Effects of Voice-Based AI in Customer Service: Evidence from a Natural Experiment

Voice-based AI systems are gradually deployed to replace traditional interactive voice response systems in call centers. However, there is little evidence on how the implementation of AI systems impacts customer behavior as well as the effect of AI systems on call center performance. Using data from a natural field experiment, we examine how the introduction of voice-based AI in a large telecommunication services call center affects call length, customers’ demands for human service, and customer complaints. We find that the implementation of AI significantly increases call length and decreases customer complaints. Although presumably the AI-enhanced service system reduces users’ efforts to transfer to human agents, we do not find a significant increase in customers’ demands for human service. In addition, regarding simple service tasks, the AI-enhanced service system reduces customer complaints for both experienced and inexperienced customers. For relatively complex tasks, customers learn from prior experience of interacting with the AI system; as this learning effect leads to the decrease of complaints. What’s more, our results suggest significant heterogeneity in the effects of the AI-enhanced service system by indicating that the AI-enhanced system has a significantly larger effect on reducing customer complaints for older and female customers as well as for the customers who are experienced in using the IVR system