Do AI assistants get Canadian city level supplier questions right, and which city domains actually won

The assistants treat Canadian cities as real places rather than as decoration, and we can show that from the citation list rather than assert it. Across 64 blind readings of ChatGPT and Gemini on Canadian buying questions on 22 October 2026, two Canadian city specific domains sat high in the source list: torontoseo.com cited in 13 of the 64 readings and ottawaseo.com in 7. Those are small, named, city level properties, cited more often than several national ones. What we have not yet done is a dedicated Canadian metro pass, six questions per city across several cities, and this page will not present a figure we did not take. What it gives instead is what the national reading establishes, and the method to run the city check yourself in an afternoon.

What the Canadian national reading establishes about cities

Host	Readings citing it, of 64	Kind
torontoseo.com	13	Canadian city specific
canadacreate.com	11	Canadian commercial
wisemedia.io	9	Canadian commercial
ottawaseo.com	7	Canadian city specific
clutch.co	12	Directory, with city pages
agencies.semrush.com	6	Directory, with city pages

Four Canadian commercial or city domains account for 40 readings between them, and two directories that carry city pages account for 18. For a Canadian supplier that is a concrete, short target list: a city specific property and a directory city page are both reachable assets, unlike a national authority position.

What right and wrong mean for a city answer

Two different questions sit behind the headline. Does the assistant understand that the buyer means a specific Canadian city? On the evidence of the citation list, yes, because city specific sources appear when the question is local. Does it name the best supplier in that city? That is unanswerable as stated, because best is the buyer's criterion rather than a fact. What is checkable, and what a buyer should actually check, is narrower: is the named firm real, is it currently trading, and is it actually in that city or metro. Those three take minutes and nobody audits the answer for you.

The Canadian city check, run in an afternoon

• Write six questions a buyer in your city would ask out loud, naming the city in the text of each. Add Ontario, Quebec, British Columbia or the province where that is how buyers speak.
• If the city is Montreal or anywhere in Quebec, run the same six in French as a separate set. A French question about a Quebec buyer is a different question and will return different sources.
• Put all of them to both engines and save the answer text, the named firms and the full source list, with the date and the engine.
• Check each named firm for three things. Real, trading, in the metro.
• Count the sources and identify the leading one for that city. That is your first target and it will differ from the national list.
• Repeat per city. Do not assume the first city generalises, because on every reading we have taken in any market the leading source changed between cities.

The cost consequence, which is the part to settle before buying monitoring

If you sell in Toronto, Montreal and Vancouver, that is three question sets, and in Montreal it is two because of the language. Do the arithmetic before you look at a plan: tracked questions, times engines, times refresh frequency, times cities, times language sets. Forty questions becomes 240 tracked questions in that configuration before any engine multiplier. Then ask a vendor which unit the price is per, a tracked question, an engine, a recorded answer or a domain, and recompute. Ask about the third city rather than the first, because the per location line is where these bills grow.

The Canadian specifics that change a city answer

• Quebec is a language market as well as a place. The Charter of the French Language, strengthened by the 2022 amendments, sets requirements for commercial advertising and public communications in the province, so a Montreal service area page in English only is both a compliance exposure and an invisible asset in French answers.
• Ottawa's catchment crosses a provincial line. A search radius around Ottawa reaches into Quebec, which changes the language question and sometimes the regulatory one for a single local business.
• Provincial rules differ for regulated trades. Licensing, insurance, legal and health advertising are provincial matters, so a city page in a regulated category carries a provincial rule rather than a national one.
• The engines read Canadian official sources. canada.ca appeared in 10 of the 64 readings, competition-bureau.canada.ca in 9, ised-isde.canada.ca in 9 and www150.statcan.gc.ca in 8. A city page that cites a provincial or federal source correctly, with a date, agrees with something the engine already trusts.

How the national reading was taken

A 198 question Canadian buyer bank in 33 batches, put to the ChatGPT and Gemini product surfaces through a data provider on 22 October 2026. Canada was named inside the text of every question, because the provider accepts gl, hl and location parameters and ignores all three, so the prompt is the only market lock that holds. Batch 33 names us in the question and is therefore a ceiling measurement, excluded from every figure. Sixty four blind readings, 32 per engine, 2,992 citation URLs across 974 distinct hosts. ChatGPT ran a live web search on all 32 of its readings and Gemini on 20 of its 32, and Gemini named a rival in 17 readings against ChatGPT's 13. One host in the raw data, ccn.web.tr, is the data provider's own fetch infrastructure rather than a publisher and is excluded from every source analysis.

Thirty three of the 64 readings named no supplier at all, which is the largest single outcome and the cheapest opening for a local Canadian business, because an unclaimed slot costs a published answer rather than the displacement of an incumbent.

What to do about your own city first

• Build or fix the service area page that names the city, the neighbourhoods and the realistic travel distance. An engine answering a city question needs a document that says the city.
• Claim the city page on the directories the engines read and make the details match your own site exactly.
• Find the city specific property that leads your metro in your own reading and work out whether it takes contributions, listings or coverage.
• Publish one counted figure about your own city market, with denominator, method and date.

If your metro is Ottawa, the catchment that crosses into Quebec changes the brief and is set out in a search radius that crosses into Quebec. If it is Calgary, sizing the market before buying the scope is the first decision, in size the market before you buy the scope.

Questions about Canadian city level assistant answers

Do the assistants know Canadian cities properly?

The citation evidence says they reach for city specific Canadian sources, with torontoseo.com in 13 of 64 readings and ottawaseo.com in 7. That is a working understanding of the metro.

Do you have per city Canadian figures?

Not yet. We have run a dedicated metro pass in another market and not in Canada, and the two cannot be mixed because they are different banks with different denominators. The method above is the one to run, and it is the same method.

Should a Montreal check be in French?

Run both as separate sets. A French question about a Quebec buyer returns different sources, and measuring it through an English set measures the wrong thing.

Can I trust an assistant to find a supplier in my Canadian city?

As a shortlist generator with verification. Check that each named firm is real, trading and in the metro before calling, because those checks take minutes.

Should I track every Canadian city I sell in?

Do the multiplication first, including language sets. Three cities with one bilingual market is four question sets, and that is the number to take into a pricing conversation rather than a plan name.

Running the city reading rather than inferring it from the national one

The national Canadian reading shows that city specific sources matter and it cannot tell you which one leads your city. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews with the city and the market named inside the question, records every cited host, reports each engine separately with its denominator, and keeps the raw answer text so a city claim can be checked against what was actually said. Outside the United States and the United Kingdom the cheapest plan that reads all five engines is Growth at 199 dollars a month with 100 buyer questions watched, read on aiknowsus.com/pricing and aiknowsus.com/llms.txt on 22 October 2026, and the question count is the line to size against your city and language count.