AGV Forklift Systems for UK Pharmaceutical Distribution

An AGV forklift is a driverless industrial truck that moves pallets under software control, navigating by on-board sensors rather than a seated operator. Across Great Britain, being struck by a moving vehicle killed 24 workers in 2025/26 — second only to falls from height — according to HSE's annual fatal injury statistics. For an Operations Director running an MHRA-regulated distribution centre, the sharper quarterly problem is evidential rather than physical. Every pallet moving from goods-in to quarantine, from quarantine to released stock, and from released stock to despatch must be accounted for. On a manual fleet that account depends on a driver remembering to scan at the right moment, and in Q3 that discipline is stretched thinnest by agency cover and holiday rotas. An AGV forklift removes the manual scan from the internal move: location history is generated by the machine that performed the move rather than reconstructed from a handheld afterwards.

Why the evidence gap opens in a regulated DC

Pharmaceutical distribution in the UK operates under wholesale dealer licensing overseen by the Medicines and Healthcare products Regulatory Agency. The expectation is simple to state and hard to sustain: at any moment you should be able to say where a batch is, what status it holds, and who moved it there. Manual handling meets that through human discipline. A driver collects a pallet, drives it, sets it down, and scans — so the record comes from a second action, separate from the move itself.

That separation is where the gap opens. Scans drift late when a queue builds at the despatch bay, and early when a driver scans on approach to save a walk back to the terminal. Neither is malicious and both are common, particularly on sites around Magna Park and DIRFT where seasonal volume is absorbed through agency labour. The result is a location record that is approximately right and occasionally wrong — and in a regulated environment, approximately right generates deviations and held stock.

There is a quieter cost alongside it. Rack-leg strikes are rarely logged as compliance events, yet still consume quality time and trigger re-inspection. The HSE's workplace transport guidance treats pedestrian–vehicle interaction as the core control problem in warehousing; in a licensed DC it also threatens the product being carried.

Lever one: automate the three fixed lanes first

The operational lever is to resist automating everything and take the three lanes that are fixed, repetitive and evidentially expensive: goods-in to quarantine, quarantine to released stock, and released stock to despatch staging. These moves have stable origins and destinations and need no operator judgement. They are also the moves quality assurance cares about most, because each is a status change as well as a transfer.

Begin by measuring how many of those moves happen per shift and how long each takes door to door. Most sites find the lane count smaller than expected and the elapsed time far longer, because manual moves are interleaved with replenishment. Automating fixed lanes converts a variable-priority task into a guaranteed-throughput one: the vehicles work the lane continuously whether or not the shift is short-handed. Keep exception handling and judgement calls with your experienced operators. The objective is not a dark warehouse; it is removing the moves whose value lies in being done identically every time.

Lever two: make the fleet manager your evidence layer

The technical lever is to treat fleet software as a compliance system, not a traffic controller. A driverless forklift that moves pallets reliably but reports nothing useful has solved half the problem. A regulated operation needs a per-move record: which vehicle, which load, which origin and destination, at what time, completed or exception-raised.

That is the job of the M4 fleet manager, which assigns missions, arbitrates traffic and writes completion events back to the systems holding stock. Vehicle-level command runs over VDA 5050, the open interface between fleet control and mobile robots, which lets trucks of different classes and different manufacture be commanded from one place. Above it, RDS robot dispatch decides sequencing, so a release event pulls stock forward rather than waiting for the next scheduled sweep. Insist on two behaviours at specification stage: every completed move posts a stock movement to your warehouse system, and every failed move raises a visible exception rather than failing silently.

Lever three: build the PUWER and ISO 3691-4 case at design stage

The regulatory lever is to write the safety case while the layout is still on paper. Autonomous trucks are work equipment, so the Provision and Use of Work Equipment Regulations 1998 apply in full — equipment suitable for the task, protection against specific hazards, and recorded inspection and maintenance. The HSE's PUWER guidance is the practical reference for what an inspector expects to see.

Alongside it sits ISO 3691-4, the safety standard for driverless industrial trucks, governing protective devices, stopping performance and behaviour around people. The design-stage questions are concrete. Where do pedestrian routes cross vehicle routes, and can they be separated rather than marked? Does the floor meet the flatness the trucks require, per TR34 industrial floor guidance? Do charging positions sit clear of escape routes? Deciding these early costs drawing-office time; deciding them after commissioning costs racking moves, and in a licensed facility a change notification too.

Lever four: size the fleet against the release queue

The commercial lever is to size against the constraint that actually delays despatch. Many fleets are specified against the busiest hour of the busiest day, producing a vehicle count nobody can defend at capex committee. In regulated distribution the binding constraint is usually not the peak — it is the queue of released stock waiting to be brought forward after a quality release, which arrives in bursts and stalls loading.

Size for that queue and the fleet gets smaller and the payback shorter. Model it: releases per day, pallets per release, the window between release and the loading slot, and the elapsed time per lane. Then decide how to hold the asset. Where the case is framed around a quarterly cost line rather than capital, fleet leasing over three, five or seven years moves spend into operating budget. Where capital is available and the flow is permanent, outright purchase remains cheaper across the asset life.

In UK pharmaceutical distribution, an AGV forklift removes the manual scan step from every internal pallet move, so location history is generated by the machine that performed the move rather than reconstructed from an operator's handheld afterwards.

What FlyWei does here

FlyWei is an independent UK systems integrator of AGV forklifts and autonomous forklift trucks. Because FlyWei is vendor-neutral, the truck specified for a quarantine-to-released-stock lane is chosen on aisle width, lift height, floor condition and duty cycle rather than on what one manufacturer happens to build. A narrow-aisle reach-truck class vehicle suits high-bay released stock; a counterbalanced class suits trailer-adjacent staging; a low-profile pallet-truck class suits dock-to-stock runs. Most regulated sites end up with a mixed fleet.

FlyWei designs the integration around the systems the site already runs. M4 holds the missions and the per-move record, RDS sequences work so a quality release pulls stock forward against the loading slot, and completions are written back so your existing warehouse system stays the record of stock. Where cart and tote movement sits alongside pallet work — sampling, returns, secondary packing — FlyWei lifting robots and AMRs handle it under the same fleet layer. Deployment is incremental: one lane proven against your own deviation and throughput data, then the next. Vehicle classes are covered in what an autonomous forklift is, and the distinction between types in autonomous forklift versus AGV.

Frequently asked questions