Automating inbound receiving documentation across 12 warehouse locations

Operating 12 warehouses for 40+ clients, the company received inbound shipments with inconsistent documentation. Some clients sent structured EDI, but most still relied on PDF purchase orders, emailed advance ship notices, and handwritten delivery receipts. Each warehouse had its own team manually entering receiving data into the WMS, leading to inconsistent item descriptions, missed line items, and a 72-hour average lag between physical receipt and system visibility. Inventory discrepancies between physical counts and WMS records averaged 4.2%.

DocumentIQ was deployed with a standardized field set: PO number, SKU, item description, quantity ordered, quantity received, lot/batch number, and expiry date (for food and pharma clients). Multi-row extraction handled POs with 50-200 line items. The team created client-specific annotation sets — since each client's PO format was consistent within itself but different across clients. The feedback loop was particularly valuable: warehouse staff corrected extraction errors during physical verification, and those corrections improved accuracy for subsequent shipments from the same client.

• 12 warehouses processing receiving docs through a single pipeline
• System visibility lag reduced from 72 hours to under 4 hours
• Inventory discrepancy rate dropped from 4.2% to 1.1%
• 93% extraction accuracy across all client PO formats after 2 weeks of feedback
• Receiving clerk headcount reduced by 30% across the network