Practical advice gives you a suggestion on what to do that will be good in some contexts, and bad in others. A causal model allows you to still still make accurate predictions in multiple unforeseen contexts, allows counterfactual modeling and thereby the modeling of interventions. There's a world of difference between the two. And I find a world of difference between people who think in terms of one or the other.