An interpretable model exposes its reasoning in a form people can work with. That might be a small decision tree, a linear model with meaningful coefficients, a rule list, or another structure where the path from input to output is inspectable.
Interpretability is especially valuable when a model output becomes a business decision. If stakeholders need to approve, challenge, or operationalise the output, a transparent model can be more useful than a black-box model with slightly better offline metrics.
