AgentReadyAI visibility appCaffeine & CommerceShopify agency
Caffeine and Commerce
By Dylan HuntJune 19th, 2026AIAgentic commerceSEO

The Shopify Global Catalog: How AI Shopping Finds and Ranks Your Products

The Shopify Global Catalog: How AI Shopping Finds and Ranks Your Products

Here is the shift in one sentence: shoppers have started asking ChatGPT, Gemini, and Shop what to buy, and those assistants do not browse your store to answer. They query a single cross-merchant index called the Shopify Global Catalog, and they recommend whatever that index can read clearly.

That changes the job. For fifteen years the goal was to get a shopper onto your product page. Increasingly, the shopper never sees your page until after the assistant has already picked your product over a competitor's. The decision happens upstream, in data you may never have thought of as marketing. This guide explains what the Global Catalog is, how an assistant actually searches it, and the concrete work that decides whether your products get surfaced or skipped.

What the Global Catalog actually is

The Global Catalog is a bundled, cross-merchant index of products from across Shopify. When a buyer tells an AI assistant "find me a daily trainer for marathon training under $150," the assistant does not visit a dozen running stores. It runs one search against the catalog and gets back a ranked list of products, each one tagged with the merchant that sells it.

This is the plumbing under Shopify's agentic storefronts, the channel that makes your products discoverable inside AI surfaces. For most merchants it is already on. You will find it under Sales channels then Agentic in your admin, where the default is to let Shopify manage participation for you. There is no separate app to install and no transaction fee beyond your normal processing rates.

So the listing part is mostly handled. The part that is not handled, and the part nobody is optimizing yet, is whether the assistant chooses you.

How an assistant searches the catalog

It helps to see the shape of a real query. When an assistant searches the catalog, it composes three distinct things:

  • A query: the literal thing the shopper wants, like marathon training shoes.
  • Context: soft signals that inform ranking and localization but do not exclude anything. This includes free-text intent like "durable for outdoor use," plus country, currency, and language.
  • Filters: hard exclusions. A result that fails a filter is dropped entirely. These cover price ranges, availability, shipping eligibility, and product condition.

That distinction between soft context and hard filters is the whole game. If a shopper says "ships to Canada" and your shipping data does not show coverage for Canada, you are not ranked lower. You are removed from the results before ranking even happens. The same goes for "in stock" and "under $40." Filters are pass or fail.

Each product that comes back carries structured fields the assistant reads directly: the selling merchant's domain, a price with its currency, a rating value, and a variant options matrix for things like size and color. When a shopper narrows to "size 10 in black," the assistant resolves that against your option data. If your variants are not modeled as proper options, that resolution fails quietly.

The rule that changes everything: an assistant only knows what your data says

This is the single most important line in Shopify's own guidance for the agents that read the catalog, and it is worth quoting plainly: never invent specs, prices, availability, or details. If the response does not say it, the assistant does not say it.

Sit with that. A human shopper on your product page will forgive a missing detail. They will read between the lines, click into a description, infer that a jacket is waterproof from the photos. An assistant does none of that. It works only from the structured fields the catalog returns. A claim that lives only in a hero image or a buried paragraph effectively does not exist to the agent making the recommendation.

This is why "AI readiness" is not a vibe. It is the concrete, checkable question of whether your product data states, in machine-readable form, the things shoppers filter and search on. Clean structured data is no longer just how you win a rich result in Google. It is the difference between being in the consideration set and being invisible.

How the catalog picks between sellers

When several stores sell the same product, the assistant does not show ten near-identical offers. It clusters them and surfaces one. That mechanic behaves a lot like the Amazon buy box, and the levers that win it are price, shipping coverage, availability, and the richness of your product data. I dug into exactly how that clustering and selection works in The Agentic Buy Box: how Shopify's Global Catalog picks between sellers, and it is the post to read if you sell anything that other stores also carry.

One subtlety worth flagging now: assistants identify a seller by its domain, not by the brand name in a product title. A listing titled with a famous brand but sold by a random reseller is treated as third-party resale, not as that brand. If you are the brand, that is an argument for owning your category data so your own listing is the clean, complete one.

Why you rank for your name and disappear for your category

The most common problem I see is a store that surfaces instantly for its own name and vanishes for the searches that actually bring new customers. The reason is simple once you see it. Your brand name is already in your titles, so name searches are trivial. Category searches like "organic herbal tea bags" or "titanium polarized sunglasses" depend on category assignment, attributes, and descriptions that an assistant can match against. When that data is thin, you win the searches from people who already know you and lose the ones from people who do not.

I ran this as a real experiment on two live brands, and the pattern was stark. The full write-up is here: why AI shopping shows your brand but skips your category.

The readiness checklist

Here is the practical version. For each thing a shopper might ask an assistant, there is a field in your store that has to carry the answer.

When a shopper asks forThe catalog filters or ranks onYour store needs
"under $40"priceAccurate per-variant pricing and currency
"ships to Canada"shipping eligibilityShipping profiles that cover the market
"in stock"availabilityReal-time inventory, synced
"size 10 in black"the options matrixVariants modeled as proper options
"well reviewed"ratingReal reviews on your product pages
"organic" or "certified"description and attributesThe claim stated in your product data

A few of these deserve extra attention:

  • Category and attributes. Map products to the Shopify Standard Product Taxonomy and fill in the attributes for your category. This is what lets you match a category search instead of only a name search. For the full field-by-field breakdown, see the product fields that decide your AI shopping rank.
  • Compliance disclosures. Agents are required to surface legal and compliance notices, things like Prop 65, allergens, age restrictions, and energy labels, proximate to the product. Put those disclosures into your product data, near the top, rather than in a footer image.
  • Structured data. The Schema.org JSON-LD on your pages is the cleanest way to hand an assistant labelled product, review, and FAQ data. If you are new to it, start with the structured data for Shopify guide, and consider an llms.txt file so assistants get a clean map of your best pages.

None of this is exotic. It is mostly product-data hygiene that you can audit today. The hard part has been knowing which gaps actually cost you visibility, which is the thing the next section solves.

How to see where you stand

You cannot fix what you cannot see, and the catalog does not send you a report card. The fastest way to find your gaps is to look at your store the way an assistant does.

That is exactly what our free Shopify AI-readiness checker does. Enter your store, and it scores the product data, structured data, and crawlability that decide whether you get surfaced, then hands you the specific fixes ranked by impact. No sign-up, nothing stored. To benchmark your standing across the queries that matter and track it over time, see how to tell if you're winning or losing in the Catalog.

The stores that win the next few years of agentic commerce will not be the ones with the loudest marketing. They will be the ones an assistant can read without guessing. The good news is that the work is concrete, it overlaps almost entirely with good traditional SEO, and most of your competitors have not started.

Doing this continuously, rather than once, has a name: Agent Experience Optimization, the discipline of keeping your catalog legible to the agents that increasingly decide what gets bought.

Browse every guide in the Shopify Catalog and AI and agentic commerce topics.

Make your store agent-ready

Get found and recommended by AI shopping assistants.

AgentReady adds Schema.org structured data, an llms.txt directory, and an AI-readability audit to your Shopify store, so ChatGPT, Perplexity, and Google can understand and recommend your products. Free for stores under 500 products.

Comments

Every comment here comes from a verified email. Write yours, confirm from your inbox, and it's live.

Loading comments…

Leave a comment

ShareXLinkedInFacebook

Written by Dylan Hunt, Founder, Caffeine and Commerce. We build Shopify stores that rank and that AI agents can read. Have a project? Get in touch.