Creating and managing inventories

Create inventories to store AI use cases and other assets.

An inventory is where you store and view AI use cases and other assets. Depending on your governance requirements, you can store all use cases and assets in a single inventory, or create multiple inventories for your governance efforts. For example, you might create an inventory for each department or business unit in your organization.

Creating inventories

You must have the manage_catalog role to create and manage an inventory. For more information, see Collaboration roles.

  1. Click AI governance > Inventories from the navigation menu to see a list of existing inventories.
  2. Click New inventory.
  3. Type a name and an optional description.
  4. Click Create.

Adding collaborators to an inventory

Inventories are meant to be collaborative so that multiple people that perform different roles can contribute to governance of key assets. To add collaborators to an inventory:

  1. From the settings menu for the inventory, click Set collaborators.
  2. Click Add collaborators to add collaborators individually, or by user group.
  3. Assign a role of Admin, Editor, or Viewer.
  4. Collaborators are added to the list for the inventory. You can remove or change the assigned access as needed.

Inventory roles

These roles apply to working with inventories.

  • Users with the Admin role can manage the inventory, add collaborators, manage access, and delete the inventory.
  • Users with the Editor role can create and edit AI uses cases in the inventory and assign assets to AI use cases.
  • Users with Viewer role can review AI use cases but cannot edit them or assign assets.

Managing assets in inventories

To view a list of inventories, click AI governance > Inventories from the navigation menu. To see the assets within an inventory, click the inventory's name.

You can also view all assets by clicking AI governance > All AI assets. The list shows all assets across all inventories. You can filter the list, for example by asset type or inventory.

From an inventory or from the All AI assets page, you can do the following tasks:

  • Import assets. For example, if you have ML models that are outside watsonx.governance, you can import them to an inventory.
  • Track an asset in an AI use case. Or untrack an asset.
  • Add tags to assets.
  • Delete assets.

For assets that you import, you can also do the following tasks:

  • If the asset is tracked in a use case, you can set the lifecycle phase of the asset.
  • You can copy an asset to a deployment space.

Importing assets to an inventory

You can import ML models and Python functions.

To import assets, do the following steps:

  1. Click AI governance > All AI assets.

  2. Click Import project/space.

  3. Select the local .zip file that contains the assets.

  4. Type a name for the asset and select an inventory. Specifying the asset name, selecting an inventory, and specifying the source of the asset

  5. Optional: Indicate the source of the asset to keep a record of where you’re importing it from.

  6. Click Import.

Tracking assets from an inventory

To track an asset in an AI use case, do the following steps:

  1. Click AI governance > All AI assets. Or click AI governance > Inventories and click an inventory.
  2. Click the options menu and click Track in AI use case.
  3. Select or create a use case. Optional: select a phase.

The asset is now tracked in the use case. If you use Governance console, the asset is synched with the Governance console.

Setting the lifecycle phase of assets in an inventory

To change the lifecycle phase of an asset, do the following steps:

  1. Click AI governance > All AI assets. Or click AI governance > Inventories and click an inventory.
  2. Click the options menu and click Set active phase.
  3. Click an option, and then click OK. For assets that you imported, you can choose any phase. For other assets, you can choose from the list of phases where the asset is available.

You can now see the asset in the use case. Notice that the asset appears under the inventory’s name.

Copying imported assets to a deployment space

You can copy an imported asset to a project or deployment space.

For example, when you’re ready to deploy a machine learning model in production, copy it to a production deployment space, and then create a deployment.

To copy an asset to a project or deployment space, do the following steps:

  1. Click AI governance > All AI assets. Or click AI governance > Inventories and click an inventory.
  2. Click the options menu and click Copy to.
  3. Select a project or a deployment space. Or create a project or deployment space.
  4. Click Copy.

A copy of the asset is now available in the deployment space.