By Steve Merrill, Founder of WRKNG Digital | August 12, 2026
Which Shopify product feed fields decide whether AI recommends you?
The Shopify product feed fields that decide whether AI recommends you are title, description, GTIN, brand, product_type, google_product_category, availability, price, image_link, condition, MPN, and Product structured data. Get these clean and machines can read your catalog and pull your products into an answer. Leave them empty and you are invisible. Here are all 12, what each one does, and the quick fix.
1. title
The title is the first thing an AI reads to match your product to a shopper's question. Stuffed or vague titles get skipped because the model cannot tell what the product actually is. Quick fix: lead with the real product name, then add brand, key attribute, and size, like "Ridge Slim Wallet, Titanium, RFID Blocking." Google's title spec is the standard to follow.
2. description
The description is where AI pulls the details that answer "does this fit what I need." Thin or keyword-jammed copy gives the model nothing to quote. Quick fix: write two or three plain sentences covering material, use, and fit, and put the important facts in the first 160 characters.
3. GTIN
The GTIN is the barcode number that ties your listing to the exact product everywhere it appears online. Without it, machines cannot confirm your item is the same one reviewed and sold elsewhere, so trust drops. Quick fix: pull the GTIN from the manufacturer and map it to a variant metafield or your feed. Google explains GTIN requirements here.
4. brand
Brand tells AI who makes the product, which is how it groups you against competitors and confirms authority. A missing brand field turns your product into a generic mystery item. Quick fix: set the brand on every product, and use your own store name for private-label goods.
5. product_type
This is your own category label, and it gives AI extra context about where the product sits in your catalog. It also helps the model understand niche items that a broad category misses. Quick fix: use a clear path like "Wallets > Slim Wallets > RFID" instead of a single vague word.
6. google_product_category
This is Google's fixed taxonomy, and it tells shopping surfaces and AI exactly what kind of thing you sell. Pick the wrong category and your product shows up for the wrong questions or none at all. Quick fix: match to the most specific value in Google's product taxonomy, not the closest broad one.
7. availability
Availability tells AI whether the product is in stock right now. Assistants avoid recommending items they think are sold out, because a dead link burns the shopper. Quick fix: keep this synced to live inventory so it flips to "in_stock" and "out_of_stock" automatically, never by hand.
8. price
Price is a hard filter. When a shopper asks for something "under $50," AI reads this field to include or drop you. A stale or missing price gets you cut before the model even looks at the rest. Quick fix: sync price and sale_price to your real store price, and include currency.
9. image_link
The image_link is the product photo the feed points to, and multimodal AI now reads images to confirm what the product is. A broken URL or a lifestyle-only shot with no clear product gives the model less to trust. Quick fix: use a clean, high-resolution main image on a plain background at a stable URL.
10. condition
Condition tells AI whether the item is new, refurbished, or used. It sounds minor, but shoppers ask for "new" or "refurbished" by name, and the model needs this field to answer them. Quick fix: set condition on every product, and default to "new" unless it truly is not.
11. mpn
The MPN is the manufacturer part number, and it works with GTIN and brand to nail down the exact product identity. When a GTIN is missing, MPN plus brand is the backup that keeps you matchable. Quick fix: add the MPN from the manufacturer to any product that has one.
12. Product structured data mapping
Your feed feeds shopping surfaces, but AI crawlers also read Product structured data on your live page. If the schema on the page does not match the feed, you send mixed signals and lose trust. Quick fix: map the same name, brand, gtin, price, and availability into Product JSON-LD, following Google's product structured data docs.
How We Chose These Fields
We picked the fields AI assistants and shopping crawlers actually read to match, verify, and recommend a product. Identity, category, and live availability got the most weight, because those are the checks a model runs before it names your store.
FAQ
Q: Which product feed fields matter most for AI search recommendations?
Title, GTIN, brand, google_product_category, availability, and price carry the most weight. Those are the fields AI reads first to match your product and confirm it is real and in stock.
Q: Do I need GTINs for AI to recommend my Shopify products?
For branded, manufactured products, yes. A GTIN ties your listing to the exact product across the web, which builds the trust an AI needs before it names you. Handmade items are the exception.
Q: How do I fix a Shopify product feed for AI search?
Start with the required fields: title, description, price, availability, image_link, brand, and GTIN. Then add google_product_category, product_type, condition, and MPN, and mirror the same data into Product structured data. Shopify's help docs cover the basics.
Q: Is the product feed different from schema on the page?
They are two copies of the same facts. The feed goes to Google Merchant Center. Product structured data sits on your live page for crawlers. Both should match, or you send AI mixed signals.
Want your Shopify products recommended when shoppers ask AI what to buy? See how we clean feeds and schema at WRKNG Digital's agentic commerce page.
