The More Realism, the Better? How Does the Realism of AI Customer Service Agents Influence Customer Satisfaction and Repeat Purchase Intention in Service Recovery

Nowadays, human customer service is increasingly being replaced by artificial intelligence (AI) customer service agents. Service recovery plays a crucial role in shaping consumer experiences and business profitability. The realism of AI agents can significantly impact users' attitudes and behaviors. However, it remains unclear how different types of realism in AI customer service agents affect customers during service recovery. Drawing on social response theory and expectation-confirmation theory, this study explores the impact of AI agents' realism on consumer satisfaction and repeat purchase behavior during service recovery, as well as the underlying mechanisms of this effect. We collected data from 784 participants in three studies. Study 1 results show that form realism and behavioral realism of AI customer service agents affect customer satisfaction in successful service recovery situations. Study 2 indicates that the interaction effect of form realism and behavioral realism only influence satisfaction customer satisfaction and repeat purchase intention in terms of failed service recovery. Study 3 further explores the mechanism of action in failed service recovery, finding that perceived warmth and competence mediate the relationship between realism and satisfaction. We systematically examine the influence of the realism of AI customer service agents on consumer satisfaction across distinct success and failure service recovery scenarios, effectively addressing a critical research void. Additionally, our findings offer valuable insights to business managers, empowering them with actionable strategies for enhancing consumer satisfaction across varied consumption contexts and effectively mitigating the repercussions of consumption failures.