Introducing RAMS AI Vision — Intelligent Warehouse Perception
Set the fleet, the shift pattern and the improvement you think is available. The model shows what those assumptions are worth over a year — so the conversation starts from your operation rather than from a headline figure.
Every value below is yours to set. RAMS supplies the measurement, not the improvement figures.
Step one
Five numbers. Everything after this is worked out from them, so it is worth getting these close.
Racked locations across the site
Trucks running on a normal day
Per hour, including on-costs
Step two
Three are on to start with, because most sites can check them against their own records. Turn on the rest only where the problem is real — a total built from problems you do not have is not a number you can take to anyone.
Impacts get logged when they happen, with the truck, the operator and the bay — instead of surfacing at the next inspection, or as a collapse.
Sensor Stack · AI CameraReported and unreported
Your assumption: 40 impacts a year at ₹25,000 each, 40% caught earlier.
The racking inspection duty stays yours. What changes is that the evidence is gathered from the air, out of hours, across every bay rather than a sample.
AirScanCounting the building from the air, out of hours, instead of walking and lifting to it.
AirScanOr the cycle-count equivalent
Your assumption: 12 counts a year over 10,000 positions at 0.4 min each, 70% released — ₹450 an hour.
Hours that go on finding a pallet the system has in the wrong place. The fix is knowing where it is, not counting again.
Location Intelligence · AirScanEveryone, not one person
Your assumption: 30 hours a week at ₹450 an hour, 50% removed.
Minutes per shift returned to each truck — from access, from not hunting for a charged machine, from not repeating a job. Enough of them add up to a truck you did not have to hire.
Sensor StackPeople and vehicles in the same space, watched in the moment rather than reviewed afterwards. The camera is a layer over your guarding, not a replacement for it.
AI CameraA void under a rack leg is cheap while it is still a void. The expensive version is the one that announces itself through the surface.
FloorScanStep three
Where a default came from published industry work, it is listed here and linked. Where it did not, it is a neutral starting point and nothing more — not research, and not a RAMS figure.
Rack repair cost
Puts average direct repair at about $750 an incident, and models roughly ten incidents a year.
Steel King — The Hidden Cost of Pallet Rack DamageImpact frequency
Reports that most impacts are never formally logged, and that large operations can see fifty a day.
Damotech — Top 5 Ways Forklifts Damage RacksIncident cost
Indirect costs of a workplace injury commonly run several times the direct claim.
OSHA — Business Case for Safety and HealthForklift claims
Puts a single forklift injury claim at roughly $38,000–$41,000 in direct cost.
Voxel — Forklift Accident StatisticsCounting effort
Describes cycle counting as continuous and labour-intensive, with 95–99% accuracy the working benchmark.
NetSuite — Inventory Cycle CountingNon-productive time
Estimates 30–40% of labour hours go on travel, waiting and searching rather than on the work itself.
UNEX — Maximizing Labor EfficiencyDefaults with no source above — inspection and counting effort, MHE minutes, the cost of a slab repair — vary too much between buildings for a published figure to mean anything. They are starting points, put there so the fields are not empty. Replace them with your own before you take the total anywhere.
An MHE productivity assessment establishes the baseline these assumptions are tested against, and states the window, the measures and the exclusions it was measured under.