Most audit vendors will show you a sample report for a site that scored well. We did something different: we ran Citeify's full audit pipeline against citeify.ai — same checkout, same pipeline, same QA gate as a paying customer — and published the result.
Overall: 41/100.
That number is the equal blend of two very different measurements:
- Readiness: 82/100. The foundations we control — structure, schema, crawlability, content quality — measured by deterministic checks against a fixed rubric.
- Measured AI Visibility: 0/100. We put ten buyer-intent questions from our own niche to ChatGPT, Perplexity, Gemini and Google AI Overviews, sampling each question three times per engine. Across 120 recorded answers, no engine cited citeify.ai once.
Why publish a zero?
Because it is true, and because it is normal. Citeify's brand is weeks old. AI engines answer buyer questions from the sources they already trust — established comparison pages, community discussion, review sites. A new brand starts invisible no matter how well its site is built. That gap between citable (82) and cited (0) is precisely what our audit exists to measure, and most site owners have no idea which side of it they are on.
What the run history shows
The more interesting story is in the run-to-run data:
- We audited the unchanged site twice in one evening: 76 and 76. Same site, same score — our readiness checks are computed from measured signals, not model mood.
- We then implemented a set of fixes our own report recommended — a proper About page with a named founder, FAQ answers rendered into the server HTML where AI crawlers can actually read them, canonical tags, a real entity graph — and re-ran it: 82.
Flat when nothing changes. Moves when something does. That is the behaviour you should demand from any scoring instrument before you trust it.
What we found (and had to fix in our own product)
Our audit flagged, on our own site, the same failures we flag for clients: FAQ answers locked inside JavaScript accordions that AI crawlers cannot expand; no named human anywhere on the site; an entity graph with no real-world profiles attached. We fixed them the same day. The audit also produced one false positive — which we traced, fixed, and recorded in the public methodology changelog. Auditing yourself is uncomfortable. It is also the fastest QA process we have ever run.
The baseline is the point
Our visibility score is 0 today, on a frozen set of questions we will re-measure on the same terms every time. When it moves, the movement will be real — and public. If you would like to know your own number, that is what we sell. The full report we generated on ourselves, warts included, is here (PDF).