Historic Meaford Hall on the main street of Meaford, Ontario.

Reviews

The Canada AI Visibility Index 2026: who ChatGPT, Gemini and Perplexity actually recommend

We measured 368 Canadian local markets with 7,360 AI answers. In 40%, the Google review leader was never mentioned by a single AI assistant.

By Lasse Pettersen · · Updated August 16, 2026

The Canada AI Visibility Index measures which local service businesses AI assistants actually recommend. In July 2026 we asked ChatGPT, Google Gemini, Perplexity and Claude the five questions a real customer asks, for 8 trades in 46 Canadian cities and boroughs: 7,360 answers across 368 local markets. The headline findings: in 87% of markets the AI favourite is not the Google review leader, and in 40% of markets the review leader was never mentioned in a single answer.

This is the first edition. The same sweep re-runs on a schedule with an identical method, so these numbers become a trendline: the first public record of how AI recommendations of Canadian local businesses shift over time.

How we built the index

For every market (one trade in one city, “plumber in Halifax”) we asked the five questions a homeowner actually types or says: what’s the best one, who should I call for this specific emergency, recommend someone reliable, who is most trusted according to reviews, and which one do locals actually use. Each question went to four models: Perplexity Sonar Pro, GPT-4o with browsing, Gemini 2.5 Flash with browsing, and Claude Sonnet 4.5 without web access as the training-memory control. Twenty answers per market, identical wording everywhere, from Victoria to St. John’s. Toronto is measured as its boroughs (Etobicoke, North York, Scarborough plus the core) because one top-20 list cannot represent a city that size.

We then counted which businesses each answer recommended, matched against the top Google Maps results for that market collected the same week. Four honesty notes on method. First, API answers approximate, not replicate, the consumer apps, which is why the identical sweep matters more than any single answer. Second, businesses whose legal name is just the service itself (“Appliance Repair”) match answers that merely use the phrase, so they are excluded from every named table and from the headline statistics. Third, a business with duplicate Google listings is counted once per answer, and the one market where no tracked business was named at all counts toward the invisibility column. Fourth, Quebec markets were asked in English, so they measure the anglophone assistant surface; a French sweep is a future edition’s job.

Like our Collingwood Review Index, the named tables below only ever name businesses that are winning. Everyone else stays anonymous and aggregated. If you want to know where your own business sits, run the free check, not the leaderboard.

The national numbers

  • 46 cities and boroughs across all ten provinces, 368 trade-city markets, 7,360 AI answers
  • 9,012 recommendations of identifiable local businesses across those answers
  • In 87% of markets (322 of 368), the business AI recommends most often is NOT the business with the most Google reviews
  • In 70% of markets (258), the AI favourite has less than HALF the review leader’s count
  • In 40% of markets (149), the review leader was never mentioned in any of the 20 answers
  • Only 43 markets have a dominant AI favourite (one business taking half or more of all recommendations), and 16 markets are wide open, with no business named more than twice

Waffle chart of 100 squares representing 368 Canadian local markets: in 40% of markets (149 of 368), the business with the most Google reviews was never named in any of the 20 AI answers for its trade and city; in 60% it was named at least once.

That first pair of numbers is the finding that should change how Canadian service businesses think about reputation. Being #1 on Google reviews is the outcome most local businesses spend years working toward. In nearly nine out of ten markets, that achievement does not carry over to the AI answer, and in four out of ten it buys literally zero AI visibility.

Horizontal bar chart of the 368 measured markets: in 87% the AI favourite is not the Google review leader, in 70% the AI favourite has under half the review leader's count, and in 40% the review leader got zero AI mentions.

Reviews still matter, but they are not the algorithm

Do not misread this as “reviews are dead.” Review volume remains the heaviest single lever in local search, as our 2026 local SEO statistics roundup documents, and most AI favourites do have healthy review counts. The finding is narrower and more useful: AI assistants are not ranking by review count. They synthesize from what the live web says about a business, and the businesses that win the AI answer are the ones the web talks about: directories that know they exist, a website that plainly answers the questions people ask, local press, and yes, a solid review base that gives the models something to cite.

A follow-up study on the Georgian Bay market shows what the models are reading when they look a business up: across 39 towns and 10 trades, only 6.6% of the citations were the recommended business’s own website. The full breakdown is in what AI reads before it recommends a Georgian Bay business.

The proof that this is a web phenomenon rather than a memory one is the control model. Across all 7,360 answers, the three web-grounded models produced 8,792 recommendations of tracked local businesses. Claude, answering the identical questions from training memory alone, produced 220, or 2.4% of the total. AI models do not “know” Canadian local businesses. They look them up. That means AI visibility is earnable, on a timescale of weeks, by changing what the web says about you. We watched it happen with our own site: three days after publishing structured local content, Perplexity began citing it, while the ungrounded models did not move at all.

Bar chart of local business mentions across all 7,360 answers by model type: the three web-grounded models produced 8,792 mentions while the training-memory-only model produced 220, or 2.4%.

Who is winning the AI answer in Canada

The strongest AI favourites in the July 2026 sweep, by how many of their market’s 20 answers named them:

BusinessMarketAI answers naming themTheir reviewsMarket review leader holds
Paysa - Taille de cèdresQuebec City, landscaping16 of 2067180
Muzzell PlumbingOwen Sound, plumbing15 of 2045171
Mr. Appliance of Red DeerRed Deer, appliance repair15 of 201041,194
Eden LandscapingRegina, landscaping15 of 208246
Electrafix NBFredericton, appliance repair14 of 20188941
Ron’s Appliance Service Ltd.Lethbridge, appliance repair14 of 2088935
Mr. Mechanic Auto Repair IncLethbridge, auto repair14 of 20289387
Service 2000Montreal, appliance repair14 of 205022,037
Mellish Appliance ServiceOwen Sound, appliance repair14 of 2085156
Sure Flow Plumbing & Heating Ltd.Saint John, plumbing14 of 2072566
Perfection Plumbing & Drain Cleaning Ltd.Saskatoon, plumbing14 of 207653,054

Look at the last two columns. Every business on this list is out-recommending a market review leader that holds 1.3 to 31 times its review count. Perfection Plumbing wins Saskatoon with 765 reviews in a market whose review leader holds 3,054. Eden Landscaping wins Regina with eight; the AI models cite its website, not its review count. The AI answer is a different competition with different rules, and right now it is being won by businesses of every size.

Closer to home on Georgian Bay, the same pattern holds: Clearview Plumbing takes more than half of Collingwood’s plumbing answers, Weldon Roofing leads the roofing answers, and Thornbury Auto owns the Blue Mountains auto repair conversation. None of them is the review leader in their market.

City-by-city results

The reference numbers for every city we measured, July 2026 edition. Each entry covers one city’s eight trades: how often the AI favourite and the Google review leader disagree, how often the review leader went completely unmentioned, and how many identifiable local businesses were recommended across that city’s 160 AI answers. The strongest winners are named in the table above; nobody is named for being absent. To see where your own business sits in your city’s answers, run the free AI visibility check. The six Georgian Bay markets (Barrie, Collingwood, Midland, Owen Sound, The Blue Mountains, Wasaga Beach) are re-measured monthly; every other city is re-measured quarterly.

Calgary · Edmonton · Lethbridge · Red Deer · Abbotsford · Burnaby · Kelowna · Richmond · Surrey · Vancouver · Victoria · Winnipeg · Fredericton · Moncton · Saint John · St. John’s · Halifax · Barrie · Brampton · Collingwood · Etobicoke · Hamilton · Kingston · London · Markham · Midland · Mississauga · North York · Oshawa · Ottawa · Owen Sound · Scarborough · St. Catharines · Sudbury · The Blue Mountains · Toronto · Vaughan · Wasaga Beach · Windsor · Charlottetown · Gatineau · Laval · Montreal · Quebec City · Regina · Saskatoon

Calgary, Alberta

In Calgary, the AI favourite is not the Google review leader in 7 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 196 recommendations of identifiable local businesses were counted across Calgary’s 160 AI answers.

Edmonton, Alberta

In Edmonton, the AI favourite is not the Google review leader in all 8 trades we measured. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 220 recommendations of identifiable local businesses were counted across Edmonton’s 160 AI answers.

Lethbridge, Alberta

In Lethbridge, the AI favourite is not the Google review leader in all 8 trades we measured. In 4 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 291 recommendations of identifiable local businesses were counted across Lethbridge’s 160 AI answers.

Red Deer, Alberta

In Red Deer, the AI favourite is not the Google review leader in all 8 trades we measured. In 4 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 291 recommendations of identifiable local businesses were counted across Red Deer’s 160 AI answers.

Abbotsford, British Columbia

In Abbotsford, the AI favourite is not the Google review leader in 7 of 8 trades. In 4 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 170 recommendations of identifiable local businesses were counted across Abbotsford’s 160 AI answers.

Burnaby, British Columbia

In Burnaby, the AI favourite is not the Google review leader in 6 of 8 trades. In 1 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 150 recommendations of identifiable local businesses were counted across Burnaby’s 160 AI answers.

Kelowna, British Columbia

In Kelowna, the AI favourite is not the Google review leader in 5 of 8 trades. In 1 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 226 recommendations of identifiable local businesses were counted across Kelowna’s 160 AI answers.

Richmond, British Columbia

In Richmond, the AI favourite is not the Google review leader in all 8 trades we measured. In 6 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 101 recommendations of identifiable local businesses were counted across Richmond’s 160 AI answers.

Surrey, British Columbia

In Surrey, the AI favourite is not the Google review leader in all 8 trades we measured. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 87 recommendations of identifiable local businesses were counted across Surrey’s 160 AI answers.

Vancouver, British Columbia

In Vancouver, the AI favourite is not the Google review leader in 7 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 222 recommendations of identifiable local businesses were counted across Vancouver’s 160 AI answers.

Victoria, British Columbia

In Victoria, the AI favourite is not the Google review leader in 7 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 171 recommendations of identifiable local businesses were counted across Victoria’s 160 AI answers.

Winnipeg, Manitoba

In Winnipeg, the AI favourite is not the Google review leader in all 8 trades we measured. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 242 recommendations of identifiable local businesses were counted across Winnipeg’s 160 AI answers.

Fredericton, New Brunswick

In Fredericton, the AI favourite is not the Google review leader in all 8 trades we measured. In 6 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 257 recommendations of identifiable local businesses were counted across Fredericton’s 160 AI answers.

Moncton, New Brunswick

In Moncton, the AI favourite is not the Google review leader in all 8 trades we measured. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 294 recommendations of identifiable local businesses were counted across Moncton’s 160 AI answers.

Saint John, New Brunswick

In Saint John, the AI favourite is not the Google review leader in 7 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 259 recommendations of identifiable local businesses were counted across Saint John’s 160 AI answers.

St. John’s, Newfoundland and Labrador

In St. John’s, the AI favourite is not the Google review leader in 7 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 300 recommendations of identifiable local businesses were counted across St. John’s’s 160 AI answers.

Halifax, Nova Scotia

In Halifax, the AI favourite is not the Google review leader in all 8 trades we measured. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 188 recommendations of identifiable local businesses were counted across Halifax’s 160 AI answers.

Barrie, Ontario

In Barrie, the AI favourite is not the Google review leader in 6 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 240 recommendations of identifiable local businesses were counted across Barrie’s 160 AI answers.

Brampton, Ontario

In Brampton, the AI favourite is not the Google review leader in 7 of 8 trades. In 5 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 120 recommendations of identifiable local businesses were counted across Brampton’s 160 AI answers.

Collingwood, Ontario

In Collingwood, the AI favourite is not the Google review leader in all 8 trades we measured. In 6 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 203 recommendations of identifiable local businesses were counted across Collingwood’s 160 AI answers.

Etobicoke, Ontario

In Etobicoke, the AI favourite is not the Google review leader in 6 of 8 trades. In 5 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 107 recommendations of identifiable local businesses were counted across Etobicoke’s 160 AI answers.

Hamilton, Ontario

In Hamilton, the AI favourite is not the Google review leader in 6 of 8 trades. In 1 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 206 recommendations of identifiable local businesses were counted across Hamilton’s 160 AI answers.

Kingston, Ontario

In Kingston, the AI favourite is not the Google review leader in 6 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 263 recommendations of identifiable local businesses were counted across Kingston’s 160 AI answers.

London, Ontario

In London, the AI favourite is not the Google review leader in all 8 trades we measured. In 5 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 93 recommendations of identifiable local businesses were counted across London’s 160 AI answers.

Markham, Ontario

In Markham, the AI favourite is not the Google review leader in 7 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 157 recommendations of identifiable local businesses were counted across Markham’s 160 AI answers.

Midland, Ontario

In Midland, the AI favourite is not the Google review leader in 7 of 8 trades. In 5 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 104 recommendations of identifiable local businesses were counted across Midland’s 160 AI answers.

Mississauga, Ontario

In Mississauga, the AI favourite is not the Google review leader in 7 of 8 trades. In 1 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 187 recommendations of identifiable local businesses were counted across Mississauga’s 160 AI answers.

North York, Ontario

In North York, the AI favourite is not the Google review leader in 7 of 8 trades. In 5 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 122 recommendations of identifiable local businesses were counted across North York’s 160 AI answers.

Oshawa, Ontario

In Oshawa, the AI favourite is not the Google review leader in 5 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 201 recommendations of identifiable local businesses were counted across Oshawa’s 160 AI answers.

Ottawa, Ontario

In Ottawa, the AI favourite is not the Google review leader in 5 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 195 recommendations of identifiable local businesses were counted across Ottawa’s 160 AI answers.

Owen Sound, Ontario

In Owen Sound, the AI favourite is not the Google review leader in 7 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 349 recommendations of identifiable local businesses were counted across Owen Sound’s 160 AI answers.

Scarborough, Ontario

In Scarborough, the AI favourite is not the Google review leader in 5 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 104 recommendations of identifiable local businesses were counted across Scarborough’s 160 AI answers.

St. Catharines, Ontario

In St. Catharines, the AI favourite is not the Google review leader in all 8 trades we measured. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 217 recommendations of identifiable local businesses were counted across St. Catharines’s 160 AI answers.

Sudbury, Ontario

In Sudbury, the AI favourite is not the Google review leader in 7 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 272 recommendations of identifiable local businesses were counted across Sudbury’s 160 AI answers.

The Blue Mountains, Ontario

In The Blue Mountains, the AI favourite is not the Google review leader in all 8 trades we measured. In 7 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 67 recommendations of identifiable local businesses were counted across The Blue Mountains’s 160 AI answers.

Toronto, Ontario

In Toronto, the AI favourite is not the Google review leader in 7 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 151 recommendations of identifiable local businesses were counted across Toronto’s 160 AI answers.

Vaughan, Ontario

In Vaughan, the AI favourite is not the Google review leader in 7 of 8 trades. In 5 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 176 recommendations of identifiable local businesses were counted across Vaughan’s 160 AI answers.

Wasaga Beach, Ontario

In Wasaga Beach, the AI favourite is not the Google review leader in all 8 trades we measured. In 6 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 174 recommendations of identifiable local businesses were counted across Wasaga Beach’s 160 AI answers.

Windsor, Ontario

In Windsor, the AI favourite is not the Google review leader in 7 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 171 recommendations of identifiable local businesses were counted across Windsor’s 160 AI answers.

Charlottetown, Prince Edward Island

In Charlottetown, the AI favourite is not the Google review leader in 7 of 8 trades. In 4 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 217 recommendations of identifiable local businesses were counted across Charlottetown’s 160 AI answers.

Gatineau, Quebec

In Gatineau, the AI favourite is not the Google review leader in 7 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 214 recommendations of identifiable local businesses were counted across Gatineau’s 160 AI answers.

Laval, Quebec

In Laval, the AI favourite is not the Google review leader in 6 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 145 recommendations of identifiable local businesses were counted across Laval’s 160 AI answers.

Montreal, Quebec

In Montreal, the AI favourite is not the Google review leader in 6 of 8 trades. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 169 recommendations of identifiable local businesses were counted across Montreal’s 160 AI answers.

Quebec City, Quebec

In Quebec City, the AI favourite is not the Google review leader in 7 of 8 trades. In 4 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 152 recommendations of identifiable local businesses were counted across Quebec City’s 160 AI answers.

Regina, Saskatchewan

In Regina, the AI favourite is not the Google review leader in all 8 trades we measured. In 2 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 279 recommendations of identifiable local businesses were counted across Regina’s 160 AI answers.

Saskatoon, Saskatchewan

In Saskatoon, the AI favourite is not the Google review leader in 7 of 8 trades. In 3 of 8, the most-reviewed business was never mentioned in any of its 20 answers. 292 recommendations of identifiable local businesses were counted across Saskatoon’s 160 AI answers.

What this means if you run a service business

Four practical takeaways from the 2026 data:

  1. Your review count is not protecting you in the AI channel. If you have spent years becoming the review leader, you have won the Map Pack, not the AI answer. Four in ten review leaders are invisible there. The only way to know your own standing is to measure it, which is what our free AI visibility check does: the same five questions, the same four models, scored for your business.
  2. The AI answer is still up for grabs almost everywhere. Only 43 of 368 markets have a dominant favourite. The median market’s top business takes just a third of the recommendations. In the AI channel, 2026 Canada looks like Google looked in 2010: early, unclaimed, and cheap to win.
  3. The inputs are fixable, and reviews are still one of them. Grounded models read the public evidence: your Google profile and reviews, directories, and whether your site answers customer questions in plain language. We break all six inputs down, with timelines, in our guide to getting your business recommended by ChatGPT. A systematic done-for-you review engine, the boring kind that asks after every job, now pays twice: once in the Map Pack, once as the citation trail AI models follow.
  4. Google now shows you part of the scoreboard. Since August 2026 the generative AI impressions report in Search Console has been live on almost every property. It counts the times your pages appeared inside AI Overviews and AI Mode. It is website data only, so it complements this index rather than replacing it.

Frequently asked questions

How did you measure AI visibility? Five buyer-style questions per trade per city (“What’s the best plumber in Halifax, Nova Scotia?”, “I need help with a burst pipe in Halifax. Who should I call?”), each asked to four AI models via their APIs in July 2026, with identical wording in every market. We recorded every local business each answer recommended and matched them against that market’s top Google Maps results collected the same week. It is a snapshot of public AI behaviour, not private data.

Which AI models are included, and what happens when new models launch? The 2026 panel is Perplexity Sonar Pro, GPT-4o with browsing, Gemini 2.5 Flash with browsing, and Claude Sonnet 4.5 as the no-web control. The panel only changes at an annual edition boundary, and when a model is swapped we run old and new side by side for a period and publish the calibration, so the trendline stays honest even as the AI landscape moves.

My business was not named. How do I find out my score? The index names market winners only; nobody is called out for being absent. The free AI visibility check runs the same sweep for your specific business and shows exactly which questions you appear in and who is being recommended instead. If you want the full picture including your review position, start with a free audit.

How can a business with 8 reviews be an AI favourite? Because grounded AI models synthesize from the whole public web, not just the review count. A business with few reviews but a clear website, consistent directory listings, and third-party mentions can out-cite a bigger competitor whose web presence is just a Google profile. That is also the opportunity: these inputs move in weeks, not years.

Why were Quebec markets measured in English? The 2026 edition asks every market the identical English questions so the numbers are comparable coast to coast. For Montreal, Quebec City, Laval and Gatineau, that measures the anglophone assistant surface specifically. A French-language sweep is planned as a companion edition.

Will the index be updated? Yes. The underlying sweep re-runs quarterly with the frozen question set, and we publish annually with the same method so editions compare cleanly. Georgian Bay markets are re-measured monthly.


Compiled by Georgian Bay SEO, Collingwood, Ontario. Local SEO and AI visibility for trades and service businesses. Free audit at https://georgianbayseo.ca, (705) 539-0398.

Want this run against your own business?

Free 15-minute audit. We pull up Google together and show you exactly where you rank, where you could be in 60 to 90 days, and the cheapest moves to get there.

Free SEO Audit On the Map in 30