A year ago, someone in Burlington who needed a furnace fixed typed "furnace repair Burlington" into Google and picked from the map. A growing share of them now ask a question instead: "Who's a reliable furnace repair company in Burlington that can come today?" They ask it in Google, which may answer with an AI Overview. They ask it in ChatGPT or Gemini. And they get a short written answer that names two or three businesses.
If you run a local business, the question is simple. Are you one of the names?
How the answer gets built
An AI assistant does not have an opinion about your business. It builds its answer from sources it already trusts: your website, your Google Business Profile, review sites, local directories and local press. If those sources agree on who you are, what you do, where you work and what people say about you, the model has something to work with. Google has said its AI features rely on the same fundamentals as regular search, with no special optimization required, and our experience matches that. If they disagree, or say very little, it moves on to a competitor it can describe with confidence.
Nobody outside Google and OpenAI knows the exact recipe, and it changes. But every test we have run comes back to the same thing: being the business the sources agree on.
The three things that matter most
Consistent facts. Your name, address, phone number, hours and service list should be identical everywhere. One old listing with a previous address or a dropped service is enough to make a model hedge. This is boring work, and it is the single highest-return item on the list.
Pages that answer the actual question. Models quote pages that say plain things plainly. A service page that says "We repair gas and electric furnaces in Burlington, Oakville and Hamilton, with same-day appointments most weekdays" is quotable. A page that says "Comfort solutions for your home" is not. Every service and every city you serve deserves a page that states what you do there in the words a customer would use.
Reviews with specifics. "Great service" tells a model nothing. "Fixed our furnace the same day, explained the problem, fair invoice" tells it exactly when to recommend you. You can shape this by how you ask. After a job, ask the customer to mention what was done and how it went.
What about the technical side
Structured data (the code that labels your business, services and reviews so machines read them correctly), a fast mobile site and clean page titles still matter. They were SEO fundamentals before AI search and they are the foundation of it now. The good news for businesses that already did SEO properly is that most of this is the same work, done more carefully.
What we tell clients
Do not wait for this to settle. The businesses showing up in AI answers today are the ones whose details were already consistent and whose pages already read like answers. Start with the facts, then the pages, then the reviews. Give it a few months, and then search for yourself the way a customer would.