Emulating Empathetic Behavior in Online Service Encounters with Sentiment-Adaptive Responses: Insights from an Experiment with a Conversational Agent

Conversational agents currently attract strong interest for technology-based service provision due to increased capabilities driven by advances in machine learning and natural language processing. The interaction via natural language in combination with a human-like design promises service that is always available, fast, and with a consistent quality and at the same time resembles a human service encounter. However, current conversational agents exhibit the same inherent limitation that every interactive technology has, which is a lack of social skills. In this study, we make a first step towards overcoming this limitation by presenting a design approach that combines automatic sentiment analysis with adaptive responses to emulate empathy in a service encounter. By means of an experiment with 112 participants, we evaluate the approach and find empirical support that a CA with sentiment-adaptive responses is perceived as more empathetic, human-like, and socially present and, in particular, yields a higher service encounter satisfaction