CIO questions · Agents & automation
When should I use multiple agents?
The short answer
When there is genuine separation of expertise, context, responsibility or policy. Creating five agents to execute five sequential prompts is not good architecture. Multi-agent systems make sense when specialized capabilities can be independently discovered, composed, evaluated and governed.
How to decide
- Real separation of expertise, context, responsibility or policy → several agents
- Sequential prompts → one agent or a workflow
- Each agent discoverable, evaluable and governable on its own
The Onni principle
Processes should depend on capabilities, not specific agents.
The architecture behind it
- Capability registry
- Discovery
- Composition
- Evaluation