This paper delves into the realm of human-aided service bots and aims to enhance consumers’ experience with online services. Through a situational experimental study, we investigate how conversation initiation (by the human agent vs. by the AI agent) affects consumers’ evaluations of human-aided service bots. The results reveal a positive influence of human agent (vs. AI agent) initiation on consumers’ service evaluation, which is mediated by perceived flexibility and expectation confirmation. Moreover, the study suggests a moderating role of consumers’ aversion to AI. These insights contribute to a deeper understanding of designing human-aided service bots and provide valuable guidance for enhancing the overall customer experience and satisfaction, ultimately benefiting both businesses and consumers.