# Discovered Assets

AI Discovery automates discovery and inventory of cloud providers, providing a centralized inventory and dashboard of AI assets, including in production and in development models, applications, datasets, and dependencies. This ensures end-to-end visibility of the AI pipeline across teams. Asset Relationships adds a relationships mapping that maps how assets are connected to each other in a top down hierarchical view.

- See [AI Discovery](/docs/products/console/asset_discovery_overview) for setting up asset discovery.


## View AI Asset Discovery

After your Databricks AI Asset Discovery integration is set up, discovered assets are added to the AI Asset Discovery page. Assets are discovered every 5 minutes.

1. In the HiddenLayer Console, click **Discovery** in the side navigation.
  - **Assets by Provider**: A pie graph that showcases a breakdown of assets by cloud provider.
  - **Asset Compliance**: A bar chart that showcases a breakdown of assets that are in compliance or non-compliance according to AI Supply Chain Security policy, over the last seven days.
  - **Assets by Type**: A bar chart that showcases a breakdown of discovered assets by the following types:
    - Model
    - Schema
    - Endpoint
    - Catalog
    - Agent



### Use View Filters

Use View Filters to help you find the AI assets you are looking for.

1. Click **View Filters**.
2. Expand the filter section to view available filters.

3. Select filters to apply to the AI Asset Discovery page.
4. Click **Show Results**.


## AI Asset Discovery Descriptions

| Data | Description |
|  --- | --- |
| Name | The name of the asset. |
| Provider | The cloud provider for the asset. Example: Databricks. |
| Service | The provider's service where the asset resides. Databricks examples: Model Registry, Unity Catalog. |
| Type | The type of asset. Example: Model. |
| Discovered On | The date and time the asset was discovered and added to the inventory list. |
| Custom Tags | User-defined labels manually assigned to an asset for organization and filtering. |
| Discovered Tags | Labels that are assigned within the cloud provider, then automatically identified and assigned to an asset during the discovery process. |
| Policy Status | Displays the AI Supply Chain Security Policy compliance status for an asset. |
| Region | The cloud provider's region. Example: eastus2 |
| Account ID | The name of the Databricks integration created earlier in [Setup Databricks AI Asset Discovery](#setup-databricks-asset-discovery). |
| Asset ID | A unique identifier for the asset. |
| Last Sync | The latest date and time the asset was verified to exist. |
| Actions | The actions available to an asset.The following actions can be taken, depending on the asset:Rescan AssetView Asset ScansView Relationships |


## Asset Relationships

Asset Relationships adds a relationships mapping that maps how assets are connected to each other in a hierarchical view.

The following Databricks assets are supported and viewed in a top down hierarchy:

**Databricks**

- Catalog
  - Schema
    - Model
      - Endpoint


The following AWS assets are supported and viewed in a top down hierarchy:

**AWS Bedrock**:

- Agent
  - Model


**AWS Sagemaker**:

- Endpoint
  - Model


To view a Relationship Mapping click on the relationship button:

HL Resources Actions
br
HL Resources Actions
br
**Important Limitation**: The relationship mapping currently displays a hierarchy from top to bottom, but not from bottom top.

- Example: Schemas with associated Models will have a relationship mapping, but an associated Catalog will not be displayed.
  - Further: Viewing a Model's relationship mapping will only display the model. To view the full relationship hierarchy, it is recommended to start at the Catalog level.


HL Resources Actions
## AI Asset Discovery and Viewing Model Scan Results

When a model asset is discovered, it is automatically scanned and it is possible to view model scan results for a discovered model asset.

For discovered models:

1. Click Actions under the HL Resources column for the target model.

2. Click **View Model Scan**.

3. The page will then be redirected to the Detection Summary of the scanned model.