What an AI Assistant Says About You.
A growing share of buyers no longer open a search box and scroll a list. They ask an assistant a plain question — and the business it names by name gets the call.
Someone needs an electrician. A few years ago they typed “electrician near me” into Google, got a list, and sorted through it themselves. Now, increasingly, they open ChatGPT, Perplexity, or the AI answer that sits at the top of a Google search, and they ask it the way they'd ask a mate: “Who's a good electrician near me, and do they actually turn up?” The assistant doesn't hand back ten options. It names one, or a small handful — and the business it names is the business that gets the call.
That is a quiet but real shift in how you get found. For twenty years the goal was to rank — to appear in a list of results the buyer then judged for themselves. A new layer now sits above that list: being the business the assistant recommends before the buyer ever sees a list at all. Different mechanics, different work, and most businesses have no idea what an assistant currently says about them.
From a list to an answer
Two terms, defined plainly, because the difference matters. SEO — search engine optimisation — is the work of being found in a list of search results, so that when someone searches, your business appears among the options they scroll through. AEO — answer engine optimisation — is the work of being the business an AI recommends by name when someone asks it a question directly. An “answer engine” is any tool that replies with a recommendation rather than a list: the AI behind ChatGPT and Perplexity, and the AI summary Google now shows above its blue links.
SEO still matters — the list hasn't gone anywhere, and plenty of buyers still use it. But when a person asks an assistant “who should I use?” and acts on the single name it gives back, the businesses further down the list were never in the running. They were never mentioned. Being findable is no longer the same thing as being recommended.
How an answer engine decides who to name
An assistant doesn't know you personally. It assembles a recommendation from what it can read about you across the public web, and it leans on a few things in particular. None of them are tricks. They're just the ingredients of a business that's easy to describe with confidence.
Consistent public information
Your name, category, location, and what you do, stated the same way everywhere the assistant can find them. If your website, your Google Business Profile, and the directories you're listed in all agree, the assistant can state it back with confidence. If they contradict each other, it hedges — or names someone clearer instead.
Reviews and what other people say
Corroboration that you exist, that you're active, and that the work is decent. An assistant reads reviews the way a person skims them: not one perfect five-star line, but a consistent picture across enough sources to trust. A business with a steady body of genuine reviews is far easier to recommend than one with almost none.
A clear description in your own words
Your own website, saying plainly what you do and who for — not buried in a slideshow or a photo with no words around it, but written out as readable text. This is the account the assistant quotes from first. If your site doesn't say it in plain language, the assistant has to guess, and it guesses conservatively.
Third-party corroboration
Directories, industry associations, local listings, and mentions elsewhere that agree with your own account. When outside sources describe you the same way you describe yourself, the assistant treats the claim as settled rather than as marketing. Agreement between you and everyone else is what turns a claim into a fact it will repeat.
Why being everything to everyone makes you invisible
Here is the part most businesses get exactly backwards. When someone asks an assistant “who's a good X near me,” the assistant is trying to match a clear category to a clear answer. It needs to slot you somewhere. A business that says plainly “we do bathroom renovations for older homes in this city” is easy to place, and easy to name when that question comes up. A business whose website promises “solutions for all your needs” and lists a dozen services with equal weight can't be placed anywhere in particular — so it's named for nothing in particular.
Vague, everything-to-everyone positioning feels safe to a human because it closes off no possibility. To a machine that needs a clear category, it reads as noise. The instinct to stay broad so you don't rule anyone out is the very thing that stops the assistant from recommending you to anyone. Clarity isn't a narrowing. It's the thing that lets you be named at all.
“An answer engine can only recommend a business it can describe in one clear sentence. If you can't say what you do and who for, it won't say it for you.”
What to actually do
Twenty minutes, and no tools you don't already have. Open three assistants — ChatGPT, Perplexity, and whatever AI answer Google shows you — and ask each of them two questions. First, the question a real buyer would ask: “Who's a good [your trade or service] in [your town or suburb]?” Second, the direct one: “What do you know about [your business name]?” Then read the answers the way a prospect would.
Note three things. Whether you're named at all when the category question is asked. How you're described when your name is put in directly. And whether that description is actually correct — the right services, the right area, the right sense of who you're for. Most owners doing this for the first time find one of three gaps: they're not mentioned, they're described vaguely, or they're described wrongly. Each gap has a plain fix.
If you're not mentioned, the usual cause is that nothing about you is clear or corroborated enough to quote — so make your own website say, in plain readable text, exactly what you do and who for. If you're described vaguely, tighten that same language to one clear category instead of a list of everything. If you're described wrongly, hunt down the inconsistency: make your name, category, address, and phone number identical across your website, your Google Business Profile, and every directory you're listed in, and keep your reviews current rather than years stale. Then wait, and check again in a few weeks — these sources don't update instantly.
One note on where you trade: the directories and local listings an assistant leans on differ a little between Australia and New Zealand, so when you audit your corroboration, check the ones that actually cover your market rather than assuming a single list works across the ditch. The principle is identical either side — the specific sources are not.
None of this is a growth hack, and none of it is about gaming a machine. An assistant recommends the business it can describe most clearly and confirm most consistently. Being that business is the whole job — the same discipline that makes you easy for a human to choose makes you easy for an assistant to name. Be the clearest, most consistent answer, and you become the one it gives.
Before you spend on being found
What does an AI say about you right now?
If the assistants either can't find you or describe you wrongly, the cause usually sits upstream — in how clearly the business defines what it does and who for. The discovery call is where we look at what the answer engines are saying today and what's creating the gap.