Zerogaps blog

Shopify collection pages: the AI category entities you're ignoring

Shopify SEOAI visibilitycollection pagesstructured dataAEO

Most Shopify owners think of a collection as a bucket. You tag some products "summer dresses," a grid shows up at /collections/summer-dresses, done. That's how it looks to a shopper. It's not how it looks to an AI assistant.

When ChatGPT, Perplexity, or Google's AI Overviews try to figure out what your store actually sells, they don't read every product page one by one and stitch a mental model together. They look for the structure you've already drawn. Collection pages are that structure. They're the layer where AI learns "this store sells running shoes, and within that, trail shoes and road shoes, and here's how they relate." If your collections are empty grids with a two-word title and no description, you've handed the AI a map with no labels.

Why collections are an entity layer, not just a grid

An "entity" is just a thing the AI can name and reason about: a brand, a product type, a category. Your homepage tells AI your store exists. Your product pages tell it about individual items. The collection page is the piece in the middle that says "this is a category, here's what it means, and here's what belongs in it." That middle layer is what lets an assistant answer "what's a good store for affordable linen shirts" instead of just "here's one linen shirt."

This matters more now than it did for plain search. Old SEO rewarded collection pages because they ranked for category keywords. AI assistants use them for something deeper: building a model of your catalog they can traverse. If you want the longer version of why this shift changes how you write everything, I broke it down in this comparison of AEO and SEO for stores. The short version: ranking is about matching a query, entity-building is about being understood well enough to get recommended.

Three things make a collection page legible to AI: real description copy, structured data, and internal links that connect categories to each other and to their products. Let's do each one in plain Shopify terms.

Write a real collection description (not a keyword line)

In your Shopify admin, go to Products, then Collections, open a collection, and you'll see a Description field with the rich text editor. Most stores leave this blank or drop in one keyword-stuffed sentence. That's the single biggest miss.

Write 80 to 150 words that actually define the category and answer the questions a buyer would ask before choosing. What is this category, who is it for, what makes choosing one product over another easier, what's the price range, what problem does it solve. You're not writing fluff. You're giving the AI the facts it needs to describe your category to a shopper in its own words.

Here's a before and after. Before: a collection titled "Yoga Mats" with no description, just a grid. After: "Yoga Mats" with a paragraph that says these are non-slip mats from 4mm to 6mm thick, the 6mm option suits people with sensitive knees, the eco line is natural rubber with no PVC, and most ship within two days. Now an AI asked "what store has thick eco yoga mats" has something real to grab. Before, it had a grid it can barely interpret.

One practical note: put the description where the theme renders it on the actual page. Most themes show the collection description near the top of the collection template. If yours hides it, that's a theme setting worth fixing, because AI reads the rendered page.

Add collection schema so AI reads it as a category

Schema is structured data: a snippet of code that labels the page so a machine knows "this is a collection of products" rather than guessing. For collection pages, the types that help are CollectionPage, ItemList (the list of products in the grid), and BreadcrumbList (where this category sits in your hierarchy).

If you're comfortable touching theme code, you can add a JSON-LD script to the collection template (usually collection.liquid or the collection section in your theme.liquid setup, depending on theme version). You'd loop through the collection's products to build the ItemList and output a BreadcrumbList that goes Home, then the parent category, then this collection. If "loop through products in Liquid" sounds like a different language, don't force it.

The non-developer route is an SEO or schema app from the Shopify App Store. Several add CollectionPage, ItemList, and breadcrumb markup to your collection pages automatically, no code. Install one, point it at your collections, and it injects the structured data for you. Either way, the goal is the same: the page announces itself as a category entity instead of leaving AI to infer it.

A quick test: if you can't tell, from the page alone, that it's a category and what sits above and below it in your store, neither can an AI.

Link collections to each other and up from products

Entities aren't useful in isolation. AI traverses relationships, so it matters that "trail running shoes" connects to "road running shoes" and both connect up to "running shoes." You build those connections with internal links and your store's navigation.

Start with your menu. In Online Store, then Navigation, structure your main menu so parent categories contain child collections as dropdown items. That nesting is a signal of hierarchy. Then, inside each collection description, link to two or three related collections in plain sentences ("pair these with our running socks" style links, pointing at the actual related collection URLs). This gives both shoppers and AI a path to walk between sibling categories.

The link people forget runs the other way: from products up to their parent collection. Many themes don't show a clear "back to category" link on the product page, so the relationship between a product and its collection is invisible to anything reading the page. Add a breadcrumb to your product template, or a simple "View all in [Collection]" link, so each product points up to where it belongs. Now AI can follow product to category to parent category, and your catalog reads as a connected tree instead of a pile of disconnected pages.

A simple order to do this in

You don't need to redo every collection this week. Pick your five most important categories, the ones you'd actually want recommended, and work through them:

Then check whether it's working. The whole point is to get recommended when a shopper asks an assistant what to buy, so see what the assistants actually say about your categories today. You can run a free AI visibility audit to see whether AI tools understand your store and which competitors come up instead of you. Fix the five collections that matter, watch how the answers change, then keep going.

Collections aren't decoration and they aren't just grids. They're the part of your store that tells AI how your catalog is organized. Spend an afternoon making them say something real, and you give every assistant a map it can actually read.

See where your store stands

Run a free AI Visibility Audit and find out if AI recommends you.

Get my free audit →

Questions store owners ask

What schema should I add to a Shopify collection page?

Use CollectionPage for the page itself, ItemList for the products in the grid, and BreadcrumbList for where the category sits in your hierarchy. You can add it as JSON-LD in your collection theme template, or use an SEO or schema app from the Shopify App Store if you don't want to touch code.

Do collection descriptions really affect how AI sees my store?

Yes. The description is often the only place that defines what a category means in plain language. AI assistants use it to understand and describe your category to shoppers. A blank or one-line description leaves them guessing, so write 80 to 150 words that define the category and answer common buyer questions.

How do internal links help AI understand my collections?

AI follows relationships between pages. Nesting collections under parent categories in your navigation, linking related collections to each other, and adding links from products up to their parent collection all give AI a path to traverse. That connected structure lets it map your catalog as a tree instead of a pile of unrelated pages.