Enterprises racing to deploy AI agents lack governance frameworks to control them,

Organisations worldwide are deploying AI agents into production environments at a pace that has outstripped the security and governance frameworks needed to control them, exposing enterprise data to a category of risk that conventional security tools were not built to address.

“These are probabilistic entities - you can’t exactly predict what they will do,” Arvind Nithrakashyap, co-founder and CTO of Rubrik told CSO in an interview. “They are not deterministic software running within the organisation. Agents are fast, they can cause 10x the damage in 1x the time of humans.”

The warning comes as enterprise adoption of AI agents accelerates sharply — and as the protocol infrastructure connecting those agents to corporate data accumulates vulnerabilities faster than security teams can track them.