The effect of the anthropomorphic design of chatbots on customer switching intention when the chatbot service fails: An expectation perspective

Abstract

An increasing number of enterprises are deploying chatbots to enhance service efficiency and reduce customer service costs. However, many customers prefer to switch to a human agent, especially when a chatbot service fails. From an expectation perspective, this study explores whether and how chatbots’ anthropomorphic design affects customers’ intention to switch from chatbots to human agents, as well as the mediating role of service recovery expectations in this relationship. Using a sequential mixed-methods approach, we initially conducted two quantitative studies, encompassing an online experiment (n = 404) and a survey (n = 338), to validate the causal relationships and the strength of the relationships between chatbot anthropomorphic design, customer service recovery expectations, and switching intention. Subsequently, in a qualitative study, we conducted semi-structured interviews (n = 10) to confirm and complement the quantitative study results. Collectively, the findings provide a nuanced understanding of the impact mechanism associated with chatbots’ anthropomorphic appearance and communication style concerning customer switching intentions, with implications for both researchers and practitioners.