Key Takeaways
- Publish claims that break when the name changes. If your sentence stays true with a competitor dropped into it, it was never a claim about your program. Specifics survive the swap and adjectives do not.
- Say the shape of the program in plain numbers. Length, format, cadence, group size, and price are the details an assistant can restate without inventing anything, and they are what a buyer asked for anyway.
- Put at least one claim somewhere you do not control. A credential on a public register, a review on a platform, a mention on a site that did not have to publish you. Self-attested and corroborated are read differently.
- Name who the program is not for. An exclusion is the fastest credibility signal you can write, because nobody padding a page ever adds one.
- Keep results specific, sourced, and honest. A described outcome with the conditions attached is usable. An unattributed transformation number reads as marketing to a buyer and gets skipped by an assistant weighing what to repeat.
Two coaching programs sit in front of an assistant. Both are good. One has a page full of warmth and transformation, the other says twelve weeks, two calls a week, eight people, built for nurses coming back from injury.
The second one gets named, and it is not because the coach is better. It is because there was something in the page an assistant could repeat without making anything up.
Here is what credibility means to a machine reading your program page, the ten-minute test that shows whether yours has any, and what to publish instead.
What makes an online coaching program credible to AI assistants?
Detail that is specific enough to be wrong. An assistant treats a claim as usable when repeating it would be a checkable statement about you rather than a pleasant sentence about coaching in general.
This is the same property search has been describing for years. Google's guidance on creating helpful, reliable content keeps returning to first-hand specifics and demonstrable experience, and warns against pages assembled to fill a slot rather than to say something. The wording is written for search, and the property it describes is what an assistant is looking for too.
Credibility, in this narrow sense, has nothing to do with how good you are. It is a question about your page, and it is answerable in ten minutes.
The swap test, and why most program pages fail it
Take any sentence from your program page and swap your name for a competitor's. If the sentence is still true, it was never about you.
Almost every line on almost every coaching page survives the swap. Personalized, results-driven, built around your goals, for busy professionals who want to feel their best. All true of you. All equally true of the other four hundred coaches in your city.
A sentence that fails the swap is doing work. Twelve weeks fails it. Eight people a cohort fails it. Shift-working nurses returning after injury fails it hard, which is why a narrow program is easier to get named for than a broad one, and a single specific client story outperforms a wall of testimonials.
The struck-through lines are the ones that stayed true after the swap. A sentence that survives being about somebody else was carrying no information about you, however well it reads. The sentences are written as illustrative examples, not quoted from any real program page.
What an assistant can actually check
Less than you would hope, and more than nothing. It cannot call your past clients or watch you coach. It can read what you published, read what other people published about you, and notice when the two agree.
That is why a credential works best when it exists on a register. Naming the body precisely turns your sentence into something a reader can go and confirm, which is the whole point, and the credentials assistants actually mention are consistently the ones with a public list behind them.
Results claims work the same way. The FTC's guidance for marketers on soliciting and paying for online reviews is worth reading once even if you have never paid for one, because the behaviors it treats as deceptive are the same behaviors that make a page read as untrustworthy to anything parsing it. Honest and legible turn out to be the same project.
Self-attested, corroborated, and independently verifiable
Three tiers, and most coaching pages live entirely in the first. Moving two or three claims up a tier does more than rewriting the whole page.
Google's overview of its AI features in Search is explicit that these surfaces are assembled from web content it can corroborate. Corroboration is a structural property of where a fact lives, not a quality of how well the fact is written.
| Tier | Example | Who published it | What it is worth |
|---|---|---|---|
| Self-attested | "Our twelve-week program is capped at eight people" | You, on your own site | Usable and quotable, as long as it is specific. The floor, not the ceiling |
| Corroborated | The same detail in a client's review, a podcast description, a partner's page | Someone else, without being obliged to | Higher. Two independent sources agreeing is the pattern that gets repeated |
| Independently verifiable | A certification on the issuing body's public register | An organization with no stake in your marketing | Highest. It stands whether or not your website exists |
| Unfalsifiable | "Personalized, results-driven coaching for busy professionals" | You, and four hundred others, in the same words | Nothing. It survives the swap test, so it carries no information |
What to publish this month
Rewrite the program page so the first screen answers four questions: how long, what format, who it is for, and what it costs. Those four are what a buyer asks an assistant, and a page that answers them can be quoted.
Then add the exclusion. One honest sentence about who should not join is worth more than three paragraphs about who should, and it is the line nobody writes to fill space.
Then move one claim up a tier. Pick the credential you already hold and check whether the issuing body lists you publicly. If it does, name the body exactly. If it does not, that is worth knowing, because the claim is currently resting entirely on your own say-so.
None of this is a rewrite of your business. It is the difference between a page an assistant can repeat and a page it has to skip, and how online coaches get recommended comes down to that difference more often than to anything about the coaching.
If you want to see which of your claims an assistant currently repeats, the free check runs 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
What makes an online coaching program credible to AI assistants?
Specific, checkable detail that would be false if it described anyone else. Format, length, cadence, group size, who it is for, and credentials that exist on a register somewhere the assistant can read them independently of your website.
Do testimonials on my own site help?
A little, and less than most people expect. A testimonial you host is something you chose to publish about yourself. The same story on a platform you do not control carries more weight because it did not have to appear.
Does having a certification make a program credible?
Only if the certification can be found somewhere other than your claim about it. Name the body precisely, and prefer bodies that publish a searchable register, because that is what turns your statement into something checkable.
Should I publish my prices?
Yes, if you want an assistant to answer questions about cost. A page with no figure cannot be quoted on cost, so the assistant either skips the question or reaches for someone who did publish one.
Do results numbers help or hurt?
They help when the conditions come with them and hurt when they float. A described outcome with the timeframe, the starting point, and the caveats reads as evidence. A bare percentage reads as advertising.
Does a bigger following make a program more credible to AI?
Not directly. Follower counts sit on platforms that assistants read inconsistently, and a large audience is not evidence about the program. Published, specific, corroborated detail beats reach on this particular job.
How does AnswerHalo tell whether a program reads as credible?
We ask ChatGPT 25 real buyer questions from your niche and record what it says about you unprompted, including which of your claims it repeats and which it ignores. What gets repeated is what came across as checkable.
Can I run the free check today?
Yes. It runs as soon as you ask for it, and the readout lands within the hour between 7am and 5pm Central. Running the swap test on your own program page takes about ten minutes and can happen now.
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 →