Getting recommended by AI is a useful first step. It is not the same thing as earning a purchase. The shopper still has questions, and many of the answers live somewhere else.

Discovery and checkout are different jobs

Digiday’s reporting on Feedvisor’s shopper survey puts a number on the distinction: 16% of surveyed shoppers started looking for products on ChatGPT, while 9% went there to buy. Only 5% said they bought immediately when an AI assistant recommended a product.

Those are separate survey responses, not a conversion funnel. But they describe a familiar behavior: use one place to build a shortlist, then go elsewhere to check the price, read reviews, and decide.

The recommendation has to survive a second look

PartnerCentric’s research found that 44% of surveyed shoppers wanted reviews before acting on an AI recommendation. Its sample was not representative of all U.S. consumers, but it gives us a useful view of the checks people make around AI-influenced shopping.

Think about a shopper asking for a carry-on suitcase for frequent work trips. AI might recommend one for durability and a generous warranty. The shopper then finds recent reviews about broken wheels, a retailer listing with different dimensions, or warranty terms that exclude the damage they care about. The recommendation got the product considered. The supporting information did not close the gap.

This is why I would look beyond whether a brand appears. What did the answer promise? Which source supported that promise? Does the product page say the same thing?

Measure the handoff, not just the mention

For an ecommerce team, AI visibility belongs alongside the rest of the buying journey. Monitor the product attributes people ask about, check the facts in the answers, and inspect the reviews and retailer pages supporting those claims.

Then keep commercial outcomes separate and measurable. AI referral traffic, product-page engagement, and purchases can help show what happens next. They do not capture every shopper who reads an AI answer and returns through another channel.

My bet is that the brands that benefit most will make the next check easy. Clear specifications, credible reviews, consistent pricing, and straightforward delivery and return terms are useful to the shopper whether or not an AI assistant sent them.