Understanding the interactivity of artificial intelligence shopping assistants (AISAs) for consumer service is essential to realize the potential of this technology. However, the effectiveness of interaction between AISAs and consumers in triggering purchasing decisions is unclear. This study examines what kind of functional characteristics are presented by the interactivity of shopping assistants and the perception of user value they elicit. It reveals the effect of interaction form (multi- vs. uni-modal) and interaction style (task- vs. social-oriented) on consumers’ intention to accept recommendations by AISA through behavior experiments, which reveals the psychological mechanism and the boundary condition required for this to occur. The results show that the interaction form and style have a similar positive effect on consumers’ intention to accept a product recommendation. In accordance with consumption value theory, the study demonstrates that the impact of interactivity on the intention to accept recommendations is mediated by functional and emotional values. Consumer characteristics play a moderating role in the influence mechanism, indicating that the effect is only operative when consumers’ need for cognition is low. These findings contribute to the theoretical study of consumer digital assistance in the context of online shopping and provide insights for marketing strategies designed to enhance purchase intention for products recommended by AISAs.