Anthropic · CCAR-P
Multi-agent systems and decomposition
CCAR-P · Lesson 3 of 18 · 14:41
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the Claude Certified Architect – Professional exam, on the exam guide version 1.0, effective July 2026. In this lesson: orchestrators and parallel subagents, when multi-agent systems pay off and when a single agent is better, the token cost of agents, precise delegation, resuming, tracing, and gradual deployments, and sequential, parallel, and run-time decomposition. Includes three original practice questions.
Preview the topics
- Harborline’s negotiation assistant works through one claim at a time, and each step depends on everything said before. A colleague proposes five subagents. What should the architect recommend?
- Harborline’s orchestrator gives three subagents the same instruction: research flood claims. Their reports overlap, and some topics are missing. What fixes this?
- Harborline reviews forty reinsurance treaties each year. Each treaty can be analyzed on its own, and a final comparison needs all forty results. How should the task be decomposed?