Where Canadian buyers start looking for a supplier in 2026, measured on 64 readings
What ChatGPT and Gemini actually put in front of a Canadian buyer on 22 October 2026, the Canadian hosts they leaned on, and what a supplier should do about each starting point.
Published by AI Knows Us · Updated 22 October 2026
Canadian buyers start in four places and only one of them is a search box. A referral, a platform or directory, a search engine, and increasingly an assistant. The fourth is the one with no evidence behind it, so here is ours. On 22 October 2026, across 64 blind readings of the ChatGPT and Gemini product surfaces on Canadian buying questions, the engines produced 2,992 citation URLs across 974 distinct hosts, named a supplier in 31 readings and named nobody at all in 33. The Canadian specific finding is in what they leaned on: four federal government properties and four Canadian city or commercial domains sat near the top of the citation list, alongside Google's own documentation.
What was measured
A 198 question Canadian buyer bank in 33 batches, put to both engines 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 stating the market in the prompt is the only lock that holds. Batch 33 names us in the question, which makes it a ceiling measurement, so it is excluded from every figure below.
| Reading | ChatGPT | Gemini | Both |
|---|---|---|---|
| Blind readings | 32 | 32 | 64 |
| Ran a live web search | 32 | 20 | 52 |
| Named a supplier that was not us | 9 | 17 | 26 |
| Named no supplier at all | 23 | 15 | 38 |
| Named us | 0 | 0 | 0 |
The two engines are reported separately and never averaged. Gemini named a rival nearly twice as often as ChatGPT in the Canadian set, while ChatGPT read the live web far more often. Averaging those two rows would produce a number that describes neither surface.
The Canadian hosts, which is where buyers actually get sent
- developers.google.com in 17 readings and support.google.com in 14, with blog.google in 8
- torontoseo.com in 13 readings and ottawaseo.com in 7, both city specific
- clutch.co in 12 and agencies.semrush.com in 6, both directories
- canadacreate.com in 11 and wisemedia.io in 9, Canadian commercial properties
- canada.ca in 10, competition-bureau.canada.ca in 9, ised-isde.canada.ca in 9 and www150.statcan.gc.ca in 8, four federal properties totalling 36 readings
- semrush.com in 9 and searchengineland.com in 6
One entry in the raw data, ccn.web.tr, is the data provider's own fetch infrastructure rather than a publisher. It is excluded from this analysis, never presented as a source, and never merged into another host's count.
That list is the Canadian answer and it differs from a generic one in two visible ways. Canadian federal properties are heavily represented, which means government data and regulator positions are part of how a Canadian buying question gets answered. And city specific domains appear high, which means a Canadian supplier's local presence is reachable through named, small, specific hosts rather than only through national authority.
The four starting points, and what to do about each
| Starting point | What the buyer does | What a Canadian supplier does |
|---|---|---|
| A referral | Asks their network, then checks the name online | Make the check easy. A current site, a Canadian address, a price position and a dated page survive it |
| A platform or directory | Compares listed suppliers | Claim and complete the profiles the engines read, clutch.co among them, and keep the price position current |
| A search engine | Types a commercial query | The familiar work, plus correct en-CA and fr-CA declarations, which is the commonest Canadian fault |
| An assistant | Asks a buying question in plain language | Publish a specific, dated answer and get cited on the Canadian hosts above, including the city specific ones |
Why the 33 readings that named nobody are the opportunity
An answer with no supplier in it is the largest single outcome in the Canadian set. An engine that answers in generalities has committed to nobody, so the slot is open and taking it costs a specific published answer rather than the displacement of an incumbent. Sort the buying questions in your category by that distinction and work the empty ones first. Where a name did appear it was usually software, 48 namings led by Semrush at 16, HubSpot at 12 and Ahrefs at 9, with six Canadian agencies named 21 times between them, Major Tom in 8 readings and Konstruct in 6.
What to do in the next thirty days
- Write the twelve questions a Canadian buyer in your category actually asks, with Canada or the city named in the text of each.
- Put each to both engines, save the answer text and the source list, and record the date.
- Mark each reading as named you, named a rival, or named nobody, and work the empty ones first.
- Count the hosts. The recurring Canadian ones are your target list and they are not a link building list.
- Re-read in thirty days with the same twelve questions and report the engines separately.
If your buyers are concentrated in one province, the structure of that market changes the answer, and British Columbia beyond Vancouver is set out in the island, the Okanagan and the rest. If you want a larger single engine source mix before building your own baseline, it is counted in the source mix behind 306 readings.
Questions about where Canadian buyers begin
Is search still the main starting point in Canada?
For most categories it remains the largest single one, and the assistants themselves leaned on Google's documentation properties more than anything else. The change is a second queue with a different gatekeeper and no reporting line in most packs.
How many Canadian buyers use an assistant first?
We measured what the engines answer rather than how many people ask, and we will not publish a share figure we did not take. Sixty four readings on 22 October 2026 is the measurement we have.
Why do Canadian government sites get cited so much?
Because they are authoritative, specific, dated and freely readable, which is the combination an engine can quote with confidence. Four federal properties appeared in 36 readings between them.
Do city specific domains really matter?
In this reading, yes. torontoseo.com appeared in 13 of 64 readings and ottawaseo.com in 7, which is higher than several national properties.
How often should a Canadian supplier re-measure?
Monthly is enough, and holding the questions, engines and denominator constant matters more than the frequency.
The Canadian queue with no reporting line
Every Canadian marketing report covers search and media. Ours found 64 Canadian readings in which we were named zero times, a rival was named 31 times and nobody was named 33 times, which is the shape of a queue most suppliers are not measuring. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews with Canada named inside the question, reports each engine separately with its denominator, records every cited host, and keeps the raw answer text. Outside the United States and the United Kingdom that is the Growth plan at 199 dollars a month with 100 buyer questions watched, the cheapest tier reading all five engines, read on aiknowsus.com/pricing and aiknowsus.com/llms.txt on 22 October 2026.