The interactive influence of service recovery agent and language style on consumer forgiveness: mediation of perceived sincerity

Abstract

This study aims to examine how the interaction between service recovery agent (human vs robot) and language style (formal vs informal) affects consumer forgiveness and investigates the mechanisms and boundary conditions of this effect, grounded in complementary fit theory. Four scenario-based experiments were conducted across hotel, restaurant, and e-commerce contexts. A total of 866 participants were recruited online. Hypotheses were tested using ANOVA, moderated mediation analysis (PROCESS), and simple-effects analysis. Forgiveness is enhanced when robots use informal language and humans use formal language, with the effect mediated by perceived sincerity. This effect is stronger for robot-informal pairings in low-severity failures or among low-loyalty customers, and for human-formal pairings in high-severity failures or among high-loyalty customers. This research makes three key contributions: (1) It challenges the dominant “consistency-is-good” logic in service recovery by showing that forgiveness is maximized through complementary fit rather than role congruence. (2) It identifies perceived sincerity as the core mechanism through which agent-language compensation translates into forgiveness. (3) It offers a context-sensitive framework that clarifies when each compensatory pairing is most effective, providing a nuanced alternative to one-size-fits-all recommendations for deploying service robots.