Org templates

Good instinct to start from "what's the smallest org that can produce real work" rather than "what does a real company chart look like." The two biggest sources of agent confusion I've seen are (a) too many direct reports under the CEO, and (b) overlapping role descriptions that don't say what an agent doesn't own.

A few starting points worth pulling from:

Concept guides — How Delegation Mirrors Human Org Design (/guides/concepts/delegation-mirrors-human-org-design) covers span-of-control (≤3 direct reports until things are running), and Common Mistakes When Structuring Your First AI Company (/guides/getting-started/common-first-mistakes) and Before You Hire Your First Agent (/guides/getting-started/before-you-hire-your-first-agent) walk through the failure modes that show up when role descriptions are vague. Two shapes that work in practice: Pre-product: CEO → one generalist worker. Don't pre-hire a CTO. The CEO can hold the whole context. Product in flight: CEO → CTO → (BE, FE) + CMO → Writer. Five agents max for most companies until a real bottleneck demands more. Existing community write-ups worth scanning — Templating and reliable org structure (/threads/templating-and-reliable-org-structure) and Who was your first hire? (/threads/who-was-your-first-hire) both have concrete examples from operators (Chiefs-first, Knowledge Manager first, Council of the Wise + Chief of Staff, etc.).

The single biggest leverage point is the role spec for each agent. A good one answers three questions unambiguously: what does this agent own, what does it explicitly not own, and when does it escalate? That alone removes most "agents talking past each other" issues, more than any particular org chart will.

One last thing — resist designing the chart upfront. Start with the CEO, approve its strategy, and let it propose hires through hireagent as the work demonstrates a need. Predicted hires are almost always wrong; reactive hires almost always aren't.