What the Bank of Canada's Own Data Says About AI Adoption

Most AI adoption statistics floating around come from somewhere with a stake in the answer: a software vendor surveying its own users, an industry association gauging member sentiment, a consulting firm promoting its next engagement. None of that makes the numbers wrong, but it's worth knowing who's counting. In late 2025, a different kind of institution weighed in.

The Bank of Canada — the country's central bank, whose job is to understand the real economy well enough to set interest rate policy — published an analysis titled "Canadian businesses' use of AI: What the evidence shows," built on its December 2025 Business Leaders' Pulse survey. The Bank has no product to sell and no membership base to please. It surveys businesses because accurately reading the economy is its actual job, not a marketing exercise. That's what makes this particular dataset worth a closer look: it's central bank economic research, not a vendor's adoption funnel.

The Headline Number: Real, Significant AI Use Is Still Rare

The Bank's survey asked businesses directly how much they're using AI in their core operations, and the results are more sobering than the "AI is everywhere" narrative suggests. Here's the full breakdown:

• 8% of businesses report using AI "significantly" in their core operations.
• 50% report using AI to a "low or moderate degree" in core operations.
• 28% are not using AI and do not plan to adopt it in 2026.
• 11% plan to start using AI in 2026.

Add those last two figures together and a striking picture emerges: 58% of Canadian businesses are either not using AI at all right now or only planning to start. Meanwhile, only 8 in 100 businesses describe their AI use as significant. The much larger group — half of all businesses surveyed — sits in a middle zone of light, exploratory use: a tool tried here, a feature turned on there, nothing that has fundamentally changed how the business runs.

This matters because it recalibrates a common assumption. Media coverage and vendor marketing can make it feel like every competitor has already rebuilt their operations around AI. The Bank of Canada's own data says otherwise: deep, committed adoption is still the exception, not the rule.

Capital Spending: Businesses Aren't Betting Big on AI Yet — But the Three-Year Outlook Shifts

The Bank also asked businesses how AI is expected to affect their capital spending — the money businesses put toward equipment, software, and infrastructure. In the near term, the answer is largely "not much." About 70% of businesses expect AI to have no material impact on their capital spending in 2026. Whatever AI adoption is happening right now, for most businesses, it isn't yet showing up as a real line item in next year's budget.

But the picture changes when businesses are asked to look further out. Looking three years ahead, nearly 40% anticipate increased AI-related capital spending. That's a meaningful gap between what businesses are budgeting for next year and what they expect to be spending on within three years — a sign that many businesses see AI adoption as a slower-building trend rather than a 2026 event, even if they aren't yet acting on that expectation.

The Employment Question: Modest Pressure, Not a Dramatic Shift

Perhaps the most closely watched question in any AI adoption survey is what it means for jobs. The Bank of Canada's findings here deserve to be read carefully, because the honest answer is more measured than headlines about AI-driven layoffs would suggest.

Looking three years out, 23% of businesses expect AI to have a negative impact on employment at their firm, while only 11% expect a positive impact. Net that out and you get modest downward pressure on hiring expectations tied to AI — a real signal, but not evidence of a wholesale disruption. Most businesses surveyed don't expect AI to meaningfully change their employment levels either way within three years.

Importantly, the Bank's data indicates this expected downward pressure is concentrated more among larger firms. That distinction matters for how this finding should be interpreted by a small business owner — the employment effect the Bank is picking up on looks less like a broad small-business phenomenon and more like something happening disproportionately at bigger organizations with larger headcounts and more repetitive, automatable roles to begin with.

The Small Business Gap Is Real, and the Bank Confirms It

One of the more directly useful findings for a small business owner is the Bank's confirmation of a size-based adoption gap: small businesses are explicitly less likely than large businesses to report using AI. This isn't a surprising result on its own — smaller firms typically have fewer resources for experimentation, less dedicated staff time to evaluate new tools, and less capital to absorb a failed pilot. But having it confirmed by the Bank of Canada's own survey data, rather than inferred from anecdote, gives it real weight.

The same size pattern shows up in the employment data too. Larger firms (20+ employees) are more likely than smaller firms to expect AI-driven employment reductions within three years — reinforcing that AI's employment effects, where they exist, are currently expected to land hardest at the larger end of the business spectrum, not among small, owner-operated businesses.

What This Actually Means If You Run a Small Business

Put the numbers together and a clear, if somewhat counterintuitive, picture emerges. Canada's central bank — with no incentive to inflate or deflate the story either way — finds that most businesses (58% combined) either aren't using AI or are only planning to start. Significant, committed AI use sits at just 8%. Capital spending on AI is mostly not happening yet in the near term, though a growing share of businesses expect that to change within three years. And the employment effects being picked up in the data are modest overall, and concentrated more at larger firms than small ones.

For a small business owner who has been putting off AI adoption because it feels like everyone else has already moved, this data says the opposite. Most of your competitors, by the Bank's own numbers, either haven't adopted AI in any real way or are still in a low-and-moderate holding pattern. That means being an early, focused adopter of one well-matched use case — like an AI voice agent that answers every call, books appointments, and follows up on leads without adding headcount — isn't catching up to a crowded field. It's still a genuine, measurable edge over a majority of businesses that, per the Bank of Canada, haven't gotten there yet.