Every AI sales call this fall runs on the same line: move now or fall behind. Canada’s central bank just published the opposite view, in plain language, and it deserves a place in your next budget meeting.
- The Bank of Canada wrote on 18 September 2026 that AI’s biggest effects “likely still lie ahead.”
- Its own August research found only 8% of Canadian businesses use AI significantly in core operations.
- A slow payoff rewards firms that buy small and stay flexible. That describes most Canadian small businesses.
The Bank of Canada published a plain-language explainer on 18 September 2026 saying nobody knows when AI’s biggest productivity gains will arrive, and that the biggest effects likely still lie ahead. For a Canadian small business, that is useful cover. It means the pressure to commit big money to AI this year is coming from vendors, not from the evidence.
TL;DR
- The Bank of Canada’s The AI transformation, dated 18 September 2026, says general-purpose technologies usually deliver their biggest applications decades after they arrive.
- It also notes that AI can do many tasks but few jobs from start to finish, and needs human oversight because of hallucinations.
- The Bank’s August 2026 business research found 8% of firms use AI significantly in core operations and 70% expect no material AI effect on 2026 capital spending.
- Contrarian read: when the payoff is slow and uncertain, the winning move for a small firm is cheap, reversible adoption.
- Watch the Bank’s 28 October Monetary Policy Report for how it describes AI and productivity.
What did the Bank of Canada say about AI?
The Bank said AI may be a general-purpose technology like the computer or electricity, but those technologies took decades to reshape the economy. It said we don’t know when AI’s biggest productivity gains will come or who will benefit most. It also said AI still needs human oversight.
The piece walks through the internet as a precedent. Businesses built websites in the 1990s, the Bank notes, but most weren’t making money from them yet. The large e-commerce and social media companies came later.
One line should make every AI salesperson nervous: the businesses that use AI best “might not even exist yet.”
The Bank also points out that big technology firms’ AI infrastructure spending is helping drive economic activity, primarily in the United States. For Canadian firms, the build-out headlines describe somebody else’s economy.
Why does it matter in Canada?
Because Canadian businesses are adopting carefully, and the central bank just described that caution as reasonable. The Bank’s own August 2026 research found most firms use AI lightly. A slow, uncertain payoff means the cost of waiting a year on big projects is low, while small experiments still pay off now.
The Bank of Canada’s August 2026 Sparks article put numbers to it. Only 8% of businesses use AI significantly in core operations. Half use it to a low or moderate degree. Seven in ten expect no material AI effect on their 2026 capital spending.
Statistics Canada adds the other end. Its third-quarter business conditions survey, released 31 August, found 52.7% of businesses have no AI plans for the next 12 months.
Put the two sources together and the consensus pitch, that Canadian firms are dangerously late, looks shaky. Most of the economy is in the same place. Early days, in the Bank’s own framing.
What does this mean for a small business budget?
It means spend on AI the way sensible firms spent on the web in 1996. Small, cheap, monthly and easy to cancel. Save the large, locked-in commitments for when the use case has proven itself in your own numbers, which the Bank suggests may take years to show up economy-wide.
Here is the practical split:
| Buy now | Wait on |
|---|---|
| AI features inside software you already pay for | Multi-year platform contracts |
| Monthly per-seat tools you can cancel | Custom builds priced as a capital project |
| One workflow with a before-and-after number | Company-wide “AI transformation” programs |
The Bank’s other point matters for how you use it. AI handles tasks well and whole jobs poorly, and it can be wrong. So the gains come from people doing the same work faster, with a human checking the output.
What should leaders do next?
Keep AI spending small and reversible, and make every experiment produce a number. Use the Bank’s own framing when a vendor pushes urgency. Revisit bigger bets once the payoff shows up in your books.
- Put every AI tool on a monthly plan until it has three months of measured results.
- Ask vendors for an exit clause before signing anything longer than a year.
- Pick one workflow per quarter, measure hours before and after, and keep only what moves the number.
- Budget for review time. The Bank flags hallucinations. Someone has to check the output.
- Stop treating AI headlines about US data centre spending as a signal for your firm. The Bank says that activity is mainly in the United States.
The skeptic’s view
Waiting has a cost. Firms that learn early build skills their competitors can’t copy quickly, and the Bank itself says AI is already saving time for some workers and businesses. A small firm that dismisses AI because the macro payoff is slow may miss the micro payoff sitting in its own inbox.
That’s right, and it’s the reason to keep experimenting. The contrarian point is about size. Learn now. Commit later.
What to watch
- 28 October 2026. The Bank of Canada’s interest rate announcement and Monetary Policy Report. Watch its wording on AI and productivity.
- Fall 2026. Statistics Canada’s replacement for the business conditions survey, and how it measures AI use.
- Your Q4 renewals. Any AI contract moving from monthly to annual terms.
Our prediction: the October Monetary Policy Report will not credit AI with a measurable lift to Canadian productivity. The Bank will keep describing it as a longer-run factor.
FAQ
What did the Bank of Canada say about AI in September 2026?
In The AI transformation, published 18 September 2026, the Bank said the biggest effects of AI likely still lie ahead, and that nobody yet knows when its biggest productivity gains will happen.
Is AI increasing productivity in Canada?
The Bank of Canada says there is some evidence AI is saving time for workers and businesses, but it does not yet know when the largest gains will arrive or who will benefit most.
How many Canadian businesses use AI heavily?
The Bank of Canada’s August 2026 Sparks article found only 8% of businesses use AI significantly in their core operations.
Should a small Canadian business wait to adopt AI?
Waiting on large, locked-in AI projects is reasonable. Small, cancellable experiments on specific tasks are still worth running now, with results measured against a baseline.
Closing analysis
A central bank telling you the payoff is years out is the best negotiating position a small business will get this year. Buy small, measure hard, and let the vendors wait for your commitment.
For Alberta firms deciding where to start, the Alberta AI Advantage blueprint sets out low-risk first steps. Operators swap notes on the AI Podcast.
Sources
- Bank of Canada, The AI transformation, 18 September 2026. www.bankofcanada.ca
- Bank of Canada, Canadian businesses’ use of AI, what the evidence shows, Sparks, August 2026. www.bankofcanada.ca
- Statistics Canada, The Daily, Canadian Survey on Business Conditions, third quarter 2026, 31 August 2026. www150.statcan.gc.ca
- Bank of Canada, Interest rate announcement and Monetary Policy Report, 28 October 2026. www.bankofcanada.ca
Related reading
- Slowing frontier AI does nothing for the 52.7% of Canadian businesses with no AI plans
- The hidden time tax in AI productivity
- AI for business in Canada, our reference hub
Disclosure
The author has no relevant financial, advisory, or board relationships with any party named in this column.
Zak Hussein writes on AI for Canadian business owners and operators. He is the founder of AI Magazine Canada and CEO of ORKA AI.
