How AI May Impact Customer-Intensive Services

Abstract

Artificial intelligence (AI) technologies are increasingly employed in customer-intensive services to automate auxiliary tasks and enhance service efficiency. We investigate a service system in which the overall service rate depends on both auxiliary and professional service times. AI effectively reduces the time allocated to auxiliary tasks. The customer-surplus maximizing service provider, like healthcare providers, decides the admission rate of customers and the professional service time spent on customers. Our findings reveal that marginal improvements in AI capability always enhance market coverage and the total customer surplus but, somewhat surprisingly, may reduce service quality and individual customer utility. Depending on the service type, individual customer utility may follow different patterns-either always increasing, initially decreasing, or initially increasing and then decreasing-before eventually increasing, as AI capability improves. Only at a sufficiently high AI capability can the service provider enhance both service quality and individual customer utility. In contrast to traditional systems without AI, AI assistance tends to lower service quality unless AI capability reaches a sufficiently high level. Furthermore, we uncover an inverted U-shaped pattern in the optimal system utilization as AI capability increases and reveal a paradox: while AI improves the efficiency of auxiliary tasks, it can actually increase the provider's workload. Our study shows that with assistive AI to improve the operational efficiency of auxiliary tasks, the customer-surplus maximizing service provider likely reduces service quality and individual customer utility, but with more customers served and boosted total customer surplus on the demand side while creating more hiring needs for labor on the supply side.

Keywords: Artificial Intelligence, Customer-Intensive Service, Consumer Surplus, Labor Market