By Steve Merrill, Founder of WRKNG Digital — October 19, 2026
How does ChatGPT choose which store to send a shopper to?
ChatGPT picks a store based on six signals it can actually read: your product feed data, structured data on your pages, third-party reviews and citations, real-time availability, clear pricing, and how often your brand shows up across the web. The cleaner those signals, the more likely your store is the one it names.
It does not flip a coin. It reads data. Most stores just hand it bad data.
I've seen this in dozens of Shopify audits this year. A store ranks fine on Google, sells great from paid ads, and gets ignored by ChatGPT. The owner thinks the model hates them. It doesn't. It just couldn't read their product page.
Where does ChatGPT get its product information?
ChatGPT Shopping pulls product details from merchant feeds, structured data on your product pages, and third-party sources like reviews and editorial coverage. OpenAI has said these results are organic and chosen to be useful to the shopper, not paid placements.
So the model is reading three things at once.
First, your feed. That's the same structured catalog data that powers product listings everywhere. Title, brand, price, availability, product ID. Second, the schema on your actual page. Third, what the rest of the internet says about you. Reviews. Blog mentions. Press.
OpenAI laid out how this works in their shopping announcement. The short version: it wants to send people to products that match what they asked for and that it can trust. Trust comes from data it can verify.
What product feed data does ChatGPT actually read?
ChatGPT reads the core fields in your product feed the same way any shopping engine does: product title, brand, description, price, GTIN or barcode, availability, and condition. Missing or messy fields make your product harder to match to a shopper's question, so it gets skipped.
This is where most Shopify stores lose. Not on branding. On blank fields.
Go look at your own feed right now. How many products are missing a brand? A GTIN? A real description instead of three words? That's the stuff the model uses to decide if your "waterproof hiking boots" are the waterproof hiking boots someone asked for.
Shopify documents how product feeds and metafields work in their product details help docs. Fill those fields out. All of them. It's boring work and it's the work that moves the needle.
Which feed fields matter most?
Title, price, and availability carry the most weight. A shopper question almost always includes a product type, a price sense, and an assumption that the thing is buyable right now. Match those three cleanly and you're already ahead of 8 out of 10 stores I audit.
Does structured data on my product page change anything?
Yes, a lot. Product structured data tells ChatGPT the price, availability, and rating of a product in a format machines read without guessing. Pages with clean Product schema get matched more confidently than pages where the model has to scrape numbers out of your layout.
Think of schema as a label you write for the robot.
Without it, the model sees a wall of HTML and tries to guess what's the price and what's the shipping cost and whether "2 left" means in stock. With it, you've handed over the facts. Google explains the exact Product markup fields in their structured data documentation, and the same markup helps across AI shopping surfaces.
Shopify themes can output a lot of this automatically. Check yours. A free schema test will tell you in 30 seconds if your price and availability are being read correctly.
How much do reviews and brand mentions matter?
They matter more than merchants expect. ChatGPT leans on third-party reviews and citations to decide if a store is legit before it recommends you. A product with real ratings and independent mentions looks safer to recommend than a store nobody on the web talks about.
The model does not want to send a human to a sketchy store. That would make it look bad.
So it checks. Do you have reviews with real ratings? Does anyone else on the internet mention your brand? Are you named in a buying guide, a Reddit thread, a comparison post? Each of those is a vote that you're real.
I made the mistake early of treating reviews as a conversion tool only. They're also a trust signal to the machine. Same review. Two jobs.
Why does availability and price clarity decide the recommendation?
ChatGPT wants to send shoppers to products they can actually buy at a price they understand. A product that shows in stock with one clear price stays in the running. A product with confusing pricing or an out-of-stock flag gets dropped, even if it's the best match on paper.
Here's what I mean.
If your availability field says "in stock" but your page says "backordered," the model sees a conflict and plays it safe by picking someone else. Same with price. One product, three prices scattered across the page, no clear sense of what the shopper pays. The model would rather recommend the store that's obvious.
Make the answer easy. One price. Honest stock status. Matching data in the feed and on the page.
How is ChatGPT different from Perplexity and Google for shopping?
ChatGPT, Perplexity, and Google all read structured product data, but they weight sources differently. Perplexity leans hard on cited web sources and shows its links openly. Google ties shopping to its own Merchant Center feed and Shopping graph. ChatGPT blends merchant feeds, page schema, and trusted third-party data into one answer.
The overlap is the good news.
Clean feed data and clean schema help you on all three. Perplexity, which publishes its approach in its company blog, rewards stores that get cited by other sites. Google rewards a complete Merchant Center feed. ChatGPT rewards the blend. Do the data work once and you show up in more than one place.
That's the whole game right now. Fix the data layer and you stop fighting each engine separately.
What should a Shopify merchant check this week?
Check five things, in order: your product feed completeness, your Product schema, your reviews, your availability and pricing accuracy, and your brand mentions across the web. Fixing these in order gives you the fastest path to getting named by ChatGPT.
- Open your product feed. Fill every empty field, especially brand, GTIN, and description.
- Run a schema test on your top product pages. Confirm price and availability are read correctly.
- Turn on reviews and collect real ones. Aim for ratings on your best sellers first.
- Match your stock and price data between the feed and the page. Kill every conflict.
- Get mentioned. Guide posts, forums, press, your own answer-style content.
None of this is fancy. It's data hygiene. And almost nobody's doing it, which is exactly why it works right now.
FAQ
Does ChatGPT use Google Shopping data to pick stores?
Not directly. ChatGPT Shopping uses its own product data from merchant feeds, structured data on your pages, and third-party sources it trusts. It reads the same kind of feed fields Google uses, but it ranks stores on its own signals.
Do I have to pay to show up in ChatGPT Shopping?
No. OpenAI has said ChatGPT Shopping results are organic and not paid placements. Your product data quality, reviews, availability, and brand mentions decide whether you get surfaced.
How does ChatGPT know if my product is in stock?
It reads the availability field in your product feed and the availability property in your Product schema. If both say in stock and your price is clear, you stay in the running. If they conflict, the model gets cautious and picks a store it trusts more.
Does my Shopify store connect to ChatGPT automatically?
Your product data becomes readable when your feed and structured data are clean, and Shopify supports syncing catalogs to shopping surfaces. ChatGPT can pull from those feeds and crawl your pages, so the cleaner your data, the better your odds of being the store it names.
How long until I show up after fixing my data?
It varies. Feed and schema changes can get picked up in days to a few weeks as crawls and feeds refresh. Reviews and brand mentions build slower. Start the data work now so the slow signals have time to stack.
Want the store that ChatGPT actually recommends? WRKNG Digital gets your Shopify catalog, schema, and brand signals ready for AI shopping so you're the store it names. See how agentic commerce works for your store.
