Your product feed is your résumé to AI. When a shopping assistant decides whether to recommend you, it's reading the feed, not browsing your pretty site. Here are the 12 fields it checks, and the three that do most of the work.
The three that matter most:
Title. Not keyword soup. A clear name with the key attributes a buyer would say out loud.
Description. Detailed enough to answer real questions. AI needs substance to recommend you with confidence, not a tagline.
Category. The right product type, mapped correctly. This is how AI knows which question you're even an answer to.
Get those three right and you're ahead of most stores. Now the rest.
Price. Accurate and current. Stale prices get you skipped.
Availability. In stock, out of stock, handled cleanly. Don't delete out-of-stock products, you'll lose the AI data. Unlist them instead.
GTIN or barcode. The universal ID that ties your product to the wider data graph.
Brand. Named clearly and consistently.
Images. Clean, multiple angles, the kind visual AI shopping can actually read.
Product attributes. Size, color, material, use case. The specifics buyers filter on.
Reviews and ratings. Trust signals the AI weighs before it puts its name behind you.
Shipping and returns. Buyers ask, so AI wants to know.
Identifiers and links that point back to a clean, fast product page.
Here's why this matters more than it used to. Around 87% of AI product recommendations trace back to Google Merchant Center data. The feed isn't a side channel anymore. It's the main one. When the feed is right, AI recommends you. When it's a mess, AI fills the gaps with guesses, and AI guesses wrong up to 40% of the time.
So the work is unglamorous and high-leverage. Clean the feed. Fill the fields. Start with title, description, category.
Want us to score your feed the way the AI reads it? That's our free AI Commerce Audit. Results in 24 hours.
