The compliance gap between a demo and a deployment
The distance from “it works on stage” to “it is legal on your floor”.
From pilot to rollout: the six gates of a warehouse AMR
What has to be true before a pilot becomes daily operation.
A successful pilot proves a robot can work. It does not yet prove a business can run on it.
A pilot is not a rollout
A pilot is deliberately forgiving: a favourite route, a supervised shift, an operator on hand to intervene. Rollout removes every one of those cushions at once, which is why the two rarely feel like the same project.
Treating a good pilot as an automatic green light for rollout is the most common way a warehouse AMR programme stalls in its first quarter of daily operation.
“A pilot tells you the robot can do the job. Rollout only happens once you know the operation can survive it doing the job without you watching.”
Between pilot and rollout sit six gates most programmes never write down.
The six gates
Route coverage across the real shift pattern, exception handling without a supervisor present, integration with the warehouse management system, spare-parts and service response, staff training beyond the pilot team, and a stable uptime figure across a full operating cycle — each is a gate, not a milestone to note in passing.
A programme that has cleared five of the six gates is not five-sixths of the way to rollout. It is still a pilot with a longer track record, and the sixth gate tends to be the one that decides the outcome.
WHAT TO TAKE AWAY
A successful pilot is necessary for rollout, but it is not sufficient on its own.
Six gates — route coverage, exceptions, WMS integration, service, training, sustained uptime — separate a pilot from daily operation.
Clearing five of six gates still means a pilot, not a rollout candidate.
Scale on evidence from a re-run gate, not on a schedule set before the pilot began.
Knowing when to scale is only half the decision — knowing when to stop matters just as much.
When to scale, when to stop
A pilot that clears every gate on a second attempt after a hard stop is a better rollout candidate than one that limped through the first attempt and was pushed forward anyway. The gate exists to be failed and re-run, not waived.
Operators across Europe, the UK, the Gulf and MENA who track this formally — gate by gate, with a real sign-off — report fewer rollout reversals than those who scale on schedule rather than on evidence.
EDITORIAL RULE
Sourced or silent. Every figure on this page carries an attributed, dated source. Where we could not source a number, we left it out rather than estimate it.
SOURCES
- 01Site commissioning logs2025–2026
- 02Operator interviewsQ1 2026
- 03N.O.N. Humanoid vetting notesFactory audits · 2026
- 04Manufacturer service SLAs2026
A QUESTION
One person answers — from the first email to the room.
EMAIL
WECHAT
nh_alliance
LINKEDIN
