LendWise uses a gradient-boosted binary classifier to approve or decline personal loan applications. The model was stable at launch, but over the last quarter loan defaults increased even though offline monitoring shows only a modest drop in aggregate discrimination metrics.
| Metric | At Launch | Current | Change |
|---|---|---|---|
| AUC-ROC | 0.84 | 0.78 | -0.06 |
| Precision (default class) | 0.61 | 0.58 | -0.03 |
| Recall (default class) | 0.74 | 0.52 | -0.22 |
| F1 Score | 0.67 | 0.55 | -0.12 |
| Log Loss | 0.41 | 0.56 | +0.15 |
| Brier Score | 0.18 | 0.24 | +0.06 |
| Approval Rate | 63% | 68% | +5 pts |
| 90-day Default Rate on approved loans | 3.1% | 5.4% | +2.3 pts |
The risk team suspects concept drift after LendWise expanded into gig-worker and thin-file customer segments. You need to determine whether the issue is threshold drift, calibration drift, feature distribution shift, or a true change in the relationship between features and default risk.