How to Prepare Your Shopify Store for AI Discovery

Shopify already puts your products in front of AI assistants. The hard part is getting picked once you are there.

Since March 2026, Shopify switches Agentic Storefronts on by default for eligible stores. Your catalogue flows through Shopify Catalog to ChatGPT, Google AI Mode, Gemini, Microsoft Copilot, Meta and the Shop app.¹ So the plumbing is done. But plumbing is not a recommendation. ChatGPT’s shopping results usually show three to five products.² If you are not in those five, you are not in the answer.

So what decides the five? The assistant reads your product pages, compares the facts, and picks the products it can describe with confidence. Your job is to make that easy.

1. Let the AI crawlers in

Check your robots.txt and your Cloudflare settings. On 1 July 2026 Cloudflare gave every customer separate controls for search, agent and training bots.³ If you block training, Cloudflare also blocks crawlers that do search and training together. Allow OAI-SearchBot, Claude-SearchBot and PerplexityBot if you want to appear in their answers.

2. Put the facts in the page, not in a script

Most AI crawlers do not run JavaScript.⁴ A spec table that loads on click, a size guide in a pop-up, a review widget that renders late: the bot sees none of it. Put the facts a buyer asks about in the server-rendered HTML of the product page.

3. Answer the questions buyers actually ask

A product page usually says what the product is. An AI answer needs to know who it is for, who it is not for, and how it compares. Write those answers on the page, in plain text, at three depths: the broad question (“a waterproof jacket”), the specific one (“a breathable waterproof jacket for hill walking”), and the exact one (brand, spec, size, use).

4. Make the structured data match the page

Structured data on its own moves very little. Ahrefs tracked 1,885 pages that added JSON-LD, and AI citations moved by about 2%, which is not a real change.⁵ It works when it repeats what the visible page already says. Use ProductGroup for variants, and never put a fact in the JSON-LD that the page does not show.

5. Fill your discovery files

Shopify theme stores now serve /llms.txt, /llms-full.txt and /agents.md by default.¹ The default text is the same short template on every store. Replace it with what makes your brand different: your range, your best sellers, your guarantees, your delivery facts.

6. Publish buying guides on your own domain

Assistants answer comparison questions every day: “best trail shoes for wide feet”, “cast iron or stainless steel”. A guide on your own site that answers one of those questions, and links to the products, gives the engine a page it can cite.

7. Test what the AI says, every week

You cannot fix what you do not measure. Ask the engines the buying questions in your category, every week, and write down who they name and which pages they cite. Then fix the gaps and ask again.

Or let Geoffy do all of it

Geoffy is a Shopify app. It writes a verified answer block onto every product page, with matching structured data, llms.txt, agents.md and buying guides, all on your own domain. It never edits your product descriptions or tags. Then it runs a weekly AI test across ChatGPT, Perplexity, Gemini and Claude, ranks you against your competitors, and shows you which pages the engines cite.

Install the app, run your first test, and see where you stand this week.

Footnotes

1. Shopify Help, “Shopify agentic storefronts” and Shopify News, 24 March 2026. Checked October 2026. ↩ ↩²

2. Profound, “ChatGPT shopping deep dive”, June 2026. ↩

3. Cloudflare blog, “Your site, your rules: new AI traffic options for all customers”, 1 July 2026. ↩

4. Vercel and MERJ, “The rise of the AI crawler”, December 2024. ↩

5. Ahrefs, “We tracked 1,885 pages adding schema”, May 2026. Neither result was statistically different from zero. ↩