How to get your Shopify store cited by ChatGPT and other AI assistants

More and more people ask ChatGPT, Perplexity, and Google's AI for shopping advice instead of scrolling through search results. When someone asks "what's a good store for X" or "where can I buy Y in the Netherlands," the assistant names a few sources. If your store is one of them, you get the visit. If it isn't, you don't exist in that conversation.

So the obvious question: how do you become one of the sources an AI actually mentions?

I've been working through this on real Shopify stores, so this is what I've found rather than the theory repeated on every SEO blog. Some of it is easy to fix yourself. Some of it isn't a technical problem at all, and the honest version of that second part is the bit most people selling "AI SEO" leave out.

First, the good news about Shopify

If you're on Shopify you're starting from a better place than you might think, and it is worth being precise about why.

AI crawlers aren't blocked. Look at any Shopify store's robots.txt and you'll notice something: GPTBot, ClaudeBot, PerplexityBot and Google-Extended are not mentioned anywhere. No named rules, no blocks. They inherit the User-agent: * rules like everyone else, which means your products, collections and blog articles are all readable by default. You are not accidentally hiding.

Your content is server-rendered. The AI doesn't need to run JavaScript to read your pages; it gets the full text immediately. Plenty of modern sites get this wrong and serve crawlers a blank page. Shopify doesn't.

You already serve an AI briefing file. Every Shopify store now responds at /llms.txt and /agents.md. Try it on your own domain. More on what to do with that below, because the usual advice about it is wrong for Shopify merchants.

Where the default Shopify blog falls short

Here's the part most merchants don't realise. The default Shopify blog is fine for humans, but it leaves a lot on the table when it comes to being understood by an AI.

The issue is structured data. When an AI reads your page it can either guess what the content means from raw HTML, or read a clean machine-readable summary that states exactly what the page is. That summary is schema. Pages with good schema are easier to parse, easier to trust, and easier to quote.

Shopify gives you some of this. Not all of it. I checked the specifics against Horizon, Shopify's own flagship theme, so these aren't guesses about some badly built custom theme.

Article schema: genuinely fine

Horizon's blog post template does this:

1<script type="application/ld+json">2  {{ article | structured_data }}3</script>

That filter emits a schema.org Article with the headline, author, dates and body. It is a real win and it means an AI can read your article structurally rather than scraping it. No complaints here.

FAQ schema: not incomplete, absent

This is the big one, and it's worse than I assumed before I went looking. Horizon ships an FAQ section that merchants can drop onto a page: a nice accordion, collapsible questions, the lot. Search the entire theme for FAQPage structured data and you get nothing. Not an empty field waiting to be filled. It simply isn't there.

So the common situation is a merchant who has done the right thing (written a genuinely useful FAQ, put it on the site, made it look good) and is emitting zero structured data about it. To a crawler it is a pile of <details> elements. Meanwhile FAQ schema is one of the easiest wins available, because assistants love direct question-and-answer pairs: it is pre-chunked, unambiguous, and trivially quotable.

Business schema: three fields

Here's Horizon's entire Organization markup, copied from the header:

1{2  "@context": "http://schema.org",3  "@type": "Organization",4  "name": {{ shop.name | json }},5  {% if settings.logo %}6    "logo": {{ settings.logo | image_url: width: 500 | prepend: "https:" | json }},7  {% endif %}8  "url": {{ request.origin | append: page.url | json }}9}

Name, a logo if you happened to set one, and a URL. No description. No sameAs.

That missing sameAs is the one that costs you. It's the array of links to your Instagram, TikTok, LinkedIn, Wikipedia entry if you have one. It is how a model connects the store in front of it to the rest of your presence online. Without it your store is a name floating on its own rather than a recognised entity with a reputation attached. Entity resolution is most of what "does the AI know who you are" actually means.

None of this requires rebuilding your store. It is schema work layered on top of what you already have. But it does require someone going in and filling the gaps, because the defaults will not do it for you.

About llms.txt, since everyone asks

You'll see blogs selling llms.txt as a must-have file you need to create. On Shopify that advice is just out of date. You already have one. Every store serves /llms.txt and /agents.md, and since the Spring '26 Edition you can customise both from your theme.

So the question isn't whether to create one, it's whether to bother editing it. My honest read: it's worth one pass and no more. No major AI provider has confirmed it as a ranking or citation signal, and it is a proposed convention rather than a proven factor. What it genuinely does is shape the source material an agent works from when it does read you, and reduce the chance of it misreading a cluttered page. That's real, but it is plumbing, not a growth lever. Fill it in once, describe your store honestly, move on.

The thing nobody selling you "AI SEO" wants to admit

Here's the honest part. Once you've fixed the technical side there is a ceiling, and it has nothing to do with schema.

The biggest factor in whether an AI cites your store is authority. Does anyone else on the internet mention you and link to you? Assistants lean heavily on sources that are referenced elsewhere, because that is the signal they have for "is this real and does anyone vouch for it." If your domain is new, has a handful of pages and almost no external mentions, no amount of perfect schema will make an assistant confidently recommend you.

The technical work opens the door. Your reputation decides whether you walk through it.

So the real playbook has two parts. Fix the technical foundation so that when an assistant does look at your store it can read and understand everything cleanly. Then get other credible sites to mention and link to you: partnerships, guest content, the right directories, press, being genuinely useful somewhere other people already read. The boring reputation work that has always mattered for SEO matters just as much here.

What to actually do this month

In the order I'd do it:

1. Add FAQPage schema to your articles and FAQ pages. Highest value per hour on this list, because the theme gives you nothing and assistants love Q&A pairs. Every article should answer real customer questions, and those questions should be marked up.
2. Complete your Organization schema. Add a description, and add every social profile to sameAs. This is the entity-resolution step and it takes fifteen minutes.
3. Check your policy pages are reachable. Since /policies/ is blocked in robots.txt, put the substance of your shipping and returns terms somewhere crawlable as well.
4. Give articles a real human author with a bio. A named person with a short credential line is a stronger trust signal than a logo. This is the E-E-A-T part, and it applies to AI citation as much as to search.
5. Edit your agents.md once. Describe what you sell and who you are, honestly. Then stop thinking about it.
6. Spend everything left on getting mentioned elsewhere. That's the lever that actually moves once the foundation is solid.

If this sounds like a lot

Most of the technical side is a few hours of focused work if you know where to look in a Shopify theme. The schema, the FAQ markup, the business details. It isn't a rebuild, it's cleanup and completion.

If you read this and realised your store is probably invisible to AI assistants right now, email hello@learnshopify.dev and I'm happy to point you in the right direction. Even if you just want to know what's missing on your own store, I don't mind taking a look.

The stores that sort this out early get the same head start early SEO gave people fifteen years ago. Worth getting right now, while most stores haven't thought about it at all.