I met with a founder recently who had just raised a pretty significant round of venture capital. As a founder, the playbook for this moment has been the same forever. Raise the round, open up a lot of reqs, and get so big so fast that nobody can touch you.
Historically, escape velocity post pmf is driven by hiring really fast.
That playbook, while exciting, has a cost beyond the salaries you pay. More people means more complexity: HR issues, management layers, cultural impact. The first big round is especially tricky. You go from a small team of trusted, battle-tested folks who are close to the metal to a much larger group, and not everyone you add will share the work ethic and risk tolerance of the early crew.
Hiring en masse, imo, has always been a necessary evil. Time consuming, with lots of risk and guaranteed to bring complexity, HR issues, and misalignment.
So when this founder walked me through the hiring plan, I asked: do you really need to hire all those people? Should some of those boxes on the org chart be filled with AI teammates instead?
If you follow my writing, you know I’m pretty obsessed with AI teammates, and manage a group of them myself. An AI teammate is a colleague with a defined role, its own Slack account, email address, and phone number. You can train it, give it feedback, and scope its tool access to its role.
But when it comes to the topic of org charts and AI, how to staff is only part of the question. The other question is whether or not you have too many levels.
Some of the sharpest thinkers on AI and work believe the org chart should flatten. Jack Dorsey has said the modern org chart will collapse into three roles: the builder, the DRI, and the player-coach. One of his points is that the logic of capping a manager at around eight direct reports goes back to the Roman army and how information moved up and down the ranks, and no longer makes sense.
AI changes the math; not just because it seems more feasible to manage large numbers of AI teammates than humans, but also because a supercharged manager should be able to manage more humans AND more AI teammates.
So org design with an eye toward AI is two questions. Which roles can a teammate fill? And which layers can you eliminate?
I think we’re about to see a lot more AI teammates on org charts. Perplexity just launched an AI teammate service. Grokbot has a credible offering. We’ve been using Skydive at Legible, and have also built our own system. AI and org design is an area to watch closely.
I do know this - if we were raising a big round, I’d be less inclined to load up on hiring. Which begs the question - what do you need to raise a big round for?