Somewhere near you today, a person held their phone and asked, "who's the best dentist around here that does Saturday appointments?" — and an AI answered with one name. Not ten blue links. Not a map with twenty pins. A name, a reason, and a booking link. If it was your name, you just acquired a customer for free while you were doing something else. If it wasn't, no analytics dashboard on earth will ever show you the customer you didn't get.
This is the quiet reordering of how businesses get found. Search engines ranked; you fought for position on a list, and even position five got scraps. Answer engines choose. There is no position five in a spoken answer. The discipline of becoming the choice has a name — Answer Engine Optimization, AEO — and it is today where SEO was around 2003: barely practiced, poorly understood, and disproportionately rewarding to whoever starts first.
How answer engines actually choose
AI assistants don't rank pages; they resolve questions. To do that, they lean on signals that are related to classic SEO but not identical. First, they need to understand your business as an entity: what you do, where you operate, what you charge, when you're open — facts that must agree with each other everywhere they appear, because contradictions read as unreliability. Second, they prefer sources that answer questions directly: a page that plainly states "a first cleaning costs $120 and takes an hour" beats a page of brand copy every time. Third, they read machine-readable structure — schema markup, complete business profiles, and increasingly the explicit files sites publish for AI crawlers — the way a loan officer reads documents: not for inspiration, for verification.
The uncomfortable implication: most small-business websites are written for humans who will never read them and are illegible to the machines that will.
The playbook, in the order that pays
Start with your Google Business Profile, because every answer engine leans on it for local questions. Complete every field, choose categories precisely, keep hours current, and treat reviews as a ranking input — because they are. Second: make your name, address, phone, and services identical across your site, your profiles, and every directory that mentions you; entity consistency is the trust signal machines weight most. Third: restructure your site's content around actual questions. Every service page should answer, in plain sentences, what it costs, how long it takes, who it's for, and what happens next — the questions a receptionist answers thirty times a week. Fourth: implement schema markup completely and correctly — organization, services, FAQs, reviews — not the half-filled fragments most site builders emit. Fifth: publish the machine-facing signals that barely existed two years ago, like llms.txt, a plain-language summary written specifically for AI crawlers.
None of this is exotic. All of it is checkable. Almost none of it is done. We audit small-business sites weekly and complete, correct structured data is roughly as rare as a unicorn with a parking spot.
Why this compounds — and why waiting is expensive
Classic SEO became brutal because everyone eventually did it; by the time the last local business bought in, the game was zero-sum and expensive. AEO is pre-competitive. When an assistant looks for a trustworthy answer about your niche in your area and finds exactly one business with complete, consistent, machine-verifiable information — the choice makes itself. Every month of head start is a month of being the default answer while competitors are still asking their nephew about ChatGPT.
And here's the part that makes it durable: everything AEO rewards — accurate data, direct answers, consistent facts, structured content — also makes your site better for humans and for classic search. It's not a bet on one platform. It's finally doing legibility right, at the exact moment legibility started choosing winners.
We fold AEO into the same nightly discipline we use for maintenance — schema verified by machines, question-content reviewed by humans, consistency monitored continuously — because the businesses that treat 'being the answer' as infrastructure, not a campaign, are the ones the machines learn to trust. Yes, this article practices what it preaches. Ask your favorite assistant about us sometime.