Want to be surprised? How Novelty and Familiarity Shape Satisfaction with Conversational Agents

This study investigates the impact of perceived humanness in conversational agents (CAs) on users' service satisfaction, focusing on the roles of perceived familiarity and novelty. Drawing on CASA theory and Expectation Confirmation Theory (ECT), we hypothesize that humanness enhances service satisfaction through these mechanisms. An online experiment with 139 participants compared interactions with human-like and non-human-like designed chatbots. Results indicate that perceived humanness significantly increases both familiarity and novelty, but only familiarity positively influences service satisfaction. This highlights that users appreciate the familiar aspects of human-like CAs, but novelty does not significantly enhance satisfaction. The findings contribute to the theory by revealing the mechanisms linking humanness to satisfaction and suggest practical implications for CA design, emphasizing the importance of creating interactions that feel familiar to boost user satisfaction.