Can ChatGPT actually see your Shopify products?
Not always. We are writing the next Genesis paper on exactly this, and the first pass says the problem sits in Shopify’s catalog layer: when a product had an official match in Shopify’s Global Catalog, ChatGPT showed an official product card in 88% of the cases we checked live. When the match was missing, or pointed at the wrong seller, the buy path broke.
What we did.
We took 40 products from 8 Shopify brands, 5 products each: bestsellers, high-ticket items, bundles, and old edge products. For every product we checked three things. Is it in the brand’s own Shopify Storefront Catalog? Does Shopify’s Global Catalog return it with the brand’s official domain attached? And when a shopper asks ChatGPT for that exact product, does ChatGPT render an official product card?
Three ways it breaks.
The failures were not random. One brand was invisible in the Global Catalog for all 5 of its products, and every one of its queries returned a non-official seller first: resellers and stale listings wearing the brand’s name. Other products were readable in the brand’s own storefront feed but never made it to the global layer. And some products had a clean official path and a retailer path sitting side by side in the same answer, which means the brand is fighting for its own sale.
In the words we use across the site: some products own the path, some leak it, and some have no path at all. This study is about finding which is which at the product level, before a single fix is proposed.
What to check on your own store this week.
Ask ChatGPT for your top 5 products by exact name.
Look at what renders: your card, a reseller, or nothing.
Check who the first seller is.
A wrong first seller means someone else is positioned to take your sale.
Watch for retailer links next to your name.
Being named and being bought from are different outcomes.
First pass of the next Genesis paper, in progress: 40 products, 8 brands, measured July 2026 against Shopify Storefront Catalog, Shopify Global Catalog, and live ChatGPT product prompts. The full study targets 200 products across 40 brands and 5 categories.
Honest caveats: only 11 of 40 products got live ChatGPT checks before ChatGPT rate-limited the session, so the 88% figure is early signal, not a conclusion. This is also correlational. It tests whether the catalog is a useful diagnostic, not whether Shopify’s Global Catalog directly powers ChatGPT shopping.
Want this run on your store?
We run real shopping prompts for your category and show you where the recommendations land: your store, a reseller, or nowhere. One form, no meeting.