Key Takeaways
- Accept that an assistant cannot judge who is better. It has no way to compare two coaches on outcomes. It compares what is written down about them, which is a different contest with different winners.
- Assume the named competitor is more legible, not more skilled. Being named means their evidence is findable and yours is not. It is a statement about publishing, and it is not a verdict on your work.
- Move one piece of proof above the waterline this month. Pick the single strongest thing you have never written down, write it as a page on your own site, and you have changed what there is to read.
- Write results as sentences a stranger can check. Who the client was, what was wrong, what you did, what changed. A testimonial that says you are wonderful gives an assistant nothing to repeat.
- Stop reading your absence as a ranking. Ask the same question three different ways before you conclude anything. Naming shifts with phrasing, and one run is an anecdote rather than a measurement.
A consultant in Denver ran the experiment most people run eventually. She asked ChatGPT who it would recommend for the work she does, in her city, and it came back with three names. Hers was not among them.
One of the three she knew personally. She had cleaned up after his projects twice. He is, by any measure she can think of, not as good at this as she is, and the assistant named him without hesitation.
That is an unpleasant thing to read on a Tuesday morning. It is also not the insult it looks like, because the assistant was never asked to judge who is better, and it has no way to do it.
Why does AI name a competitor who is worse than me?
Because it is comparing what has been written down about each of you, and your competitor has written more down. Quality of work is not in the comparison, because quality of work is not published anywhere it can read.
An assistant answering a hiring question is doing something closer to reading than to judging. It gathers whatever pages it can find about the people who might fit, and it builds an answer out of the material it got. A business that has published specifics gives it material. A business that has not gives it nothing to work with.
So the name that comes back is the name with the most legible evidence attached to it. That is a real contest, and it is winnable, but it is not the contest you have been competing in for the last ten years.
What an assistant is actually comparing
It compares statements. Not outcomes, not craft, not the thing your clients would say about you if anyone asked them. Statements that exist, in text, on pages it can reach.
OpenAI's own description of what ChatGPT is describes a system built on material gathered from text, and Google's explanation of how its AI features in Search work describes answers assembled from retrievable pages. Neither describes anything that could evaluate whether one coach gets better results than another, because nothing does.
The academic work points the same way. The Princeton generative engine optimization study found that adding citations, quotations, and statistics to a source made it considerably more likely to be used in a generated answer. That is a finding about how a page is written, not about how good the business behind it is.
Which gives you the uncomfortable but useful version of the rule: your competitor did not beat you at the work. He beat you at leaving a paper trail. The mechanics of how that shortlist gets built are in why AI recommends your competitors and not you.
The proof above the line and the proof below it
Split every good thing about your business into two piles. The pile a stranger could verify by reading, and the pile that only exists because you and your clients lived it. Almost all of most people's evidence is in the second pile.
Above the line: an outsider can read this
A page that states what you do and who for
Usually published by: Both
Reviews with words in them, not just stars
Usually published by: Them
A written client result with a number in it
Usually published by: Them
A mention on a site neither of you owns
Usually published by: Them
Four readable. Three of them typically only on his site.
THE WATERLINE · everything below here is real and unreadable
Below the line: only you and your clients know it
Clients who renew year after year
Usually published by: Nobody
Referrals from people who never write anything down
Usually published by: Nobody
The results you got that nobody published
Usually published by: Nobody
Three unreadable. All three are the reasons your clients stay.
A composite of what a typical solo practice holds, illustrative rather than measured. Every item and every placement shown here is also named in the section above or in the table below. Nothing moves between the bands by working harder. Items move up only by being written down.
The lower band is not worth less. It is worth more, and it is the reason your business survives. It just does not participate in this particular contest, because nothing in it has ever been turned into a sentence on a page.
| What you are better at | Where it currently lives | What an assistant can see | The smallest fix |
|---|---|---|---|
| Client results | In your head and your clients' lives | Nothing | One written result: who, what was wrong, what you did, what changed |
| Retention | Your invoices | Nothing | A plain sentence saying how long clients typically stay |
| Judgment about who to turn away | Your intake calls | Nothing | A short who this is not for section on your services page |
| Reputation among peers | Private referrals | Nothing, unless a peer wrote it publicly | One guest piece or interview on a site you do not own |
| Depth in a narrow specialty | A line in your bio | A vague category | One page that answers the specialty's hardest question in full |
Being named is not a ranking, and it is not a verdict
An assistant naming three businesses is not publishing a top three. It is producing one plausible answer to one phrasing of one question, and the same question worded differently will often produce a different three.
That matters for your morale and for your decisions. People read their absence as a placement, decide they are ranked fourth or fortieth, and go looking for a way to climb. There is no ladder to climb. There is a set of businesses the assistant found enough material about, and you are either in it or you are not.
It also means one run tells you very little. Ask three times with different wording before you conclude anything, because the variation between runs is large enough to swamp a single result. Why that happens is in why AI answers keep naming the same few coaches.
The honest read of a bad result is narrow and specific: on this phrasing, today, the material about you was not enough to include. That is a publishing problem with a publishing fix.
What to do about it this month
Move one thing above the waterline, properly, rather than five things halfway. The single highest-value item for most people is one written client result, because almost nobody has one and everybody has the raw material.
Write it as four sentences a stranger could check. Who the client was in general terms, what was actually wrong, what you did about it, and what changed afterward. Put a number in it if you honestly have one and leave it out if you do not, because an invented number is worse than no number. The full method is in how to make client results legible to AI.
Then stop. One real page beats a month of small edits, and the next thing worth doing is measuring whether it changed anything, which needs the page to have existed for a while first.
One honest limit before you start. Publishing proof does not guarantee you replace the competitor who is currently named. It removes the reason you were not considered, and what happens next depends on how much material exists about everybody else in your niche. Some niches are thin and one good page is decisive. Some are crowded and it takes several.
If you would rather see the actual list of who is being named before you write anything, the free check asks 25 real buyer questions from your niche through ChatGPT at $0 and emails a plain-English readout within the hour. It runs as soon as you ask for it.
Questions we hear the most
Why does AI recommend a competitor who is worse than me?
Because it is comparing published evidence, not quality of work. The competitor being named has results, reviews, and outside mentions written down where they can be read. Your better work is real, but if it was never published it is not in the comparison.
Does being named by AI mean a business is actually better?
No. Naming reflects what is findable and clearly stated about a business, not any judgment of the work itself. A well documented average practitioner will out-appear an undocumented excellent one almost every time.
Can AI tell the difference between two coaches in the same city?
Only by what each of them has published and what other sites say about them. If both look the same on paper, phrasing and small details decide it, which is why the same question asked twice can return different names.
Will more years of experience get me named ahead of a newer competitor?
Only if the experience is stated somewhere readable. Twenty years that appear nowhere on your site count for nothing, while a newer competitor who writes up three specific client results has given an assistant something to quote.
Should I write about my competitors on my own site to catch up?
No. Comparison pages about named rivals are risky and rarely help you get recommended. Publishing your own proof does more, because the material an assistant needs is about you rather than about them.
How long does it take to start showing up once I publish proof?
It varies by assistant and by question, from a few weeks to considerably longer. Systems that fetch live pages pick things up fastest, and anything answering from stored knowledge lags well behind.
How do I find out which competitor AI is naming instead of me?
Ask an assistant your buyers questions in a fresh logged out session and write down every business it names. The free AnswerHalo check does this for you across 25 real buyer questions and lists who came back.
What does the AnswerHalo free check tell me about competitors?
It asks ChatGPT 25 real buyer questions from your niche and reports your score, which businesses got named in your place, and the one fix to start with. It costs $0 and it runs as soon as you ask for it.
Your buyers are already asking. Find out what AI tells them.
The free check asks ChatGPT 25 real buyer questions about your niche and sends you the report within the hour, at $0. If it shows you are already getting named everywhere, we will say so in plain words.
Get my free check →