Someone visiting Brampton opens ChatGPT and types: “where should I eat Indian food tonight?”
They get three names. Not ten blue links, not a map with twenty pins — three names and a sentence about each. Then they pick one and book.
If you’re not in those three, you didn’t rank badly. You didn’t appear at all. And there’s no page two to be on.
That’s the shift worth understanding, because it’s already happening to your customers.
The bit most people get wrong
There’s a common assumption that AI assistants read your Google rating and rank you by stars.
Mostly, they can’t. Google’s reviews are Google’s data, and they don’t hand it to competitors.
So you’re really dealing with two different systems that work in different ways:
Google’s own AI — the summaries appearing above search results — does read your Google Business Profile. Your rating, your review text, your hours, your categories. That’s Google using its own data.
ChatGPT, Perplexity, Copilot — these browse the open web. They see what other sites say about you: your own website, directories, Yelp, food blogs, local news, “best Indian in Brampton” listicles. They generally cannot see your Google star rating directly at all.
Same question, two completely different sources. Which means being brilliant on Google and invisible everywhere else gets you into one of those answers and not the other.
So what actually gets you named?
Nobody outside these companies knows the exact recipe, and anyone selling you a guaranteed formula is inventing it. But the mechanics point clearly in one direction:
1. You have to be described somewhere in words. An AI can only recommend what it can read. If the only place your restaurant exists in detail is a Google profile and a one-page site with a PDF menu, there’s very little for a language model to work with. A page that actually says “all-you-can-eat North Indian buffet in Brampton, halal, open late, good for large groups” gives it something to match a question against.
2. Your details must agree everywhere. Different phone numbers or opening hours across your site, Yelp and old directory listings don’t just confuse customers — they make you a low-confidence answer. Systems that summarise prefer sources that corroborate each other.
3. Review text matters more than review score. Here’s the part that surprises people. When an assistant explains why it picked a place, it paraphrases what customers wrote. Not “4.4 stars” — “reviewers consistently mention the butter chicken and say the buffet is good value for families.”
Those sentences come out of review text. Which means the words your customers use are quietly becoming your marketing copy, whether you’re involved or not.
4. Recent beats old. A business with steady recent reviews reads as open and thriving. Reviews that stop two years ago read as a business that might not exist any more.
Where replying comes in
If review text is what gets quoted, then your replies are part of the record too.
A profile where an owner answers thoughtfully — naming problems, describing fixes — reads very differently to a machine summarising it than a wall of unanswered complaints. Same star rating, different story.
That’s not a trick. It’s the same reason a human reader trusts one and not the other.
The uncomfortable part
Traditional search was forgiving. Ranking eighth still put you on the page. Someone scrolling might still find you.
An AI answer names three businesses. There is no eighth place, and no scrolling. The gap between being included and being left out is total, and it’s decided by information you may never have thought of as marketing — a directory entry from 2019, whether your site says what you actually serve, what your customers happened to write last month.
Which brings it back to the free thing you’re ignoring
Follow all of that to its conclusion and it lands in one place: your Google Business Profile.
Not because AI reads your star rating — most of it can’t. Because your profile is doing four jobs at once, and it’s the only one of them you fully control:
It’s the direct input to Google’s own AI. Google runs the summaries sitting above search results for a large share of local queries, and it builds them from Business Profile data. Your categories, your hours, your review text. No intermediary, no guessing.
It’s where your description originates. Directories, aggregators and “best of” round-ups routinely take their cut of your details from what’s published about you on Google. Get it wrong there and the error propagates to the sources the other assistants read.
It’s where the review text lives. Those are public, quotable sentences describing your business, written by people who aren’t you. They’re the raw material for “reviewers consistently mention…” — and your replies sit right underneath them.
It’s the freshness signal. Recent reviews, recent photos, current hours. It’s how anything reading you decides whether you’re a going concern or a business that closed in 2023.
So the answer to “how do I get recommended by AI?” turns out to be unglamorous. Look after the free listing you already have. It was worth doing when it just fed the map. Now it feeds the map, the AI summary above it, and half the sources everything else is reading.
What to do about it
None of this needs a new budget:
- Claim and complete your Google Business Profile properly — categories, hours, services, real photos. This is the highest-leverage hour you’ll spend.
- Keep reviews coming, and keep replying. You’re feeding the only description of your business you don’t write yourself.
- Say plainly on your own site what you are, where, and who for. Cuisine, area, hours, halal or not, groups, parking. Text on a page — not only in an image or a PDF.
- Fix your details everywhere they appear. Same name, address, phone. Old listings included.
The takeaway
You can’t optimise your way into an AI’s answer, and anyone promising otherwise is guessing.
What you can do is make sure that when something goes looking for a good Indian restaurant in your town, it finds a business that is clearly described, consistently listed, and recently talked about.
Every one of those traces back to the same free profile most owners set up once and never opened again. It was already the cheapest marketing you had. It’s now also the thing that decides whether a machine mentions your name at all.
Not sure what the web currently says about your business? I’ll go and look, and tell you what I find — including the parts you won’t like.