Technical SEO for AI Search: Making Your Shopify Store Readable to ChatGPT, Claude & Perplexity

Search has changed. Twice.

The first shift was Google adding AI Overviews. The second is users skipping Google entirely and asking ChatGPT, Claude, Perplexity, and Gemini for product recommendations.

Both depend on the same thing: machines understanding your store at the data layer, not the visual layer.

This post is how I get a Shopify store from "invisible to AI" to "well-cited by AI."

1. JSON-LD on every commercial page

If you do nothing else, do this. AI crawlers and traditional search both rely on Schema.org JSON-LD. On a Shopify store you want, at minimum:

• Product with offers, prices, availability, brand, and aggregate rating.
• BreadcrumbList matching your collection hierarchy.
• Organization sitewide with logo, sameAs, and contact info.
• FAQPage for common product, shipping, and return questions.
• WebSite with a SearchAction so engines surface your search.

I emit all of this from Liquid using product metafields as the source of truth. No third-party app, no JS injection. The schema is in the HTML on first paint.

2. Add a public llms.txt

llms.txt is the emerging standard (llmstxt.org) for telling LLMs what your site is about, who runs it, and where the canonical content lives. Think of it as a robots.txt for meaning instead of crawl rules.

A good llms.txt for a Shopify store includes:

• Brand name, location, contact.
• A short, factual description of what you sell and to whom.
• Your top collections and key product categories.
• Sizing, shipping, return policies in plain text.
• Links to canonical pages.

Put it at /llms.txt and a longer version at /llms-full.txt. Reference both in your sitemap.

3. Allow AI crawlers explicitly

Many stores accidentally block GPTBot, ClaudeBot, PerplexityBot, Applebot-Extended, and Google-Extended through restrictive robots.txt rules they don't even know about (often inherited from older themes or apps).

Audit your /robots.txt. If you want to be cited, allow these explicitly:

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot Allow: /

User-agent: PerplexityBot Allow: /

User-agent: Google-Extended Allow: /

User-agent: Applebot-Extended Allow: / ```

This won't make you appear; it removes a barrier that prevents you from appearing.

4. Site architecture humans and machines both like

LLMs cluster by topic. So does Google. Your collection structure should mirror how a customer would describe what you sell.

• Each collection page should answer one question. ("Men's linen shirts" — not "New arrivals.")
• Internal links should connect collections that belong together.
• Avoid orphan PDPs.
• Make breadcrumbs match your collection hierarchy. JSON-LD breadcrumbs included.

5. Hreflang done right

If you sell internationally, you need hreflang. If you have it but it's wrong, you have a problem. Common mistakes:

• Self-referencing hreflang missing from each language version.
• x-default missing.
• Mismatched canonicals (canonical points to one URL, hreflang points to another).

Fix these before doing anything else. Bad hreflang doesn't just hurt SEO — it confuses AI summaries that try to pick the right version for the user.

6. Make your most-asked questions easy to find

LLMs love to quote FAQs. Concrete, factual, well-structured Q&A on shipping, returns, sizing, and care will get cited disproportionately often when users ask "is X store reliable?" or "how does X handle returns?"

Mark up these pages with FAQPage schema and link to them from the footer.

7. Watch the right metrics

You can't measure AI citations the way you measure clicks. What you can measure:

• Branded query volume in Search Console (a leading indicator that AI is referring users).
• Direct traffic from no-referrer, especially mobile (often a chat or voice referral).
• Mention monitoring — set alerts for your brand on Reddit, X, and via tools like Mention or Brand24.
• Test queries: every two weeks, ask the major chat models the questions your customers would ask. See if you're cited. If not, ask why and fix the underlying data.

Bottom line

AI search isn't a separate channel. It's the same SEO discipline taken seriously: clean data, fast pages, well-structured content, explicit permissions, and a real description of who you are and what you sell.

A Shopify store that is well-built for AI search is also a Shopify store that converts better, ranks higher in classic SEO, and is easier to operate. Win on all three at once.

This is part of how I work as a Shopify Operator. If your store needs a technical SEO and AI-readiness audit, [let's talk](/contact).