Hierarchical Delegation
Hierarchical delegation pattern for multi-agent systems: orchestrating specialized agents with a coordinator for complex tasks.
Last updated
After this section you can
- Write a delegation brief that a worker can execute without the lead’s context, including an output contract
- Choose a model per role and estimate the token cost of a team against a single agent
- Tell router, planner and orchestrator apart, and prevent the coordination failures of parallel agents
13
Hierarchical Delegation: Multi-Agent Teams
An orchestrator splits a job into briefs and hands each one to a worker agent with its own context window, tools and model. It is the most powerful way to spread a big task across agents, and the most expensive way to be right.
Key idea
Multi-agent is a context-management strategy. Each worker burns its own window on the reading and returns a short report; the orchestrator keeps only decisions. You pay for it in tokens, several times a single agent, so use it for broad, parallel, high-value work.
One lead holds the decisions. Each worker reads a lot in its own window and hands back a little.