Don't Neglect the User! - Identifying Types of Human-Chatbot Interactions and their Associated Characteristics

Interactions with conversational agents (CAs) become increasingly common in our daily life. While research on human-CA interactions provides insights into the role of CAs, the active role of users has been mostly neglected. We addressed this void by applying a thematic analysis approach and analysed 1000 interactions between a chatbot and customers of an energy provider. Informed by the concepts of social presence and social cues and using the abductive logic, we identified six human-chatbot interaction types that differ according to salient characteristics, including direction, social presence, social cues of customers and the chatbot and customer effort. We found that bi-directionality, a medium degree of social presence and selected social cues used by the chatbot and customers are associated with desirable outcomes in which customers mostly obtain requested information. The findings help us understand the nature of human-CA interactions in a customer service context and inform the design and evaluation of CAs.

Automated summary

Analyzing human-chatbot interactions in a customer service context revealed six different interaction types. The study found that bi-directionality, moderate social presence, and selected social cues contribute to desirable outcomes, where customers mostly obtain the requested information.