It’s 8pm on a Tuesday. Your hot water system has just blown. You pick up your phone and ask, “who’s the best emergency plumber near me right now.”
A year ago, you would have got ten blue links. You’d scroll the first three. Pick the one with the highest review average. Call.
Tonight, you don’t get ten links. You get one name and a phone number. Maybe a one-line reason why. You don’t scroll anything. You just call.
That’s the part of the internet most small business owners haven’t noticed yet. The list became the answer. The answer is one business, named.
* * *
The customer didn’t disappear. Their behaviour changed.
When ChatGPT, Claude, Perplexity, and Google’s AI Overview answer a question about a local business, they don’t return a directory page. They cite. They name a business by name, give a short reason, and stop. The customer never sees the second result. There isn’t a second result.
This is what Answer Engine Optimisation is. AEO is the work that decides which business gets named when a real person asks a real question of an AI assistant. It’s not a new layer on top of SEO. It’s a different game. SEO got you into the top ten. AEO gets you cited as the answer.
Being on page one used to be enough. Today, page one might never load.
It matters most for the businesses that already do good work. The cafe with the line out the door on a Saturday morning. The mechanic who’s been on the same street for twenty years. The chiropractor whose patients drive in from the next suburb. These are the businesses with the strongest reputation per dollar of marketing spend. They’re also the ones least likely to be cited by an AI assistant, because their digital presence was set up before any of this existed.
* * *
Why aren’t they the ones being named? Five reasons, usually.
The Google Business Profile is half done. 56% of profiles are incomplete or unoptimised (BrightLocal). Missing categories, missing hours, no description, no posts in the last year. The AI assistants read Google before they read anything else, so a half-finished profile is the first place a business loses.
The website has no schema markup. Schema is the invisible layer of structured data that tells an AI exactly what a page is about. Without it, the assistant has to guess. With it, the assistant has a clean answer to lift. Most small business sites have no schema at all, or have whatever default the platform added on day one and never updated.
Some businesses don’t have a website at all. They live on Facebook or Instagram. That’s fine for everyday updates and showing off the work, but AI assistants don’t read social pages the way they read websites. There’s no schema layer behind a Facebook post. The content is locked behind logins, infinite feeds, and platforms the assistant can’t reliably crawl. Without a real website, an AI assistant has nothing structured to lift, and a social-only business is effectively invisible to the answer.
The business details aren’t consistent. The address on the website doesn’t match the address on the Google profile. One says “Suite 4, 12 Main Street”, the other just “12 Main St”. The phone number on the website starts with 07, the one on Google starts with +61 7. The trading name on an old directory listing is the previous owner’s. AI assistants are trained to be cautious. When the signal is muddy, they won’t name you.
And the content doesn’t answer the question. Most small business websites describe what the business is. They don’t answer what customers ask. “About us” doesn’t help the AI when someone asks “who fixes leaking hot water systems after hours in Carrara.”
The result is what Varoo’s own research finds across thousands of audits: under 1% of Australian small businesses are cited by AI search assistants today (Varoo research).
* * *
The reassuring part: none of this is mysterious. AI assistants aren’t ranking businesses by some hidden taste. They’re reading the data that’s already public, deciding which business they can describe with confidence, and naming that one.
Fix the five things above and the chances of being cited rise sharply. It’s a structure problem, not a popularity problem. The mechanic who’s been on the same street for twenty years can absolutely be the name ChatGPT returns. A real website has to exist. The schema on it has to be in place. The Google profile has to be alive. The details have to match across the web. The content has to answer the question.
That’s not the kind of work most owners want to do on a Sunday afternoon. It also isn’t the kind of work that pays off in a single weekend. It’s structural, it compounds, and it has to be kept current as the AI assistants update their citation criteria.
* * *
The shift has happened. The customers asking AI assistants for local recommendations aren’t a future segment. They’re tonight’s customers. The plumber whose phone rang at 8pm wasn’t chosen because they had the best Google reviews. They were chosen because the assistant could read them and was confident enough to name them.
If your business only lives on Facebook or Instagram today, the first step isn’t to fix the structured data. It’s to have a website at all. That’s where the structured data goes. Without it, every other piece of advice above is academic.
If you want to see how your business reads to an AI assistant right now, tell us who you are and we’ll search Google and audit what we find. About two minutes.
That tells you where you stand. The work to fix it is a separate conversation.