By Steve Merrill, Founder of WRKNG Digital | September 24, 2026
What structured data fields matter most for AI shopping recommendations?
The 8 structured data fields that matter most for AI shopping recommendations are Product name, Offer price, Availability, AggregateRating, Review, Brand, GTIN, and Product description. Get those right in your Product schema and AI assistants can actually read your product, compare it, and recommend it. Skip them and you're invisible.
I've had a hard time getting store owners to care about this. Schema feels like plumbing. But an AI shopping agent can't see your pretty product page. It reads the structured data. So here are the 8 fields that decide whether you show up.
1. Product name
The name field is the first thing an AI reads, and it has to match how people actually search. "Organic Cotton Crew Neck Tee - Charcoal" beats "The Weekender" every time, because the AI matches your product to a shopper's intent through this field. Per the Schema.org Product spec, name is required, so there's no excuse to leave it thin or clever.
2. Offer price and priceCurrency
Price lives inside the Offer object, and it needs both a number and a priceCurrency like "USD." AI assistants filter and rank on price constantly, so a missing or malformed price gets you dropped from comparison the moment budget comes up. The Schema.org Offer type spells out exactly how to nest price under the product.
3. Availability
The availability field tells the AI if the item is in stock, using values like https://schema.org/InStock or OutOfStock. This one matters more than people think, because no AI wants to recommend a product a shopper can't buy. Keep it synced to real inventory, not hard-coded to InStock, or you'll send buyers to a sold-out page and burn trust.
4. AggregateRating
AggregateRating holds your average star rating and review count, and it's one of the strongest signals AI uses to decide what's worth recommending. A product with a 4.7 across 300 reviews wins over one with no rating data at all, because the AI reads social proof through this field. Only mark up ratings that are real and visible on the page. Faking it violates Google's product structured data guidelines.
5. Review
Individual Review objects give the AI the actual words customers used, which is different from the star average. This is where an assistant pulls language like "runs small" or "held up after a year," and it feeds those details straight into recommendations. Include the reviewer, the rating, and the review body so the AI has something to quote.
6. Brand
The brand field ties your product to a named maker, and AI shopping increasingly filters by brand when a shopper asks for one. Leave it blank and your product floats free with no identity, which hurts you on any branded query. Set it as a proper Brand object with a name, not just loose text.
7. GTIN and MPN
GTIN (the barcode number) and MPN (manufacturer part number) are unique product identifiers that let AI match your exact item across the whole web. These are what Google Merchant Center uses to connect your listing to the same product on other sites, which drives price and review comparison. If your products have real GTINs, include them. It's the cleanest way for an AI to know your tee is the same tee it saw somewhere else.
8. Product description
The description field is the plain-language summary the AI reads to understand what the product actually is and who it's for. Write it for a person, cover material, fit, and use case, because vague descriptions give the AI nothing to match against. This is the field that turns "a shirt" into "a midweight organic cotton tee for everyday wear."
How We Chose This List
We ranked these fields by how often AI shopping assistants pull them when building product recommendations, based on the required and recommended properties in the Schema.org Product spec and Google's product markup docs. Fields that gate visibility or drive comparison came first.
FAQ
Q: What structured data type should Shopify products use?
Use Product schema in JSON-LD with a nested Offer object. That's the format Schema.org defines and Google reads, and it covers all 8 fields above.
Q: Do I need GTINs for AI shopping recommendations?
Not required, but they help a lot. GTINs let AI and Google Merchant Center match your exact product across sites, which powers price and review comparison. If your products have real barcodes, include them.
Q: Does structured data guarantee my product gets recommended by AI?
No. It makes your product readable and eligible, which is the price of entry. Ratings, price, and availability still decide whether the AI picks you over a competitor.
Q: What happens if my schema has wrong or fake data?
You risk getting ignored or penalized. Google's guidelines require marked-up data to match what's visible on the page, and AI assistants drop products with data that doesn't line up.
Want your Shopify store's product data built so AI assistants can actually read and recommend it? That's the work we do. See how WRKNG Digital sets up agentic commerce for Shopify stores.
