The four verdicts

Every question the AI Visibility Tracker measures comes back as one of four states, per engine. Three of them are familiar. The fourth is the useful one.

What each one means

CITED

Your domain appeared in the answer’s source list. The engine did not just know about you — it pointed at your page.

What to do

Nothing, except keep it. Re-run periodically; citations erode as competitors publish.

MENTIONED

Your brand was named in the answer text, but without a link to you. The engine knows the name and is not using your page as a source.

What to do

This is usually a citability problem rather than an awareness one. Audit the page that should be the source and see what is missing.

ABSENT

Neither named nor cited. A competitor took the answer.

What to do

Look at which pages were cited instead, then run a Gap Analysis against them. That turns an abstract loss into a concrete list of differences.

OPEN

Nobody owns the answer yet. The engine had no settled source — not you, not a rival.

What to do

Write the page. You are not out-competing anyone, you are being first, and that is the cheapest visibility available to you.

Why OPEN matters most

Most visibility tools report three states: you are cited, you are mentioned, or you are not there. That framing assumes every answer is contested and your job is to take it from someone.

Run a real prompt set and you find that is often not true. For a meaningful share of genuine buyer questions, the engines have no established source at all — the answer is assembled loosely, cites something tangential, or wanders.

An unclaimed answer is far cheaper to win than a contested one. There is no incumbent to displace and no authority gap to close. You publish the page that deserves to be the source, and you are first.

It also compounds. Once an engine has a reliable source for a question, it tends to keep using it until something better appears.

Reading a verdict properly

Verdicts are per engine, not per question

The same question can come back CITED on one engine and ABSENT on another. That is normal and it is useful: it tells you the problem is with how a particular engine sources answers, not necessarily with your page.

One pass is not a pattern

AI answers vary between runs. A domain cited once in three passes is noise; one cited three times out of three is a real position. This is why runs are multi-pass by default.

Absence is reported as absence

If an engine could not be retrieved during a run, it is excluded and said so — not counted as a zero against you. A missing measurement and a bad result are different things, and the report keeps them apart.

We do not guarantee that fixing a page will change a verdict. Citation depends on factors beyond the page, including off-domain authority we do not measure. What the loop gives you is the diagnosis, the fix, and an honest re-measurement.

Common questions

What is the difference between cited and mentioned?

Cited means the engine listed your domain as a source for its answer. Mentioned means it named your brand in the answer text without pointing at your page. Cited is the stronger signal; mentioned usually indicates a citability problem rather than an awareness one.

Which verdict should I act on first?

OPEN, then ABSENT. Open questions have no incumbent, so they are the cheapest to win. Absent questions need you to out-compete a page that is already being cited.

Why did the same question give different verdicts on different engines?

Because the engines source answers differently. Writing structure helps across all of them, but authority, recency and sourcing vary per engine. That variation is information, not error.

Can a verdict change without me doing anything?

Yes. Competitors publish, engines update, and sources move. That is why re-running the same prompt set on a cadence matters more than any single run.

See your own verdicts

A ten-prompt run across six engines is 15 credits. Fifteen land free on signup.

Start free — 15 credits