Trust-first Shopping Agents Win
Stuart Kearney, co-founder of Vetted, on AI agents in shopping
The winning shopping product is usually the one that removes risk, not the one that shows the most options. Amazon trained users to expect the same basic flow every time, clear reviews, fast fulfillment, familiar returns, and a merchant they already trust. Google Shopping widened supply by pulling in many sellers, but that also pushed trust work back onto the shopper. That is the gap specialized shopping agents are trying to close.
-
Vetted is built around the messiest part of commerce, research before checkout. The product tries to narrow choices, compare items, and route buyers to a merchant only after the hard decision is mostly made. That mirrors the idea that trust is earned upstream, before payment details are entered.
-
Google has kept improving Shopping, using its Shopping Graph and AI powered shopping home, but the structure still depends on aggregating listings from many merchants. Google has also had to add trust badges and merchant scorecards, which shows that selection alone is not enough without predictable service quality.
-
The same reliability problem shows up in AI checkout. Perplexity limits instant purchase flows to compatible merchants and selected products, and falls back to merchant checkout when it cannot support the full transaction path. In shopping, every broken stock check, return mismatch, or failed order quickly damages repeat behavior.
The next wave of shopping winners will look less like open ended search engines and more like controlled trust layers. They will combine recommendation quality with tighter merchant coverage, cleaner product data, and more predictable fulfillment. The company that makes buying feel routine and safe across categories will own the default starting point.