Skip to main content

Security

Security brings together the controls that protect your organization's AI usage, from a posture overview down to the individual checks that inspect content in real time.

Posture and enforcement

The Security Center summarizes your current posture and points you at the settings that improve it. Radar scores that posture and turns risks and warnings across your systems into a prioritized, actionable feed. Guards are the enforcement layer: they inspect content at runtime and build time and decide whether to allow, warn, redact, require approval, or block.

Identity risk

User Risk Score puts a number on how risky each user is, imported from CrowdStrike Falcon Identity Protection or pushed in from your own systems. Guards can be narrowed to a score range, and users above a threshold can be refused outright.

Secrets and discovery

The Vault stores the credentials your MCP servers need, kept out of plain configuration. AI Discovery finds the AI tools already in use across your organization, including shadow AI, so you can bring them under governance.

On this page