Problem definition: Despite rapid advances in artificial intelligence, the adoption and effective use of customer service chatbots remain slow relative to their capabilities. This paper explores the behavioral reasons for these adoption hurdles. Methodology/results: We use incentivized online experiments to study chatbot uptake. The results of these experiments are threefold. First, people respond positively to improvements in chatbot performance; however, the chatbot channel is used less frequently than expected-time minimization would predict. A key driver of this underutilization is reluctance to engage with a gatekeeper process (i.e., a process with an imperfect initial service stage and possible transfer to a second expert service stage—a behavior that we term gatekeeper aversion). Second, we find that gatekeeper aversion can be further amplified by an additional hurdle—algorithm aversion. Third, we find that chatbot adoption decreases when stakes are higher and when the human/algorithmic nature of the server is manipulated with more realism. Managerial implications: We use an illustrative case to show how the behaviors identified in our experiments affect optimal technology investment and staffing levels and how failing to anticipate these behaviors can lead to suboptimal decisions and higher realized costs. More broadly, our results suggest that adding a chatbot to a service system requires rethinking the entire service process, including technology investment, staffing, and queueing policies. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2024.1141 .
Kagan et al. (Tue,) studied this question.