Abstract: The traditional manual method is mostly used for the quality inspection of outbound voice,and it is time-consuming,laborious and inefficient to judge customer satisfaction through listening to recordings. Speech emotion analysis of service hotline complaints is the key link in operator customer service and customer perception quality inspection. This paper proposes an intelligent voice emotion quality inspection analysis technology to realize the intelligent service operation for home broadband outbound call voice. By mining the value of complaint voice files and conducting quality inspection based on the recognized customer emotion and semantic information,it is possible to know the potential dissatisfaction tendency of customers in advance and take targeted measures to continuously improve users’ service experience.

Key words: emotional computing, intelligent quality inspection, natural language understanding, robot process automation