Reading the AI visibility screen

The AI visibility screen is a grid of prompts against models, two rates, a list of what was cited, and a history of recent sweeps. Each part keeps a distinction that would be easy to collapse and expensive to lose.

The grid has six states, not four

• Named and cited: your brand in the prose and your site linked.
• Cited: a link to you, without your name written out.
• Named: your name, without a link.
• Not named: the model answered and you were not in it. This is a finding.
• No answer: the call failed. This is not a finding.
• Waiting: the sweep has not reached that cell yet.

The last two are kept apart from “not named” because a sweep where every call failed would otherwise render exactly like a sweep that measured your market and found you absent — an outage and a finding, which a reader has to be able to tell apart from the grid alone.

Two rates, two denominators

The mention rate and the citation rate are both divided by the answers that actually came back rather than by the number of calls attempted, and each counts its own denominator because the two read different records. A sweep with nothing to divide by reads as a dash and never as a rate of nothing.

A mention is a guess and a citation is a fact

A brand whose name is an ordinary word matches prose that is not about it, and nothing fixes that in general. What keeps it contained is that the sentence around every mention is stored, so you can read it and decide in a glance; that a link is matched by a rule for addresses rather than by a rule for words; and that the two are never averaged into one number, because averaging a fuzzy signal into a precise one gives you a figure that is neither.

Where you came among names

When an answer names several brands, the screen records where yours came among the ones you asked about — your own domain and the rivals you track. It is not a place among every brand the model named. Track one more rival and your own position in that list can move without the model having said anything different, which is why the column says it is a place among the brands you track.

A rate over several sweeps

The headline figure is a rate over the last 5 sweeps rather than a verdict from the newest one, for the reason the setup article gives: one sweep is a sample. The Dashboard’s AI mentions figure reads the newest finished sweep against the one before it, which is a different and narrower question, and the per-model bars beneath it carry the window.