The use of artificial intelligence (AI) service robots is on the rise. With service frontlines gradually shifting to human–robot interactions, the question of whether AI robots should be humanlike or machinelike has emerged. While many firms use robots that resemble humans in their appearance and actions, others use machinelike robots, assuming that very humanlike robots may lead to uncanny valley effects. There is no consensus on whether the anthropomorphism of service robots facilitates or constrains consumers’ experiences. To fill this gap, this article examines when and how service companies should use anthropomorphic AI service robots.