The rapid deployment of Artificial Intelligence technologies in Indian banking—including Chabot, fraud detection, robo-advisory, and biometric authentication—has exposed a persistent trust-adoption gap between institutional capability and customer willingness. This study develops and tests an integrated framework synthesizing the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT) to investigate relationships among AI-driven banking features, customer experience (CX), trust, perceived risk, and adoption behavior. Data from 389 banking customers across metropolitan, semi-urban, and tier-2 Indian cities were analyzed using Partial Least Squares Structural Equation Modeling (SmartPLS 4.0). Seven hypotheses were tested via bootstrapping (5,000 subsamples), complemented by mediation decomposition and Importance-Performance Map Analysis (IPMA). AI features strongly enhance CX (β = 0.68, p < 0.001), which predicts adoption (β = 0.54, p < 0.001). Trust operates dually—as direct predictor (β = 0.47, p < 0.001) and moderator amplifying the CX–adoption relationship (β = 0.18, p < 0.01). Perceived risk independently suppresses adoption (β = −0.21, p < 0.001), confirming its distinctness from trust. The model explains 58.3% of adoption variance (R² = 0.583; Q² = 0.412). Serial mediation yields a total indirect effect of β = 0.810. IPMA identifies trust as the highest-priority intervention area. The study contributes to technology adoption theory by demonstrating TAM-UTAUT complementarity and provides actionable recommendations for banks, policymakers, and technology providers....