AI Visibility puts the questions your client's customers actually ask to six answer engines, on a schedule, and records every reply — whether the business was named, who else was, and which model said so on which day.
A paid add-on, per account, on top of the plan. Agency partner rates differ from list, so the portal shows the price for the account rather than this page guessing at it.
And for the first time in fifteen years, nobody in this industry can see through it.
For fifteen years the question a local business paid you to answer was “where do we come in the map pack?” It had a good answer. You could pull it, screenshot it, and argue about it with a number in your hand.
Some of that demand now goes somewhere else first. A homeowner opens ChatGPT, a property manager asks Perplexity, somebody types the query into Google and reads the paragraph above the links instead of the links. In every one of those cases a machine writes a short answer and names two or three businesses in it, and your client is either in that paragraph or they are not.
We are not going to quote you a percentage for how much of local search has moved. Nobody has one you could check, and the number people repeat changes with whoever is selling something. What you can check is what the engines say. That is a smaller claim and a far more useful one, and it is the only thing this add-on does.
Asking ChatGPT yourself proves almost nothing. Your session is personalised, your location is attached, your history is in it, and the answer you get at 9am is not the one your client's customer got last Tuesday. One look is an anecdote.
A fixed set of questions, put through the API rather than a logged-in session, on a schedule, with every reply kept. One observation is an anecdote; fifty of them on the same question is something you can put in front of a client.
Most owners have typed their own business into an AI at some point, and plenty have been unpleasantly surprised. This is the report for the thing they already looked up once at 11pm and then had nowhere to put.
Every reply is kept whole, and these are what is read out of it.
Named, not named, or could not check — three outcomes, not two. A provider outage or a refusal is recorded as its own thing and kept out of the denominator, because somebody else's bad afternoon is not your client's bad month.
First named, third named, last named. This is a position inside one paragraph of text — it is not a place in a result, and the product will not print it as one anywhere.
The rest of the paragraph, kept. Over a few months this is the part agencies find they use most: it is the only place you get to read who the engines put beside your client, in your client's own city, in their own words.
Engine and date on every single observation. Models change underneath you without announcing it, and a figure you cannot date is a figure you cannot defend six months later.
It is not a ranking, a score or a grade, and you will not find one anywhere in the product — not on the dashboard, not in the PDF, not in the monthly email. There is nothing to rank. An answer engine writes a fresh answer every time it is asked and writes a different one for different people, so a “position” would be a number we made up.
What you get instead is a citation rate: how many of the answers we recorded named the business, printed beside the two counts it came from, every time. A percentage that arrives without its numerator and denominator is a percentage somebody is hoping you will not examine.
Each question goes to every engine on the account's plan. Same questions, same week, same wording.
The answer Google writes above the blue links. The one a client is most likely to have already seen, and the one most likely to be seen without anybody meaning to look.
Answers with its sources attached, which makes what it drew the name from unusually easy to read.
The one a homeowner opens by name. When somebody says they asked an AI, this is usually the AI they asked.
Google's assistant, asked separately from the AI Overview — the same question can come back differently from the same company.
Used heavily at work, which is where a property manager, a GC or a facilities lead does their shortlisting.
On the list for the only reason any of them are on it: people ask it, so it is worth knowing what it says.
Which engines an account is asked is set by its plan and is the same for every account on that plan — so two clients on the same plan are directly comparable, and neither of them has a quietly different list behind their report.
A starting set is suggested for each client from the cities that business actually works in and what people search for before they find it — so the first draft is already about their trade and their towns.
Reword any of them. Switch any of them off. Add your own. This is the part worth an hour per client, because the questions are the report.
Every active question goes to the engines on the plan, weekly, through their APIs rather than a logged-in session. Nobody has to remember to run anything.
The first answers arrive on the next scheduled run, not the moment the add-on is switched on — worth saying to a client before they open the screen on day one looking for something.
Every reply is kept: named or not, where in that answer the name fell, who else was in it, and which model answered on which day.
It lands on a dashboard in the portal, and on the monthly scorecard the client already gets — with the sampling caveat printed inside the section rather than in a footnote.
Every agency in the pitch has a rankings slide. Almost none of them can say a word about what the answer engines are doing, and every owner has wondered. Turning up with twelve weeks of recorded answers is a different meeting.
It is an add-on against a single account, so you can put it on the three clients who will care, prove it for a quarter, and then have a much easier conversation about the other nine.
Partner rates apply where your agency has one. The portal shows the rate for each account rather than making you work it out.
Every answer keeps the other businesses it named. Read a quarter of those and you know which three competitors the models have decided are the obvious answer in that city — which is a list worth having whatever you do about it.
An answer counts only if the name in it matches the name on the account or one of its saved variants. Clients trade under more names than they think — the legal entity, the name on the trucks, the one people actually type.
Save all of them first. It takes ten minutes per client and it is the difference between a real measurement and a month of false misses.
Each figure is a sample: one question put once to one model through its API on the date shown. Answer engines reply differently to different people, in different sessions and as their models change, so this is evidence of what an engine said when it was asked — not a position, a ranking or a score.
That paragraph is printed in full on the dashboard, in the exported PDF and inside the monthly scorecard — not shortened, not moved to a footnote, not dropped once the numbers start looking good. It is on this page for the same reason: if the caveat is only comfortable to print when the result is bad, it was never a caveat.
Used the way it is meant to be used, that limitation is not much of one. You are watching a set of questions over time, under the same conditions, and asking whether the business is turning up in the answers more often than it did last quarter. That is a real, checkable thing. It is simply not a ranking, and we would rather sell you the real thing than a better-sounding one.
Tell us which clients you want it on and we will quote the rate for your book, set the name variants up with you, and get the first questions written.
