AI, Applied Benchmarks | Wave 3
How Canadian B2B Tech Decision Makers are catching up on AI maturity while falling further behind on the operating stack that makes agentic AI work.
Findings from Wave 3 of Georgian + NewtonX AI, Applied Survey. Canadian Tech Decision Makers, n=72. Survey May 2026.
In Wave 2 of our AI, Applied Benchmarks Report published in June 2025, we reported that only 7% of Canadian B2B software companies operated as Runners (the most AI-mature tier in Georgian’s Crawl, Walk, Run framework), and 48% of Canadian Tech Decision Makers called absence of technical talent their top barrier to getting AI into production.
In Wave 3, the Canadian Runners share more than doubled to 17%, with more than 56% (up from 31% a year ago) of Canadian Tech Decision Makers saying they’ve moved beyond AI experimentation (Joggers + Runners).
But one gap widened: Canada is now 23 points behind the rest of the world in company-wide implementation of agentic AI (the gap was only 13 points in Wave 2).
Three numbers that frame the Wave 3 Canadian story: AI maturity rising sharply while the agentic operating stack falls further behind the rest of the world.
| Indicator | Gap / value | Detail |
|---|---|---|
| Runner-tier share | 7% → 17% | Canada, Wave 2 → Wave 3 (+10 pts) |
| Past experimentation (Joggers + Runners) | 31% → 56% | Canada, Wave 2 → Wave 3 (+25 pts) |
| Agentic implementation gap vs ROW | 13 → 23 pts | Company-wide agentic AI; gap widened |
Canada’s Runner share grew from 7% in Wave 2 to 17% in Wave 3, narrowing the gap to Rest of World Runners (24%).
| Period | Canada | Rest of World |
|---|---|---|
| Wave 1 (Nov. 2024) | 5% | 12% |
| Wave 2 (Jun. 2025) | 7% | 17% |
| Wave 3 (May 2026) | 17% | 24% |
Wave 3 (May 2026): Canada 17 percent, Rest of World 24 percent, gap 7 points.
| Wave | Canada | ROW | Gap (ROW - Canada) |
|---|---|---|---|
| Wave 1 (Nov. 2024) | 5% | 12% | 7 pts |
| Wave 2 (Jun. 2025) | 7% | 17% | 10 pts |
| Wave 3 (May 2026) (selected) | 17% | 24% | 7 pts |
Canada’s AI Runner share grew from 7% in Wave 2 to 17% in Wave 3, reducing the gap to Rest of World (ROW) Runners (24%). The Canadian companies that have moved past experimentation (Joggers and Runners combined) climbed from 31% to 56%, a 25-point gain.
The talent crunch the Wave 2 data highlighted appears to have flattened in Wave 3. When asked about their blockers in acquiring or upskilling technical talent, Canadian Decision Makers report the same pattern as their global peers, except when it comes to losing AI talent to Big Tech, where they are significantly less worried than ROW.
Share of Tech Leaders selecting each as a top-2 blocker to acquiring or upskilling technical AI/ML talent. Wave 3 (May 2026). Canada n=72; ROW n=180. Ordered by Canada-vs-ROW gap: only competition from Big Tech diverges materially; all other blockers sit within ~5 points of ROW.
Only competition from Big Tech / startups diverges materially: 17% Canada vs 30% ROW. Every other blocker sits within ~5 points of the rest of the world.
| Measure | Canada | ROW |
|---|---|---|
| Competition from Big Tech / startups | 17% | 30% |
| Limited candidate pipeline | 44% | 40% |
| Lack of time / resources for upskilling | 40% | 37% |
| Internal processes (hiring, approvals) | 22% | 20% |
| Employer brand not strong enough | 21% | 19% |
| Budget constraints | 47% | 46% |
Automated coding solutions may have helped Canadian companies increase their AI maturity and manage their tech talent challenges in the last year. Last year, 47% of Canadian Tech Decision Makers were piloting automated coding tools versus 38% globally. Those pilots now appear to have converted to production.
Automated coding adoption (e.g. Claude Code, Cursor, Copilot in production systems) by Canadian Tech Decision Makers climbed from 36% to 71%, putting Canadian engineering teams at parity with the global rate of 66%.
Share of Tech Leaders with automated coding tools (Claude Code, Cursor, Copilot) in pilot vs production. Wave 2 (Jun. 2025) → Wave 3 (May 2026). Production bars rise to parity as pilot bars fall, showing pilot-to-production conversion.
Canadian automated-coding adoption in production climbed from 36% to 71%, reaching parity with the global rate (66%) as Wave 2 pilots converted.
| Measure | Wave 2 (Jun. 2025) | Wave 3 (May 2026) |
|---|---|---|
| Canada: in production | 36% | 71% |
| Rest of World: in production | 35% | 66% |
| Canada: pilot / testing | 47% | 19% |
| Rest of World: pilot / testing | 38% | 21% |
Share of Tech Leaders using citizen-developer vibe-coding tools (Replit, Bolt, Lovable), which debuted in Wave 3 (May 2026). Canada tracks the global rate at both stages.
Vibe coding (Replit★, Bolt, Lovable: citizen-developer-friendly tools that debuted in Wave 3) also reached parity: 44% Canada vs 48% globally in production.
| Indicator | Gap / value | Detail |
|---|---|---|
| In production | 44% vs 48% | Canada vs ROW |
| Pilot / testing | 31% vs 27% | Canada vs ROW |
The 23-point company-wide agentic AI implementation gap (44% vs 67%) is the largest single Canada-vs-ROW operational gap in the Wave 3 dataset.
Tech Leaders reporting agentic AI already implemented and expanding, or in the process of implementing. Wave 2: Canada n=100, ROW n=226. Wave 3: Canada n=72, ROW n=180.
Canada’s rate of agentic implementation increased from 22% in W2 to 44% in W3 (+22% WoW) while the ROW rate increased from 35-67% (+32%). The 13 point gap that Canada saw vs ROW in W1 has now increased to a 23 point gap in W3.
| Period | Canada | Rest of World |
|---|---|---|
| Wave 2 (Jun. 2025) | 22% | 35% |
| Wave 3 (May 2026) | 44% | 67% |
The deployment that does exist in Canada skews toward individual users rather than company-wide rollout. 32% of Canadian Tech Decision Makers say their organization has agentic AI in the form of individual users but no company-wide adoption, twice the rate of the rest of the world at 16%.
Tech Leaders reporting agentic AI in individual-user form only (no company-wide adoption) versus full company-wide implementation. Wave 3 (May 2026). Canada n=72; ROW n=180. Canada’s deployment skews to solo pilots at roughly 2× the ROW rate.
| Measure | Canada | ROW |
|---|---|---|
| Individual users only (no company-wide adoption) | 32% | 16% |
| Company-wide agentic AI implemented | 44% | 67% |
We expected to see signs that Canadian Tech Decision Makers were more skeptical about agentic AI. The data doesn't support that assumption. Canadian Tech Decision Makers want the same things from agentic AI as global peers, accept the same autonomy levels, worry about the same threats and use the same guardrails.
The pilot-to-production timeline shows this pattern: 35% of Canadian Tech Decision Makers take 7-12 months to ship an AI feature from pilot to production, versus 19% globally. Only 62% of Canadian Tech Decision Makers ship pilot-to-production in under six months, versus 75% globally. Canadian companies have the pilots, but are slower to ship them.
Time from pilot to production for AI features among Tech Leaders. Wave 3 (May 2026). Canada n=72; ROW n=179 (excludes “unsure”). Canada has the pilots but is slower to ship them.
| Indicator | Gap / value | Detail |
|---|---|---|
| Ship in under 6 months | 62% vs 75% | Canada vs ROW |
| Take 7–12 months to ship | 35% vs 19% | Canada vs ROW, nearly 2× |
| Ship in under 3 months | 17% vs 28% | Canada vs ROW |
Share of Tech Leaders by time from pilot to production for AI features. Wave 3 (May 2026). Canada n=72; ROW n=179 (excludes “unsure”).
Only 62% of Canadian Tech Decision Makers ship in under six months, vs 75% globally. The 7-to-12-month band is nearly 2× ROW (35% vs 19%).
| Measure | Canada | ROW |
|---|---|---|
| Less than 3 months | 17% | 28% |
| 3 to 6 months | 46% | 47% |
| 7 to 12 months | 35% | 19% |
| More than 12 months | 3% | 6% |
Canadian companies are roughly on par with the rest of the world in terms of the number of infrastructure components being used as well as foundation model types and AI techniques. The agentic infrastructure deficit is in two specific components, both associated with a higher rate of implementation of agentic AI. In our view, this pattern of infrastructure adoption is consistent with Canada's agentic implementation being at an individual rather than organizational level.
Canada is roughly on par with the rest of the world on the number of infrastructure components, foundation model types, and AI techniques used. The agentic deficit concentrates in two specific components, both linked to higher rates of agentic implementation. Wave 3 (May 2026).
| Measure | Canada | ROW |
|---|---|---|
| Large reasoning models (LRMs) | 32% | 49% |
| Durable workflow engines | 11% | 27% |
Canadian Tech Decision Makers also flag that integrating with legacy systems is a top barrier to agentic implementation (22% Canada vs 11% globally cite it as a top-3 agentic barrier).
Canadian Tech Decision Makers report AI is making code changes more likely to fail in production at twice the global rate (17% vs 8%).
Three operational-friction signals where Canadian Tech Leaders trail the rest of the world by roughly 2×. Wave 3 (May 2026). Canadian Tech Leaders want the same things from agentic AI as global peers, and the gap is in execution.
| Indicator | Gap / value | Detail |
|---|---|---|
| AI code changes more likely to fail in production | 17% vs 8% | Canada vs ROW, roughly 2× |
| Legacy-system integration a top-3 agentic barrier | 22% vs 11% | Canada vs ROW, roughly 2× |
| Take 7–12 months from pilot to production | 35% vs 19% | Canada vs ROW, nearly 2× |
We believe there is an opportunity for Canadian companies to build and ship new agentic products. Canadian Tech Decision Makers currently appear to be building new AI products at a significantly lower rate than their global peers (19% vs 31%). Since Wave 2, there has been a shift away from enhancing existing products with AI with Tech Decision Makers preferring either a mixed strategy (new products + enhancement of existing products) or using AI to predominantly build new products.
Share of Tech Leaders by how AI has affected their product strategy. Wave 3 (May 2026). Canada n=72; ROW n=180. Canada under-indexes on predominantly building net-new AI products (Canada +5 pts from Wave 2; ROW +12 pts).
Canada under-indexes on predominantly building net-new AI products (18% vs 31% ROW) and over-indexes on a mixed strategy.
| Metric | Predominantly new products | Equally new + enhancing | Predominantly enhancing existing |
|---|---|---|---|
| Canada | 18% | 64% | 18% |
| Rest of World | 31% | 54% | 15% |
We believe the next phase of AI value creation is increasingly going to be about net-new AI-native products and the agentic systems that power them. In our view, Canadian tech companies appear to be behind the global trajectory on this next phase. They’re past experimentation and ready to move from pilot to deployment.
We will be watching the Canadian benchmarks next year to see whether Canadian companies follow the global trend and turn agentic experiments into agentic systems or whether the infrastructure deficit and the slower shift to net-new product strategy creates a structural lag for Canadian tech companies.