India, March 11 -- As enterprises race to operationalise agentic AI, automation is no longer confined to efficiency gains or isolated use cases. Autonomous agents are beginning to operate inside security operations, identity systems, and enterprise workflows, observing signals, making decisions, and acting at machine speed. This shift promises scale and resilience, but it also introduces new risks that traditional governance and security models were never designed to handle.

Ganesan Karuppanaicker, Chief Technology Officer at Birlasoft, argues that the real challenge of agentic AI lies not in capability, but in control. As multi-agent systems become embedded across enterprise environments, organisations must rethink security, trust, and ...