
Artificial intelligence (AI) models demand as a probability range rather than a single forecast, then sizes each buffer against the real shape of uncertainty for that reference. It folds in lead-time variability, seasonality and recent demand shifts, and recomputes often enough that buffers track reality instead of lagging it. Cover concentrates where risk and business impact are highest and drains away from stable items, which is what modern AI-driven demand planning brings to inventory decisions.