How Perceived AI-Driven Personalization Is Associated with Online Purchase Intention in Tourism E-Commerce: A Moderated Sequential Mediation Model of Perceived Relevance and Trust in AI Recommendations

This study examines how perceived artificial intelligence (AI)-driven personalization is associated with online purchase intention in tourism e-commerce. It develops a moderated sequential mediation model in which perceived recommendation relevance and trust in AI recommendations represent distinct evaluative pathways, while perceived AI response creativity and AI recommendation transparency condition different relationships within the proposed model. Data were collected from 871 consumers with recent experience using personalized recommendations on AI-enabled tourism platforms or travel assistants. Respondents were recruited through experience-based purposive sampling and screened to ensure direct and recallable experience with AI-generated travel recommendations. The proposed relationships were examined using partial least squares structural equation modeling. Perceived AI-driven personalization was positively associated with perceived recommendation relevance, trust in AI recommendations, and online purchase intention. Perceived recommendation relevance and trust in AI recommendations were also positively associated with online purchase intention, while perceived recommendation relevance was positively related to trust in AI recommendations. Both constructs showed significant separate indirect associations between perceived AI-driven personalization and online purchase intention, and the theoretically ordered sequential indirect pathway was also statistically supported. Because the direct association remained significant and operated in the same direction as the indirect pathways, the observed statistical pattern was consistent with complementary partial mediation rather than evidence of a causal mediation mechanism. Perceived AI response creativity positively moderated the association between perceived AI-driven personalization and perceived recommendation relevance, whereas AI recommendation transparency positively moderated the association between perceived recommendation relevance and trust in AI recommendations. The study distinguishes perceived AI-driven personalization from perceived recommendation relevance by treating the former as consumers’ assessment of adaptive recommendation tailoring and the latter as their judgment of recommendation–decision fit. It further conceptualizes perceived recommendation relevance and trust in AI recommendations as distinct organism evaluations reflecting cognitive fit and reliance-oriented judgment, respectively. Creativity and transparency are positioned as relationship-specific boundary conditions rather than temporal stages, providing a more differentiated account of the theoretically ordered statistical pathways linking perceived AI-driven personalization with online purchase intention in tourism e-commerce.