Is your AI working? Prove it.
Magnitude North is a senior AI advisory firm working with companies in Canada and the US. We measure what your AI costs and what it returns, and we make those numbers defensible to your board, your auditors, and your regulator. Then we name the one thing to fix first.
- Measured. A real number on what AI costs and returns. One your CFO can sign off on.
- Defensible. Governance that satisfies the people who can say no: boards, regulators, insurers, and procurement teams.
- Senior. Two principals do the work themselves. No juniors, no pyramid.
01 · Where most companies are
The bill is coming due
Most companies adopted AI without outside help. Statistics Canada found that only 30% of large businesses using AI brought anyone in, and 11% of small ones. Now they're finding out what going it alone costs.
You can't prove it's working.
Over 80% of AI projects fail, twice the rate of ordinary IT projects. When RAND asked the experts why, 84% pointed to leadership and problem definition, not the technology.
Source: RAND
Nobody can say what it returns.
98% of organizations now actively manage AI spend, up from 31% two years ago. The FinOps Foundation sums it up: “Is your AI providing value? No one can answer that question yet.”
Sources: FinOps Foundation, State of FinOps 2026 · Deloitte CFO Signals
Your employees are ahead of your policies.
36% of Canadian workers used generative AI on the job this year, against 19% of businesses that have formally adopted it. That gap is shadow AI, and it now has a number.
Source: Statistics Canada, Labour Force Survey March 2026 · Q2 2026 business survey
The deadlines now have dates.
OSFI E-23 binds every federally regulated Canadian financial institution on May 1, 2027. US AI-in-hiring laws are already live in Texas and Illinois, with Colorado's following on January 1, 2027. The EU AI Act reaches any firm whose output is used in the EU.
Sources: OSFI · state statutes · Official Journal of the EU
All four are the same problem: activity without measurement, and adoption without governance. Someone has to close the gap between what leadership believes and what the data shows. An outsider can say it plainly. That's the job.
02 · How we help
Four engagements
Every engagement starts with a fixed-price diagnostic: the number, the gap, and where to start. The readout is yours whether we implement it or not.
AI Value & Cost Audit
For leadership teams that can't answer “what is AI returning?” We inventory your spend, instrument usage, and deliver a number your CFO can sign: what it costs, what it returns, what to cut, what to scale.
- Duration
- Three to four weeks
- You get
- A spend and usage inventory. A cost-and-return number with the method behind it. A ranked list of what to cut, keep, and scale. A one-page memo written for your board.
- Price
- Fixed, quoted on the call
Model Governance & Compliance Readiness
A model inventory, a risk framework, and policies that hold up when your board, your auditor, or your regulator asks. HIPAA-bound AI use in healthcare, where we have done the work. New rules are arriving, from US state AI-in-hiring laws to ISO/IEC 42001 in enterprise procurement, and the first job is knowing which ones reach you.
- Duration
- Four to six weeks
- You get
- An inventory of every model and AI system in use. A gap assessment against the rule that applies to you. Policies and a risk framework your auditor can read. A remediation plan split into what must be done before the deadline and what can wait a quarter.
- Price
- Fixed, quoted on the call
Data & Foundations Readiness
Roughly 80% of enterprises say data constraints are what hold AI back. The returns live in the foundation, not the model, so that is where we look first.
Sources: Cloudera Data Readiness Index 2026 · Statistics Canada
- Duration
- Three to four weeks
- You get
- A data readiness diagnostic. The constraints ranked by what they block. A costed roadmap an implementation team can work from.
- Price
- Fixed, quoted on the call
AI Engineering Effectiveness
You rolled out AI coding tools. Did it work? 84% of developers use them, and 29% trust them. We instrument the rollout, measure honestly, and fix the standards, review practices, and reporting. We ship software with these tools ourselves.
Source: Stack Overflow Developer Survey 2025
- Duration
- Four to six weeks
- You get
- An instrumented baseline of tool usage and outcomes. A clear verdict on whether the rollout worked. Updated standards, review practices, and reporting. A monthly readout your CTO can take to the executive team.
- Price
- Fixed, quoted on the call
What we don't do: rent out headcount, build chatbots, or say yes to all seven of your agent ideas. Part of the job is scoping the one workflow that is ready, and saying no to the other six.
Canadian company? A diagnostic is the first step of a BDC LIFT plan, and the work may qualify for an NRC IRAP contribution. Here is what's actually available.
03 · Why us
Not the Big 4. Not a freelancer platform.
The principals do the work.
No leverage pyramid, no juniors learning on your invoice, no unreviewed AI output shipped under our name. Two senior practitioners with decades of shipped software between them, and both are in every engagement.
We're not selling build hours.
AI collapsed the price of execution, and we'll be the first to say so. What stays valuable is judgment, measurement, and accountability. Those are the things you can take to a board.
You keep everything we build.
Frameworks, inventories, dashboards, prompts, training material. No proprietary lock-in, and no retainer you can't walk away from. If the readout says you don't need us, it says so.
04 · Fit
A good fit if
- You want senior judgment directly. We take a small number of engagements at a time, usually with companies of roughly 20 to 500 people. If you need a team on site, we'll say so.
- You're already using AI. Our clients aren't asking “should we?” They're asking “is it working, and can we defend it?”
- You're in Canada or the US. We're Canadian, and we've worked under HIPAA and data-residency rules. Half our work is in the US.
05 · The first call
What happens on the call
Thirty minutes. No deck, no pitch, and nothing to prepare.
- Ten minutes on what you're running. What you've deployed, what it costs, and who is asking the hard questions: your board, your auditor, your regulator, or your own CFO.
- Ten minutes on the gap. Where the number is missing, where the governance is thin, and which deadline applies to you.
- Ten minutes on the first move. We name it, and it's yours whether we work together or not.
- If we're not the right fit, we say so on the call and point you somewhere better.
Book a 30-minute call Or email [email protected]. We reply within one business day.
06 · The principals
Who you're hiring
Matt Inglot
Principal · Strategy, governance & deliveryMatt founded Tilted Pixel and has spent twenty years delivering client systems and running lean, senior-only engagements. At Magnitude North he leads strategy, governance, and delivery: the accountability side of the practice.
Jason Hanley
Principal · Measurement & engineeringJason has run Syllogistic Software for over two decades, building and maintaining production systems for businesses that depend on them daily. At Magnitude North he leads engineering effectiveness and cost instrumentation: the measurement side of the practice.
AI should sharpen your judgment, not replace it. We automate the work, never the accountability.
07 · Questions
Before you book
What do you need from us?
Three to five 30-minute interviews with the people who own the spend, the models, and the outcomes, plus the invoices and usage exports you already have. A diagnostic needs no access to your systems.
How much of our team's time does it take?
Plan on four to six hours across your leadership team over the whole engagement. We do the work between the interviews, not during them.
What about confidentiality?
An NDA before we see anything. Your data never goes through an AI tool you haven't approved. The readout is yours, and we don't name clients without permission.
Why not the Big 4, or our vendor?
The Big 4 bring a pyramid and a twelve-month program. Your vendor has a product to sell. We bring two principals, a fixed price, a few weeks, and nothing to sell you afterwards.
Why not hire someone in-house?
Hire if the work is continuous. A diagnostic is a point-in-time need, and a senior AI lead takes six months to a year to find and ramp. If you do hire, they inherit the instruments we leave behind.
What if the readout says we don't need you?
Then it says so, in writing, and you keep it. Some first calls end there. That is the point of a diagnostic that isn't tied to an implementation.
Do you implement, or only advise?
Both. Every engagement starts with a diagnostic, and implementation is a separate, scoped decision you make after reading it. We ship software with the same tools we assess.
Remote or on-site?
Remote by default, across Canada and the US. On-site for kick-offs and board sessions when it matters. We're based in Alberta and British Columbia.
Is there government funding for this?
In Canada, sometimes. BDC LIFT lends companies with over $1 million in revenue the money to adopt AI, and NRC IRAP makes non-repayable contributions to smaller firms. We've written up what's actually available. We don't write applications or take a percentage.
Ready to know where you actually stand?
Book a 30-minute call. We'll look at what you're running, find the gap, and tell you where to start.
Book a 30-minute call Free. No deck, no pitch, just a working conversation.