Why AI Shopping Shows Your Brand but Skips Your Category

I ran a small experiment that I want to share, because the result is the single most common agentic-discovery problem I see and almost nobody is looking at it.

I took two real Shopify brands, a premium eyewear company and an organic tea company, and queried Shopify's Global Catalog the way an AI assistant does. First I searched their names. Both came back instantly, sitting at the top with their products cleanly described. Then I searched the way a new customer actually shops: "titanium polarized sunglasses," "premium aviator sunglasses," "organic herbal tea bags," "turmeric ginger tea." On those generic category searches, both brands were almost entirely absent. The eyewear brand surfaced for exactly one query, and only because the product name itself described the product. The tea brand showed up for nothing.

This is the branded-versus-category gap, and once you see it you cannot unsee it.

The two kinds of shopping queries

When someone already knows your brand, they ask for it by name. You will almost always show up for that, because your store is the authoritative source for your own products. That traffic is real, but it is demand you already earned somewhere else. The assistant is just the new doorway.

The queries that grow a business are the category ones. "A good daily trainer for marathon training." "Indestructible sunglasses under $300." "Caffeine-free tea for sleep." The shopper has no brand in mind. Whichever products the assistant surfaces are the ones that get considered, and the brand that was invisible never entered the conversation. This is exactly where most stores lose, and they lose quietly, because nothing tells you about the sale that went to someone else.

Why brands disappear on category queries

Two causes, and they compound.

The first is brand-only product naming. A product called "The Hook" is meaningful to people who already know the brand and meaningless to a machine matching "titanium sunglasses." The catalog reads your title literally. If the category word is not in the title, and often not in the description either, there is nothing for a category search to match. Clever product names are great for a brand and terrible for discovery, and most stores never reconcile the two.

The second is missing structure. Shopify's Global Catalog lets assistants filter by product category using a standard taxonomy. If your products are not assigned a real taxonomy category, you can be filtered out of category-scoped searches before relevance is even considered. A free-text product type is not the same thing as the structured category the filter reads.

There is a quieter third factor underneath both. Shopify uses its own models to enrich your catalog, inferring attributes and features from your text. When your descriptions are thin, that inference runs on almost nothing and gets it wrong or leaves it blank. Rich, specific copy is not just for humans anymore. It is the raw material the catalog uses to understand and rank you.

The fix is unglamorous, which is why it works

Put the category in the title. "The Hook" becomes "The Hook Titanium Polarized Sunglasses." You keep the brand name and you give the machine the noun it is matching on. Do this across the catalog and you stop being invisible on the searches that matter.

Assign a real taxonomy category to every product, not just a product type. This is what keeps you eligible for category-filtered searches.

Write descriptions that actually state the attributes: material, use case, who it is for, what makes it different. This feeds both the assistant reading your page and the catalog inferring your facts.

Cover availability, condition, and shipping clearly, because the same searches filter on those too, and a product the assistant cannot price or ship gets dropped.

None of this is speculative or tied to a single protocol. It is the same structured, complete, machine-legible data that helps you in Google today and in every AI surface that recommends products. The difference is that an assistant reads it more literally than a person ever did, so the gaps cost you more.

Where we land

Ranking for your own name feels like winning, and it is the easiest thing to measure, which is why so many stores stop there. The growth is in the category queries, and that is precisely where brand-only naming and missing structure leave most stores off the list. The good news is that the fix is concrete and entirely in your control.

This is the work we do for the stores we run, and it is what AgentReady scores and maintains automatically: a Catalog Readiness check that flags brand-only naming, missing taxonomy, and thin data before they cost you the searches you never see. You do not have to guess why the assistant skipped you. You have to be the store it understands well enough to recommend.