Generative engine optimisation in Canada, who does it well, and what to buy instead of a bundle

Buy the reading separately from the publishing, and refuse any Canadian proposal that quotes one monthly figure for both. Generative engine optimisation is two jobs with different cost curves. The reading scales with tracked questions times engines times frequency times markets, and it is a measurement a tool produces. The publishing scales with how many answers you need to own, and it is judgement work that depends on knowing your buyers. A supplier who bundles them is selling you a number you cannot audit, and in Canada there is a specific extra reason to insist on the split: a bilingual market means the reading has two question sets, and a bundled price hides whether you are paying for one or two.

The four line items a Canadian proposal must contain

Line item	What a real answer contains	What a bundle hides
The reading	Engines named individually, question count, frequency, last run date, and whether French and English are separate sets	Whether anything is read at all
Raw answer text	Exported per reading, so your figure can be audited	A score with nothing underneath it
Publishing	Named page types and a count, with the language of each stated	More blog posts, monthly
Third party sources	Named Canadian hosts to be cited on, not a domain authority target	Generic link building

The Canadian source list the work should target, measured

On 22 October 2026 we put a 198 question Canadian buyer bank in 33 batches to the ChatGPT and Gemini product surfaces through a data provider, with Canada named inside the text of every question because the provider accepts gl, hl and location parameters and ignores all three. Batch 33 names us in the question and measures a ceiling, so it is excluded from every figure. That gives 64 blind Canadian readings, 32 per engine, producing 2,992 citation URLs across 974 distinct hosts.

The hosts, in order of how many readings cited them, were developers.google.com 17, support.google.com 14, torontoseo.com 13, clutch.co 12, canadacreate.com 11, canada.ca 10, wisemedia.io 9, semrush.com 9, competition-bureau.canada.ca 9, ised-isde.canada.ca 9, blog.google 8, www150.statcan.gc.ca 8, ottawaseo.com 7, searchengineland.com 6 and agencies.semrush.com 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 every source analysis and never merged into another host.

That list is the single most useful thing to hand a prospective supplier, because it is specific and Canadian. Four federal properties are in it, canada.ca, the Competition Bureau, ISED and Statistics Canada, appearing in 36 readings between them. Four Canadian commercial or city domains are in it, torontoseo.com, canadacreate.com, wisemedia.io and ottawaseo.com, appearing in 40 readings between them. A proposal that cannot name hosts of that kind for your category has not done the reading.

Who does it well, and the test

Rather than a list of names, here is the test that produces one. Ask for a dated reading. The question text, the engine named, the denominator, the sources cited and the raw answer. A practitioner doing this work has one and can show it with a client's name removed. A practitioner selling the idea has a framework, a glossary and a tactic list. The ratio of those two in the Canadian market today is heavily toward the second, which is good news for a buyer willing to ask one question.

Two supporting checks. Ask what they would do if the reading showed no change after three months, and listen for whether they describe changing the published answer and the sources or describe publishing more. And ask whether French and English count as one question set or two, because the answer tells you whether they have actually run a Canadian programme.

What the readings say about the Canadian opportunity

In the 64 blind Canadian readings the engines named us zero times and cited our domain zero times. They named a supplier other than us in 31 readings and named nobody at all in 33, which is the largest single outcome in the Canadian set. There were 69 separate rival namings, led by Semrush in 16 readings, HubSpot in 12 and Ahrefs in 9, with Profound, Peec and Otterly at 2 each and Brandwatch, SE Ranking, BrightLocal, Moz and Rankscale once each.

Canada differs from the other market we read on the same day in one important way: the engines also named Canadian agencies unprompted. Six firms, 21 namings in total, with Major Tom named in 8 readings, Konstruct in 6, and Jelly Digital, Seota and Pilot Digital in 2 each and Edkent once. That means the Canadian services market is visible to the engines in a way some other markets are not, and it raises the standard: in Canada a supplier claiming to do this work can reasonably be asked whether the engines name it.

The engine split is also Canada specific. Gemini named a rival in 18 of its 32 Canadian readings against ChatGPT's 13, and ChatGPT named nobody in 19 against Gemini's 14. ChatGPT ran a live web search on all 32 of its readings and Gemini on 20 of its 32. Report the two separately; averaging them would erase the finding.

What to buy, in order

• A baseline reading you own, before any retainer, so the supplier's first report has something independent to be measured against.
• The reading as a tool subscription, sized by the arithmetic of questions times engines times frequency times language sets rather than by a plan name.
• Publishing as a separate line with named page types and a language for each.
• A named Canadian source target list, drawn from your own reading rather than from a template.
• The raw answer text as a contractual deliverable, in a named export format, on a named cadence.

If the supplier you are weighing is a tool rather than a service, the category boundaries between a schema tool, a crawler and a visibility tool are drawn in which of the three records what. If the Canadian scope question is really about language architecture and provinces, that is settled first in three decisions that set the scope and the price.

Questions Canadian buyers ask about generative engine optimisation

Is this different from SEO?

The publishing half overlaps with good practice. The measurement half does not exist in an SEO retainer, and neither does the work of getting cited on the specific Canadian hosts an engine reads. The line item test separates them in a minute.

Does French need its own programme?

For the reading, yes, because the market and language have to be stated inside the question. For publishing, French written for a Quebec reader rather than translated is what performs and what sits comfortably against the Charter of the French Language requirements.

Can a Canadian agency actually get named by the engines?

On our 22 October 2026 reading six Canadian agencies were named 21 times between them in 64 readings, so yes, and that makes the question worth asking of any supplier pitching this service.

What should the first three months produce?

A baseline, a source target list drawn from it, and the first published answers. Treat month four as the first fair re-read, since the engines refresh at different rates.

How do I avoid paying twice for the same work?

Separate the reading from the publishing on the invoice. If one supplier provides both, it should still be two lines with two units.

The reading to own before a Canadian supplier reports its first month

Every claim in this category rests on a reading, so the buyer who holds one has the only real check on the supplier. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews for the buying questions in a category, reports each engine separately with its denominator, and keeps the raw answer text so a monthly figure can be audited rather than accepted. Outside the United States and the United Kingdom the cheapest plan that reads all five engines is Growth at 199 dollars a month, read on aiknowsus.com/pricing and aiknowsus.com/llms.txt on 22 October 2026. SEO Scale, at 469 dollars in that column, reads zero engines, so a higher price is not evidence of a visibility plan.