Unveiling Practices and Challenges of Machine Teachers of Customer Service Conversational Systems

Abstract

This paper describes a set of qualitative interventions which aimed to unveil the challenges of teaching conversational machines used in the automatic customer service. The study aimed to understand the activity of mapping information into conversational systems platforms to create chatbots (text or voice-based) to attend end-users in conjunction with call centers. We interviewed eleven domain experts with non-machine learning skills responsible for curating the content of the chatbots. Additionally, we conducted four design workshops with experienced curators to understand deeper their challenges when teaching novice curators. We list a group of challenges and opportunities for improving the machine teacher's practices and supporting decision making.