Feature importance from tree models helps practitioners:
Select an answer to reveal the explanation.
Short Explanation and Infographic
Importances guide debugging and pruning — not automatic causal proof.
Full explanation below image
Full Explanation
Importances are useful but can be biased (correlated features, encoding). Not pure causal estimates.