Beyond Identity Launches Platform to Secure AI Agents
Security Boulevard, Tuesday, June 16th, 2026
Beyond Identity launches Ceros, a platform that secures enterprise AI agents by enforcing policy at the process level.
Beyond Identity launched its Ceros platform, announced at Identiverse 2026, enabling security teams to run AI agents securely by continuously enforcing policies at the process level within per-session authorization envelopes.
Teams can control which agents may launch, restrict the tasks and tools agents can invoke, and degrade or terminate violating sessions while generating alerts. Ceros logs every session including user and device, monitors external components like cloud-hosted LLMs and local programs, and lets administrators define per-agent and per-tool access controls.
Managed agents use device-bound passkeys that cannot be moved, and the platform defends against prompt injection and API key theft.