I'm building my AI Chief of Staff

I’m finally building my AI Chief of Staff, and I’m documenting it here to find and share the best practices for multi-agent orchestration.

In my first few weeks at Zero, I’ve been building with AI basically all day. At any one time, I typically have 5 Claude Code sessions chugging along. While this has led to incredibly fast iteration to impact, it’s also led me to consistently context switching in order to keep my agents chugging along in the right direction.

So this week, I’m taking a step back and building out an AI Chief of Staff - one agent I interact with the most to delegate, plan context, plan token usage, and advance my projects without me having to bounce between context every 5 minutes.

I’m curious if anyone’s had success with this and if so how they built it?