“Skills gap” gets used as a catch-all for anything that slows AI adoption down. Statistics Canada’s Canadian Survey on Business Conditions actually measures a narrower thing — which businesses changed training or staffing because of AI, and which didn’t — and that data draws a sharper picture than the phrase usually implies.
Key takeaways
When people say Canada has an “AI skills gap,” they usually mean one of two different problems: not enough people who can build or evaluate AI systems, or not enough people inside an ordinary business who know how to use one well. Statistics Canada’s Canadian Survey on Business Conditions measures the second kind directly, by asking businesses whether AI use changed their training or staffing practices at all — and the answer is a useful corrective to a debate that otherwise runs on impression.
In the second quarter of 2026, among businesses that reported using AI to produce goods or deliver services over the preceding 12 months, 44.4% said they made changes in training or staffing practices due to AI use. Of that group, 32.0% provided AI-related training for existing employees and 21.6% did the same for existing executives. Those are not small proportions, but they are also not universal — well over half of AI-using businesses made no training or staffing change at all, which is itself informative: for a large share of adopters, using AI has not yet required retraining anyone.
The StatCan breakdown by employee count shows the real fault line. Among AI-using businesses with 100 or more employees, 68.1% reported training existing employees and 51.7% reported training existing executives; 32.8% hired staff specifically for AI-related skills, and 30.2% brought in external consultants or vendors. Among AI-using businesses with 1 to 4 employees, the same figures were 24.0%, 15.6%, 8.2% and 10.7% respectively. A large firm facing a skills gap can absorb it — through internal training budgets, dedicated hires or outside help. A four-person firm facing the identical gap has none of those levers available at the same scale, which is a structural difference, not a motivation difference.
A separate StatCan study of AI adoption and productivity found that the businesses most likely to adopt AI are not simply the largest ones — they are the ones that already had certain capabilities in place. Firms using data analytics are 15.0 percentage points more likely to adopt AI than firms that do not, and firms using advanced robotics are 8.1 percentage points more likely, alongside R&D activity, cloud computing and ICT training for employees as further predictors. That is the mechanism behind the skills gap, not just a description of it: a business that has never invested in data analytics or staff ICT training is starting an AI rollout from a different position than one that has, independent of how many people it employs.
When StatCan asked AI-using and non-using businesses what limits their use of AI, “lack of skilled workers” was not a flat, economy-wide number — it concentrated. Information and cultural industries reported it as a barrier at 20.9%, and manufacturing at 12.0%, well above the reported barriers for cost (10.6% economy-wide) in some sectors and below it in others. That unevenness matters for anyone trying to size the problem for their own business: the skills constraint that shows up in national commentary may not be the constraint your own industry actually faces, and the only way to know is to look at the industry-level breakdown rather than the headline figure.
A related StatCan analysis of employment trends from November 2022 to December 2025 — the period generative AI moved from novelty to mainstream tool — found that employment generally grew regardless of potential occupational exposure to and complementarity with AI, though job growth varied across worker characteristics; younger employees and those less educated generally saw weaker job growth over the period. Coding-intensive professions grew at a similar rate to other jobs overall, but that growth concentrated among workers aged 30 to 49, while employment among coding professionals younger than 30 stagnated. The authors are explicit that this is not proof AI caused the pattern: “it is unclear whether more recent trends reflect the advent of AI, other economic factors such as labour market adjustments after the COVID-19 pandemic, rapid demographic shifts, recent trade tensions with the United States or a combination of factors.” Treat the finding as a documented pattern worth watching, not a settled causal claim about AI and entry-level hiring.
Canada’s national AI talent programme is often discussed as if it were a response to generative AI. It is not — the Pan-Canadian Artificial Intelligence Strategy launched in 2017, built around $125 million through Budget 2017, the Canada CIFAR AI Chairs programme, and the three national AI institutes: Amii, Mila and the Vector Institute. Its second phase, backed by Budget 2021 and the 2024 Fall Economic Statement, organizes federal effort around three pillars — Commercialization (helping businesses use AI), Standards (advancing AI-related standards) and Talent and Research (growing Canada’s AI talent pool). None of these figures are a market-size or ROI number; they describe programme funding for research and training infrastructure, not what any individual business should expect to spend or save.
Reading your own gap against the data
A useful exercise is not to ask “does my business have an AI skills gap” in the abstract, but to place your business on the two axes StatCan actually measures: size (which determines whether internal training, a dedicated hire or an outside vendor is realistically available to you) and industry (which determines whether “lack of skilled workers” is a barrier your peers report at 20% or closer to zero). A four-person retailer and a forty-person software firm are not facing the same gap even if both call it “the AI skills gap.”
Related: AI in Canadian small business, and how Canadian businesses actually use AI.
How an organization decides whether it is actually ready to commit to an AI plan — skills included — is covered on the strategy and roadmapping hub.
No. StatCan’s barrier data shows “lack of skilled workers” reported at 20.9% in information and cultural industries but only 12.0% in manufacturing (Q2 2026 survey) — a real gap, but one concentrated in specific sectors rather than spread evenly across the economy.
The mechanics differ sharply. Large businesses (100+ employees) reported training existing staff at 68.1% and using external consultants at 30.2%; the smallest businesses (1–4 employees) reported 24.0% and 10.7% respectively (StatCan, Q2 2026). The skills constraint is similar; the resources available to close it are not.
No — the Pan-Canadian Artificial Intelligence Strategy dates to 2017, five years before ChatGPT-era tools existed, and was built around research infrastructure (the CIFAR AI Chairs and the Amii, Mila and Vector institutes) rather than business adoption specifically.
A short call is enough to place your business on the size and industry axes that actually predict the gap.
