A year ago, someone looking for a local business opened Google, typed something like "best family dentist near me" or "emergency plumber in Toronto," and worked down a page of blue links. Today, a growing number of them open ChatGPT or Gemini instead and ask a full question: "Who's a good dentist in North York for a nervous patient who needs a root canal fast?" A few seconds later they have three names and a reason for each. They never scroll ten links. They get a shortlist of three, and if your business isn't on it, you were never in the running.
That shift is not a prediction. In BrightLocal's 2026 Local Consumer Review Survey, 45 percent of consumers said they had used an AI tool like ChatGPT, Gemini, or Perplexity to find a local business in the past year. The year before, that number was 6 percent. AI has quietly become the third most-used way people discover local businesses, ahead of Yelp and TripAdvisor. And they are not just experimenting: 42 percent said they trust an AI recommendation as much as they trust online reviews.
So there is one question every owner should be asking. When a potential customer describes their problem to an AI, does it ever say your name?
What this means: your website now has to do more than win over a human visitor. It has to be clear enough that a search engine or an AI can tell, on its own, when you are the right answer to a stranger's question.
There is a new layer on top of SEO, and it has a name
If "SEO" already felt like a black box you paid someone to manage, brace yourself, because there is now a layer sitting on top of it. The industry is still arguing about what to call it. Two terms have stuck. Answer Engine Optimization, or AEO, is the broad idea of showing up in any AI-generated answer. Generative Engine Optimization, or GEO, is the narrower goal of getting named and cited by chatbots like ChatGPT and Claude. You will also hear "AI SEO." They mostly mean the same thing.
Here is the reassuring part, and it comes straight from Google, not from an agency trying to sell you something. In its official guidance for AI search, Google writes that "the best practices for SEO continue to be relevant because our generative AI features on Google Search are rooted in our core Search ranking and quality systems." In plain terms: this is not a teardown of everything you have done. It is a new layer on the same foundation. If you have been doing the fundamentals well, you are already partway there.
Why this matters for your business, specifically
There are really two ways an AI can find you, and the gap between them is the whole story.
The first is by name. Ask ChatGPT about your business directly and it will usually say something flattering. This is the check most businesses pass, and it is reassuring, but it only helps the people who already know you exist.
The second is by need. The customers you actually want are the ones who do not know your name yet. They describe a situation ("I run a Mississauga bakery, I failed a health inspection, who can help me fix it?") and let the AI pick. This is the check most small businesses fail, and it is where new clients are quietly won or lost. You show up when someone already knows to look for you, and you vanish at the exact moment a stranger is deciding who to call. (If you have met the jargon before, the industry calls these branded and non-branded visibility. By name and by need is the same idea in plainer words.)
How AI discovery actually works (the part most articles skip)
Strip away the jargon and an AI answer is assembled from sources the machine can read, can trust, and can match to the exact question. Get those three right and you are in the running. Miss any one and you are invisible. It is worth walking through each, because this is where the real work lives, and where most of the advice you will read online quietly goes wrong.
Can it read you? An AI can only recommend what it can actually parse. Google's documentation on its AI features is blunt: "There are no additional requirements to appear in AI Overviews or AI Mode... You don't need to create new machine readable files, AI text files, or markup." What you do need is unglamorous. Pages that can be crawled, with your important information written out as real text rather than locked inside an image or a PDF, connected by clear internal links. If a machine cannot read it, a machine cannot repeat it.
For the technically minded. The three big AI systems each run their own web crawlers, and here is the trap: block the wrong one and you go invisible to that AI. Google just needs your pages crawlable in the normal way. Anthropic runs Claude-SearchBot (for search) and Claude-User (which fetches your page live when someone asks Claude about you), and its own documentation says disabling them "may reduce your site's visibility." OpenAI runs OAI-SearchBot, which it says is "used to surface websites in search results in ChatGPT's search features," opt it out and you will not appear in ChatGPT's answers. The common mistake: a business blocks "AI bots" wholesale in its robots.txt to keep its content out of training, and accidentally locks itself out of AI recommendations at the same time. You can allow the search and answer crawlers while still blocking the training ones. If you have never looked at your robots.txt for this, that is step zero.
Can it trust you? This is where the newest research gets interesting. A 2026 MIT study led by Sinan Aral ran 24,000 searches across 243 countries and found that AI search leans on far fewer sources, with less variety, than traditional search does. It concentrates its answers on a much narrower set of sources. For a small business, that finding cuts two ways. It is harder to break in, but once you are established as a trustworthy, consistent source, you tend to keep getting named. Your real credentials, your genuine reviews, and business details that stay consistent everywhere they appear are what earn that trust.
Can it match you? An AI recommends the source that most clearly answers the specific question asked. If someone asks about hiring a foreign worker and your services page buries that topic inside a giant "work and study" page alongside eight other things, the machine struggles to connect you to that question. The business whose page answers that one question in plain language gets the mention.
Notice what is not on that list: tricks. Google's own guidance is unusually blunt about the tactics agencies love to sell, filing them under a heading it literally calls "a few things you can ignore for Google Search": the llms.txt files, the content "chunking," the robotic AI-rewrites, the heavy schema markup. Google says it does not use any of them and you do not need them, "you don't need to write in a specific way just for generative AI search," and there is "no special schema.org structured data that you need to add." Schema still earns its place when it accurately describes what is already on the page, it just does not buy you a spot in AI answers. The freeing, slightly uncomfortable truth is that being genuinely helpful and clearly organized beats every gimmick on the market.
One honest caveat, because it matters. Google's guidance describes Google's own AI features. ChatGPT, Claude, and Perplexity are separate systems with their own methods. But the one platform that has actually told us how its AI picks content says to do the fundamentals well, and the independent research points the same way. When the clearest evidence and the outside research agree, that is the direction to walk.
Audit yourself this afternoon
You do not need to hire anyone to find out where you stand. You can get a rough read in under an hour.
The fastest route is our free AI-visibility self-check: type in your business name and city, and it builds the exact prompts below for you, split into the same two checks. If you would rather understand what it is doing, here is the manual version.