Chatbots at the frontline: Unveiling antecedents of customers’ willingness to accept chatbot intervention in service recovery

Abstract

This study proposed a model to explain customers’ willingness to accept chatbot intervention in service recovery. Grounded in the realism maximization theory and the service robot acceptance model, it investigated whether the incorporation of realism in chatbot messages (through message relevance and message warmth) can enhance service recovery outcomes. An online survey study was conducted recruiting 319 participants in the United States who experienced interactions with chatbots, and SPSS PROCESS Macro bootstrapping was used to test the hypotheses. The results confirmed that message relevance and message warmth negatively affect uniqueness neglect, which in turn positively affects perceived chatbot artificiality, and negatively affects perceived chatbot rapport, and perceived chatbot intelligence. The results also confirmed the mediating role of uniqueness neglect, perceived chatbot rapport, and perceived chatbot intelligence. This research contributes to the understanding of how chatbot realism can improve service recovery. It also offers actionable insights for managers to craft relevant and warm chatbot messages to build chatbot capabilities that emphasize rapport and intelligence to enhance service recovery outcomes.