Generative AI-Powered Personalization and Consumer Purchase Intention in E-Commerce

Generative AI allows e-commerce platforms to personalise not only which products are shown but the descriptions, images, offers and messages through which they are presented, tailored to each consumer in real time. Personalisation of this depth raises relevance, but it also raises consumers’ sense of being manipulated, and the net effect on purchase intention depends on which response dominates and on whether the platform discloses what it is doing. This paper models purchase intention as the outcome of two mediating responses, perceived relevance and perceived manipulation, both driven by the degree of generative personalisation, with disclosure reducing manipulation perceptions. A simulation of 900 consumers shows that personalisation raises purchase intention through relevance (standardised indirect effect 0.72) and lowers it through manipulation (−0.50), so that the net gain is less than a third of the gross relevance benefit. Disclosure changes the balance: the personalisation–disclosure interaction is positive (0.11), and full personalisation raises purchase intention from 0.28 to 0.36 without disclosure but from 0.32 to 0.50 with it. The paper argues that generative personalisation should be disclosed, that platforms should manage manipulation perceptions as deliberately as they optimise relevance, and that consumers’ privacy dispositions set a ceiling on how far personalisation can go. Design and research implications are derived.