AI guest personalization as driver of hotel performance and hotel guest behavior: A dual-model analysis

Artificial intelligence (AI)–driven personalization is reshaping hotel operations, yet evidence on its financial effects and guest acceptance mechanisms remains limited. This study adopts a dual-perspective design combining hotel-level performance modeling with guest-level behavioral analysis. Study 1 examines revenue dynamics using Interrupted Time Series (ITS), SARIMA forecasting, and TOPSIS, comparing AI-driven and non-AI-driven hotels based on publicly documented implementations of AI personalization. Findings indicate heterogeneous short-term responses but more consistent differences in long-term revenue stability and predictability. Study 2 develops a Diffusion of Innovation–based structural model extended with trust in AI as a mediator, using survey data from 1396 hotel guests. Relative advantage and compatibility increase trust and adoption intention, whereas complexity reduces them; trust partially mediates these effects. Multi-group analysis reveals model differences between business and leisure guests. Overall, the study provides complementary evidence on revenue dynamics and adoption mechanisms of AI personalization in hospitality.