Infrastructure for Warehouse AI - Intelligent Slotting & Picking

Intelligent slotting and pick path optimization require formally documented warehouse operating rules: storage constraints by product class (hazmat segregation, temperature zones, weight limits per shelf), slotting criteria (velocity, pick frequency, ergonomic rules), and labor productivity standards. These must be current and findable so the AI generates recommendations that comply with safety regulations and physical constraints. When the system recommends relocating a fast-mover to a forward pick zone, it must query documented capacity and hazmat storage rules—not rely on supervisor judgment.

Slotting optimization and pick path learning require systematically captured pick history by SKU and location, order wave data, and labor productivity by picker and route. WMS captures pick completions with timestamps, enabling velocity calculations per SKU. Template-driven capture of physical product dimensions and handling requirements at receiving ensures the slotting algorithm has complete constraint data. Systematic capture of actual pick times per route segment feeds the path optimization model.

Pick path optimization and dynamic slotting require a formal ontology: Location entities with physical dimensions and zone attributes, SKU entities with velocity class and handling constraint properties, and OrderWave entities linked to required pick locations. Without explicit relationships—Location.Zone.CONTAINS.Slot WITH Constraint: SKU.WeightClass AND SKU.StorageTemp—the AI cannot generate valid slotting recommendations that respect physical and regulatory constraints. Graph-based location adjacency maps are required for optimal path routing.

Slotting recommendations and pick wave generation require API access to WMS (current inventory positions and location data), OMS/ERP (incoming order volumes and wave schedules), and labor management systems (staff availability). The AI must query current slot occupancy before generating slotting change recommendations—recommending a relocation to an occupied slot creates execution failures. API access to WMS enables real-time slot availability verification without manual position reports.

Slotting parameters and velocity classifications must update when product mix shifts, new SKUs are introduced, or seasonal demand patterns change. Event-triggered maintenance ensures that when a new product line is added, the slotting system incorporates its velocity and physical characteristics before the first receiving cycle—not after three months of suboptimal placement. Outdated velocity classifications cause the AI to slot slow-movers in prime picking positions, increasing travel time for high-frequency picks.

Warehouse AI requires API-based connections between WMS (location master, pick history, inventory positions), ERP/OMS (order volumes and product master), and labor management systems (productivity rates and shift schedules). The slotting and picking optimization assembles product velocity from WMS pick history, incoming order patterns from OMS, and physical constraints from WMS location master. Point-to-point API connections between these systems enable the AI to generate recommendations without manual data consolidation.