Do AI assistants favour Canadian suppliers when the buyer is in Canada, and what our readings actually show

For services, yes. For software, no. That is the honest shape of the answer and both halves come from the same reading. On 22 October 2026, across 64 blind readings of ChatGPT and Gemini on Canadian buying questions, the engines named six Canadian marketing agencies unprompted, 21 namings in total, and named no overseas agency at all. In the same readings they named software products 48 times and most of those products are not Canadian companies. So the locality preference is real and it is category specific: when the question calls for a service supplier the engines reach for Canadian firms, and when it calls for a product they reach for whatever product has the best published evidence, wherever it is from.

The two halves, as measured

	Service supplier questions	Software and product questions
Who was named	Six Canadian agencies, 21 namings	Eleven products, 69 namings
Named by name	Major Tom 8, Konstruct 6, Jelly Digital 2, Seota 2, Pilot Digital 2, Edkent 1	Semrush 16, HubSpot 12, Ahrefs 9, Profound 2, Peec 2, Otterly 2, Brandwatch 1, SE Ranking 1, BrightLocal 1, Moz 1, Rankscale 1
Overseas suppliers named	None	Most of them
What that implies	Locality matters when the supplier has to be local	Locality does not matter when the product travels

How the reading was taken

A 198 question Canadian buyer bank in 33 batches, put to both engine surfaces through a data provider. 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 therefore measures a ceiling rather than a market, excluded from every figure. Sixty four blind readings, 32 per engine, producing 2,992 citation URLs across 974 distinct hosts. Thirty three readings named no supplier at all, the largest single outcome.

The comparison that explains why

We ran the same bank for the United States on the same day. There the engines named zero local agencies in 62 blind readings while naming software 48 times. Same bank, same engines, same date, and the local services half of the answer behaved completely differently between two neighbouring markets.

The explanation is in the citation lists rather than in the engines. Canada's leaned on Canadian sources: torontoseo.com in 13 readings, canadacreate.com in 11, canada.ca in 10, wisemedia.io in 9, competition-bureau.canada.ca in 9, ised-isde.canada.ca in 9, www150.statcan.gc.ca in 8 and ottawaseo.com in 7, alongside developers.google.com in 17, support.google.com in 14 and clutch.co in 12. The American list had no local firm properties near the top at all. So locality preference is a property of what is published and citable in a market, not a property of the assistant, and Canada happens to have more attributable published material about its own service suppliers. These are two separate measurements with their own denominators, reported side by side and never merged. One host in the raw Canadian data, ccn.web.tr, is the data provider's own fetch infrastructure rather than a publisher and is excluded from every source analysis.

The engine split, which is also Canada specific

Gemini named a rival in 18 of its 32 Canadian readings and ChatGPT in 13, while ChatGPT ran a live web search on all 32 of its readings and Gemini on 20 of its 32. In the American pass the more rival naming engine was the other one. That reversal is the clearest argument against a single blended visibility score across engines or across markets.

What a Canadian supplier should do about it

• If you sell a service, the locality advantage is real and claimable. Six Canadian firms hold 21 namings. The barrier to joining them is attributable published material rather than size.
• Publish in the register the engines quoted. A dated fee position in Canadian dollars with its provincial tax treatment, a service specification, and a counted figure about your own market with a denominator.
• Target the Canadian hosts in your own reading. City specific domains and federal data properties were cited heavily, and they are reachable in a way national authority is not.
• If you sell a product, locality buys you nothing here. Compete on published specification, published price and documentation, which is what the named products did.
• Run a French reading separately if Quebec is a market. A French question returns a different source pool, and the Charter of the French Language makes French a requirement for selling pages in that province anyway.

How to test your own position

• Write twelve buying questions a Canadian buyer in your category would ask, with Canada or the province named in each.
• Put each to both engines and save the answer text, the named suppliers and the full source list with the date.
• Classify each named supplier as Canadian or not, and as service or product. The mix is your answer.
• Count the readings that named nobody and work those first.
• Report the engines separately with their own denominators.

If the comparison you want is the other direction, what a foreign supplier has to do to be credible in Canada is set out in the five checks a Canadian buyer runs. If you want the five gates an engine applies before naming anybody, they are in the five gates.

Questions about locality in Canadian assistant answers

So do the engines prefer Canadian suppliers?

For services, on our evidence yes, with six Canadian firms named 21 times and no overseas agency named. For software, no, since most of the 69 product namings went to non Canadian products.

Why did Canada get local firms named and the neighbouring market not?

Because the published, citable evidence differed. The same bank on the same day named zero local agencies in the other market. That is a statement about what each market publishes rather than about the engines.

Does being Canadian help a Canadian software company?

Not on this evidence. Compete on published specification, price and documentation, which is what the named products did regardless of nationality.

Should I target Canadian sources specifically?

Yes, from your own reading. In ours, four Canadian commercial or city domains and four federal properties were cited in 76 readings between them, which is a short and reachable target list.

Can I combine the Canadian and other market figures?

No. Two banks, two denominators. Report them side by side with their own labels.

Reading your own locality rather than inferring it

Whether locality helps you in Canada turns out to depend on your category and on what is published about it, which only a reading can tell you. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews with Canada named inside the question, records who was named and what each answer cited, reports the engines separately with their denominators, and keeps the raw answer text so a classification can be checked rather than assumed. 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.