Delinea Launches Runtime Authorisation for AI Agents
Delinea has unveiled new runtime authorisation capabilities for AI agents, expanding its identity security platform to enforce policies on agent actions before they are executed. The company says it is the first platform to provide runtime policy enforcement inside active sessions, addressing a growing security gap as AI agents gain autonomous access to enterprise systems.
As AI agents increasingly interact with production databases, SSH hosts, Kubernetes clusters, cloud environments and Model Context Protocol (MCP) servers, Delinea argues that existing identity security solutions typically verify access only when a session begins or audit activity after the fact. According to the company, these approaches fail to control what AI agents actually do once connected.
“The security industry spent years solving credential theft, but AI agents have introduced a new problem: authorized access doing unauthorized things,” said Art Gilliland, CEO at Delinea. “You can have perfect credential hygiene and still have an agent tear through your production environment in milliseconds. Recording what happened or revoking access after the fact doesn’t stop that – enforcing policy on the action as it runs does. The perimeter has moved to inside the session, and no one has figured out how to address that until now.”
The new runtime authorisation capabilities provide a unified policy engine across multiple protocols, enabling organisations to apply consistent security controls and maintain a complete audit trail regardless of how AI agents connect to enterprise resources. The platform also supports just-in-time credential injection, ensuring AI agents never retain standing credentials. Access is scoped to specific tasks and credentials are automatically revoked once those tasks are complete, reducing the risk of credential exposure.
Delinea’s platform evaluates every tool call within an AI session individually, allowing, blocking or escalating requests for human approval before execution. The company says this enables organisations to contain the potential impact of compromised or misbehaving AI agents by enforcing policies on each action rather than only at session initiation.
The platform also differentiates between human-driven and agent-driven connections, applying policies specifically designed for AI agents before access is granted. Every database query, SSH command and tool invocation is linked to a named identity, providing organisations with a comprehensive audit trail for investigation, compliance and accountability.
The new runtime authorisation capabilities are available immediately as part of the Delinea Platform, allowing organisations to extend Zero Standing Privilege principles to AI agents without requiring additional infrastructure.
