Expert or partner: The matching effect of AI chatbot roles in different service contexts

Abstract

Anthropomorphizing AI chatbots has become a widely adopted strategy to enhance customer-chatbot interactions. However, prior research has largely overlooked the role of social anthropomorphism, particularly how assigning different social roles to AI chatbots influences customer acceptance. To address this gap, this research investigates the impact of specific social roles across various service contexts on customer acceptance and the mechanisms underlying this effect. Through four experimental studies conducted in both field and laboratory settings, the findings consistently reveal a significant matching effect between AI chatbot roles and service contexts on customer acceptance, as well as the mediating roles of perceived competence and perceived warmth. Specifically, in utilitarian-dominant services, customers preferred expert (vs. partner) chatbots because they were perceived as more competent. Conversely, in hedonic-dominant services, customers favored partner (vs. expert) chatbots because they were perceived as warmer. These findings contribute to the understanding of customer acceptance of AI chatbots by highlighting the influence of various AI roles in different service contexts, and offer practical implications for companies to enhance the effectiveness of AI chatbots through role-matching strategies.