Tourists are increasingly relying on artificial intelligence (AI) recommendations for travel decisions, but rigid content severely limits recommendation persuasiveness. This paper examines the matching effect and mechanism between AI role type and destination personality in the tourism recommendation process, drawing on perceived fit theory, through three situational experiments. Results show that servant AI enhances tourists’ information adoption when recommending competent destinations, while partner AI better motivates decision willingness for warm destinations. Perceived usefulness and perceived emotional support play the mediating roles in this process. The tourist decision focus preference (desirability vs. feasibility) is an important boundary condition of this interaction effect. Theoretically, this paper proposes a “role-situation” adaptation framework for tourism AI recommendations, providing a novel perspective on the effects of AI anthropomorphism. Practically, this study offers a new direction for optimizing and upgrading the tourism AI recommendation system.
Zeng et al. (2026) studied this question.