If you run a business, you’ve heard the AI advice by now: Get your data house in order first.
You don’t have time for that. And you don’t need too.
14% of Canadian organizations say they “use” AI.
2% report ROI.
That gap isn’t a data problem. It’s the advice we keep giving:
❌ “Get your data house in order first.”
❌ “Lock down privacy and security across the org first.”
❌ “Become AI-ready before you start.”
It sounds responsible. It’s actually why so much “use” turns into so little return.
78% of non-adopting firms can’t even see how AI applies to their business.
That’s not a data problem.
That’s a workflow imagination problem.
The “data house in order” narrative is bogus. It creates a readiness trap. 12 to 18 months frozen before the first useful win.
What’s old is new.
Start with the business outcome. Revenue, margin, cycle time, customer experience, conversions, gm/week… Whatever business outcome you need to improve.
Pick the workflow that drives it. Not the org. Not the data lake.
One workflow.
Map it and then rewire it with the team.
Put assistive or agentic AI right at the constraint.
Scope data quality, privacy, and governance to that workflow. Proportional.
Not enterprise-wide.
A meeting-notes-to-CRM agent doesn’t need a data governance council.
An agentic workflow touching regulated systems of record absolutely does. And you’ll know, because you started with the workflow.
Data and governance don’t disappear. They move from universal gate to process-specific decision.
The companies winning with AI right now aren’t the ones with the cleanest data.
They’re the ones who understood their constraint and rewired around it.
Carney’s “AI for All” strategy wants business adoption to go from 12% to 60% by 2034. A $200B GDP bet.
If we want 60%, we have to stop gatekeeping adoption with fictious prerequisites that don’t apply to most use cases.
Outcome. Workflow. Rewire. Scoped readiness.
That’s the order. In that sequence, AI pays back – and fast.
In any other order, it becomes another organizational change that fails.
Entrepreneurs are too busy for the data first approach. And it is overkill to say you need it in place first. Sure you will need it later (and if you have it in place, great!) but most organizations we work with have a bunch of barely baked back end systems held together with duct tape and a few good people.
Data Integrity is not the necessary first step.
