AI shopping/answer engines (ChatGPT, Perplexity, and similar) increasingly decide what to recommend based on the depth and content of a product's customer reviews rather than just its listing copy or ads — pulling reviews from public sources like Reddit, Yelp, and Google, not only from the seller's own site.
This creates a causal chain: more (and more substantive) reviews, seeded across multiple public platforms, raises the odds of being surfaced by an AI shopping assistant. Per Stackline, ChatGPT users alone make over 84 million shopping queries per week, and brands that repeatedly surface in AI recommendations tend to share a deep well of customer reviews — Fireclay Tile's CEO credits reviews as "1 million percent" driving their AI visibility.
A platform can also block this signal from competitors: Amazon has quietly blocked OpenAI's bots from crawling Amazon content, including reviews, so ChatGPT-style assistants cannot see Amazon reviews at all. This inverts Amazon's usual review-depth moat and forces brands to duplicate review-gathering effort off-Amazon (Reddit, Yelp, Google, brand's own site) to be visible to AI shopping engines.
Tactically, brands are timing review requests for substantive content rather than raw volume — e.g. Paco (dog food) offers $20 off after a customer's third order in exchange for a review, waiting 2 weeks after a one-time order or three full reorder cycles for subscriptions before asking.
Related: GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), Amazon Review Acquisition Tactics Risk Spectrum.