Research blog · 26 October 2026

Can AI assistants read online shops? 27% of 1,365 shops leave out key product facts

When someone asks ChatGPT, Gemini or Perplexity what to buy, the assistant compares products by a few plain facts: name, price, currency and stock. It reads them from product data (schema.org Product) on each product page.

What we found

• 21% of shops give AI assistants no product data at all.
• 27% have no product data or leave out the price or stock.
• 28% have at least one real problem for AI shopping (blocked agents, an unreadable page, or missing product facts).

By platform

• Other or custom-built: 51% of 304 shops leave out key product facts.
• WooCommerce: 30% of 416 shops leave out key product facts.
• Shopify: 10% of 502 shops leave out key product facts.
• Wix: 0% of 90 shops leave out key product facts.

What to do

• WooCommerce: WooCommerce writes product data itself unless a theme removed it; our plugin puts it back and adds brand, barcode, shipping and returns.
• Custom-built shops: one template in the product page template, filled from the product records, covers every product.
• Other platforms: look for a structured-data or rich-snippets setting or app that outputs schema.org Product with price, currency and availability.

How we counted

1,365 online shops with a product page we could check, checked by Khanyo's research crawler up to 26 October 2026. Each shop counts once (its latest check). We read the page the way AI crawlers receive it (no JavaScript) and the shop's robots.txt as RFC 9309 requires; nothing is estimated. Full method: khanyo.app/how-it-works

Check your own shop

The check takes about a minute and shows exactly which product details AI assistants can read on your shop, and the fix: khanyo.app/