Model connections

Model connections act as the bridge between your automations and the power of AI. They allow you to connect to pre-trained AI models from hyperscaler vendors such as Amazon Bedrock, Google Vertex AI, Azure OpenAI, OpenAI and others.

These models provide the intelligence for AI Skills, enabling your automations to perform tasks like understanding language, analyzing documents, and making predictions.

The Model connections page is accessible from the left navigation under AI. It provides a centralized view of all generative AI model connections configured within your tenant. These connections are used within AI Skills for sending prompt queries and obtaining model responses.

The landing page displays a list of available model connections.

You can create a new model connection directly from this page using + Create Model connection. You need view or modify Model connections permissions to modify or perform other actions on existing connections through the options available for each list item (indicated by the vertical ellipsis icon). For more information on roles and permissions, see Rollen und Berechtigungen.

Key aspects of Model connections

  • Connecting to AI: Model connections establish a link between Automation 360 and the chosen AI model. This allows your automations to access the model's capabilities.
  • Powering AI Skills: AI Skills, the reusable AI components in AI Agent Studio, rely on Model connections to function. The connection determines which AI model will be used to execute the skill.
  • Centralized management: Model connections are created and managed within the Control Room, providing a central place to configure and control your AI integrations.