By Steve Merrill, Founder of WRKNG Digital — August 5, 2026
What structured data makes ChatGPT recommend a Shopify product?
ChatGPT recommends products backed by valid Product schema with a complete Offer object: name, brand, price, priceCurrency, availability, and aggregateRating. That schema has to match your Google Merchant Center feed. When both agree, an AI assistant can name your product, quote your price, and tell someone it's in stock.
Miss those fields and you become a product it describes but never picks.
I've run this on client stores. The pattern is boring and consistent. The store with clean, complete schema gets named. The store with a half-filled feed gets skipped, even when the product is better.
Which schema fields does an AI assistant actually read?
Start with the Product type on Schema.org. This is the map every parser follows. Fill these and you cover most of what matters.
- name. The exact product title. Match it to your feed title.
- brand. Your brand name as a Brand object, not plain text.
- sku and gtin. The identifiers that let an AI tie your page to a real catalog entry.
- image. A direct URL to a high-res product image.
- description. Plain, specific, no fluff.
Then the part most stores get wrong. The Offer object.
An Offer needs price, priceCurrency, and availability. Set availability using the real value, https://schema.org/InStock, not the word "yes." Add priceValidUntil so the price doesn't read as stale. This is the block that turns a description into a recommendation.
Last, add aggregateRating with a ratingValue and reviewCount, plus at least one review. A 4.7 with 300 reviews gives an AI a reason to choose you. A product with no rating gives it a reason to move on.
One more field earns its place: gtin. It's the global identifier that ties your listing to a known product across catalogs. When an assistant sees a matching gtin in your schema and your feed, it trusts the record. When the gtin is missing, your product looks like a floating page with no proof behind it. I've watched stores add gtins to 40 SKUs and start getting quoted within a month.
How do I set up the schema step by step?
Here's the order I use on every audit.
- Add Product JSON-LD to every product page. One script tag, in the head or body. Include name, brand, sku, gtin, image, description, and the offers object.
- Fill the Offer fields. price, priceCurrency, availability, priceValidUntil. All four.
- Add ratings. aggregateRating plus a review object pulled from your real reviews.
- Match your Merchant feed. Make the schema and the feed say the same thing.
- Validate. Run every page through Google's Rich Results Test and clear the errors.
Don't skip step four. It's where most of the damage hides.
Why does my Google Merchant feed matter for ChatGPT?
Because AI shopping answers lean on structured product data, and your feed is a big source of it. If your page schema says $49 and your feed says $59, you've handed the assistant a conflict. Conflicts get resolved by dropping you.
Line up these attributes in your Google Merchant Center product data spec with the schema on the page:
title. Same wording as your schema name.gtinandbrand. Identical to the page.priceandavailability. No gaps, no mismatches.product_typeandgoogle_product_category. So the assistant knows what shelf you're on.
We fixed a feed-to-schema price mismatch on a home goods store last quarter. Three of their top SKUs had been getting ignored in shopping answers. Once the two sources agreed, those products started showing up by name.
Does Shopify add this structured data automatically?
Some of it. Most Shopify themes output a basic Product block, and the platform documents its approach in the Shopify structured data docs. But basic isn't complete.
Themes routinely miss gtin, a proper brand object, priceValidUntil, and aggregateRating. The theme thinks its job is done. The AI assistant sees a thin product and passes.
So check it yourself. Pull up a live product URL, run it through the Rich Results Test, and read what's actually there. Don't trust the theme's word for it.
What does a complete product setup look like?
Full Product schema. A complete Offer with price, currency, availability, and a valid-until date. Real ratings. A Merchant feed that matches every one of those fields.
That's the whole game. There's no secret tag. AI assistants reward the store that filled in the boring fields and made its sources agree. Same principle OpenAI describes for how models pull from clear, well-formed sources in its web search tool docs.
The stores that win here aren't the biggest. They're the ones that did the setup and validated it.
Frequently asked questions
Does ChatGPT read Shopify structured data directly?
It reads it through the sources it pulls from. When ChatGPT browses or uses a shopping index, it parses your Product and Offer schema plus your Google Merchant feed. Clean schema makes your product easy to quote. Missing schema makes it easy to skip.
What is the single most important schema field for AI recommendations?
The offers object. Without price, priceCurrency, and availability, an AI assistant can describe your product but can't recommend it with confidence. Availability set to InStock is what lets it say buy this now.
Do I need reviews in my structured data?
Yes. aggregateRating and review objects give an AI a reason to pick you over a competitor with no ratings. A 4.7 with 300 reviews beats a blank product almost every time.
Will Shopify's default theme add this schema for me?
Partly. Most themes output basic Product schema but miss gtin, brand, priceValidUntil, or aggregateRating. Check your live pages in the Rich Results Test. Don't assume the theme handled it.
Get your product schema fixed the right way
If your products aren't showing up when people ask ChatGPT what to buy, the schema is usually the reason. We audit it, fix it, and match it to your feed.
See how WRKNG Digital gets Shopify products recommended by AI assistants.
