← All territories
Territory 03 · Carrier audit

Territory · Substrate signal

03The demand-signal seam.

Google AI Search delivered +10% basket lift upstream. Afresh decides fresh, Gateway decides promo. The seam between basket-demand signals and the multi-modal supply plan is unowned.

Upstream signal · downstream decision · disconnected.

Customer demand · POS · Google AI search · loyalty
CADDIE · multi-source demand reconciliation · planning-layer feedback
Afresh (fresh demand)
Gateway (promo demand)
Blue Yonder Enterprise Supply Planning

The basket-demand signal that drove +10% lift never reaches the master plan. Caddie closes the loop.

Upstream wins · downstream gaps.

+10% basket lift · Google AI search

Albertsons used Google AI search to drive 10% bigger grocery baskets. Customer-facing AI is shipping; the supply-side feedback loop is not yet closed.

CX Today

Afresh + Gateway · domain-narrow

Afresh handles fresh; Gateway handles promo. Each is excellent in its lane. The cross-domain demand signal that moves between them is a planner-on-Teams reconciliation today.

4 Big Bets

Bengaluru build mandate

New Global Tech & Innovation Center charter spans cloud, security, data, AI; demand-signal reconciliation is plausible early scope.

BusinessWire

Three moves · one signal loop.

Multi-source demand sensing

Ingests POS + loyalty + search + promotion signal alongside Afresh + Gateway baselines, reconciles to the planning forecast at product-week-region grain.

Cross-domain anomaly detection

Flags products where fresh demand and promo demand are moving in opposite directions, surfaces drafted reconciliations.

Planning-layer feedback

Closes the loop back into Blue Yonder so the master plan moves on the new signal, not after the fact.

Where we have shipped this shape.

88-91%

Forecast accuracy across deployed customer base

8-15%

Logistics cost savings measured across deployments

~99%

Estimated delivery accuracy across multimodal deployments

Questions, fence, timeline.

  1. Does the +10% basket lift signal feed back into the supply plan today, or does it stop at the customer-experience layer?
  2. How are cross-domain demand conflicts (fresh up, promo down on the same product) reconciled today?
  3. Of the 22 distribution centers, which ones see the most demand-signal volatility and would benefit most from a closed-loop reconciliation pilot?

We would touch

  • Multi-source demand-signal ingestion + reconciliation to Blue Yonder.
  • Cross-domain anomaly detection (fresh ↔ promo).
  • Drafted reconciliation actions for planner approval.
  • Closed-loop feedback into the master plan.

We would not

  • Afresh replacement (fresh forecasting).
  • Gateway replacement (promo replenishment).
  • Blue Yonder Enterprise Supply Planning replacement.
  • Customer-facing search / personalization (Google territory).
90-day proof · one banner · one product line

First lane: one banner + one product line instrumented for cross-domain signal reconciliation, plus a 6-month historical backtest showing where the demand-signal gap cost forecast accuracy.

← Previous02 · Demand-signal seam All territories Next →04 · Vanning + load building