AI customer service: Task complexity, problem-solving ability, and usage intention

Abstract

Artificial intelligence (AI) in the context of customer service, we define as a technology-enabled system for evaluating real-time service scenarios using data collected from digital and/or physical sources in order to provide personalised recommendations, alternatives, and solutions to customers’ enquiries or problems, even very complex ones. We examined, in a banking services context, whether consumers preferred AI or Human online customer service applications using an experimental design across three field-based experiments. The results show that, in the case of low-complexity tasks, consumers considered the problem-solving ability of AI to be greater than that of human customer service and were more likely to use AI while, conversely, for high-complexity tasks, they viewed human customer service as superior and were more likely to use it than AI. Moreover, we found that perceived problem-solving ability mediated the effects of customers’ service usage intentions (i.e., their preference for AI vs. Human) with task complexity serving as a boundary condition. Here we discuss our research and the results and conclude by offering practical suggestions for banks seeking to reach customers and engage with them more effectively by leveraging the distinctive features of AI customer service.