Approve your first objective

An objective is how a request becomes work your team actually runs, rather than a single answer in chat.

The lifecycle

1. Proposed — your Chief of Staff turns your request into a scoped objective and shows it to you. Nothing has run yet.
2. Approved — you accept it. A CEO agent decomposes it into two to seven tasks and delegates each to a department chief.
3. Running — tasks move across the board. Agents work; some pause for your approval.
4. Delivered — finished work lands as versioned deliverables you can read, revise or revert.

Steps

1. In Chat, read the proposal card. Check the scope, not the wording.
2. If the scope is wrong, say so in a sentence — it re-scopes rather than starting over.
3. Click Approve.
4. Open Objectives and click into it to watch the task board.
5. Use the Deliverables tab once work starts landing.

You're done when the objective shows tasks on its board and at least one has started.

Steering a running objective

You are not locked in once it's running. You can pause it, retry a failed task, stop it entirely, or read the narrative — a plain-English account of what happened and why, which is usually faster than reading the task log.

Gotchas

• Running an objective needs a funded plan. Chatting is free; scoping and running work is not. If you meet that wall, the card that appears takes you straight to the fix — see Fund your AI.
• An objective with no tasks yet is still planning. Give it a moment before assuming it stalled.
• Deliverables are versioned. Revising one does not lose the previous version, so you can ask for changes freely.