The Claude audit that actually works — and the one that doesn't
Hey,
Last week I ran revenueexperts.ai through a technical SEO/AEO audit twice.
Once with Claude.ai. Once with Claude Code.
The Claude.ai version told me robots.txt was inaccessible, llms.txt might not exist, and it couldn't confirm schema markup was present. The Claude Code version fetched the real files and found: robots.txt correctly configured, llms.txt present but auto-generated by Yoast (a problem — I'll explain), and 6 H1 tags on the homepage when there should be exactly 1.
Same site. 3-minute gap between runs. Completely different findings.
This matters for your AEO work because the files Claude.ai can't read — robots.txt, llms.txt, schema markup in <head> — are exactly the signals that tell AI engines whether to trust and cite your content. If your audit can't see them, you're flying blind.
Here's what's actually happening and what to do about it.
🔍 This week's AEO insight: the llms.txt problem nobody's talking about
What changed: Yoast SEO v27.1+ now auto-generates llms.txt files. This sounds helpful. In practice it creates a new problem.
Auto-generated Yoast llms.txt files use ad-copy brand descriptions: "Looking to boost B2B revenue with AI automation? We can help you accelerate sales, improve lead generation, and drive sustainable growth."
That's not how AI systems want to be briefed about who you are. It's what a human would skim on a landing page and close. AI engines parsing llms.txt are looking for entity definitions — clear, factual, structured descriptions that answer: what is this company, what does it do specifically, who leads it, what's its expertise domain.
Here's the difference:
Yoast auto-generated (bad):
"Looking to boost B2B revenue with AI automation? We can help you accelerate sales, improve lead generation, and drive sustainable growth"
Manually curated (what you want):
"Revenue Experts AI is a B2B consulting firm specializing in AI operating system deployments, Answer Engine Optimization (AEO), and multi-agent revenue automation. Founded by Elizabeta Kuzevska. Primary expertise: n8n workflow architecture, RAG pipeline deployment, AI search visibility strategy for mid-market B2B companies."
The second version gives AI engines named entities (company name, founder name, specific methodology names), domain classification (B2B consulting), specific capabilities (n8n, RAG, AEO), and audience (mid-market B2B). That's what gets you cited accurately when someone asks ChatGPT or Perplexity about your category.
Why it matters now: Every site running Yoast SEO v27+ has this problem active right now. The file exists (so it looks fine), but the content is telling AI engines you're a generic growth marketing service rather than whatever you actually are.
What to do about it: Check whether you have an auto-generated llms.txt at yourdomain.com/llms.txt. If it reads like ad copy, replace it. The correct format is on the llms.txt specification page — or use the SEO/AEO Claude Code skill (below) to generate a custom one from an audit of your actual site.
📊 The numbers
H1 tags matter more for AEO than most people think. Multiple H1s on a page create ambiguity for AI systems about what the page's primary topic is. Ambiguous pages get cited less often because AI engines favor clear, authoritative answers on specific topics. One H1 per page isn't just an SEO rule — it's an AI clarity signal.
Bing Webmaster Tools submission affects ChatGPT visibility because ChatGPT's web search uses Bing as its primary index. Most sites submit only to Google Search Console. If you haven't submitted to Bing Webmaster Tools, ChatGPT web search has limited visibility into your content regardless of your Google rankings.
FAQPage schema is the single highest-ROI AEO investment for most sites. Google's own data shows FAQ schema dramatically increases AI Overview appearance rate for pages already ranking in the top 20. If you have pages in positions 1–20 without FAQPage schema, that's the gap to close first.
🛠️ Quick win of the week: audit your llms.txt in 3 minutes
Step 1: Open a browser and go to yourdomain.com/llms.txt
Step 2: Read the brand description at the top (the line after the > symbol). Does it read as an entity definition or as marketing copy?
Step 3: If it reads like marketing copy, open a text editor and write a replacement — one sentence that includes: company name, what you do specifically, founder name, primary methodology or expertise terms, and target audience.
Step 4: Add a proper content hierarchy below it. List your most important pages by category (Services, About, Blog, Resources) with one-sentence descriptions of what each covers.
Step 5: Save the file as llms.txt and ask your developer to replace the existing file at the domain root. Or use the SEO/AEO Claude Code skill to generate the full file automatically from an audit — it produces a complete, properly formatted llms.txt ready to deploy.
Expected result: AI engines that crawl your llms.txt will have accurate entity information about your company. Over 4–8 weeks, citations in AI engines should reflect your actual expertise rather than a generic category.
Test at revenueexperts.llmauditpro.com — the 5-category AEO audit runs in 60 seconds and includes a Technical Accessibility score that flags llms.txt issues specifically.
📚 Learn more
AI SEO Blueprint — 39 technical modules covering the complete technical SEO and AEO implementation stack. Starts with crawlability and bot access, covers llms.txt strategy, schema implementation by type, and ends with a full monitoring system for tracking AI search visibility over time. The module on FAQPage schema alone has generated measurable citation increases for students within 2 weeks of implementation.
Until next week,
Elizabeta Kuzevska Co-Founder, Revenue Experts AI revenueexperts.ai | onlinemarketingacademy.ai Connect on X: @ekuzevska · Connect on LinkedIn