Connect to any LLMs, including open-source, On-Premises, or custom models to Document Automation using the Model connections feature in AI Agent Studio.

Prerequisites

Ensure that you have the required license and permissions. For detailed information, see Create and manage Model connections.
Task Licenses, roles, and permissions
View Model connections
Create Model connections
  • Attended bot runner license
  • AAE_Basic role
  • View Model Connections and Manage Model Connections permissions
(Optional) Create credentials and lockers All the permissions in the CREDENTIALS option

You will need these permissions if you want to store and use your credentials securely.
Important: If you are updating to v.39 or later release and using BYOK for generative AI providers in your existing learning instances, then perform the following steps:
  1. Create Model connections for the generative AI provider and then use it in the learning instance.
  2. Open the Extract data action of the learning instance and set None in the Additional settings option.

With Model connections managed in the Control Room, you can access these connections without having to configure them every time. See Model connections.

You can create Model connections and use them in learning instances created in Document Automation.

Procedure

  1. Create a Model connection. See Create and manage Model connections.
  2. Use the created Model connection in a new learning instance or an existing learning instance.
    • New learning instance
      1. Click Create new.
      2. On the Create Learning Instance configuration screen, click the Generative AI-driven data extraction drop-down menu.
        Note: It might take some time to display the Model connections in the drop-down menu.
      3. Select the created custom model in the Custom model connections section.
      4. Configure the required fields to create the learning instance. See Create a learning instance in Document Automation.

      Selecting Model connections when creating new learning instances

    • Existing learning instance
      1. Open the learning instance.
      2. Click the Learning Instance configuration (Learning instance configuration icon) option.
      3. On the Edit properties configuration screen, click the Generative AI-driven data extraction drop-down menu.
        Note: It might take some time to display the Model connections in the drop-down menu.
      4. Select the created custom model in the Custom model connections section.
      5. Click Modify.

      Selecting Model connections when editing new learning instances

    Note:
    • If you are using OAuth token in Model connection, the access token is set to expire in an hour by default and the refresh token is set to expire as per the user configuration. Ensure to replace the expired tokens to keep the Model connection active.
    • You will not be able to delete a Model connection that is being used by the current version of learning instances. To delete such a Model connection, you must first remove the association from the learning instances and then delete the connection.
    • If a Model connection was recorded in the version history of the test mode and the Model connection was deleted, an error is displayed when you try to restore to the version that was using the Model connection.