Shopify Catalog: how your products end up in ChatGPT without you knowing

part 2 · visibility in AI engines

This happened in March 2026, and it affects every Shopify store. The good news: the feed is already flowing. The bad news: what flows into it is exactly what sits in your admin right now.

Shopify launched Catalog and enabled it by default in every store. Your products are already being passed to ChatGPT, Copilot, Gemini and Perplexity — without you knowing.

Catalog does not index your website. It takes the data from your admin and normalises it for AI queries. If the fields are filled in, the product makes it into answers. If they are not, it drops out in favour of a store whose fields are.

In the previous article we said that more than half of shopping queries now end in an AI answer rather than a visit to a website. And we said that getting into that answer is a story of its own.

Here is that story.

01 · what happened

the quiet launch of shopify catalog

In December 2025 Shopify quietly launched a service called Shopify Catalog, with no fanfare. By the end of March 2026 it was enabled by default in every store on the platform. Not “you can now connect it”, but “it is already connected, with no action from you”.

Shopify took your product catalogue — title, description, price, availability, images, category, variants, reviews — and started passing it directly to partner AI engines. Today there are four of them. Tomorrow there will almost certainly be more.

chatgptagent query
shopify catalogdata normalisation
your productin the AI answer

02 · how it works

the shopper asks — catalog answers

When someone asks ChatGPT about a coffee machine, ChatGPT does not index your website. It asks Catalog: “Which suitable products exist in Shopify stores?” Catalog returns several options with current prices and availability. ChatGPT includes them in its answer.

If your product made that shortlist, you are in the answer. If it did not, you are not.

the selection mechanism that matters

Catalog does not pick products at random. It picks them by data quality. It takes the catalogue, runs it through its own language models, normalises and categorises it — and decides how well your product matches a specific query.

The quality of that decision depends on exactly one thing: what sits in your store’s admin. Not on how attractive your site is. Not on your advertising budget. Not on how long you have been in business.

03 · the gap

“fine for humans” ≠ “fine for machines”

Take a concrete example. A clothing store with a thousand products, each with ten to twenty size variants. On the website everything is tidy: good photography, clear buttons, decent descriptions.

But open the admin and look at the variants: half have no barcode. A third are categorised as “Other”. The descriptions are written for people — “this shirt will become your favourite choice for warm evenings” — with no mention of material, composition, fit, country of manufacture or sleeve length.

For a shopper that is fine. They look at the photo, read the short description and decide for themselves. For Catalog it is a disaster. Asked to “find me a white linen shirt with long sleeves in size L under €80”, Catalog cannot tell whether your shirt qualifies. It does not have the data.

So it shows someone else’s. The one where every field is filled in.

This is the point at which most stores lose customers. Not on the website. Not in Google’s results. In Catalog — the very place they do not know exists.

04 · the good news

catalog invents nothing

It takes exactly what you gave it. Which means visibility is entirely in your hands — you only need to know which fields to look at.

Shopify has a dedicated tool in the admin called Catalog Mapping. It lets you configure which of your store’s fields are sent to Catalog. You can use the standard description, or a separate metafield with more precise copy. You can send the standard category, or a custom one. Most store owners do not even know the tool exists.

That is genuinely good news. Visibility in AI engines for a Shopify store is not a matter of “hacking the algorithm”. It is a matter of “tidying up data that is already being sent”. The difference is enormous.

In the next article: the seven specific fields worth looking at this evening. No specialists, no agencies. Straight in the admin.

an external audit based on public data

Want to know what your Shopify store is sending to Catalog right now? We run a check across four agent channels, review your markup structure and cross-check the fields. You get two reports: a one-page summary for the client side and a detailed technical one with a list of specific fixes.

request an audit →

series · visibility in AI engines

· · ·

more articles

· bankpay

Portmone via the Payment Apps API: BankPay and IBAN payments in Shopify Checkout

Portmone as a native Shopify Checkout payment method through the Payment Apps API: BankPay (IBAN-to-IBAN transfers with no chargebacks), fees of up to 2%, an...

read →
· payment apps api

WayForPay via the Payment Apps API: instalments inside Shopify Checkout

WayForPay as a native Shopify Checkout payment method through the Payment Apps API: monobank instalments (2–25 payments) and PrivatBank instalments (2–12) ri...

read →
· liqpay

LiqPay via the Payment Apps API: Privat24 right inside Shopify Checkout

LiqPay by PrivatBank as a native payment method in Shopify Checkout through the Payment Apps API: paying with Privat24 without entering card details, the 2.7...

read →
free briefing · estimate in 24 hours · no obligation
get
a project estimate →
we reply within one business day · the estimate is free