Data Shows That Infrastructure Is Everything
StatCan’s study on AI adoption and productivity in Canadian firms found that overall, AI adopters show a 16.8% productivity premium. Here’s the critical nuance, though: that figure falls to 10.2% controlling for pre-existing productivity, and to a statistically insignificant 5.1% once complementary capabilities (data analytics, cloud, ICT training) are controlled for.
“This underscores that the productivity premium observed earlier may not be attributable to AI alone,” said researchers Jiang Li and Huju Liu. “Instead, it reflects the critical role of complementary investments in enabling AI’s productivity-enhancing potential.”
I mentioned earlier that nearly two-thirds of business leaders report a lack of cohesion between their people, processes, and tech. Dig deeper into the data and you start to find some culprits: In most organisations (89%), fewer than half of roles have been updated to reflect AI capabilities. Meanwhile, 37% of time saved by AI is being offset by time spent on rework, and 1.5 weeks’ worth of employee time per year is spent fixing AI outputs. You can easily see where that productivity premium is being lost.
Overcoming this all-too-common pitfall requires a thoughtful approach designed to diagnose and address three critical gaps:
- The role gap: Jobs haven't been redesigned around what AI and agents now do.
- The skills gap: Training isn't reaching the people absorbing the most rework.
- The governance gap: No one owns the blended workforce end to end.