The AI Visibility Score
Most tools count how often you show up in AI answers, then leave you guessing. Primacy Score measures the two things that matter, Discovery, whether AI finds you when nobody says your name, and Reputation, how it speaks about you once someone does, behind one certified Primacy Score: where you stand, why, and exactly what to do about it.
Not sure yet? The readiness check is free and takes a minute, it tells you whether AI engines can find and read your site at all, before you measure what they say about you.
Primacy Score · Sample / Cycle 12
The two reads behind it
52.7
The search a new customer runs
84.6
Once AI knows you
Discovery by engine
+ Copilot Search & Grok included with every Score · reported beside it, never blended
The question a board, a client, or an investor will eventually ask. Have the answer ready.
The standard
Asked many times.
One AI answer is a lucky screenshot. We ask each buyer question repeatedly and publish the range that produces, so the number is a measurement, not a moment.
Every major AI tool.
ChatGPT, Claude, Gemini, Google AI Mode, Google AI Overviews, and Perplexity. Shown for each, on every plan. Nothing hidden.
Versioned and published.
Our method is open and version-stamped, so your Score holds up over time and against anyone else's.
Use your search tools for
Day-to-day search operations, keyword tracking, and site health, the work platforms like Semrush and Ahrefs are built for.
Use Primacy Score when someone asks
“Does AI actually recommend us, and if not, why?” The one question a search dashboard can’t answer cleanly. Primacy Score can.
The Invisible Shortlist
A buyer with a problem no longer starts with ten blue links. They ask an AI assistant, get two or three names, and decide who to call. That answer is the market, as far as they know.
The answers change run to run.
Ask the same question three times and you can get three different lists. One check tells you almost nothing.
Every engine judges you separately.
Strong in ChatGPT and absent in Gemini means a share of your market simply never sees you, depending on nothing but which assistant the buyer opened.
Your dashboard cannot see any of it.
Traffic, form fills, and lead scores measure the buyers who reached you. They say nothing about the ones routed elsewhere by an answer you never saw.
And when you’re absent from the answer, someone else is in it. Primacy Score shows you who AI recommends instead of you, by name, by question, by engine.
You cannot manage what you cannot measure. Until now, nobody was measuring this honestly.
The Instrument
Primacy Score runs the questions your buyers actually ask, across all six major AI engines, and scores what comes back, sampled repeatedly, because these systems do not answer the same way twice, and the report states the depth used on each engine. The Score you track is built on two reads that are never blended into each other, so it always tells you where to work.
Leads the report · Discovery Score
The search a new customer runs.
Scored only on questions that don’t name you, the unbranded and competitor-comparison prompts where a stranger actually shops. All six components below feed it. This is the number that measures new demand: can you get into the room when nobody says your name?
Reported beside it · Reputation Score
Once AI knows you.
Scored on questions that already name you, from the quality components only: citation, prominence, and sentiment. Appearing when your own name is in the prompt earns nothing, it’s the floor, not an achievement, and failing to appear is flagged as a red-line finding, never averaged away.
All six components feed Discovery. Reputation reads citation, prominence, and sentiment (reweighted 40/40/20) on the questions that name you. Your certified Primacy Score reports with both reads beside it.
Yesterday this was invisible. Today it is a Score you can manage.
Methodology
A score is only as good as its method, so ours is public.
Every
line
you sell is an answer surface, and your question set covers all of them. Sized to your category, product lines, and competitors during setup.
All
8
engines on every certified Score: ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews, Google AI Mode, Copilot Search, and Grok.
Repeated
sampling on every question, because one answer is an anecdote and a rate is a measurement.
Question count is determined by coverage, not by a package. Each product line or service you sell is an answer surface, and each surface needs questions across the awareness, comparison, and branded stages to be measured honestly. Your set follows your business rather than a fixed number, and is set and locked with you during setup. Certification asks whether the complete set ran, not how large it was. Larger catalogs add expansion packs. Setup recommends your number and shows you why.
And what we ask is published too, not just how often. Depth, margin of error and methodology version are all disclosed, but the first thing a sharp challenger asks is not how many times you asked. It is what you asked. Your set is designed against a published framework, locked with you and versioned, and every cycle states on its own report whether its configuration was frozen and can be re-derived. A tight range around a badly chosen question set is a precise wrong answer, and nothing else in this category discloses its inputs at all.
It is the difference between “we asked ChatGPT and we looked fine” and a number you can put in front of a board, an investor, or a skeptical sales leader.
Who it’s for
Private equity & portfolio operators
One method, every portfolio company.
A portfolio-wide read on a channel none of your portcos are measuring, comparable across companies because the method is fixed rather than reinvented per engagement. It is a diligence and value-creation signal, and it is cheap relative to the decisions it informs. The purchase is private by nature: no badge, no public tell, no positioning cost.
Agencies & consultancies
Build your own for operations. Buy Primacy Score for proof.
You can build a mention tracker, and for steering your own optimization work it is the right tool. But an agency reporting on its own performance with its own instrument is marking its own homework, and at renewal it is worth nothing. Primacy Score is the measurement you put in front of a client when the question is whether the retainer moved anything. You already know the instinct: you don't self-report search rankings either.
Marketing leadership
Walk in with the answer, and the trend.
Your board will eventually ask whether AI recommends you. Walk in with the answer instead of a guess, and track it quarter over quarter like any KPI that matters.
Practices & local business
See whether it says your name.
Your next patient or client asked AI before they called anyone. See whether it says your name or the competitor's across town, on the questions people actually ask before they book.
Brands & D2C
AI answers are the new shelf.
See how much of it you own, which products get recommended by name, and who's taking the recommendation when it isn't you.
Running this across a portfolio or a client book? Start a conversation. Volume engagements are handled directly, with one report standard across every property.
Sizing
Coverage is determined by what you sell, not by a package. Give us your domain and we’ll read your site, count your answer surfaces, and show you the math.
Question set sizing
No email. We read your site and show the math.
Pricing
One certified measurement, and a subscription that keeps it current. Extended coverage and extra questions are add-ons on a Score, so the certified number always means the same thing.
Your set follows what you sell: every product line or service is an answer surface. Run the sizing tool below and it picks your number, or set it here.
Primacy Score
$749 one-time
Credits in full toward your first year of Standing if you subscribe within 90 days.
Continuous certification
Primacy Standing
$349 per month
Your plan runs through the end of its term. Turn off renewal whenever you like and everything stays active until then. No auto-renew surprises.
A certified Score is certified because it achieved a stated precision, not because a number of samples was bought. That is what makes it hold up in the room where it matters, a diligence process, a board review, an agency renewal, and those rooms are private, which is the point. Nobody displays this. They use it.
Where the engines disagree with each other, we deepen the sampling on exactly those questions at our own cost until the disagreement resolves. In the rare case one stays contested, we say so rather than certifying around it: you receive your Score with its full stated range and a credit toward a deeper re-run.
Both doors read the same locked question set, so you never pay for setup twice and every measurement is comparable to the last. Start with a Score and it credits in full toward a subscription within 90 days.
Running several brands?
Agencies and private equity portfolios measure every property on the same instrument, so a number means the same thing in company A as it does in company B. That is what twelve separate vendors cannot give you. Volume rates, one report standard, and a single view across the portfolio.
Talk to us about portfolio measurementThe two halves of your visibility, and they mean different things. Discovery is measured only on questions that don't name you: can AI find you when a new buyer shops? Reputation is measured on the questions that do: are you cited, framed well, and recommended once you're in the room? A brand can be loved but unfindable, or found but poorly spoken of, and a single average would hide both stories. So your certified Primacy Score always reports with the two reads beside it, never blended into each other: the Score gives you the trackable headline, the reads tell you exactly where to work.
Because the measurement is real. Every reading asks your buyer questions repeatedly across six engines, which costs actual compute. A free version would have to guess, and a guess defeats the purpose of a score.
Then you found out before your competitors did, and before your board asked. A low score with a clear gap map is the start of a fixable plan. Not knowing is the only unmanageable outcome.
Ask the same question three times and you can get three different answers. A spot check is one roll of the dice. Primacy Score asks every question repeatedly across all six engines and turns the noise into a rate, with the uncertainty that remains printed beside it.
Every certified Score measures all eight: the Core Six (ChatGPT, Gemini, Claude, Perplexity, Google AI Overviews, Google AI Mode) plus two extension surfaces, Copilot Search for B2B and enterprise categories and Grok for consumer and practitioner-influenced categories. The extension surfaces are reported beside your Score and never blended into it, so the certified number means the same thing for every brand that holds one. The roster is governed by published criteria and reviewed on the methodology page.
Your question set follows your business rather than a fixed number: each product line or service you sell is an answer surface, and the set covers every one of them. Certification asks whether the complete set ran, not how large it was. Each product line or service is an answer surface, covered across the awareness, comparison, and branded stages. Catalogs with many lines add expansion packs of 30, up to 300 questions in all; beyond that we quote the engagement rather than sell it self-serve. Setup recommends your number and shows the reasoning.
Yes, that is half the point. Competitors appear in the same AI answers you are measured against, so every Score includes your position against 5 named rivals: which buyer questions they win, which you win, and who owns the recommendation space on each engine. Monitoring plans, coming soon, will track how that position moves.
No. There are two ways to buy: a Primacy Score on its own, a point-in-time certificate you can put in a board paper or an RFP response; or Primacy, at $349 a month on a twelve-month term, which includes the certified Score and recertifies it quarterly, so every year delivers four certified Scores measured on the same locked question set. If you start with the Score and subscribe within 90 days, the $749 credits in full toward your first year. Agencies and portfolios: talk to us.
Yes. Every Score starts from your category, your competitors, and your applications, and you can tune the question set during setup before anything runs. Agencies and portfolios: per-client tuning at volume is available. Talk to us.
Sample Report
Download a complete sample report: the headline Score, the Discovery and Reputation reads behind it, the six components, the per-tool breakdown, the competitor comparison, and the ranked action plan. Built on a fictional company, so it is exactly what you get without naming a real client.
Email required. We’ll send a few short notes to help you read it. Unsubscribe anytime.
The instrument pointed at it
A growing share of your demand now forms in AI answers you can’t see. Primacy Score is the instrument pointed at it.