Comparative Effects of AI-Driven Recommendations and Amazon Live Video on Online Cosmetic Purchase Decisions

This research examines the relative impact of AI-based personalization and Amazon Live Video in consumer decision-making within Amazon.in in the cosmetic category. On a sample of 350 respondents we have examined the perceptions during two high involvement stages of purchase, discovery and validation. Reliability Analysis confirmed that both AI Personalization (Cronbach’s α = 1.000) and Video Engagement (α = 0.99) scales have excellent internal consistency. A t-test (one sample) showed that there were significantly positive perceptions of AI (mean = 3.55, p < 0.001), but constrained by static imagery. On the contrary, Live Video rated higher (mean = 4.11) and had significant effects on trust and confidence to purchase (p < 0.001). Regression analysis indicated that there were perfect predictive relations in each construct, which indicated uniformity of responses. The hypothesis testing proved that although AI will increase efficiency, Live Video will have more persuasive capacity- especially in shade check and live demonstration. The comparative analysis (H₃) revealed that there was a steady preference towards video during discovery and validation (mean preference = 2.61, p = 0.061), and there was no decline in involvement. Results indicate that the dynamic, human-created video content fills in sensory loopholes created by algorithmic recommendations and, therefore, it is a more influential factor in highinvolvement cosmetic purchases.