With a 480% increase over the last 6 months in Fortune 500 companies, AI agents are by far the fastest growing identity type in the enterprise. At the same time, they often end up with more access permissions than the task requires, and the security review that's supposed to catch that can lag well behind deployment.
An over-permissioned agent surfaces sensitive content to anyone who prompts it, whether that person should see it or not. Microsoft Agent 365's registry confirms an agent exists and flags unusual behavior but it can't tell you what that agent is configured to reach, which is the gap that Cyera closes.
Cyera Agent Guardian now delivers data security and posture risk signals for agents directly into Microsoft Agent 365, including all aspects of agent configurations, capabilities, connected tools, access paths to sensitive data, drift from policy, and more. These signals show up on the Agent Registry page in the Microsoft 365 Admin Center, alongside Microsoft's own runtime detections, giving admins one view of agent risk signals that includes both Cyera's data context and Microsoft's behavioral detections.
Complementary Intelligence for AI Agent Risk
Understanding AI agent risk requires more than one dimension of visibility. Microsoft Agent 365 provides organizations with a centralized control plane to discover, govern, and monitor AI agents across their environment, combining signals from Microsoft security services to identify unusual behavior and operational risk.
Cyera Agent Guardian complements that view with deep data intelligence. It continuously evaluates each agent’s data access, permissions, connected tools, posture, policy compliance, and exposure to sensitive data, helping organizations understand not only what an agent is doing, but what it’s capable of accessing.
Together, customers gain a unified view of AI agent risk that combines Microsoft’s behavioral insights with Cyera’s data security context. Every finding includes both operational and data context, allowing data and security teams to investigate, prioritize, and remediate risk from a single workflow. The result is faster, more informed decisions and greater confidence to scale AI across the enterprise.
"Agent 365 provides organizations with a control plane for discovering, governing, and managing AI agents across their environment. Through its integration with Agent 365, Cyera delivers risk signals that appear directly in the Agent Registry alongside risk signals from Microsoft, adding visibility into what data each agent can access and where it's exposed. This gives joint customers a single place to view and investigate agent risk signals across both Microsoft and partner security solutions."
— Nirav Shah, CVP, Agent 365 & Agent Cloud
How it works
Cyera's detections are policy-driven. When a policy triggers, Cyera opens an Issue and pushes a matching risk event to the Agent 365 Risk Service through Microsoft's API in near real time.
Authentication runs through Microsoft Entra ID using OAuth 2.0 client credentials, Microsoft's standard model for third-party risk collectors. Every request is token-authenticated and scoped to the customer's tenant.
Each event follows the Agent 365 third-party risk schema:
Cyera’s policy severity maps to riskLevel (low/medium/high). Policy name and description become riskEvidence. Issue status drives status. Findings land in the Admin Center ready to act on.
Looking ahead, Cyera and Microsoft will close the loop end to end: Cyera will create new signals with POST, update status with PATCH, and remove resolved risks with DELETE. When an Issue is remediated in Cyera, the Agent Registry will update automatically, reflecting current state instead of a stale snapshot. That is the full lifecycle both teams are building toward over the next few months.
Turn it on today
The integration is available for Cyera customers in preview. Reach out to your Cyera account team to enable it for your tenant. If you are not a customer yet, book a personalized demo today.
Learn more about the Cyera and Microsoft partnership.
Get your AI Security Assessment
.png)
.png)
.svg)