A service script recommendation method and system based on dynamic tag matching

Abstract

The application discloses a service script recommendation method and system based on dynamic label matching, and relates to the technical field of electronic information processing. The method responds to a target user service request, obtains multi-dimensional feature data of the target user, and generates a label set containing core labels and moderate labels; determines a task label group in combination with a service scene and the label set, searches a preset strategy library through a preset double-layer similarity matching algorithm, and matches a target service strategy; after successful matching, a recommended script is generated and pushed to an agent. When no matching is performed, a complete service strategy is generated by calling a large model in combination with a knowledge base, and after being audited, the complete service strategy is stored in the strategy library and a recommended script is generated. Case references and historical experience information are simultaneously pushed, and script supplementary content is updated in real time. The method improves the personalization and adaptability of script recommendation, reduces operation cost, and optimizes service experience and agent efficiency.

Images