Delinea Introduces Runtime Authorization for AI Agents

Delinea announced new runtime authorization capabilities for AI agents, claiming to be the first identity‑security platform that can enforce policies on actions inside an active session rather than only at connection time. The feature adds a unified policy engine that works across databases, SSH hosts, Kubernetes clusters, cloud consoles and Model Context Protocol servers, allowing organisations to block, allow or require human approval for each tool call an AI agent makes.

CEO Art Gilliland said the solution addresses a growing security gap as AI agents gain autonomous access to enterprise systems. It also provides just‑in‑time credential injection, automatic revocation after task completion, and a complete audit trail that distinguishes between human‑driven and agent‑driven connections, aiming to prevent authorised access from being misused.