Retailers across India have turned to algorithmic recommendation engines, dynamic pricing, and chatbot-driven service to keep shoppers from drifting to rival platforms, yet whether these tools actually translate into durable loyalty remains contested among practitioners and scholars alike. This study examines how artificial intelligence-enabled personalization shapes customer retention in e-commerce, with perceived value, satisfaction, and loyalty positioned as the mechanisms through which that influence travels. Data were gathered from 365 online shoppers in Tiruchirappalli district, Tamil Nadu, using a structured questionnaire administered through both physical and digital channels. Exploratory factor analysis confirmed a five-factor structure explaining 71.46 percent of total variance, and confirmatory factor analysis verified acceptable convergent and discriminant validity across all constructs. Structural equation modelling, conducted using AMOS, indicated that personalization exerts a significant positive effect on perceived value (β = 0.52, p < .001) and a smaller direct effect on satisfaction (β = 0.31, p < .001), while perceived value itself strongly predicts satisfaction (β = 0.46, p < .001). Satisfaction, in turn, drives both loyalty (β = 0.58, p < .001) and retention directly (β = 0.27, p < .001), and loyalty further strengthens retention (β = 0.49, p < .001). The model accounted for 51 percent of the variance in retention. These findings suggest that personalization works mainly by enhancing how customers value an experience rather than by satisfying them outright, a nuance that complicates the assumption common in industry reports that more data-driven customisation automatically yields stickier customers. Practical implications for e-tailers operating in Tier-II Indian markets are discussed, along with the boundaries of what algorithmic personalization can reasonably be expected to achieve.