Zenity, the AI security and governance platform purpose-built for AI agents, today expanded its security coverage across the Microsoft Ecosystem. Many enterprises deploy autonomous agents across this environment at their greatest operational scale. Security teams now gain unified visibility and governance for agents operating in GitHub Copilot, Microsoft Copilot Studio, and Agent Builder in Microsoft 365 Copilot. Furthermore, this protection covers Microsoft Copilot Cowork, Microsoft Foundry, and Claude Enterprise in Microsoft 365. Zenity also collaborates with Microsoft to contribute eligible runtime-security findings through Microsoft Sentinel. Consequently, this partnership enriches dynamic agent risk calculations across supported Microsoft Copilot Studio and Microsoft Foundry agents.
During a single workday, one employee’s tasks often pass through several different AI agents. For example, a coding agent proposes code changes inside GitHub. Meanwhile, a Copilot Studio agent executes a core business process. Additionally, Cowork sends emails on the employee’s behalf while Claude Enterprise builds financial forecasts in Excel. Each platform records this activity under separate identities and distinct data formats. Therefore, security teams must know who acted, what the agent did, and whether an analyst needs to investigate. Raw event feeds provide partial details, yet they frequently lack the context necessary to prioritize real security risks.
“Microsoft’s agentic capabilities now span a large portion of its portfolio and every one of those agents can act on company data,” said Ben Kliger, CEO and co-founder of Zenity. “Zenity is laser-focused on securing how AI agents act inside the enterprise. Today we’re expanding our coverage of the Microsoft AI ecosystem so organizations can delegate work to AI with confidence.”
Unified Alerting and Visibility Across the Microsoft Ecosystem
This expanded coverage across the Microsoft Ecosystem delivers immediate, practical benefits for enterprise security operations teams:
- Actionable Alerts in Microsoft Sentinel: Security teams now receive high-fidelity alerts directly inside Microsoft Sentinel instead of raw-log exports. Thus, analysts triage AI agent issues alongside their regular incident queue. For supported agents, eligible Zenity findings contribute runtime security evidence to Microsoft-calculated dynamic agent risk while preserving Zenity attribution. Microsoft combines these partner alerts with native signals, exposure data, and blast-radius risks to generate one dynamic risk level per agent. Eventually, Microsoft will surface this exact risk assessment inside Agent 365, Microsoft Defender, Microsoft Entra, and Microsoft Purview.
- Safe Adoption of Claude Enterprise: Organizations can deploy Claude Enterprise in Microsoft 365 with strict security controls active. Specifically, Zenity provides telemetry, security findings, and inline prevention across Office apps so security teams can safely approve its usage.
- Clear Attribution for Coding Agents: Security analysts can now trace GitHub Copilot coding-agent activity back to a specific person. Zenity provides telemetry, findings, and inline prevention for GitHub Copilot inside GitHub Enterprise. Even when enterprise settings obscure actor names, Zenity leverages Microsoft Entra ID to resolve true user identities.
Securing Citizen Developers and Core AI Infrastructure
Beyond developer tools, Zenity protects every layer of the Microsoft Ecosystem for business users and infrastructure engineers:
- Visibility for Citizen Developers: Agents built by citizen developers and IT teams remain fully visible. Zenity delivers telemetry and limited inline prevention for agents running on the Microsoft Copilot harness in Microsoft Copilot Studio. Moreover, Zenity extends telemetry and findings to custom agents that employees build using Agent Builder in Microsoft 365 Copilot.
- Monitoring Delegated Employee Actions: Security teams now gain deep visibility into agents acting directly on behalf of staff. Newly available Microsoft Copilot Cowork transcripts allow Zenity to generate accurate telemetry and findings regarding user-delegated tasks.
- Coverage on Modern AI Infrastructure: AI infrastructure teams stay fully protected on Microsoft’s newest agent platform. In fact, Zenity has completely migrated its integration to support the new Microsoft Foundry Agent Service.
“Microsoft Agent 365 is the observability platform for AI, helping organizations turn shared visibility into informed management, governance and security decisions,” Ray Smith, corporate vice president, Agent 365 Product, Microsoft. “Our initial integration with Zenity will bring runtime security evidence from supported Microsoft AI agents through Microsoft Sentinel to enrich dynamic agent risk alongside Microsoft’s native security signals, helping customers better understand agent activity and prioritize investigations.”
As Microsoft continuously advances its autonomous capabilities, Zenity actively extends security coverage across the Microsoft Ecosystem. Through Microsoft Sentinel, eligible Zenity findings enrich Microsoft-calculated dynamic agent risk alongside native runtime alerts and posture signals. Ultimately, this unified approach across the Microsoft Ecosystem gives customers one clear risk level per agent to prioritize critical investigations.
To explore how Security Operations Centers (SOC) play a crucial role in defending against modern cyber threats, read our latest SOC News.