By Steve Merrill, Founder of WRKNG Digital | August 28, 2026
ChatGPT doesn't rank products the way Google ranks pages. It reads what it can find about a product, matches it to the exact question a shopper asked, and recommends what fits. If your data is thin, you're not in the conversation.
I've run more than 40 AI-readiness audits on Shopify stores this year. Most of them never surface in ChatGPT. Not because the products are bad. Because the machine can't read them.
Here's how the decision actually works.
How does ChatGPT choose which products to recommend?
ChatGPT recommends products it can pull clean, structured data about and connect to a shopper's intent. It leans on product feeds, structured markup, and partner catalogs. When those signals are missing, your store gets skipped.
OpenAI has been building shopping into ChatGPT, including product results and a checkout flow with partners. Shopify confirmed a connection that lets merchant catalog data reach ChatGPT. That pipe only carries what's actually in your data. Blank fields carry nothing.
Think about it like a salesperson. Price. A real title. Materials. Sizes. Who it's for. If a shopper would ask it, the model needs it written down somewhere it can read.
What signals does ChatGPT weigh most?
The strongest signal is complete, specific product data. Title, description, price, availability, brand, category, and attributes like material, gender, age group, color, and size. The more of these you fill with real detail, the more shopper questions your product can answer.
Most stores stop at a title and two sentences. "Blue Jacket" tells the model nothing. "Men's Waterproof Insulated Winter Jacket, Recycled Nylon" tells it who it's for, what it does, and what it's made of.
Google publishes a product data spec that maps closely to what AI systems want. Read the Google Merchant product data specification and treat it as your floor, not your ceiling.
Does structured data change what ChatGPT sees?
Yes. Product schema in JSON-LD is the machine-readable version of your product, and it's the fastest signal for AI to parse. Name, description, brand, offers, price, and availability all need to be filled in, not left as an empty shell.
Schema.org's Product reference lists every field. Your theme may output part of it. Most output a skeleton with empty values. I've seen valid markup with no price inside it. Useless.
Then there's the boilerplate problem. When 90% of your page HTML is navigation, popups, and app scripts, the model has to dig for the 200 words that matter. Dense, substantive product copy wins.
How does intent matching work on the shopper's side?
ChatGPT reads the shopper's phrasing, not just keywords. "A waterproof jacket for winter hiking under $200" is three constraints. If your data doesn't state waterproofing, season, and price, you can't match all three.
This is why attribute density matters more than clever copy. Every specific fact you add is another question you can answer. Every blank is a question you lose.
Here's the bottom line: you win the recommendation by being the most readable match, not the loudest brand.
How do you check if ChatGPT can see your products?
Ask it. Describe a shopper looking for exactly what you sell and see whether your store comes up. Then add a material, a price range, a use case, and watch where you drop out.
Do this for five real buyer questions. If you never appear, that's your baseline. Fix your feed and schema, wait for a recrawl, and run the same test to see the gap close.
This is the same read we give stores in a WRKNG audit. What AI sees, what it says, and where your data lets you down.
Frequently Asked Questions
Does ChatGPT rank products like Google does?
No. ChatGPT matches products to the intent behind a shopper's question using structured data and feeds, rather than ranking pages by backlinks and keywords. Complete, specific product data is what puts you in the running.
Do I have to pay to be recommended by ChatGPT?
No. There is no paid gate to be recommended organically. ChatGPT recommends products it can read and match, so complete data and clean schema get you in regardless of store size.
How fast will changes show up in ChatGPT?
It varies. Feed and schema changes need to be recrawled and synced before AI systems reflect them, which can take days to a few weeks. Fixing the data is a prerequisite, not an instant switch.
Is this different from SEO?
It overlaps but isn't identical. SEO targets search rankings. Getting recommended by ChatGPT targets machine-readable product data and intent matching. Good structured data helps both.
Want to know if AI assistants can actually find and recommend your store? Get a free AI-visibility read on your Shopify store at WRKNG Digital. We show you exactly what ChatGPT, Perplexity, and Google AI see when they look at your products.
