Create an Azure DevOps project: This is the foundational step. You need a project within your new Azure DevOps organization to house your pipelines and settings.

Install the Machine Learning extension for Azure Pipelines: This extension is crucial. It adds the "Azure ML Model Deployment" task and, more importantly, enables the Model Artifact trigger. This allows the Release Pipeline to listen to the Azure ML Model Registry and trigger automatically when a new version is registered, satisfying the automation requirement.

Create a service connection: Before the pipeline can interact with your Azure resources (the ML Workspace), you must authorize Azure DevOps to access your Azure subscription via a Service Connection (specifically an Azure Resource Manager connection).

Create a release pipeline: Finally, you configure the Release Pipeline. You set the Artifact source to the Azure ML Model (enabled by the extension) and define the deployment tasks using the service connection.

Microsoft Learn - Set up MLOps with Azure DevOps: "To use Azure Pipelines with Azure Machine Learning, you need to install the Machine Learning extension... You also need a service connection to your Azure subscription... Then you create a release pipeline to deploy your model." Source: Use Azure Pipelines with Azure Machine Learning - Prerequisites & Steps.

Microsoft Documentation - Azure Machine Learning extension for Azure DevOps: "The extension allows you to trigger a release pipeline when a new model is registered in the Azure Machine Learning workspace." Source: Azure DevOps Marketplace - Machine Learning extension.