Exploring the effect of empathic response and its boundaries in artificial intelligence service recovery

Abstract

Artificial intelligence (AI) is increasingly applied to customer services, but failures inevitably occur from time to time. Empathy-based response through AI is a service recovery approach, but its effectiveness and boundary conditions remain unclear. In this study, we draw on the theory of system usage to investigate the conditions in which AI empathic responses can be an effective remedy, and explore the differences in the effectiveness of empathic responses from AI and human agents in service failures. The results of three scenario-based experiments reveal that the recovery of AI services through the use of high-empathy responses can increase users’ intentions to continue using the service. Our findings also suggest that the need for human interaction increases the impact of empathic responses on continued usage intention, and that this effect is reduced if the task is deemed urgent. We also find that the recovery effect from the empathic responses provided by both AI and human agents is similar. This article focuses on the issue of continued use when AI services fail, and we demonstrate the effectiveness of AI empathic responses in service recovery and reveal their boundary conditions. This provides insights into how to effectively use AI empathy in service recovery and provides a theoretical framework that companies can draw on when addressing AI service failures.