The Control Gap: The Thesis
- Jonathan Olsen
- 6 days ago
- 2 min read
Every big firm is now selling the same pharma launch. Agentic. Sense-and-respond. "Launch as a learning system." They are all describing the same thing, and none of them is in the room when it fails.
The short version:
The big-firm launch point of view has converged, and the platforms behind it are real.
Launches still drift anyway, because the gap was never structure. It was control.
The question isn't which agentic platform to buy. It's who holds the loop closed when the market moves.
If you read IQVIA, ZS, BCG, and Deloitte back to back, the launch point of view has converged on one picture: real-time data, agentic course-correction, and a launch that learns as it goes. The diagnosis is right and the platforms are real, so this isn't a complaint about the technology. It's about what the consensus leaves out.
The frameworks already exist, and teams already follow them, and launches still drift while those same teams find themselves reacting a beat too late. ZS put a number on it this year: only about 40% of AI pilots ever reach scaled deployment. What stands between a launch team and the outcome usually isn't ambition, it's execution.
I find it telling that the loudest realists in this whole conversation are the AI champions themselves. Eli Lilly's Diogo Rau came from Apple to run one of the most aggressive AI agendas in pharma, and he is also the first to wave off the idea that AI rescues a timeline on its own. The people closest to the technology are the ones telling you it won't save you.
The real problem was never the absence of structure. It's the absence of control. Structure is a plan and a dashboard, while control is a person turning a signal into a decision before the window closes, and a platform can show you the signal but it cannot make the call, own the trade-off, or move a cross-functional team by Friday.
So the question for your next launch isn't which agentic platform you should buy. It's who is holding the loop closed when the market moves, and whether they actually have the authority to act.
This is the first in a short series on the difference between a launch that's instrumented and a launch that's controlled.
On your last launch, what surfaced quickly, and how long did it take you to actually act on it?
Comments