Causal Analysis utilizing CausalML

We often talk about correlation vs causation in theory but while implementing Data Science solutions towards solving business problems not much influence is given to validating causation amongst independent and dependent features. Conventional Machine Learning methods identify patterns in existing data to make predictions and they always result in retrieving some underlying patterns even if they are not real and factitious. The assumption is these patterns are same in training, testing, validation data and deployed environments. However, if these patterns change for some reason, ML models fail. The reasons could be numerous like a distribution shift, external unexpected factor etc. Causal Machine Learning helps by defining treatment condition with respective control data to make causal inferences guiding machine learning models to pay attention to cause and effect relations.