One of my AI experts only ever names problems. She is built to mark problems, never fix them or decide, and is told never to invent one. Here is what that looks like on a real page, with three failures she caught in one pass.
My AI team works while I sleep. For four days in September it did not, and nothing told anyone. This is what it takes to trust a night shift that only exists while the laptop is awake.
Four AI experts, each with one job, one memory, and a list of things they are never allowed to do. The never list turned out to matter more than the job.
I came home from a conference in Amsterdam and committed a plan to an empty repo. Three weeks later my terminal has a chief of staff, a team of experts, and a rulebook written in incidents.
If you use an AI coding agent daily, you have a toolbox of prompts and commands you've never once cleaned up. Here's the refactor, how to spot when yours needs it, and what happens when a tool outgrows the repo.
My terminal has an org chart now. Fable does judgment, Opus ships, Sonnet handles the small stuff, Haiku does the summaries, and the cover art is outsourced.