Most companies are running their AI strategy like a waterfall project. And it's going to cost them.
I've spoken to a number of organisations over the past six months about how they're approaching AI. The pattern is almost always the same. They form a committee, draft a strategy document, plan a phased rollout, and aim for a big coordinated launch. Meanwhile, the technology shifts three times before the first phase is complete.
These are the same companies that claim to run agile. They operate sprints for their day-to-day delivery, but the moment something this significant arrives, they default straight back to waterfall thinking. Big plan, long timeline, late delivery.
AI doesn't reward that approach. It rewards small experiments, fast feedback, and constant adaptation. It rewards the company that gives a team two weeks to test a use case and report back, not the one that spends six months building the perfect AI policy before anyone touches the technology.
This is where 20% time stops being a Google-era novelty and becomes a genuine competitive advantage. The companies that will pull ahead aren't the ones with the best AI strategy deck. They're the ones creating space for their people to experiment, learn, and feed those learnings back into how the organisation works.
The gap between Company A and Company B won't be who adopted AI first. It will be who learned to adapt fastest.
If you're leading a team and your AI approach involves a committee, a twelve-month roadmap, and a phased rollout, I'd challenge you to try something different. Pick one team, give them a real problem, let them experiment for a fortnight, and see what comes back. That single experiment will teach you more than any strategy document.