By Steve Merrill, Founder of WRKNG Digital | August 30, 2026
Which product feed fields do AI shopping assistants actually read first?
Title, description, price, availability, brand, and a few key attributes. Get those right and you're a candidate. Leave them blank and you're invisible.
I've audited more than 40 Shopify stores this year. The ones AI recommends fill these fields with real values. The ones it skips leave half of them empty.
Here are the nine that carry the most weight, in the order a model reaches for them.
1. Title
The title is the first fact a model reads, so make it describe the product, not your brand slogan. Lead with what the item is, then key attributes like material or size.
A title of "Waterproof Merino Wool Hiking Socks, Men's" beats "The Adventurer." One is a match, the other is a guess.
2. Description
The description is where you answer the shopper's real questions in plain text. State what it is, who it's for, and why it fits, without filler.
Models quote descriptions directly, so write it to be quoted. Vague copy gives them nothing to lift.
3. Price
Price has to be a real, current number in the feed, not just baked into an image. Shoppers filter on it and models won't recommend what they can't price.
Keep sale and regular pricing accurate too. A stale price is a trust problem the model will avoid.
4. Availability
Availability tells the model whether recommending you is safe. In stock, out of stock, or preorder needs to be current and correct.
An agent won't send a shopper to a sold-out product. Ambiguous availability is a reason to pick a competitor.
5. Brand
Brand ties your product to an entity a model can recognize and cross-reference. Missing brand data makes you harder to place and easier to skip.
Use consistent brand naming across your feed, schema, and page so nothing looks mismatched.
6. GTIN or MPN
A unique identifier like GTIN or MPN lets a model match your product to the wider catalog and verify it's the real thing. The Google Merchant product data specification treats these as core fields for a reason.
Without an identifier, your listing floats alone and gets less trust.
7. Product category
Category places you in the right consideration set. Get it wrong and you show up for the wrong requests, or not at all.
Use the standard taxonomy values rather than a made-up label so the model can slot you correctly.
8. Attributes (color, size, material)
Attributes are what shoppers filter and ask for, so each one is a question you can answer. Color, size, and material turn a generic listing into a precise match.
Fill them with real values, not blanks. Every filled attribute is a reason the model can pick you.
9. Image link
A clean, working image link matters because AI shopping surfaces increasingly show product images beside the answer. A broken or low-quality image weakens an otherwise strong listing.
Make sure the URL resolves and the image matches the current product, not a discontinued variant.
Further reading
Frequently Asked Questions
Which feed field matters most for AI recommendations?
The title and description do the most work, because models read and quote them first. But they only help if price, availability, and brand back them up, so treat the core fields as a set, not a ranking.
Do I need GTINs for every product?
For most catalog products, yes, a GTIN or MPN helps a model verify and match your item. For truly custom or one-of-a-kind products, provide MPN and complete attributes instead.
Will filling these fields help my Google Shopping too?
Yes. These are the same fields the Google Merchant product data specification asks for, so completing them helps your Shopping listings and your AI visibility at the same time.
How do I find which fields are empty?
Export your product feed and scan for blanks in title, description, price, availability, brand, identifier, and attributes. Empty cells are your fix list, ranked by how many products they affect.
Want to know if AI assistants can actually find and recommend your store? Get a free AI-visibility read on your Shopify store at WRKNG Digital. We show you exactly what ChatGPT, Perplexity, and Google AI see when they look at your products.
