Omdia Research: AI Agents & Identity Security: Risks, Controls, and Insights

The Omdia study (commissioned by BeyondTrust) surveyed 400 IT and cybersecurity professionals in North America to understand how enterprises are adopting, managing, and securing AI agents.

Teams face mounting pressure to help accelerate AI projects, while ensuring secure and well-governed AI agent deployments. The Omdia study found that, on average, organizations are in the middle of 22 distinct AI agent projects, many involving multiple agents. AI agents, which act autonomously or semi-autonomously, substantially expand the attack surface and pose unique security challenges, as compared to human and machine identities. Their delegated permissions, dynamic behaviors, and interactions within multiagent systems further increase oversight complexity.

• Where and how AI agents are being adopted across the enterprise, spanning IT operations, data analytics, security, DevOps, IAM, customer support, and more

• The most concerning risks and security gaps for IT / security practitioners

• Why identity security is critical to governing agentic AI, and accelerating its secure, enterprise-scale AI adoption

Download the report now to gain clear insight into how IT / security practitioners perceive and are addressing the evolving agentic AI landscape.