A Pharmaceutical Research study trained Random Forest and XGBoost classifiers on particle size, shape, surface energy, and silica dry-coating data to predict powder blend flowability category. The accuracy holds up as a screening tool and breaks down in specific, predictable places.