When Customers Ask AI About You
AI assistants don't show search results — they give an answer. How the new front door works, what being the answer depends on, and how to become recommendable.
Ask ChatGPT to recommend a family dentist in your town. Really — try it. Use your own profession, your own town.
You'll get something unsettling. Not a list of links. An answer. Two or three practices, named directly, each with a sentence of reasoning. "Well-reviewed for nervous patients." "Offers evening appointments." "Practising for over twenty years." A shortlist, pre-argued, delivered with the calm confidence of a well-informed friend.
Now the question that actually matters. Were you in it? And if you weren't — do you even know why?
Get Found by AI
The Distribution Audit
Is This Where You're Leaking?
You ask ChatGPT about your profession in your town, and you're not in the answer? That’s distribution leaking here — and rarely the only place.
Five minutes · scores your whole distribution
Cluster tool
The Get Found by AI Audit
Everything AI assistants read about you, scored with fixes ranked.
Shows whether AI names you
What actually changed
It's tempting to file this under "SEO, but newer." That misses what actually happened.
AI assistants aren't replacing search. They're replacing the act of evaluating search results. For twenty years, being findable meant appearing in a list of ten blue links that a human would then judge. Scanning. Comparing. Clicking. Forming an impression. That judging step was where your website, your reviews, and your reputation actually did their work.
The assistant now does the judging itself. It reads the pages. It weighs the reviews. It cross-references the mentions. Then it hands over a conclusion. The customer never sees the list. There is no page two of an answer. Either you're one of the names, or the conversation about who to hire happened entirely without you.
Put the two eras side by side, and the difference is one sentence. Search rewarded being listed. Assistants reward being safe to recommend.
The classic assistant query is already urgent — "plumber near me who can come out tonight." But it lands hardest on professional practices. The searches that matter most are exactly the high-stakes, research-heavy questions people now hand straight to an assistant. "Best assisted living near me." "Accountant for cross-border tax." "Family lawyer who does fixed fees."
What being the answer depends on
So what does being recommendable by AI actually rest on? Three things.
A site machines can actually read. Facts on open pages, no tricks. An off-domain record that tells one story. Reviews, listings, mentions that agree. And substance worth citing. Real answers to the questions your clients actually ask.
Nobody can guarantee a recommendation. These three things move the odds. The odds are all there is.
Two mechanisms turn those into an answer. What the model already knows about you, from its training. And what it looks up the moment someone actually asks. The first shifts slowly. The second responds to your work within weeks. That's where everything in this cluster acts.
Each leg has its own guide.
Your site, made legible. Assistants read like fast, literal-minded researchers. They need your what, who, where, and how-much, stated plainly, on pages that render without tricks. The good news: it's the same clarity that converts human visitors too. How AI decides who to recommend explains the machine. Making your site legible to AI is the checklist.
The story told about you elsewhere. Assistants are consensus machines. Before naming you, they cross-reference what the rest of the web says. Your reviews. Your Google profile. Your directory listings. The occasional mention. If those agree, you're a safe answer. If they contradict each other, or barely exist, you're a risk the assistant quietly skips. What AI reads about you elsewhere covers making the record agree.
Substance worth citing. When your pages genuinely answer the questions your clients ask, they become the exact material assistants quote — discovery you never pay per click for. This asset isn't AI-specific. It's the same compounding content that feeds ordinary search, covered in content that compounds over in Get Found. What's new is where it gets quoted.
And one more shift is close behind. Soon the assistant won't just recommend. It will act. Can an agent book you? looks at that wave — AI agents trying to enquire and book on a customer's behalf, and whether your site would even let them.
The honest part
Two things every practice owner deserves to hear plainly.
First. Nobody can guarantee an AI recommendation. Anyone selling "guaranteed AI rankings" is selling smoke. Assistants show sources, not rankings. You can see who they cited. Never why those names won. And the answers vary by phrasing, day, and user. What you can actually do is move every checkable factor in your favour, and verify the trend. The honest limits of AI visibility draws the line between real work and snake oil.
Second. You can check this yourself, today, for free. No tool, no consultant required. The check is the next step, at the bottom of this page.
What's in this cluster
Each guide takes one part of becoming the answer an assistant gives:
How AI decides who to recommend — assistants are consensus machines: what they read before naming a business, and why agreement beats cleverness.
Making your site legible to AI — assistants read like fast, literal researchers: the checkable list, plus the five-minute legibility test.
What AI reads about you elsewhere — your site is one witness. The assistant interviews several: reviews, listings, and mentions must tell one story.
Is AI recommending you? — the fifteen-minute monthly ritual: ask the assistants what your customers ask, log the answers, watch the trend.
Can an AI agent book you? — the next wave after recommendations: agents that enquire and book for a customer, and the honest timeline for caring.
The honest limits of AI visibility — nobody can guarantee an AI recommendation: what can and can't be promised, and the snake-oil patterns to recognise.
Questions practices actually ask
Is this actually happening, or is it hype? It's happening. A meaningful and growing share of buying research now runs through assistants, especially for high-stakes services. What is hype is anyone claiming they can control the answers.
Is this just SEO with a new name? They overlap at the foundations. Clear pages. Real answers. Consistent listings. They diverge at the endpoint. Search rewards being listed. Assistants reward being safe to recommend. The consensus about you matters far more than any keyword ever did.
Should I block AI crawlers from my site? For a business whose whole aim is to be known, mostly no. Blocking the crawlers that fetch and cite pages is opting out of AI answers entirely. There are three different kinds of AI visitor, and the nuances are covered in the cluster.
How long does this take? The site-side fixes are days of work. The off-domain consensus builds over months. Reviews. Listings. Mentions. Start the slow part first, and let it accrue while you fix the fast part.
Your next step
You already ran the query once, at the top of this page. The whole practice is just that, made regular. Fifteen minutes a month. Ask the assistants the questions your customers would ask, and log what comes back. Most of your competitors will never keep that log. Is AI recommending you? turns it into a ritual. And the ritual into a trend you can actually act on.
And when you want the instrumented version, that's exactly what the Get Found by AI Audit does. Every checkable factor scored. With the findings explained. Same check, with instruments.