As enterprises continue deploying autonomous AI agents at scale, cybersecurity leaders are confronting new security and governance challenges. Traditional controls for managing human and machine identities are not designed for autonomous, decision-making systems with elevated permissions and access—and identity is emerging as the control plane for securing these new digital workers. This paper examines the current state of agentic AI adoption compared to actual controls deployed, according to the 104 financial and technology security leaders surveyed.
This paper explains why, defining what agentic AI is and is not, as well as addresses:
- Rate of adoption: how Agentic AI is moving from pilot to production and 6 key types of agents enterprises are already adopting
- Risks and mitigations: how autonomy amplifies familiar threats and introduces novel risks that organizations must mitigate
- Shifts in the security stack: why identity, governance, and data are becoming the new foundations for securing autonomous AI
- Governance and workforce implications: how accountability and workforce models are evolving in light of agentic AI
