By Steve Merrill, Founder of WRKNG Digital | August 11, 2026
The product feed fields that decide whether ChatGPT recommends your Shopify products are GTIN, a full description, category, core attributes, price, availability, and reviews. AI reads these fields to match your product to a shopper's question. Leave them blank and you're invisible. Here's what each field does and why it matters.
Most stores fail this. When we ran 2,400 products through our audit tool, only 11% had the structured data needed to be recommended by ChatGPT. The feed is where that fails, and where you fix it.
Why Does the Product Feed Decide AI Recommendations?
ChatGPT doesn't browse your store the way a shopper does. It reads structured product data and trusted sources, then recommends what it can clearly understand. Your feed is that structured data. If it's thin, the AI has nothing to match against.
Think of it like this. The feed is the resume. AI hires the product with the complete resume, not the one with blanks.
Which Feed Fields Matter Most?
Seven fields carry most of the weight: GTIN, title, description, product category, core attributes, price, and availability. Each answers a question an AI might get from a shopper. Missing any one narrows the questions you can win.
Fill them in this order and you cover the majority of shopping queries an assistant would field.
How Important Is the GTIN?
Very. The GTIN is the global identifier that lets AI and shopping systems know exactly which product you're selling. Without it, your product is harder to verify and match. Google's Merchant Center docs treat GTINs as a core identifier for a reason.
If your products have GTINs and you're not including them, that's a fast win. Add them.
What Makes a Description AI Can Use?
A description that answers real questions. Specific details on material, use, fit, and features give AI the language to match your product to a query. Vague marketing copy gives it nothing.
Write for the question, not the vibe. "Waterproof to 50 meters" beats "built for adventure" when an AI is matching a shopper who asked for a waterproof watch.
Do Attributes Like Color, Size, and Material Matter?
They matter a lot. Shoppers ask AI for specific things: a blue size-10 running shoe, a gold hypoallergenic earring. AI can only answer if those attributes exist in your data. Blank attributes mean you lose specific, high-intent queries.
These are the queries closest to a purchase. Missing them is the most expensive gap.
How Do Price and Availability Affect Recommendations?
AI leans toward products it can confirm are available at a clear price. Accurate, current price and availability make your product safer to recommend. Stale data makes it risky, and AI avoids risk.
Keep these in sync with your real inventory. A recommendation that leads to an out-of-stock page is a recommendation AI learns not to repeat.
Where Do Reviews Fit In?
Reviews are a trust signal. AI is more comfortable recommending a product with real reviews than one with none. They don't live in the feed itself, but they feed the same decision. Build them with a post-purchase review request and add review markup to your pages, in line with Google's product schema guidance.
How Do You Fix a Thin Feed?
Audit first. Find which products are missing which fields, then fill the highest-impact gaps first: GTIN, attributes, and descriptions. On a big catalog, use enrichment rules with human review. On a small one, do it by hand. Either way, measure whether AI can find the product afterward.
FAQ
What is a product feed?
A product feed is the structured file of your catalog data, one row per product, holding fields like title, GTIN, price, and attributes that AI and shopping systems read.
Does Shopify create the feed automatically?
Shopify can export a feed, but the default often leaves attributes blank and lacks full schema. That's exactly what AI reads first, so the default usually isn't enough.
Which single field should I fix first?
Complete product descriptions and add missing GTINs. Those two do the most to make a product understandable and verifiable for AI.
How fast can feed fixes change AI recommendations?
AI can start reading improved products within weeks as systems re-crawl. Trust and citation growth then compound over the following months.
Want to see which feed fields are keeping your products out of AI answers? Get a free AI commerce audit at wrkngdigital.com/agentic-commerce-landing-page.
