Tip You can set up this integration in the unified integrations experience (in preview). This feature is in public preview and applies to metadata, AI, SAP, and lineage integrations. In one guided flow, you select an Edge site, create the connection, add the integration, and configure it. To get started, in the banner on the Integration configuration tab, click Try the new Integrations page. For more information, go to About the unified integrations experience (in preview).
Azure AI Foundry is an enterprise AI operations platform that lets you build, train, and deploy generative AI models and AI agents. When you integrate Azure AI Foundry, you integrate the metadata of your AI models and AI agents from Azure AI Foundry to the Collibra Platform. The resulting assets are listed below:
- AI Base Model refers to base models used in your Azure AI Foundry workspace.
- AI Model Deployment refers to a deployment of an AI model in Azure AI Foundry
- Azure AI Foundry Model Version refers to a specific implementation of the model logic and parameters that serves as the audited model version before it is activated in Azure AI Foundry
- Azure AI Foundry Agent refers to an Agent created in the Azure AI Foundry Agent Builder.
- AI Agents are only available in the following regions:
- australiaeast
- eastus
- eastus2
- francecentral
- japaneast
- norwayeast
- southindia
- swedencentral
- uksouth
- westus
- westus3
- AI Agent Version refers to the version of an Azure AI Foundry agent. Azure AI Foundry agents always have a single version.
- AI Agent Tool refers to a tool type that an agent version can invoke.
- Azure AI Foundry Project refers to an Azure AI Foundry resource that manages the development, deployment, and management of AI models and agents.
- AI Monitor refers to the automated tracking parameters and alerting thresholds used to observe a deployment’s real-world performance and detect deviations from expected behavior.
- AI Endpoint refers to the access point for external systems.
For more information about Azure AI Foundry, go to the Azure AI Foundry documentation.
Supported data sources
The Azure AI Foundry integration has three independent tabs in the integration settings. You can configure one or more.
-
You can configure Edge connections and capabilities without an active AI Governance license. However, AI Governance must be enabled to harvest AI model and agent metadata, ingest corresponding AI assets in Data Catalog, and access the dashboards and features necessary to visualize and govern your AI landscape.
- You can integrate Azure AI Foundry only via Edge.
- To integrate Microsoft Copilot Studio or Microsoft Agent Builder assets, you need Collibra version 2026.09 or newer.
Azure AI Foundry
Use this tab to integrate AI models, deployments, projects, and related assets from Azure ML workspaces. For AI agents ingested from Azure AI Foundry, the integration also collects operational trust metrics from Azure's continuous evaluation feature (pass/fail) and Azure Monitor (token counts).
Microsoft Copilot Studio
Use this tab to integrate Copilot Studio agents, including GitHub Copilot-powered agents, from one or more Dataverse environments. These agents are ingested as the base AI Agent asset type, not the Azure AI Foundry Agent subtype used for Foundry agents.
Only Dataverse environments in the Microsoft public (commercial) cloud are supported. Sovereign clouds such as US government (GCC High and DoD) and China aren't supported.
Microsoft Agent Builder
Use this tab to integrate Agent Builder agents, built in Microsoft 365 Copilot, from your tenant. The integration finds all agents in your tenant, both published and draft. These agents are ingested as the base AI Agent asset type, not the Azure AI Foundry Agent subtype. This tab doesn't ingest Copilot Studio agents.
Only Microsoft public (commercial) cloud is supported. Sovereign clouds such as GCC, GCC High, and DoD aren't supported.
Helpful resources
Check out the A human-centered intro to AI integrations course in Collibra University.