Getting Your Shopify Catalog Into ChatGPT and Perplexity Shopping (2026)

There is a new discovery surface for commerce, and it does not look like search. When a buyer asks ChatGPT or Perplexity for a product recommendation, the assistant answers with specific items and links, and the buyer decides without ever seeing a ranked results page. Shopify's mechanism for getting your products into that answer is Shopify Catalog, and the change that matters for the kind of brands we work with is that since the Winter 26 edition, Catalog explicitly covers headless Hydrogen storefronts on Oxygen. A headless build used to mean opting out of the surfaces a themed store received automatically. Catalog closes that gap: your headless products can now be discovered by AI shopping tools the same way a standard storefront's can.

This is worth treating as its own channel, with its own requirements, separate from both traditional SEO and the in store AI agent work. It is not the same thing as putting an assistant inside your own storefront, which is the Storefront MCP story we covered in the in store AI agent for cart and checkout. Catalog is about being found by assistants that live somewhere else entirely.

	Traditional search SEO	AI shopping via Catalog
Where the buyer is	Your site, after a results page	Inside ChatGPT or Perplexity
What ranks	Pages, links, content	Structured products and their attributes
Signal you control	Keywords, backlinks, page speed	Product data quality and completeness
Headless storefront support	Yours to build	Now covered through Catalog on Oxygen
Failure mode	Low ranking	Not represented in the answer at all

How the Catalog channel actually works

The mental model to drop is the results page. In search, the buyer sees ten links and picks one, and your job is to be one of the ten. In AI shopping, the assistant reads structured product data, reasons over the buyer's stated need, and returns a short list of specific items. There is no page of blue links. Either your product is in the set the assistant considers, with enough structured detail to be matched to the query, or it is absent from the conversation. Absence is the real risk, and it is silent: you do not see a low ranking, you simply are not in the answer.

Shopify Catalog is the pipe that makes your products eligible to be in that consideration set for the assistants that integrate with it. Because the products flow from your Shopify product data, the quality of that data is the entire game. This is a different discipline from writing meta titles. The assistant is not reading a marketing page, it is reading fields: title, product type, attributes, variants, price, availability. Thin or inconsistent product data that a human shopper would forgive because the photos look nice is exactly the data an assistant cannot use, because it has no photos to fall back on and no patience to infer what you left blank.

What your product data has to look like

The practical work is unglamorous and it is where headless brands often have the most room to improve, because a headless build tends to push merchandising detail into the frontend and leave the underlying product records thinner than they should be. A few things decide whether Catalog can represent you well.

Complete, structured attributes. An assistant matching "waterproof hiking boots for wide feet under 200 dollars" needs the width, the waterproof rating, the category, and the price as structured fields it can filter on, not as a sentence buried in a description. Every attribute a buyer might constrain on is an attribute that has to exist as data. For most catalogs this means using metafields and Shopify's standard product taxonomy consistently rather than leaving attributes implicit in prose.

Consistent product types and taxonomy. Assistants reason over categories. If half your boots are typed as "footwear" and half as "shoes" and some as nothing, you have fragmented your own catalog and made yourself harder to match. Consistency across the catalog is worth more than richness on a few hero products.

Accurate price and availability. A recommendation that sends a buyer to an out of stock or mispriced product is the fastest way to lose the assistant's trust and, more importantly, the buyer's. Because Catalog flows from live Shopify data, keeping the source of truth accurate is the same work that keeps your storefront accurate, which is a point in favor of not fragmenting your data across systems.

Why this changes the headless calculus

Going headless has always carried a discovery tax. You trade Shopify's built in themes and their automatic surfaces for full control of the frontend, and part of that trade was giving up channels that themed stores got without effort. Catalog covering Hydrogen on Oxygen removes one of the larger items on that list at exactly the moment AI shopping is becoming a real discovery channel. A headless brand can now compete for the AI shopping surface without abandoning the architecture that gave it control of experience and performance. We laid out the broader shift in agentic commerce and headless storefront discovery in ChatGPT and Perplexity, and Catalog is the concrete Shopify mechanism underneath that shift.

The strategic read is that product data has quietly become a discovery asset, not just an operational one. For years the argument for clean, structured, well modeled product data was internal: better filtering, better merchandising, fewer support tickets. Now that same data is what determines whether an assistant can put you in front of a buyer. The teams that invested in a proper product data model are positioned for this channel with no extra work. The teams that treated product records as an afterthought have discovery debt they did not know they were carrying. Where this sits in the wider architecture is something we mapped in the agentic commerce stack for 2026.

FAQ

What is Shopify Catalog in the context of AI shopping? It is the channel that exposes your Shopify products to AI shopping assistants such as ChatGPT and Perplexity, so those tools can recommend and link to your products inside their own interface. Since the Winter 26 edition it covers headless Hydrogen storefronts on Oxygen.

Do I have to be on a themed Shopify store to use it? No. The change that matters for headless brands is that Catalog now supports Hydrogen storefronts running on Oxygen, which were previously outside this kind of surface.

Is this the same as putting an AI agent in my store? No. That is Storefront MCP, an assistant inside your own storefront. Catalog is about being discovered by assistants that live in ChatGPT or Perplexity, not on your site.

What is the single biggest thing that determines whether it works? The quality and structure of your product data. Assistants read structured attributes, types, price, and availability, not marketing prose, so complete and consistent product records are the deciding factor.

How is this different from SEO? There is no results page. Instead of ranking a page among links, you are trying to be in the set of products an assistant considers and recommends, which is decided by data completeness rather than keywords and backlinks.

When this applies to your stack

If you run a headless Shopify Plus build, AI shopping is a discovery channel you can now compete in, and the work to win it is mostly product data work: a consistent taxonomy, complete structured attributes through metafields, and an accurate live source of truth for price and availability. That is architecture, not marketing, and it is the kind of headless commerce data modeling and integration work we do at Contra Collective. If you want a review of whether your catalog is ready to be represented well in ChatGPT and Perplexity shopping, that is a good place to start.