video1.0<iframe src="https://www.loom.com/embed/81cca5f0ad69456fb7ccead173950bc1" frameborder="0" width="1280" height="960" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>9601280Loomhttps://www.loom.com9601280https://cdn.loom.com/sessions/thumbnails/81cca5f0ad69456fb7ccead173950bc1-3200cf637c09650d.gif304.157This Loom demonstrates Spellwether, a loan performance intelligence engine for multi-outcome prediction using competing risk survival, anomaly detection, macro scenarios, and explainability. It reviews synthetic data quality (92 out of 100) with 0.98 basic exception detection and shows data intelligence outputs including defect-free and mean quality scores, missing data by rates, validation rules, and train versus test divergence. For prediction, it compares a baseline and improved logistic model with seven features, reporting a ROC AUC of 0.87, and visualizes time-to-event results for prepayment, default, and competing risks over a 12 month horizon. It also covers anomaly detector coverage and exception handling, stress simulation by segment with downloadable reports, a local explainability layer with example cases, and a copilot and loan explorer with audit trail and reproducibility checks.