The data from 200+ AEO Checker scans - what's actually broken, and why technically solid websites are invisible to AI.

We built AEO Checker to test how visible websites are to AI engines - ChatGPT, Perplexity, Google AI Overviews, Gemini, and others.
After 200+ real scans, here's what the data shows:
Average score: 61.8 / 100 Only 7 websites scored above 90. Zero websites scored below 31.
That last one surprised us. Nobody is completely broken. But almost nobody is actually good.
Here's the distribution:
18% scored 31–50 - significant gaps, likely invisible to AI
59% scored 51–70 - the "mediocre zone" - partially visible, inconsistently cited
20% scored 71–90 - solid but fixable gaps
3% scored above 90 - actually AI-ready
What's breaking most sites:
The pattern across 200+ scans is consistent. It's rarely one thing - it's the combination:
Schema markup exists but is incomplete. Organization schema missing sameAs. Service schema missing. FAQPage schema absent even when a FAQ section visually exists on the page.
Bot access is quietly blocked. GPTBot, PerplexityBot, ClaudeBot - accidentally disallowed in robots.txt. The site looks fine to humans and Google. AI crawlers never get in.
Content isn't structured for answers. Paragraphs are too long. No direct answer below headings. AI engines can't extract a citable response.
Entity clarity is weak. The business exists on their website. Nowhere else consistently. AI builds a "blurry" picture and skips the citation.
The uncomfortable truth:
A site can have a 95+ Lighthouse score, rank on page one of Google, and still score 55 on AI visibility. These are different optimization targets. Most websites were built for the old standard.
We built AEO Checker because we needed a fast baseline before client implementations. It's free - you get a score across 4 categories - schema markup, content structure, AI crawler access, and citability - with specific recommendations
→ Test your site: aeo.inity.agency
Curious what score people here get. Drop it in the comments.
1
16
2