AI search engines like ChatGPT, Claude, and Perplexity now handle 900 million weekly searches. When someone asks "what's the best payment API?", the AI doesn't return 10 blue links — it gives one answer, citing one or two sources.
The question isn't whether your site ranks anymore. It's whether AI can understand your site well enough to recommend it.
We audited 10 major websites for AI Engine Optimization (AEO) readiness. We checked for llms.txt, structured data, FAQ schema, content clarity, and overall AI-friendliness. The results were revealing — and one finding was downright ironic.
What We Checked
For each site, we evaluated five core AEO signals:
- llms.txt — Does the site provide a structured guide for AI agents?
- Structured Data — Schema.org markup (Organization, Product, FAQ, etc.)
- FAQ Schema — FAQPage markup for direct answer extraction
- Content Clarity — Clean hierarchy, clear descriptions, no JS-only rendering
- AI Citation Potential — Would an AI model likely cite this site?
The Results: 10 Sites Audited
Stripe
stripe.com
✓ llms.txt (+ llms-full.txt)
✓ Rich structured data
✓ Clear product descriptions
✓ Comprehensive docs
Key finding: Stripe is best-in-class. Their llms.txt provides a structured overview of every product with direct links. They also offer
llms-full.txt — complete documentation in a single file specifically for AI consumption. This is what AEO-readiness looks like at scale.
GitHub Docs
docs.github.com
✓ llms.txt with APIs
✓ Dedicated LLM endpoints
✓ Clean markdown output
✓ Search API for agents
Key finding: GitHub goes further than anyone. Their llms.txt doesn't just list pages — it provides dedicated APIs for AI consumption: a Page List API, Article API (returns clean markdown), and Search API. They've built infrastructure specifically for AI agents to consume their docs. The gold standard.
Salesforce
salesforce.com
✓ llms.txt present
✓ Rich structured data
✓ Agentforce content
~ Focused on AI/agent topics
Key finding: Salesforce's llms.txt is strategically focused on their Agentforce product line — AI agent courses, agentic architecture. They're using llms.txt as a marketing channel to position their AI products for recommendation by other AI systems. Smart move.
Shopify Dev
shopify.dev
✓ llms.txt present
✓ Structured data
✓ Developer docs
~ Dev platform only (not main site)
Key finding: Shopify's developer platform has llms.txt. Worth noting: it's on
shopify.dev, not shopify.com. The main commerce site likely doesn't have one. This is a common pattern — developer tools adopt first, marketing sites lag behind.
Vercel
vercel.com
✓ llms.txt present
✓ Docs linked
~ Basic format
✓ Clear product positioning
Key finding: Vercel has llms.txt with a clear pitch: "the AI Cloud." Interesting positioning choice — the llms.txt file itself frames Vercel as an AI platform, not just a hosting provider. They're using AEO to reinforce their brand narrative.
Linear
linear.app
✓ llms.txt present
✓ Clean structure
✓ Docs with markdown links
~ No full-doc file
Key finding: Linear's llms.txt is clean and well-structured with docs sections (Getting started, Concepts, etc.). A strong implementation for a SaaS product — exactly what AI needs to understand and recommend a project management tool.
Notion
notion.so
✓ llms.txt present
~ Minimal content
~ Brand description only
✗ No doc links
Key finding: Notion has llms.txt but it's minimal — just a brand description ("50 million users...") without documentation links or structured product info. It's a start, but not leveraging the format's full potential.
Anthropic 🤦
anthropic.com
✗ No llms.txt (404)
✓ Good structured data
✓ Strong docs (docs.anthropic.com)
~ Great content, no AEO file
The irony: Anthropic — the company behind Claude, one of the most popular AI models — doesn't have an llms.txt file. Their main site returns a 404. This is like a locksmith with a broken lock. Their docs at docs.anthropic.com are excellent, but the main brand site misses a basic AEO signal. Even AI companies haven't fully adopted AEO practices.
HubSpot
hubspot.com
✗ No llms.txt (404)
✓ Rich structured data
✓ Massive content library
~ Strong SEO, weak AEO
Key finding: HubSpot is an SEO powerhouse — arguably the best content marketing machine in SaaS. But they don't have llms.txt. This represents the gap between traditional SEO excellence and AEO readiness. Having great content isn't enough if AI agents can't easily parse and recommend it.
Nike
nike.com
✗ No llms.txt (404)
~ Basic structured data
✗ Heavy JS rendering
✗ Not AI-agent friendly
Key finding: Nike's site is built for visual browsing, not AI consumption. No llms.txt, heavy JavaScript rendering, and minimal structured data beyond e-commerce basics. When someone asks an AI "what running shoes should I buy?", Nike's site isn't helping itself get cited. Traditional brands are the furthest behind in AEO.
The Scorecard
| Site | llms.txt | Schema | AEO Level |
|---|---|---|---|
| Stripe | ✓ + full | ✓ | Excellent |
| GitHub Docs | ✓ + APIs | ✓ | Excellent |
| Salesforce | ✓ | ✓ | Good |
| Shopify Dev | ✓ | ✓ | Good |
| Vercel | ✓ | ✓ | Good |
| Linear | ✓ | ✓ | Good |
| Notion | ~ minimal | ~ | Basic |
| Anthropic | ✗ 404 | ✓ | Missing |
| HubSpot | ✗ 404 | ✓ | Missing |
| Nike | ✗ 404 | ~ | Not Ready |
Key Takeaways
1. Developer-facing companies lead. 7 out of 7 tech/developer platforms had llms.txt. Non-tech brands (Nike) and even marketing-heavy SaaS (HubSpot) are behind.
2. Having llms.txt isn't enough. Notion has one but it's minimal. Stripe and GitHub show what "excellent" looks like: full documentation, dedicated APIs, structured product descriptions.
3. SEO ≠ AEO. HubSpot dominates traditional SEO but scores poorly on AEO. The skills are different. Great content needs AEO infrastructure to be cited by AI.
4. Even AI companies aren't ready. Anthropic builds Claude but doesn't have llms.txt on their main site. The cobbler's children have no shoes.
5. llms.txt as marketing. Salesforce and Vercel use their llms.txt to push specific narratives ("the AI Cloud", "Agentforce"). The file isn't just technical — it's a brand positioning tool for AI audiences.
What Should You Do?
If your site doesn't have llms.txt, you're in the same boat as Anthropic, HubSpot, and Nike. Here's a quick priority list:
- Add llms.txt — Even a basic one helps. Describe what your company does, link to key pages.
- Check your structured data — Schema.org markup (Organization, Product, FAQ) helps AI parse your content.
- Ensure content renders without JavaScript — AI crawlers can't always execute JS. Server-side rendering matters.
- Add FAQ schema — AI models love pulling from FAQPage markup for direct answers.
- Audit your AI visibility — Ask ChatGPT and Claude about your product category. Are you being cited?
Want to know where your site stands?
Get a complete AEO audit with specific, actionable fixes — not just another dashboard to stare at.
Check Your AEO Score →