The news: Google’s Gemini AI recommends only a thin slice of online stores, and most retailers never surface at all, per Recomaze.

• In an analysis of 9,720 ecommerce stores, 60% won none of their six purchase-intent queries.
• Stores were named in just 14% of 58,320 tests.
• Recommendations were scattered across 50,287 brands, with focused, single-category stores, like Patagonia, appearing most often.

Recomaze’s tests show AI recommendations pooling among a handful of catalogs engines can read.

• Etsy, which ranked No. 1, drew just 1.3% of recommendations, with the top 100 brands taking 11% combined.
• 71% of winning brands appeared exactly once across the 58,320 tests.
• Home & Living ranked least visible, with 74% of its stores absent from results.

Why this matters: AI assistants now decide which stores a shopper ever sees, and the traffic they send is large and growing. Traffic from AI sources to US retail sites grew 393% YoY in Q1 2026, and those visits converted 42% better than non-AI traffic in March, per Adobe Analytics.

Consumer demand is moving toward genAI. We forecast that more than half of genAI users will use AI engines for shopping this year, and 34% of US internet users will shop on AI platforms by 2030. Right in the middle, by 2028, we expect 60% of genAI users and 32.8% of US internet users will use Gemini.

Size no longer guarantees a spot in the results. “The stores that get recommended are not the biggest ones. They are the ones [who have] product information an engine can actually read and trust,” said Recomaze founder and CEO Delian Coroamă, per The Next Web.

Recommendations for retailers: Surfacing in AI shopping results comes down to boosting appearances across community forums and feeding engines data they can read and trust.

• Structure core product data, including clear titles, attributes, and specifications, so engines can match your catalog to specific queries.
• Write product descriptions that spell out use context—if the product is treadless boots, name the conditions they suit, from workwear to city nights, and where they fall short, like snowy climates.
• Add alt text to images and captions to videos, including customer review photos and clips, giving engines more to parse. Details matter.
• Include author names and dates on blog posts, work in relevant keywords without stuffing, and refresh posts at least twice a year, since engines favor current content.
• Tailor content for Google search as well as Gemini, since the two surface different results. SEO and generative engine optimization (GEO) are equally important.