See full schedule

Multi-agent AI systems can draft entire features overnight, but they cannot establish policy, evaluate risk, or accept responsibility for a failed deployment. In this talk, JetBrains CEO Kirill Skrygan dissects the systemic failure modes of unconstrained agent swarms. He outlines a pragmatic blueprint for enterprise governance: combining real-time developer supervision, automated verification pipelines, and provenance to turn volatile AI output into reliable, enterprise-grade software.

Taming the agent swarm

Speakers