Service voice intelligent search and analysis system and method

Abstract

A voice search and analysis system comprises a voice transcription module, an index establishing module, a voice search module and a test analysis module. Keywords contained in voice, positions where the keywords appear and other characteristic information are searched for. Analysis can be effectively performed on massive voice recording files and audio files through voice analysis, knowledge exploration in the voice and rapid search on the voice files are achieved, and the problems that voice data are difficult to search for and analyze are solved.

Description

Technical field [0001] The invention relates to a voice search and analysis system and method, and belongs to the field of electronic information technology. Background technique [0002] Telecom, mobile and other call centers have a large amount of customer service recording data, which contains a large amount of useful information such as customer needs. From the perspective of consumer behavior, every interaction between a customer and a customer service agent can give feedback about his current consumption situation and the next consumer behavior trend. Customers’ preference for the product, focus of attention, and psychological expectations will be fed back to the company through agent representatives. However, the voice data containing a large amount of useful information such as customer needs is difficult to search and analyze, just like having a voice treasure but no key to open the treasure; [0003] At present, the data used by the call center for analysis mainly comes...

AI Technical Summary

Problems solved by technology

Due to the limitation of data sources, there are obvious deficiencies in the following aspects: (1) It can only passively see the distribution of call reasons listed in the knowledge base, and more in-depth data such as customers' in-depth call appeals, problems they want to solve, and focus of attention are received. Factors such as hotline pressure and the quality of front-line personnel have not been shown, and comprehensive marketing analysis that combines these customer behavior characteristics with customer age groups, APRU values, package types, business preferences, and market activities cannot be carried out; (2) In some current customer marketing methods, such as using a dedicated model for predictive dialing and selecting a number segment for outbound marketing through a certain strategy, due to various reasons, the marketing success rate is only 3%

In the face of the vast majority of unsuccessful marketing calls, how to accurately locate the cause, how to analyze and improve from the perspective of customers, and for successful marketing calls, how to analyze the success points and copy and promote, currently lacks a better voice Search and analysis methods, unable to obtain structured information from recordings for analysis and improvement

Images