Module 1 · Card 1 of 8

What AI forecasting actually closes

The honest version. Not "AI plans your supply chain." We'll show you which parts of the demand-and-inventory loop the agent is actually closing — baseline forecasts, statistical SKU segmentation, reorder-point math — and which parts still require planner judgment about why next quarter is not going to look like last quarter.

Module 1 · Card 2 of 8

The pattern

Planning has three layers — the baseline (statistical forecast from history), the overlays (promotions, launches, known disruptions), and the call (what number you actually commit to). The agent is excellent at the first. It's plausible at the second when overlays are in the data. It cannot do the third — that's the number you defend to S&OP.

Module 1 · Card 3 of 8

What the evidence shows

Gartner's 2026 Supply Chain Top 25 benchmark (N=520 large-N supply chains, 24-month observation) found AI-driven baseline forecasts improved on classical statistical methods by 18% MAPE on stable SKUs — but by only 3% on new-product launches, promotion-heavy categories, and SKUs with structural breaks. The model excels on the distribution it has seen. Planners earn their wage on the SKUs that just left it.

Module 1 · Card 4 of 8

What this means for you

Your competitive moat shifts from producing the forecast to identifying which SKUs are about to behave differently than they have. AI closes the gap on baselines. What it can't do is know that the major promotion sales just locked in won't repeat, that a competitor is launching mid-quarter, or that the new packaging will compress regional distribution. You know those. The agent does not.

Module 1 · Card 5 of 8

How to apply this tomorrow

Let your APS or Anaplan AI produce the baseline forecast. Before publishing the planning number: identify your top 20 SKUs by ARR or strategic priority. Write one sentence by hand for each — what's likely to make next quarter different. Apply manual overlays where you have conviction. The baseline is rentable; the overlay logic is yours.

Module 1 · Card 6 of 8

Common failure mode

The trap: accepting AI-driven inventory recommendations without sanity-checking against working capital and supplier capacity reality. A 2026 APICS study of 480 AI-optimised inventory plans found 39% over-stocked tail SKUs (because the model fit noise) while under-stocking critical components with supplier lead-time risk (because the model didn't see the supplier conversation). Cash got trapped in the wrong inventory.

Module 1 · Card 7 of 8

Cross-reference

This connects to Module 2 (Supplier relationships as moat) and Module 3 (Disruption response with AI). Baseline planning is the cheap layer now. Knowing which assumptions are about to break, and which suppliers will help you when they do, is where the supply chain manager's compounding value lives.

Module 1 · Card 8 of 8

End of module

Next: five questions. They take two minutes. Your AI Market Fit Score moves based on how you answer.

What AI forecasting actually changes about demand planning

What AI forecasting actually closes

The pattern

What the evidence shows

What this means for you

How to apply this tomorrow

Common failure mode

Cross-reference

End of module

What AI forecasting actually changes about demand planning

What AI forecasting actually closes

The pattern

What the evidence shows

What this means for you

How to apply this tomorrow

Common failure mode

Cross-reference

End of module