Anthropomorphic service robots in service failure: The roles of warmth, competence, and failure severity in shaping complaint intentions

As service robots become more prevalent in frontline service delivery, understanding their role in handling customer complaints is critical. Drawing on the stereotype content model, this study examines how robot anthropomorphism influences customers' willingness to complain, mediated by perceived warmth and competence. Two independent scenario-based experimental studies were conducted (Study 1: humanoid robot, N = 354; Study 2: non-humanoid robot, N = 351). Using a combination of structural equation modelling (SEM) for testing the overall theoretical structure and Hayes' PROCESS macro (Model 4 and Model 15) for estimating mediation, moderation, and conditional indirect effects, we examined how perceived anthropomorphism shapes complaint intentions under varying levels of service failure severity. Results show that humanoid robots foster higher complaint intention through warmth and competence perceptions. Failure severity moderates the mediator–outcome paths, producing distinct conditional effects across mild versus severe failures: competence drives complaint intentions in severe failures, whereas warmth is more salient in mild ones. These patterns are consistent for humanoid robots but less stable for non-humanoids. By integrating stereotype content theory with service recovery research, this study advances understanding of human–robot interactions in hospitality and provides actionable guidelines for deploying service robots in complaint handling.