BlueVoyant, a cybersecurity company, has introduced its Microsoft Agent 365 Security Deployment Service. The service is designed to help enterprises discover, govern and secure AI agents within their Microsoft environments. As organisations continue to adopt AI agents, many are reviewing how existing cybersecurity measures apply to these technologies.
The increasing use of AI agents, including platforms such as Microsoft 365 Copilot, Copilot Studio and Azure AI Foundry, has created additional governance and security considerations. These agents, which can be classified as non-human identities, may be able to access sensitive data, interact with external tools and perform autonomous actions. Some organisations may not have a complete inventory of their AI agents or fully developed identity and access policies, which can increase security risks associated with AI agent deployment.
The BlueVoyant Microsoft Agent 365 Security Deployment Service is intended to establish Microsoft Agent 365 as a centralised framework for AI governance. The service integrates with Microsoft's security infrastructure, including Microsoft Entra and Microsoft Purview, to enable AI agents to be managed using identity and security controls similar to those applied to human users.
The deployment is structured as an approximately 90-day engagement during which BlueVoyant's specialists work within the client's Microsoft tenant. Key features of the service include:
Agent Inventory and Registry: Visibility into AI agents, including their ownership, accessible data and operational tools, as well as the identification of AI agents deployed without IT oversight.
Identity and Access Controls: Implementation of Conditional Access and Identity Protection policies through Microsoft Entra, applying authentication controls to AI agents.
Threat Detection and Response: Deployment of Microsoft Defender capabilities for AI, integrated with Defender XDR to support threat detection and response.
Data Protection Policies: Extension of Microsoft Purview Data Loss Prevention (DLP) and risk management controls to AI agents to help manage data access and sharing.
Client-Owned Configuration: Configuration remains within the client's Microsoft environment, with BlueVoyant providing deployment expertise while the client retains ownership and control after the engagement.
According to BlueVoyant, the service is designed to integrate with existing Microsoft deployments, including Microsoft 365 E7 and Agent 365, providing a structured approach to AI agent security and governance.
The Microsoft Agent 365 Security Deployment Service is now available. The service requires customers to have Microsoft E5 or E7 licensing together with the Agent 365 add-on.
