Inside Veza’s AI Agent Security for Microsoft Copilot Studio: A Technical Deep Dive

Veza’s AI Agent Security for Microsoft Copilot Studio automatically discovers published bots through its existing Microsoft integration and represents each as a first-class AI agent identity in a unified Access Graph. It maps bots’ inbound access controls, including Dynamics 365 users, application users, teams, and Dataverse security roles, alongside their outbound connections to topics, actions, knowledge sources, external APIs, Power Automate flows, AI models, enabled capabilities, and public MCP servers. This identity-first model is intended to let security and IAM teams trace what each bot can access, who can invoke or manage it, and how bot permissions and connected tools create risk across Microsoft cloud environments. Veza supports graph and natural-language queries, access reviews, risk analysis, and alerts for events such as broad role assignments, new MCP connections, unapproved model changes, or enabled web browsing and file-upload capabilities. The offering is generally available to existing Veza Microsoft integration customers and focuses specifically on bots built in Copilot Studio rather than Microsoft 365 Copilot.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.