Not All Praise Is Equal: How AI Shopping Assistants' Praise Target Shapes Consumer Behavior

ABSTRACT

Generative AI has led many e‐commerce firms to adopt AI shopping assistants that advise consumers much as human salespeople do. In offline retail, salespeople often praise either the consumer or the product to encourage a purchase. Although praise is among the most common persuasion tactics in selling, its effects in AI‐mediated commerce have drawn little attention, in contrast to tactics such as empathy. This paper examines how the target of an AI shopping assistant's praise (the consumer versus the product) affects consumers' intention to purchase the recommended product. Across three studies, we find that praising the consumer rather than the product lowers purchase intention, and that perceived manipulative intent largely accounts for this effect: when the assistant praises the consumer, consumers infer that it is trying to manipulate them, which lowers their intention to buy. Consumers' AI literacy moderates the effect of praise target on perceived manipulative intent, such that consumer‐directed praise raises perceived manipulative intent more strongly among consumers with higher AI literacy. Together, these findings show that praise can backfire in AI commerce, extend the Persuasion Knowledge Model to AI‐mediated shopping, and offer practical guidance for firms deploying AI shopping assistants.