Questions and answers
The questions people actually ask about AI and local business - answered from our own measurements
Every question below is one Google itself returned in a People Also Ask block on a search we ran, or one we put to an assistant ourselves and recorded the answer to. Every answer either carries a figure we measured or says plainly that we have not measured it.
Whether AI recommends you
How do I get ChatGPT to suggest my business?
Nobody can promise you that, and anyone who does is selling something they cannot deliver. What we can tell you is how the slot works. On local questions the assistant answers with a short business list, and across the categories we have measured that list comes from a pool of five or six businesses per question - four or five shown at a time.
The pool is small and it is not fixed. Of 253 names we saw across 45 repeated questions, 119 came back in every reading and 134 did not. So the honest goal is not “get into ChatGPT”. It is: be in the pool, and hold your place in it.
The one thing every named business had in common: of the 34 we could check against Google Maps, 34 run their own live site. Having a site is not the bar - having one an assistant can quote is.
Google returned this question in a People Also Ask block on 1 search we ran.
Is there a service that checks whether ChatGPT recommends my business?
We put this exact question to ChatGPT ourselves, in a business owner's voice, and recorded what came back. It does not deflect into advice: nine of the ten questions we asked returned a named shortlist of vendors with prices and a “best for” column.
Two things about that shortlist are worth knowing before you trust one. It is local - we ran the same ten questions framed in Austin and in Miami and roughly 3% of the names overlapped. And it moves: on the service-intent surface, two readings of one identical question ten minutes apart shared an average of 15% of their named companies across 19 readings.
So yes, such services exist, we are one of them, and the list you get depends heavily on where you are and when you asked.
A question we put to an assistant ourselves and recorded the answer to.
How do I find out if AI is recommending my competitors instead of me?
Ask the question your customers ask, in your city, and read who is named. That is exactly what our free scan does on one question, and what the report does across the full set for your trade.
One warning from our own data, because it is the mistake everyone makes: a single check is not an answer. 134 of the 253 names we saw did not survive a re-read of the same question. Ask once on the wrong morning and you will conclude the opposite of the truth.
A question we put to an assistant ourselves and recorded the answer to.
Who does AI visibility audits for small local businesses?
When we asked an assistant this, it named a different set of companies in each city we tried. Austin returned Austin firms, Miami returned Miami firms, and one name appeared in both.
That is the most useful thing we can tell you about choosing one: this is a local market as the assistant sees it, not a national one. Ask in your own metro and you will get your own answer - and if you want the measurement rather than the shortlist, that is what we sell.
A question we put to an assistant ourselves and recorded the answer to.
AI and search, honestly
Can ChatGPT do SEO?
This was the single most repeated question in everything Google returned to us - it appeared in the People Also Ask block on nine separate searches we ran.
It can write things for you. It cannot make a search engine rank you, and it is not what we do. We do not sell SEO, we do not audit your Google rankings, and we do not measure them - so we will not tell you what they are.
What we measure is different and narrower: whether an assistant names your business when someone asks it for one, and what it quotes when it does.
Google returned this question in a People Also Ask block on 9 separate searches we ran.
Is SEO dead now with AI?
No, and our own data argues against the question rather than for either answer. Google's AI Overview appeared on 23 of the 25 category searches we measured, so the AI layer is real. But the citations inside it are not the ranking: of 127 citations we recorded, 62 - 49% - came from domains that were not in the organic top ten at all.
That is the useful finding. Two different things are being decided, by two different mechanisms, on the same page. A ranking report cannot tell you about the second one.
Google returned this question in a People Also Ask block on 3 separate searches we ran.
Is GEO replacing SEO?
We tried to find the lever that decides AI citation and failed three times, on our own data, and the failures are more informative than a slogan.
Page structure does not explain it: across 112 cited pages and 128 uncited ones from the same searches, every signal - schema, FAQ markup, headings, list count, number density, length - was flat or worse on the cited side. Cited pages carried less schema, not more. Matching the wording of the cited sentence does not explain it either: 38% across 26 pre-registered comparisons, which is a coin flip. Nor does per-vertical site architecture - four in five companies in this market already build it.
So: anyone selling you a GEO checklist is selling the three things we measured and found to do nothing. What is left, and what we do, is measurement.
Google returned this question in a People Also Ask block on 1 search we ran.
Is Claude or ChatGPT better for lawyers?
Google surfaced this question inside a law-firm marketing search, which tells you the audience is asking it.
We have not measured a like-for-like comparison across assistants for any single trade, so we are not going to answer it. What we have measured is one assistant in depth, and when we run a report we record which surface each reading came from and never pool them - because they behave differently and averaging them would hide exactly what you are paying to see.
Google returned this question in a People Also Ask block on 1 search we ran.
Being invisible, and what it costs to fix
Why is my business not publicly visible?
This is the owner's own phrasing and it came up on three separate searches we ran, so it is worth answering plainly even though half of it is outside what we do.
The Google side of it - your Business Profile, verification, suspensions - we do not measure and will not diagnose. What we can tell you is that the same three inputs that decide Google visibility also decide whether an assistant names you: the Google card, the review filter, and what your own site gives a model to quote. On the last of those we have a number: of the 34 named businesses we could check, 34 run their own live site.
If your question is really “why does AI never mention me”, that is the one we measure, and the scan is free.
Google returned this question in a People Also Ask block on 3 separate searches we ran.
How long does it take for my business to show up on Google Search?
We do not know, and we are not going to guess.
We looked for an empirical anchor and there is none in our data: the publication-date field was empty on all 1,935 answers we collected, so we have no measured time-to-citation for any surface. Anyone quoting you a number for this is quoting a feeling.
What we do instead is measure at fixed intervals after a change and publish what comes back, whichever way it goes.
Google returned this question in a People Also Ask block on 2 separate searches we ran.
How much should I pay for local SEO?
We cannot tell you, because we do not sell local SEO and do not measure it. This question appeared on five of the searches we ran, so it clearly matters - but an honest answer from us has to start by saying it is not our market.
What we charge is on the page: a one-off report is $99 ($199 in the higher-ticket categories - the report is the one price that depends on your trade), monitoring is $99 a month or $79 a month with our badge in your footer, reworking your site is $499 ($399 with the badge) and a new site is $999 ($799 with the badge). Execution - building or reworking the site - is invoiced by request, not sold from a checkout.
The one piece of pricing advice our own data supports: whatever you buy, insist that it is repeated. 134 of 253 names did not survive a re-read of the same question, so a single measurement - ours or anyone's - describes a moment, not a position.
Google returned this question in a People Also Ask block on 5 separate searches we ran.
Where these questions come from
Every question on this page is one Google itself returned in a People Also Ask block on a search we ran - 161 were captured across those searches - or one we put to an assistant ourselves on our own gate rig. None of them was invented to fill a page.
We run this measurement for a living. The free scan is the same instrument on one question, and the report is the full set for your trade. If a figure on this page looks wrong to you, tell us - every one of them is recomputed from raw readings by a single script, and we will run the measurement again and publish what it returns.