Tip
Casey Winters talks about founder bottlenecks: “When I was at Grubhub, we acquired a competitor. Even though they were largely in the same phase as us, they still operated like they were founding the company. Everything was going through the founder. Intuitively, I was like, ‘Oh, this is why we’re acquiring you, and not the other way around.’ Everything was not moving as quickly because everything had to go through the same person.”
Turns out AI organizational scaling works the same way.
You’re the VP of AI at a 200-person company. Every AI initiative still comes through you for approval—prompt reviews, model selection, vendor decisions, budget allocation. Your team submits proposals. You review them, make the call, send them back. Projects that should take 2 weeks take 6 because everything queues waiting for your review. Your calendar is 40 hours of AI decisions every week.
Meanwhile, your competitor hired a similar AI leader six months ago. They’ve shipped 3x as many AI features. How? Their AI VP delegated. Set frameworks for when teams can pick models (latency/cost/accuracy trade-offs), established prompt review guidelines, gave teams budgets and autonomy. They’re unblocked.
Younger leaders think delegation means losing control or quality dropping. They haven’t watched enough org structures over decades to see the pattern: founder/leader bottlenecks kill velocity. You have. You’ve seen it at three companies—the brilliant technical leader who couldn’t let go, became the constraint. The visionary CEO who needed to approve every decision, watched competitors ship faster.
This recognition—knowing when you’re the bottleneck—comes from seeing the pattern repeatedly. You know what “everything goes through one person” looks like in the metrics: long cycle times, high meeting load for leaders, low experimentation volume. That pattern recognition lets you delegate before you become the constraint.
The hard part isn’t the frameworks. It’s recognizing you’ve hit the phase shift. What got you to this point (hands-on technical decisions, deep involvement, setting the standard) is now what’s slowing you down. Experienced leaders recognize that shift because they’ve lived through it. Younger leaders are still in the phase where being hands-on is serving them.
Context
Casey Winters led growth at Grubhub, Pinterest, and was CPO at Eventbrite. At Grubhub, he watched a competitor get acquired specifically because their founder wouldn’t delegate—everything bottlenecked through one person while Grubhub scaled by distributing decision-making.
For experienced executives managing AI scaling, this isn’t theoretical—you’ve watched enough organizational bottlenecks over decades to recognize when hands-on leadership becomes a constraint. That pattern recognition comes from living through the full arc of scaling from 10 to 200+ person organizations multiple times.