Based on self-expansion theory, this study examines the influence mechanism of artificial intelligence interaction types (partner-style vs. servant-style) on customer experience. Through multi-scenario experiments and a field experiment, the direct, mediating, and moderating effects of interaction types on customer experience were analyzed. The results revealed that compared to servant-style interaction, the partner-type interaction can stimulate a higher level of perceived control, warmth perception and customer experience. Moreover, perceived control and warmth perception play a mediating role in the impact of artificial intelligence interaction types on customer experience. Meanwhile, service types and self-perceived uniqueness play a moderating role in the impact of interaction types on customer experience. Specifically, for hedonic services and customers with high self-perceived uniqueness, partner-style interaction has a more pronounced effect on perceived control and warmth perception. However, for functional services and customers with low self-perceived uniqueness, the difference between the two interaction types is not significant.