Reading Canada's draft AI for All strategy as a delivery partner.
Eight pillars, one honest read: what the draft gets right, where the delivery gap is, and what we'd build first if we were one of the partners on the other side of the procurement form.
01 · What the draft actually proposes
The draft federal strategy — reportedly titled AI for All — is the most concrete national AI plan Canada has put on paper. Stripped to its load-bearing pieces, it commits to eight things:
- Free AI literacy training at population scale by 2031.
- Lifting business AI adoption from ~12% today to over 50% by 2030.
- Hundreds of millions added to the AI Compute Access Fund so SMEs can use Canadian compute.
- Sovereign infrastructure — a public supercomputer, Canadian AI data centres, a secure national data platform, a Health Sector Data Space.
- An expanded Canadian AI Safety Institute and a new Canada Trusted AI Certification program.
- AI “missions” in health, energy, natural resources, agriculture, transportation, manufacturing, robotics, and government services.
- Up to 90,000 youth AI work placements through SWPP, Canada Summer Jobs, Skills for Success, Mitacs ADOPT, and AI+X.
- New privacy, online-safety, transparency, and election-protection laws, including watermarking and child-data protections.
The framing is the right one for Canada: adoption first, sovereignty second, safety as the wrapper. It treats AI literacy and SME readiness as infrastructure work, not communications work. That alone is a meaningful shift.
02 · Where the delivery gap is
We've worked with hospitals, R1 universities, and global non-profits long enough to know what happens between a strategy document and a deployed system. Three honest observations on the draft:
Literacy at scale needs delivery partners, not curriculum
Free AI literacy is the right ambition, but the binding constraint is delivery capacity. Libraries, school boards, and community organizations rarely have spare technical staff. A national curriculum without a national delivery model becomes a PDF nobody opens. The strategy needs a bench of Canadian-owned partners who can actually run cohorts, train trainers, and measure the lift in plain English.
SME adoption ramp is steeper than the slide suggests
Going from 12% to 50% business AI adoption in seven years means moving the middle of the market — not the early adopters. Those SMEs don't need more AI hype. They need a five-axis readiness assessment, a 90-day plan, and someone who will tell them when AI is the wrong answer. The promised federal readiness tool is welcome; the field also needs partners who can sit across the table and walk an owner through the result.
Safety, trust, and certification need enforcement, not just programs
The draft is strong on intent for the AI Safety Institute, the Trusted AI Certification, and modernized privacy law. The open question is how the enforcement teeth land in procurement. The fastest way to make “safe AI” real is to require it in federal RFPs — and to fund the audit work that lets organizations actually meet the bar.
03 · What we'd build first
If we were scoping the first 12 months of delivery under this strategy, here is the order we'd run it:
A national AI Readiness baseline for SMEs
Before any AI adoption funding flows, every applicant SME runs a fixed-scope readiness assessment — five axes, scored against a benchmark. That single instrument turns the 12 % → 50 % target into a measurable curve and gives program officers something defensible to fund.
A literacy delivery bench Canadian institutions can call
Train-the-trainer kits for libraries, school boards, and community orgs, delivered by a vetted bench of Canadian-owned partners. Bilingual, measurable, and accountable on adoption — not enrolment.
Sector pilots in health and government services
The two missions with the highest ratio of public benefit to deployment risk are health (Health Sector Data Space, clinical operations) and government services (citizen-facing workflows, internal automation). Both benefit from senior, audit-grade delivery from day one.
A certification-prep audit pattern
The Trusted AI Certification will only matter if organizations can get ready for it without hiring a Big-4 audit team. A reusable, lightweight audit pattern — sized for hospital IT, university registrar offices, and municipal CIOs — is the unlock.
04 · Who should be partnering now
The draft will move. What won't move is who delivers the work: hospital CIOs, university CTOs, municipal IT directors, ministry innovation leads, and the program officers who will write the cheques. If you're in one of those seats, the most useful thing you can do this quarter is to identify your two or three preferred Canadian delivery partners before the RFPs land.
We're one of them — senior-led, Canadian-owned, based in Toronto and Ottawa, with a track record of shipping products at scale and a service stack that already maps to the strategy's pillars.