A multicenter derivation-and-validation study paired with a prospective randomized trial reported that an explainable AI ensemble can predict multiple diabetic foot ulcer outcomes and improve outpatient recommendation quality. The approach emphasizes interpretability alongside performance, aiming to help clinicians triage risk and tailor follow-up strategies for patients with complex ulcer trajectories. If validated further in broader settings, the model could reduce variability in clinical decisions where outcome prediction is currently dependent on heterogeneous documentation and clinician judgment.