Before optimisation, someone has to walk the floor and count the steps.
Warehouse economics are decided by layout, slotting and labour sequencing long before any system is chosen. A picker walking an unnecessary twelve metres per line, forty times an hour, is a cost that no software licence addresses.
We map the physical work first and then decide what the system should enforce. Applying AI to routing or forecasting before the layout is right optimises a bad path very efficiently.
What the floor shows.
Slotting set once, never revisited
Fast movers sit at the back because that is where they were placed when the range was different.
Double handling on receipt
Inbound freight is put down and picked up again because the receiving process assumes a layout that changed.
Picking strategy mismatched to order profile
Single-order picking on a profile of many small orders, or batch picking where accuracy matters more than speed.
Labour planned by headcount, not by wave
The same crew size regardless of the order profile that day, so the shift is over- or under-resourced by lunchtime.
Accuracy fixed by counting more
Cycle counts increase instead of fixing the process that generates the discrepancy.
Routes built by habit
Delivery runs follow the same geography they did five years ago, with drop density never recalculated.
What the engagement actually includes.
Walk and measure
On the floor, on shift: distance per pick line, travel versus touch time, congestion points and dock-to-stock. Measured, not estimated from a system report.
Re-slot to the order profile
Velocity-based slotting matched to the real order mix, with a re-slotting cadence so it does not drift back within a season.
Fix the receiving process
Remove double handling by aligning the receiving flow to the current layout, which is usually the fastest single gain available.
Choose the picking strategy
Single, batch, zone or wave — selected against the actual order profile rather than inherited, and enforced by the system.
Labour to the wave
Crew size and zone allocation planned against the day's order profile, so the shift is resourced to the work rather than to the roster.
Routing and drop density
Delivery runs rebuilt around current customer geography and order frequency, with cost per drop attributed to the customers that drive it.
Floor first. System second.
The floor decides the economics; the system enforces what you decided. These are the components involved.
Ranges observed on Al Jawad engagements. Your targets are agreed in assessment, before the work starts.
Do we need a WMS first?
No. Layout, slotting and receiving flow are decided before any WMS matters, and most of the first-phase gain comes from them. The system then enforces what you decided.
Where does AI actually apply?
Slotting optimisation, wave planning and route sequencing, once the physical process is right. Applied before that, it optimises a bad path efficiently.
How long does the floor work take?
Two to three weeks of measurement across shifts, including a peak day. Estimating from system reports instead is how most projects get the profile wrong.
Can you work across multiple sites?
We prove the redesign on one site first. Standardising operating definitions across sites usually matters more than the software consolidation that follows.
Start a conversation.
Choose the one that fits where you are. None of them is a sales call. Each is an advisory conversation calibrated to a specific question.