Towards Better Customer Experience: Models for Automatic Detection of Potential Misuse or Mistakes in Customer Care Process

Abstract

In this study we explore AI methods and models for intelligent customer support solutions that could help automatically supervise the support case handling process and detect potential mistakes or misuse in case handling that may lead to an unwanted customer experience. We developed different types of proof of concept classification models for automatic prediction of case status based on its conversation/activity history. We also conducted LDA topic modelling analysis to help obtain insights into the characteristics of different types of customer care cases.