We ran our own audit tool on our own homepage. SEO: 94/100. AI-citability (GEO): 41/100 — full breakdown inside.
We built a tool that scores sites on classic SEO and AI-search citability (GEO), then generates page-specific fixes and re-verifies they actually worked — instead of just listing suggestions nobody checks. Before pitching anyone, we pointed it at ourselves. Full public report: https://kinetixseo.com/seo-check/QQCGinF2dD
SEO: 94/100 (A) — content quality 95, technical SEO 92, on-page 89, structured data 100, AI search readiness 100, images/video 100, local SEO 100, trust/compliance 85.
GEO: 78/100 (moderate) — the category spread is the real story: authority/brand 91, structural readability 91, technical accessibility 80, multi-modal 100... and citability 41, the single heaviest-weighted GEO dimension. That's what actually decides whether ChatGPT/Perplexity/Gemini quote you back — and it's dragging the whole score down despite everything else scoring 80+.
16 fixable issues found, the notable ones:
Scaled/low-value content risk (high severity) — low readability, repetitive paragraphs.
Insufficient citation-ready passages and stat density — not enough concrete, sourced numbers for an AI system to quote confidently.
Brand name inconsistent across title, meta description, and H1.
No first-hand experience language ("we tested," "we found") anywhere on the page.
URL path not descriptive enough, missing social profile links, no analytics installed, Core Web Vitals not yet measured.
A technically excellent site (schema 100, AI-search-readiness 100) can still be nearly invisible to AI citation on the one dimension that matters most. Closing that gap — and proving the fix worked instead of just claiming it — is the actual product.
Running your own analysis normally costs €29 for a one off scan (generates the fixes and re-scans to confirm they worked). Reply here with your site and what you'd want checked, and I'll send a code to run one free — just want honest feedback on what it gets right or wrong.
The gap between technical SEO and citability is the interesting finding
here, most people assume schema + structured data automatically buys
you AI citations, but "well-marked-up" and "quotable" turn out to be
different properties. Makes sense in hindsight: an AI model quoting you
back needs an actual sourced number or first-hand claim to lift, not
just a page that's crawlable.
The "no first-hand experience language" finding is the one I'd have
missed entirely, hadn't considered that "we tested" / "we found" phrasing
specifically signals citability versus generic third-person claims
making the same point.
Running it on your own homepage before pitching anyone is the right
call too, a tool that scores its own maker's site honestly (94 not 100)
is more credible than one that would've conveniently returned a perfect
score.
The fun part is, it is impossible to score 100 percent across the board. What is good for AI might not be good for humans to read. Local SEO is very low for a SaaS software.
Yes many things can be done, should be done and reworking texts to suit both humans and AI is where it gets interesting. I will post an update as soon as the new texts are in.
Really interesting audit. The GEO score caught my attention, especially since the technical SEO numbers are already so strong. Would be interesting to see how much the score improves after fixing those citation issues.
Well the numbers are in:
https://kinetixseo.com/seo-check/tM4RgtpaAo
I fixed all the pages and the scan multiple times to get to better results.
The problem is/was the homepage. Yesterday I redesigned it to be more logical and CI citation went down. So tomorrows scan will not be as good as yesterdays.
Making your site or page score higher only matters if your content is good, so make no mistake. Fixing your scores is important but it must go hand in hand with good content.