AI adoption sits in a strange place. Everyone has the tools. Almost nobody has the returns. Over the last few weeks I have spoken with dozens of business leaders, and the picture is remarkably consistent: teams stretched thin between shipping products, running R&D, keeping customers happy and driving sales. The real work sits in the “I would like to, but I just don’t have the bandwidth” category.
The Copilot trap
So they start where most companies start. They buy a batch of Copilot licences and hope productivity flows to the bottom line. Most confess it simply adds cost. The deliverables are noticeably better — SOPs, pitch decks, marketing content — but better documents do not become financial gain. Why? People now finish in six hours what used to take eight. The working day is still eight hours, and the extra two quietly fill with other things.
This is not just anecdotal. McKinsey finds 88% of companies use AI, but only 39% see any EBIT impact. BCG is starker still: only 5% create substantial value at scale.
The expendable booster
So is the Copilot phase pointless? Not quite. Think of the first fifty years of space exploration. We accepted as normal that enormously expensive first-stage rocket boosters would be built, used once, and left to crash into the ocean. Nobody questioned it until SpaceX proved you could land them and reuse them.
The Copilot phase is the expendable booster. It gets you off the ground. The real value comes when you stop accepting the waste and start redesigning the mission, which is where AI agents and genuine workflow transformation begin.
Intentional beats patient
What separates the companies that get there? BCG found struggling companies spread thin across six or more use cases, while leaders focus on about 3.5 and generate twice the ROI. McKinsey’s high performers are three times more likely to pursue transformative workflow redesign rather than incremental improvement. This is exactly the gap between buying tools and rebuilding how work is done, and it is why we start every engagement with the process, not the model.
The companies that win are not the most patient. They are the most intentional. How many of your AI tools are actually changing how work gets done, rather than just making existing work faster?