Skip to content
Control Center

Run an orchestration

This guide shows you how to turn one big goal into a whole-team plan that runs itself. See Orchestration for how it works.

When to use an orchestration

Section titled “When to use an orchestration”

Reach for an orchestration when the ask is bigger than one agent and you don’t already know the exact steps. For a single task, message an agent in a channel; for a process you can spell out by hand, build a pipeline template.

Step 1: Ask for the plan

Section titled “Step 1: Ask for the plan”

Ask your orchestrator (usually the CEO agent) for the goal, in plain language:

Research the competitive landscape for an offline-first meeting recorder and write a one-page positioning doc.

The orchestrator returns a structured proposal: the roles it needs, the work broken into sub-tickets, an optional research and discussion phase, a synthesis step that produces the deliverable, and a budget.

Step 2: Review the proposal

Section titled “Step 2: Review the proposal”

Read the proposal before approving. In particular, check:

  • Roles: each is either an existing agent or a new hire the orchestration will create on approval. Swap any role for an agent you trust, or change a hire to an existing agent.
  • Sub-tickets: the work DAG. Drop anything you don’t want.
  • Budget: the hard spending ceiling for the whole orchestration. Tighten it if you want a guardrail.

Because the proposal is a typed object, the app can validate it (well-formed roles, declared output schemas, and so on) before anything runs.

Step 3: Approve

Section titled “Step 3: Approve”

Approve the plan. This is the only decision you make. From here a deterministic materializer (a pure function, no LLM) turns the proposal into a real pipeline and creates the scaffolding:

  • hires any new roles (and records them)
  • groups the roles into a team
  • files the sub-tickets under a project
  • starts the pipeline run that drives it all

The status moves to approved, then executing.

Step 4: Watch it run

Section titled “Step 4: Watch it run”

The sub-tickets execute as pipeline steps; if you enabled a discussion round, each role posts a position first. Progress surfaces where you already look:

  • the parent ticket the orchestration opened against
  • the shared channel the team talks in
  • the pipeline run detail, with its usual per-step timeline and cost rollup

Once the sub-tickets are done, the status moves to synthesizing and the synthesis step produces the deliverable onto the parent ticket, then completed.

Stop it early

Section titled “Stop it early”

If the plan is going sideways, cancel it. Cancellation tears down the in-flight work. A failed orchestration (a hard error, every sub-ticket failing, or the budget exceeded) lands in failed with an error message.

Section titled “Related guides”