// GLOSSARY -- AGENTIC AI

What is an Agent Swarm?

1 min read Updated

An agent swarm is a collection of AI agents that collaborate through decentralized, emergent coordination patterns — inspired by biological swarms — rather than centralized orchestration.

WHY IT MATTERS

The term 'swarm' in AI agents draws from swarm intelligence in biology (ant colonies, bird flocking). OpenAI's experimental 'Swarm' framework (now succeeded by the Agents SDK) popularized the concept of lightweight, composable agent teams.

In practice, agent swarms emphasize simplicity and composability. Each agent is a self-contained unit with instructions, tools, and the ability to hand off to other agents. There's no complex central orchestrator — coordination emerges from individual agent decisions.

For financial applications, swarms present interesting challenges. A swarm of trading agents might collectively drain a budget without any single agent violating its individual limits. Swarm-level controls are needed alongside individual agent policies.

HOW POLICYLAYER USES THIS

PolicyLayer provides fleet-level spending controls for agent swarms. Individual agents have their own limits, while a shared budget pool ensures the swarm's collective spending stays within bounds — preventing uncoordinated overspending.

FREQUENTLY ASKED QUESTIONS

How is a swarm different from a multi-agent system?
A swarm emphasizes decentralized, emergent coordination (agents decide locally). A multi-agent system is the broader category, which includes centrally orchestrated systems. All swarms are multi-agent systems, but not all MAS are swarms.
When should I use a swarm architecture?
When tasks are naturally decomposable, agents need to be independently deployable, and you want simplicity over complex orchestration. Swarms work well for customer service routing, parallel research, and distributed monitoring.
What was OpenAI's Swarm framework?
An experimental framework demonstrating lightweight multi-agent patterns using agents and handoffs. It was educational/experimental and evolved into the production-focused OpenAI Agents SDK.

FURTHER READING

Enforce policies on every tool call

Intercept is the open-source MCP proxy that enforces YAML policies on AI agent tool calls. No code changes needed.

npx -y @policylayer/intercept
github.com/policylayer/intercept →
// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.