Intelligent sales customer service question answering method and system, computer device, and storage medium

CN122925548APending Publication Date: 2026-10-16GUANGZHOU JIANZHI INFORMATION TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202611068251.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-17
Publication Date
2026-10-16

AI Technical Summary

Technical Problem

这种技术方案存在以下显著缺陷：首先，系统仅能处理字面匹配的查询请求，无法理解用户潜藏购买意向的模糊表达；其次，FAQ系统的扩展性较差，面对产品线更新或价格调整时，需要人工逐条更新知识库，维护成本高昂且响应滞后；再者，此类系统缺乏上下文理解能力，无法处理多轮对话中的指代消解和信息关联，导致用户体验碎片化

Benefits of technology

[0016]Compared to existing technologies, the intelligent sales customer service Q&A method, system, computer equipment, and computer-readable storage medium provided in this application significantly improve the accuracy and timeliness of sales customer service Q&A by combining a large language model with a professional knowledge base through retrieval-enhanced generative algorithm (RAG) to process user sales Q&A, and by employing a confidence-layered human-computer collaboration mechanism. Furthermore, by dynamically determining the autonomous response boundary of user sales Q&A through multi-dimensional confidence assessment, the system automatically responds to user sales Q&A based on the generated reply content when the overall confidence level is greater than or equal to a predefined threshold. When the overall confidence level of the response is less than the predefined threshold, it intelligently triggers human intervention to process the user sales Q&A, significantly improving the service quality of sales customer service Q&A and optimizing the user service experience.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The application relates to an intelligent sales customer service answering method and system, computer equipment and a storage medium. The method comprises the following steps: receiving natural language input by a user from a client device for sales answering; calling a large language model to perform deep semantic analysis on the natural language to generate a user query; using a multi-task identification algorithm to perform intent identification on the user query to obtain an intent identification result; performing knowledge retrieval on a pre-constructed knowledge base according to the intent identification result to obtain a knowledge segment; calling the large language model to generate reply content for sales answering based on the knowledge segment; using a weighted fusion algorithm to calculate the comprehensive confidence of the reply content; if the comprehensive confidence is greater than or equal to a predefined threshold, automatically answering according to the reply content; and if the comprehensive confidence is less than the predefined threshold, transferring to manual customer service to process the sales answering of the user. The method can significantly improve the accuracy and knowledge timeliness of sales answering, and optimize the user service experience.
Need to check novelty before this filing date? Find Prior Art