Forced chatbot interactions – investigating the role of customer expectations in assigning attributions for chatbot failure

Abstract

Customers are increasingly being required to interact with chatbots to self-serve, sometimes encountering no other option for service delivery. Such instances lead to customers being forced to interact with chatbots. We argue that when chatbot interactions result in service failure, forced interactions will have a distinct impact on the way customers attribute responsibility for service failure. Experimental research was employed to test the research framework in a customer service setting. Our study found that forcing customers to interact with chatbots, results in more negative responsibility attributions of controllability, stability and causality towards the company, when that service fails. The study also found support for the mediating effect of disconfirmation of expectations. Forcing customers to interact with chatbots results in a more negative disconfirmation of expectations, which in turn leads to stronger attributions of controllability, stability and causality. Such attributions are moderated by the type of failure, such that customers make stronger attributions for technology failures than for process failures. The findings from this experimental research offer a more substantive understanding of blame attributions in settings in which failure is inevitable, in the process helping to better understand the implications of AI technologies, and to find ways to reduce negative reactions to such technologies.