to the user selection actually selected by the user, and updating the implicit task as a new subtask into the routing link.
6. The intelligent processing method for customer service according to claim 5, wherein the assigning each sub-task to a corresponding matched question-answer service for processing based on the execution order indicated by the routing link further comprises: after receiving the confirmation instruction of the user, judging whether an unplanned event exists in the confirmation instruction of the user, wherein the unplanned event refers to that the content fed back by the user does not contain any user selection or contains a new intention which is not covered by the problem content; if an unplanned event exists, the execution operation of the routing link where the subtask corresponding to the unplanned event is terminated, new problem content is generated based on the unplanned event, a new routing link is generated for the newly generated problem content, and each subtask in the new routing link is distributed to a corresponding matched question-answer service for processing according to the execution sequence indicated by the new routing link.
7. The intelligent processing method for customer service according to claim 3, wherein the question types at least comprise complex reasoning questions, and the question-answering service matched with the complex reasoning questions is an Agent service; the method distributes each subtask to the corresponding matched question-answer service for processing based on the execution sequence indicated by the routing link, and further comprises the following steps: And if the evaluation result is unstable, the corresponding processing result is sent to a user for confirmation, and a subsequent execution flow is determined based on confirmation feedback of the user.
8. An intelligent processing system for customer service, which is characterized by comprising, A question processing request receiving module (201) for receiving a question processing request of a user, the question processing request including at least a question content; A user question analysis and identification module (202) for analyzing the question content and identifying the question type to which the question content belongs; The question-answering service matching processing module (203) is used for matching corresponding question-answering services for the question contents based on a plurality of preset question-answering services and the identified question types, and routing the question contents to the question-answering services obtained by matching for processing, wherein the question-answering services are configured to respectively process the question contents with different complexity; And the reply content generation feedback module (204) is used for acquiring reply content generated by the matched question and answer service and feeding back the reply content to the user so as to enable the user to know.
9. A customer service intelligent processing device comprising a memory and a processor, the memory having stored thereon a computer program capable of being loaded by the processor and performing the method of any of claims 1 to 7.
10. A computer readable storage medium, characterized in that a computer program is stored which can be loaded by a processor and which performs the method according to any of claims 1 to 7.

Description

Intelligent processing method, system, equipment and storage medium for customer service Technical Field The application relates to the technical field of intelligent customer service, in particular to an intelligent customer service processing method, an intelligent customer service processing system, intelligent customer service equipment and a storage medium. Background In the field of enterprise customer service, traditional manual customer service modes have long faced a number of challenges. Firstly, the service efficiency is low, the average response time of the artificial customer service is as long as three to five minutes, and the queuing phenomenon of the user is serious in the peak business period, so that the customer experience and satisfaction are continuously reduced. Secondly, due to the fact that professional level and experience of the manual customer service are different, the service quality is uneven, the error rate is high and reaches 15-20%, and brand image and customer trust of enterprises are seriously affected. In addition, in order to provide uninterrupted service, enterprises need to input a large amount of manpower and material resources, so that the operation cost is high, and the cost of single service is usually three to five times that of an intelligent customer service system. To address the above problems, intelligent customer service systems have been developed. However, existing intelligent customer service systems still have significant drawbacks. Existing solutions typically employ a single question-answer model to handle all user requests. This approach is adequate for simple standard problems (e.g., inquiry hours), but the processing power of a single model is just a fly-over for complex problems that require specialized knowledge to be retrieved. The system can recover complex questions by standard answers to influence user experience, or can transfer all the complex questions to manpower, so that the operation cost can not be effectively reduced, and the system needs to be improved. Disclosure of Invention In order to realize targeted processing of problems with different complexity, ensure user experience and reduce human intervention frequency at the same time, the application provides a client service intelligent processing method, a system, equipment and a storage medium. In a first aspect, the present application provides a method for intelligently processing a customer service, including: Receiving a problem processing request of a user, wherein the problem processing request at least comprises problem content; Analyzing the problem content and identifying the problem type to which the problem content belongs; Based on a plurality of preset question-answering services and the identified question types, matching corresponding question-answering services for the question contents, and routing the question contents to the question-answering services obtained by matching for processing, wherein the question-answering services are configured to process the question contents with different complexity respectively; and obtaining reply content generated by the question and answer service obtained by matching, and feeding back the reply content to the user so as to enable the user to know. By adopting the technical scheme, the problems of single processing and incapability of targeted processing of the problems of different complexity of the traditional customer service system are overcome by automatically analyzing the types of the problems and routing the problems to the question-answering service which is most suitable for processing the complexity of the types. The simple questions are quickly led to the light service, the complex questions are distributed to the heavy service, and the optimal configuration of system resources is realized, so that the response time is shortened as a whole, and the accuracy of the answer is improved due to 'professional matching'. Optionally, the matching the question content with the corresponding question-answer service based on the preset question-answer services and the identified question type, and routing the question content to the question-answer service obtained by matching for processing includes: If the identified question type is a single question type, matching a corresponding question-answer service for the question content, packaging the corresponding question content into a message, and processing the message through a message queue routing to the question-answer service obtained by matching to obtain answer content; if the identified question type is a plurality of question types, the question content is disassembled into a plurality of subtasks, corresponding question and answer services are respectively matched for each subtask, each subtask is distributed to the matched question and answer services for processing, and processing results returned by the question and answer services matched corresponding to each