By Steve Merrill, Founder of WRKNG Digital | September 7, 2026
What product feed fields do AI shopping agents read before recommending you?
AI shopping agents read nine fields first: title, GTIN/MPN, brand, price, availability, condition, product category, image, and description with attributes. Get these clean and matched across your feed and your page, and the agent can trust your product enough to put it in front of a buyer.
Here's the thing. The agent never sees your store the way a human does. It reads data. If the data is thin or conflicts, you don't get recommended. You get skipped.
Below are the nine fields, in the order agents lean on them.
1. Product Title
The title is the first thing an agent matches against what the shopper asked for. Google's Merchant Center spec puts the title at the front of the feed for a reason. Front-load the brand, product type, and key attributes like size, color, and material, because that's the text the agent scores for relevance.
2. GTIN and MPN
The GTIN is the global barcode number. The MPN is the manufacturer part number. These let an agent confirm your product is the exact item a shopper wants and compare it across every seller carrying it. Google's product data spec treats GTINs as a strong quality signal, and feeds missing them for branded goods often get demoted.
3. Brand
Brand tells the agent who makes the product. Shoppers ask for brands by name, so the agent filters hard on this field. Leave it blank or stuff it with junk and you drop out of every branded query.
4. Price
Price is a ranking and trust field at the same time. The agent compares your price against other sellers of the same GTIN, and it checks that the feed price matches the price on your product page. When those two numbers disagree, the agent flags the listing and can pull it. Match them exactly, currency included.
5. Availability
Availability tells the agent whether the item is in stock right now. An agent will not recommend something a shopper can't buy, so out-of-date stock status gets you dropped fast. Keep this synced in real time, because a stale "in stock" is worse than an honest "out of stock."
6. Condition
Condition says new, refurbished, or used. It sounds small. It changes everything about how the agent presents and prices your item, and a wrong value becomes a trust failure that can suppress the whole listing.
7. Product Category (Google Product Category / GPC)
The Google Product Category maps your item to a shared taxonomy the agent already understands. It's how the agent knows a "shell" is a jacket and not seafood. Google publishes the full taxonomy list, and picking the most specific category improves how often you match the right query.
8. High-Quality Image
Agents and multimodal models read the image to confirm it matches the title and category. A clear, high-resolution shot on a plain white background lets the model verify the product and hand it to the shopper with confidence. Blurry, watermarked, or placeholder images get disqualified in the spec.
9. Description With Attributes
The description is where the agent pulls the details a title can't hold: fit, materials, use case, specs. Write it with real attributes stated plainly, not marketing fluff, because the agent extracts facts from it to answer specific shopper questions. Back it with Schema.org Product markup on the page so the facts in your feed and the facts on your page say the same thing.
How We Chose This List
These nine come straight from the fields AI shopping agents pull from the Google Merchant Center product data spec and Schema.org Product markup, ranked by how hard agents filter on each one. We weighted the fields that break a recommendation when they're wrong ahead of the ones that only fine-tune it.
FAQ
What product feed fields do AI shopping agents read first?
Title and identifiers first, then brand, price, availability, condition, category, image, and description. Title and GTIN/MPN come first because they let the agent match your product to what the shopper actually asked for.
Do I need a GTIN for AI shopping agents to recommend my product?
For branded products, yes. A GTIN lets an agent confirm your product is the exact item and compare it across sellers. Without it, many feeds get demoted or dropped.
How does structured data help AI shopping agents?
It gives the agent machine-readable facts instead of guesses. When price, availability, and identifiers match across your feed and your page, the agent trusts the data enough to recommend it.
Does image quality affect AI product recommendations?
Yes. The agent reads the image to confirm it matches the title and category. A clear, high-resolution image on a plain background helps it verify the item and pass it to the shopper.
Sources worth reading: the Google Merchant Center product data specification, the Google Product Category taxonomy, Schema.org Product markup, and Shopify's guide to product feeds for Google.
Your product data is what the agent trusts. Fix these nine fields and you stop getting skipped. Want us to audit your feed and get your catalog agent-ready? See how WRKNG Digital gets your store recommended by AI shopping agents.
