AI-Powered Personalized Recommendations and Purchase Intention in Online Clothing E-commerce: A Serial Trust-Transfer Mechanism

AI-powered personalized recommendation systems are increasingly used in online retail, but their influence on purchase intention may depend on how consumers interpret recommendations and form trust. Drawing on the Stimulus–Organism–Response model, multidimensional trust, and Trust Transfer Theory, this study examines how perceived personalization (PP), perceived accuracy (PA), and perceived usefulness (PU) relate to purchase intention through competence trust (CT), benevolence trust (BT), and trust in recommended clothing products (TRP). A structured cross-sectional online survey of Chinese consumers with recent online clothing-purchase and personalized-recommendation experience yielded 500 valid responses from 512 collected questionnaires. Data were analyzed using SPSS 27.0 and AMOS 26.0. All six hypothesized direct paths were positive and significant: PP was associated with BT; PA and PU with CT; CT and BT with TRP; and TRP with purchase intention. Bias-corrected bootstrapping with 5,000 resamples supported all three serial indirect pathways from PP, PA, and PU to purchase intention through the corresponding system-trust dimension and TRP. The findings conceptualize the Organism component of S-O-R as a staged trust process and extend Trust Transfer Theory from AI systems to recommended physical products. They also suggest that platforms should combine relevant recommendations with credible product information.