You can integrate AI models and agents from different data sources in Data Catalog. These integrations are only available via Edge.
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.
Supported AI integrations
Collibra offers out-of-the-box integrations that ingest machine learning (ML) model and AI agent metadata from the following AI platforms:
- Anthropic AI
- AWS Bedrock AI
- AWS SageMaker AI
- Azure AI Foundry
- Azure ML
- Databricks AI
- Gemini Enterprise Agent Platform
- MLflow AI
- OpenAI
- SAP AI Core
- Snowflake Cortex AI
Beyond ingesting metadata, specific integrations also support AI model traceability and operational trust metrics, using data from the source platform.
AI model traceability
What Collibra can trace back to an AI model differs by integration. The integrations surface the relations that the source platform reports; Collibra itself doesn't compute lineage. The following table is scoped to links between an AI Model Version or AI Model Deployment asset and its data, not the structural relationships between AI assets themselves.
- Training data attaches to the AI Model Version asset.
- Output data attaches to the AI Model Deployment asset.
- The AI Model Deployment asset links to the AI Model Version asset it deploys.
| Integration | AI Model Version → training data | AI Model Deployment → output data |
|---|---|---|
| Anthropic AI | Not available | Not available |
| AWS Bedrock AI | Yes, for custom (fine-tuned) models only, linked to a Storage Container asset representing the training data S3 bucket | Yes, for custom-model and imported-model deployments, linked to the S3 location configured for model invocation logging in that deployment's region |
| AWS SageMaker AI | Not available | Yes, to a Storage Container asset with a File asset per S3 location |
| Azure AI Foundry | Yes, for fine-tuned models, from a Storage Container asset (with File assets) holding the fine-tuning dataset. The files include validation and training files | Not supported, the AI Model Deployment asset exists but does not have an output data link |
| Azure ML | Not supported, the AI Model Version asset exists but does not have any relations | Not available, no AI Model Deployment asset exists in this integration |
| Databricks AI |
Yes, linked to the Delta tables used as dataset inputs for the model version's source MLflow run. Notebook-based (code) datasets are linked instead to a Storage Container asset containing a File asset representing the notebook |
Yes, linked to the Delta table that Databricks uses as the endpoint's inference table |
| Gemini Enterprise Agent Platform | Yes, linked to a Storage Container asset (with File assets) found in the model's training pipeline | Yes, linked to the BigQuery table where the endpoint logs requests and responses |
| MLflow AI | Yes, only for datasets whose source is a Delta table | Not available, no AI Model Deployment asset exists in this integration |
| OpenAI | Not available | Not available |
| Yes, from a dataset artifact represented as a Storage Container asset | Yes, to a resultset artifact represented as a Storage Container asset | |
| Snowflake Cortex AI | Yes, from an already ingested Snowflake table | Not supported, the AI Model Deployment asset exists but does not have an output data link |
To enable these relations, you need to configure a field in the integration settings. For the specific setting for each integration, go to that integration's own documentation.
Operational trust metrics
Operational trust metrics track how an AI agent performs in production over time, using judge pass/fail rates and token consumption metrics.
Like attributes, these metrics are stored internally. Unlike attributes, they aren't accessible through the Collibra Platform APIs. They power the Agents dashboard, the Quality tab on AI Agent assets, and the Operational Health theme on the AI Trust Score.
To explore these metrics, in Settings, under AI Command Center, switch on Operational trust for AI agents.
Supported integrations:
| Integration | Operational trust metrics | Token consumption metrics |
|---|---|---|
| AWS Bedrock AI | Yes, pass/fail/error counts per AI Endpoint (agents) | Yes, per AI Endpoint (agents) |
| Azure AI Foundry | Yes, daily trust/quality score from Azure's continuous evaluation per AI Agent Version, only for agents built through Azure's Agents API on Foundry project resources | Yes, per AI Agent Version |
| Databricks AI | Yes, pass/fail/no-assessment counts, Knowledge Assistant and Multi-Agent Supervisor agents only | Yes, Multi-Agent Supervisor agents only |
| Gemini Enterprise Agent Platform | Yes, pass/fail counts post automatically once synced | Yes, requires the agent's own code to self-instrument a custom metric |
For more information, go to Operational trust for AI agents and AI monitoring for AI agents.
For the specific metrics, coverage, and limitations for each integration, go to that integration's own documentation.