What is a multi-agent system and when should you use one?
LLM & AILevel: IntermediateMulti-Agent Systems
What is a multi-agent system and when should you use one?
Hints
- Identify the stage: pretraining, adaptation, generation or evaluation.
- Name the inputs, constraints and a case where the approach fails.
Answer
A multi-agent system has multiple specialized agents collaborating on a shared task, each with a defined role.
Use multi-agent when:
- Task requires parallel execution (research multiple topics simultaneously)
- Different subtasks need different capabilities (one agent for code, one for data analysis)
- Task is too complex for a single agent's context window
- Checks and balances are needed (critic agent reviews generator agent's output)
Avoid multi-agent when:
- Single agent can accomplish the task — multi-agent adds coordination overhead
- You need deterministic, auditable output — more agents = more unpredictability
Source and licence: JustInCache & contributors — Awesome AI Interviews · Apache-2.0