what is the ai visibility score (and what it hides).
every ai monitoring tool now sells you a score. before you subscribe to one, it is worth understanding what that number is actually made of, because the answer is less flattering than the dashboards suggest.
an ai visibility score is a composite number produced by running a set of prompts through engines like chatgpt, perplexity, and google ai overviews and counting how often your brand shows up in the answers. it measures appearance frequency. it does not measure whether anyone holding a budget saw you, and that difference is the entire problem with treating it as a metric.
where the score comes from
the ai visibility tool market has exploded: there are now well over 200 tools tracking ai mentions in some form. nearly all of them work the same way. they take a basket of prompts about your category, run them on a schedule, log which brands each engine names, and compress the result into one number. you appeared in 12% of answers. your score is 61 out of 100. up 4 points since last month.
that is a real measurement. the tools are not lying to you. the compression is the problem.
the averaging problem
a composite score averages queries that have nothing in common commercially. being named when someone asks "what is content marketing" and being named when someone asks "who should i hire to run content marketing" are not the same event. one of those questions has budget behind it. the other never will. a score that averages them treats a definition lookup and a purchase decision as equal appearances, which means you can watch your score climb while your presence on the five queries that precede actual spending stays at zero.
a visibility score tells you that you appeared. it will not tell you whether you appeared in front of anyone holding a budget, and those two facts are worth wildly different amounts.
the inverse also happens, and it is worse. a brand that owns the three queries buyers type right before they spend can carry a mediocre composite score, panic about it, and start optimizing for volume across queries that will never produce a dollar. the score made them worse at the thing that was working.
why citation is still worth chasing
none of this means ai visibility does not matter. the click data says it matters more every month. seer interactive's analysis found that brands cited inside ai overviews earn roughly 120% more organic clicks per impression than uncited brands on the same queries. and ahrefs' march 2026 analysis of 863,000 keyword serps found only 38% of ai overview citations now come from pages ranking in google's top 10, down from 76% seven months earlier. rank and citation have decoupled. citation is winnable separately, and it pays.
the argument is not against measuring. it is against measuring an average when the value lives in a handful of specific places.
what to measure instead
replace the composite with a short list you can hold in your head:
- your money queries. the five to ten questions a buyer types right before spending in your category, phrased the way a buyer phrases them, not the way your industry does.
- named or not, per engine. for each money query: does chatgpt name you, does perplexity name you, does google's ai answer name you. yes or no beats a percentage.
- who gets named instead. the competitor list ai actually returns is your real competitive set, and it rarely matches the one in your head.
- which sources the answers cite. those pages are where the model formed its opinion. that list is your work queue.
four columns, checked monthly. no subscription required, and every row is something you can act on, which is the part a score never gives you.
the honest role of a score
a composite score has one legitimate use: trend direction across time on a fixed prompt set. if you hold the basket constant and the number moves, something changed and it is worth finding out what. as a headline metric to report or optimize, it hides more than it shows. score the queries that carry money. ignore the rest.
the free scan on the audit page checks how ai engines read your site, in about 60 seconds, no email asked.
one email when the next field note drops.
no course pitch, no daily emails. the notes, when they exist.
sources: seer interactive, ctr and citation analysis of ai overview queries (2026) · ahrefs, analysis of 863,000 keyword serps and 4m ai overview urls, mar 2 2026 · trustmary, ai search visibility tools landscape (2026), 200+ tools tracked. last verified: aug 10 2026.