By Steve Merrill, Founder of WRKNG Digital | October 3, 2026
Nine product feed fields decide whether AI can recommend your Shopify products: title, description, identifier, brand, price, availability, product type, attributes, and image. Get these right and AI can read you. Leave them blank and you're invisible to the machine doing the recommending.
I keep coming back to one number. We audited 2,400 products and only 11% had the feed data AI needs. The other 89% weren't beaten. They were unreadable. Here's the nine fields that fix that.
1. Product title
Write titles the way buyers search, not the way your brand team names things. Include the main attribute and use case. "Stainless Steel Insulated Water Bottle, 32oz" beats "The Summit." AI matches words. Give it words that match intent.
2. Description
Lead with what the product is and who it's for, in plain language. AI pulls descriptions to match queries, so bury the poetry and front-load the facts. First sentence does the work.
3. GTIN / identifier
Global trade item numbers let AI connect your product to the same item across the web, which builds trust and comparison data. Google's GTIN guidance explains why identifiers matter. Missing GTINs weaken every downstream match.
4. Brand
Set the brand field explicitly on every product. AI uses it to build an entity profile of your store. Inconsistent or blank brand data confuses that profile and costs you recommendations.
5. Price
Machine-readable, accurate, and matching your storefront. AI filters by price constantly ("under $50"). If your feed price is wrong or missing, you get filtered out of the queries you'd otherwise win.
6. Availability
In stock, out of stock, preorder. Keep it current and consistent with your live page. Stale availability is the fastest way to teach AI your data can't be trusted. Shopify's inventory tools keep this honest.
7. Product type and category
Use a specific category path, not a vague one. "Home > Kitchen > Water Bottles" tells AI exactly where your product lives. Generic categories blur the match and push you down.
8. Custom attributes
This is where recommendations are won. Material, color, size, compatibility, weight, use case. The more specific attributes you provide, the more buyer queries you can match. Blank attributes are missed sales. That's the whole point.
9. Image link
A working, high-quality image URL with descriptive context. AI shopping surfaces increasingly show images, and a broken or missing link drops you from visual results entirely. Check that every product resolves.
How We Chose This List
These nine come from the feed specs AI shopping channels actually read, cross-checked against the gaps we find most on live Shopify stores. Attributes and identifiers are the usual failures.
FAQ
Q: Which feed field matters most for AI recommendations?
Custom attributes and product type. They let AI match your product to specific buyer intent, which is where recommendations happen.
Q: Does Shopify fill these fields automatically?
It fills the basics but leaves many attributes and identifiers blank. Those gaps are where AI loses the thread.
Q: Where does AI read my product feed?
From your structured feed, Product schema, and channels like Google Merchant Center. Keep all three consistent.
Q: How often should I update my feed?
Price and availability in real time. Attributes whenever products change. Stale feeds lose trust fast.
Want to see which feed fields your products are missing? Get a free AI commerce readiness audit at WRKNG Digital's agentic commerce page.
