The effect of robot anthropomorphism on revisit intentions after service failure: a moderated serial mediation model

Purpose The purpose of this study is to examine how robotic anthropomorphism and personalized design may affect consumers' reactions to brands after service failure. Design/methodology/approach This study conducted two studies based on cognitive appraisal theory and artificial intelligence device acceptance theory. Study 1 explored the mechanisms by which the type of anthropomorphic design of the service robot (humanoid robot/nonhumanoid robot) influenced revisit intention after service failure through a one-factor between-subjects design based on a restaurant dining scenario. Study 2 was based on a hotel check-in scenario and explored the moderating effect of robot personalization design on the above mechanisms through a 2 (anthropomorphic design: humanoid robot/nonhumanoid robot) × 2 (personalized design: self-name/no name) between-subjects design. Findings Study 1 shows that consumers have higher performance expectations for nonhumanoid robots, leading to a higher tolerance for service failure, which in turn generates higher revisit intentions. Study 2 shows that consumers' performance expectations are significantly enhanced after custom naming of humanoid robots, so the serial mediation mechanism for the effect of robot anthropomorphic design on revisit intention does not hold. Originality/value This study extends the research of artificial intelligence device acceptance theory in the field of service failure and exploratively proposes an intervention mechanism for the negative effects of the anthropomorphic design of service robots.