Organizing the AI Agents

Industry and academic leaders debated how AI agents will be organized, from open cross-company protocols to private company-managed systems.

Speakers said an open “internet of agents” is likely over time, but enterprises will probably start by testing agents inside controlled environments.

The discussion stressed that observability, orchestration, trust, and governance will be essential for managing agent swarms at scale.

The outcome could shape enterprise workflows, interoperability, and how much control organizations keep over autonomous AI systems.