Any business whose buyers use AI tools to research providers before making contact. In 2026 that covers most B2B firms, professional services, healthcare and financial services, and a growing share of consumer businesses. The test is not your sector. It is whether your buyers ask an assistant before they ask you.
The short version
- If your buyers research before they contact you, they are probably researching in an assistant.
- Considered purchases need it most. Impulse and pure-local purchases need it least.
- High-value, low-frequency, comparison-heavy buying is where AI shortlisting bites hardest.
- The honest test takes two minutes and is at the bottom of this page.
- Sector matters less than buying behaviour. Two firms in the same sector can have very different exposure.
Why some businesses need this and others do not
The useful question is not “is my industry affected”. It is how much research happens before a buyer contacts you, and where that research now takes place.
Buying that involves comparison, risk and a shortlist has moved furthest. A finance director choosing an accountancy firm, a general counsel picking outside advisers, a procurement lead sourcing a supplier: all of them build a shortlist before anyone in your business hears about it. That shortlist increasingly starts with a question put to an assistant, and it returns two or three named options rather than ten links.
Buying that involves little research has moved least. Nobody asks ChatGPT which corner shop to walk into. Emergency and proximity-driven purchases still run on maps and phone calls.
The middle ground is where most of the confusion sits. A local business with a genuine specialism is more exposed than a generalist, because the specialist query is exactly the kind an assistant answers well. “Emergency plumber near me” is a maps query. “Which UK plumbers are certified for unvented cylinder work in listed buildings” is an AI query, and it is worth considerably more.
There is also a second-order effect worth naming. Even when a buyer finds you conventionally, many now ask an assistant about you before making contact. If the answer it gives is thin, wrong or missing, that costs you a conversation you never knew you were in.
How to check it yourself
Two minutes, no tools:
- Open ChatGPT, Perplexity or Gemini.
- Ask the question your best-fit buyer would ask, in their words rather than yours. Not “GEO agency” but “who should I talk to about getting my law firm cited by AI”.
- Ask it a second time as “tell me about [your company name]”.
If the first question names competitors and not you, you are being excluded from shortlists. If the second returns something vague or wrong, you have an entity problem rather than a content problem. Either way you now have evidence rather than a hunch.
What to do about it
- Establish the baseline properly before spending anything. The free AI Visibility Score does the technical half in about a minute.
- Work out which queries you can realistically win. Specific beats broad, always. A specialism you genuinely own is worth more than a category term you do not.
- Fix the entity first. If a model cannot confidently say who you are, nothing else compounds.
- Only then invest in content and authority. Doing it in the other order is the most common way to waste a year.
If the answer to the test above is “we already appear, consistently, for the queries that matter”, then you do not need this yet. That is a legitimate outcome and worth knowing.