Use what-if counterfactuals when you need to: Provide solutions to users and determine what they can do to get a desirable outcome from the model. Use causal inference when you need to: Identify the features that have the most direct effect on your outcome of interest. https://learn.microsoft.com/en-us/azure/machine-learning/concept-causal- inference?view=azureml-api-2 https://learn.microsoft.com/en-us/azure/machine-learning/concept-counterfactual- analysis?view=azureml-api-2