Rubrik launches Agent Identity to control AI agent access in real time

Rubrik has launched Agent Identity, a new AI-based capability designed to manage and control AI agents' access and permissions at runtime, addressing what the company describes as a growing blind spot in enterprise agent deployments.

Agent Identity is delivered as an expansion of Rubrik's existing Agent Cloud platform and introduces just-in-time, scoped tokens that replace broad static credentials - minting short-lived access per tool call rather than granting standing permissions.

Agent Identity sits alongside three other pillars of the Rubrik Agent Cloud platform: Agent Observability, which monitors agents and model context protocol (MCP) servers at runtime; Agent Runtime Security, which uses Rubrik's SAGE governance framework to enforce policies and detect errors; and Agent Rewind, which the company said can undo destructive autonomous actions.

The architecture works by routing every MCP tool call through three checkpoints before execution - behavioural analysis via SAGE, access policy enforcement, and identity verification - blocking unauthorised write or modification actions before they occur.

"Agents are no longer just synthesising information, they are acting on behalf of employees and using the access models we built for humans. Static credentials were never designed for autonomous actors," said Dev Rishi, general manager of AI at Rubrik.

"Agent Identity lets enterprises decide who can do what with agents and enforces it at each tool call, at the moment of action, with scoped, short-lived access."

Agent Identity integrates with Okta and Microsoft Entra ID, extending existing enterprise identity structures to machine actors without requiring new directories. Rubrik said the approach uses On-Behalf-Of federation to scope access by user and group to specific MCP servers and tools.