Public Preview: New Agent Loop capabilities in Azure Logic Apps

Announcing a new set of capabilities for Agent Loop in Azure Logic Apps, now in public preview. These updates expand how organizations can build, secure, and deploy agentic workflows across their enterprise environments enabling more flexible AI orchestration, governance, and collaboration. With these new enhancements, you can: * Use and manage MCP servers and tools: Leverage out-of-the-box Model Context Protocol (MCP) servers, bring your own, or even create custom MCP servers with user authentication for on-behalf-of (OBO) scenarios extending agentic workflows with secure, policy-driven access to enterprise APIs and systems. * Enable document-level access control: Strengthen governance as agents leverage knowledge sources like Azure AI Search. Using users’ existing Access Control permissions, agents now have scoped access to documents and data ensuring compliance and data security by design. * Bring your own model: Integrate any model into your agentic workflows, including locally hosted models or those behind private network boundaries. This flexibility allows organizations to maintain control over their model choices and data privacy requirements. * Deploy agents directly to Microsoft Teams: Publish conversational workflows to Microsoft Teams, allowing users to interact with agents natively within Teams for contextual, conversational experiences. Learn more: * [About building agentic workflows](https://learn.microsoft.com/en-us/azure/logic-apps/agent-workflows-concepts). * [Get started](https://azure.github.io/logicapps-labs/docs/logicapps-ai-course/overview/) with Agent Loop in Azure Logic Apps.