Why there is no ranking to show you, and what can be measured instead.
An answer engine does not publish a ranking. There is no position 3, no impression count, no log you can request. So the honest question is not "where do we rank" — it is "when a buyer asks, does the answer name us".
That is measurable, by asking. The platform puts real buyer questions to real assistants and records which brands the answers name. What comes back is a sample of what those engines said on that day, to those questions. It is evidence, not a leaderboard.
So nothing here promises a citation, a mention or a position, and nothing here can. A run tells you what happened; it does not tell you what will.
One more line runs under every screen in this course: a measurement that failed is excluded, never printed as a zero. An engine outage is not your absence, and rendering it as 0% would be the single most convincing lie this plugin could tell.
2.2 The citation scorecard
Set up a scorecard and read what each engine actually answered.
Fill in the brand, the category and who your buyer is. The platform writes buyer questions from that, asks the engines, and reads the answers back — here, inside WordPress. Readiness predicts; this measures.
The engine table has five columns and every one of them is a fact rather than a score. Engine. Answered by, the models that replied. Mentioned you in, a share. Answers counted, the denominator that share was taken over. Web search, on or off, because an engine allowed to browse and an engine answering from memory are not the same instrument and their numbers should never be pooled without saying so.
Questions that name your brand or a competitor lead the engine, so they are counted separately from questions that name nobody. Both figures are on the page; neither is folded into the other.
Failed engine calls are named and excluded, with the count that survived printed beside them. And when a competitor was named where you were not, the answer is stored word for word — open one and read what your buyer actually saw.
2.3 Which crawlers read this site
See which AI crawlers visited. Counted here, sent nowhere.
The crawler observatory is entirely local. It matches the user agent a visiting crawler declares against a vendored list of AI agents, and counts the hit. No account, no network call, nothing sent.
A declared user agent is a claim, not a proof — anything can send any string. The observatory reports what was declared and does not pretend to have verified it.
Counts are kept for 90 days; the table on the screen shows the most recent 14. Older days fall off on their own.
Two things can send this data onwards, and each is its own opt-in with its own sentence. Forwarding to a collector URL you type in. Sharing to the AEOSearch dashboard, which sends the per-day agent hit counts and nothing else — the paths crawlers requested never travel.
2.4 Rank, and what it is not
Positions the platform measured, with nothing averaged and nothing filled in.
The rank screen shows positions the platform measured for keywords tracked on your account, with the date of that measurement. It is a classic search-engine position, not an answer-engine one.
The plugin does not measure a position, estimate one, or fill a gap. A keyword the platform has no reading for shows as missing, which is a different fact from a bad position and is displayed as one.
Nothing on this screen is averaged into a single number. There is no rank score, because a rank score would hide exactly the movement you opened the screen to see.
2.5 The change digest
What changed between two good audits, and why an outage never counts.
A digest entry is written when two successive successful audits differ — what improved, what regressed, what appeared. No difference, no entry. An empty digest means the last two runs agreed.
Two successful audits. A failed run is not compared against anything, so an outage never manufactures a regression.
The digest reports the difference and stops there. It does not tell you why the number moved, because the plugin does not know why — only that the platform measured something different.