Key Takeaways
- Describe the situation, not the service. Buyers type the mess they are in. A page named after your offering cannot match a question phrased as a problem.
- Say who it is not for. Exclusions read as confidence to a human and as precision to a machine, and they are the fastest way to stop matching the wrong questions.
- Put the qualifiers in sentences. Industry, location, format, and stage have to be readable as text. A logo wall and a pricing graphic carry none of it.
- Use one label consistently everywhere. Three different self-descriptions across your site, your profiles, and your listings read as three weak signals instead of one strong one.
- Check where you actually landed. Ask an assistant to describe your business back to you. The category it picks is the one it will use when a buyer asks.
Ask an assistant what your business does. Whatever it says back is the category it has filed you under, and that filing decides which questions you are eligible to answer.
Most of the time the filing is not wrong so much as lazy. It is the broadest true thing that could be said about you, because the broadest true thing is what your own pages say.
This is how to write a description that lands you somewhere useful, how far to narrow it before it starts costing you, and how to check where you ended up.
How do I describe my business so AI understands it?
Write the situation your client is in, who they are, and what changes, using the words they would use. That is the description an assistant can match to a real question, and it is almost never the description on a services page.
Google's guidance on creating helpful content keeps returning to a single test: does the page demonstrate first-hand knowledge and answer a real question completely. A category label is not an answer to anything, which is why a page built around one has so little to offer a system trying to assemble a reply.
The practical version is short. Stop writing what you are. Start writing when someone should call you.
Why a broad description leaves you unmatched
Because a broad description is true of thousands of people, and an assistant picking three names has no reason to pick you from a crowd of identical entries. Vagueness does not make you eligible for more questions. It makes you the wrong shape for all of them.
This is the part that feels backwards. Every instinct says a wider net catches more. In a matching system it works the other way, because the system is not choosing the most capable provider. It is choosing the one the available text most clearly describes as fitting the question asked.
- Coach Too broad to match anything
- Business coach Thousands of others sit here Where vague wording parks you
- Coach for founders of small firms Now a real audience
- Coach for founders managing their first managers A situation, not a service
- Coach for first-time managers at Denver nonprofits Where the question actually lands Where the question lands
Each step down is one more fact stated in plain words. Nothing about the business changed between the top of this list and the bottom, only how much of it was written somewhere readable. The example is illustrative, drawn to show the shape of the descent rather than to report a result.
Note that the bottom entry still contains the top one. Someone asking the broad question can still reach a narrowly described business, because the broad words are inside the narrow description. The reverse never works.
Finding the right altitude for your description
Go as narrow as you can while staying honest about work you would happily take tomorrow. That line is usually two or three steps below where people stop, and the fear of narrowing costs far more than the narrowing does.
A useful test: read your description and ask whether a stranger could tell you apart from your three closest competitors. If they could not, you have not described yourself, you have described your industry. The same problem seen from the buyer's side is in why ChatGPT cannot find your business.
| How it is written | What an assistant can file it under | Questions it can match | What it cannot do |
|---|---|---|---|
| "Executive coach" | A very large general bucket | Only the broadest category questions | Survive a shortlist of three |
| "Executive coach with 20 years of experience" | The same bucket, with a credential | The same questions | Distinguish you, since everyone says it |
| "Leadership coach for nonprofit directors" | An audience | Questions naming that audience | Match a question about a specific moment |
| "Coach for directors stepping into their first CEO role" | A situation | Questions describing that situation | Match on place, which many buyers include |
| "Coach for first-time nonprofit CEOs in Denver" | A situation and a place | Broad and narrow questions both | Nothing you would actually miss |
The last row matches everything the first row matched, and more. That is the whole argument for going narrow, and it is why the perceived tradeoff mostly is not real.
The five things that have to be stated in words
Who you work with, what situation they are in, what you actually do, where you do it, and what changes afterward. Each has to appear as a sentence, because an assistant cannot read a layout, a logo, or an implication.
That last constraint catches a lot of good websites. Design carries meaning to people and nothing at all to a machine. A row of client logos says "enterprise software" to any human who looks at it, and says nothing whatsoever to a system reading the page as text. If it matters, write it.
Credentials sit in the same category. They help, but only once they are stated somewhere readable rather than displayed as a badge image, which is the distinction drawn out in what credentials AI assistants mention. Structured details are worth adding too, since the shared vocabulary that sites use to state facts about a business, published at schema.org, exists precisely so machines can read them without guessing.
How to check which category you actually landed in
Ask an assistant to describe your business, then ask it who your business is for. Do it several times across a few days, and write down the words it reaches for. Those words are your current category, whatever your website intended.
Read the answers for two things. First, how specific they are: a reply that stays at "business consulting" means nothing narrower was available to it. Second, whether they are current, because assistants often describe the work you were doing three years ago if that version of you is better documented.
Do the same check on a tool that shows its sources. Google's notes on AI features in Search describe answers that carry links to the pages behind them, and those links tell you which of your pages, or whose, produced the category you were given.
Then rewrite one page rather than the whole site. Pick the situation you most want to be called about, write the page that answers it completely, and use the same description on every profile you control. Repeating one clear sentence in five places beats five different clever ones. The anatomy of that page is in the one page AI quotes about your niche.
If you would rather see the gap measured than guess at it, our free check asks ChatGPT 25 real buyer questions from your niche at $0 and emails a plain-English readout within the hour, including the questions you are missing entirely. It runs as soon as you ask for it.
Questions we hear the most
How do I describe my business so AI understands it?
Describe the situation your client is in, who they are, and what changes, in the words they would use. An assistant matches a question to text, so the description has to read like an answer to that question.
Why does AI put my business in the wrong category?
Usually because your own wording is the most general thing available about you. If your site says you are a consultant and nothing narrower, the general bucket is the only place there is to file you.
Is it risky to describe myself too narrowly?
Less risky than most people expect. A narrow description still matches broad questions, because the broad words are inside it. A broad description cannot match a narrow question, because the specifics are simply absent.
Do I need different wording for different AI tools?
No. One clear description, stated the same way everywhere, works across all of them. What varies between tools is how quickly they notice a change, not what they are looking for.
Should my job title change, or just my website copy?
Start with the copy. The sentence that describes who you help and when they call you does the work, and it is much easier to change on every page and profile than a title you have built a reputation on.
How do I know which category an assistant currently puts me in?
Ask it. Type your business name and ask what the business does and who it serves, several times across a few days. The category it reaches for, and how confident it sounds, is your starting point.
How does the AnswerHalo free check help with this?
It shows you which buyer questions you match and which you miss, which is the practical version of this question. It runs 25 real questions from your niche at $0 with a readout within the hour. It runs as soon as you ask for it.
What does the paid audit add beyond the free check?
The $500 audit runs 125 checks across five AI tools and arrives by this time tomorrow. It names the competitors taking the answers you want, the sources behind them, and a ranked list of wording fixes.
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 →