Share
Report
59We're setting up AI-powered demand forecasting in Dynamics 365 Supply Chain Management for a retail distribution business. We have about 18 months of clean transactional history but the forecast accuracy is still quite low — especially for seasonal products and new SKUs with limited history.
Has anyone found a sweet spot for how much historical data the model needs before it becomes reliable? Also wondering if anyone has had success with manual baseline adjustments for seasonal spikes versus letting the model learn on its own over time.
Categories:
I have the same question (0)
All responses (
Answers (
