Receiving through dispatch on one stock ledger
A warehouse runs on where things are. We make every movement a scanned, timestamped record made by the person who moved it, so the system and the floor never disagree.
Pickers keep a second list because the system cannot be trusted.
Stock is found in a location the system says is empty.
Dispatch cut-offs are missed because picking is released one order at a time.
A count means shutting the building for a weekend.
Inbound and putaway
Receipt against purchase order or ASN, inspection hold, and directed putaway to a bin that exists.
Picking and packing
Wave, batch and zone picking released against dispatch cut-offs, with pack verification.
Dispatch
Load building, carrier manifests, labels and proof of dispatch back to the order.
Counting and audit
Cycle counting by velocity and value, with variance investigation as routine work.
Included in every WMS scope
Location and bin naming scheme Barcode and label standard Device and network survey Putaway and pick strategies Wave and cut-off configuration Multi-client stock separation Count and variance policy Floor training and go-live support
The seams we design first
What AI does in a warehouse, specifically
It looks at movement history and suggests. A supervisor still releases the work.
Slotting suggestions
Fast movers proposed for closer bins based on real pick history, with the travel saving stated.
Short-pick prediction
Lines likely to short flagged at wave release, while there is still time to source them differently.
Count targeting
Counts aimed at the locations where movement history says the ledger is most likely wrong.
Labour forecast
Expected work by shift from the open order book, so the roster is set on demand rather than habit.
ERP implementation checklist
Thirty-one decisions that determine whether a rollout lands, most of which come before anyone configures anything. Readable in full, no email required.
Tell us where your stock count drifts.
That answer usually names the movement that is not being recorded.