Some early AMR discussions revolve around purchase price, nominal payload, speed and whether the system needs an internet connection. Those questions are legitimate, but they are insufficient for an investment that changes the way people and materials move through a facility.
In my professional discussions across different European environments, I have observed that technically mature evaluation processes tend to move quickly toward material-flow evidence, integration responsibility and long-term support. This is an observation about project approaches, not a general judgment about any national market. The useful distinction is between asking what a robot is and asking what a complete operating system must deliver.
01 / Describe the current process before the target technology
Document where goods originate, how movements are requested, what happens during peaks and which handovers rely on informal instructions. Identify critical routes and occasional jobs. Measure representative cycle times rather than assuming that the best route on a drawing represents the whole shift.
A factory can be well organized yet still rely on experienced operators to resolve missing labels, blocked paths and changing priorities. Record these workarounds as engineering inputs rather than treating them as edge cases.
02 / Define interfaces with people, vehicles and software
Clarify the relationship between AMRs, forklifts, pedestrians, manual carts, doors, elevators and loading stations. For WMS, MES and ERP, define which system creates a mission, updates its priority, acknowledges success and handles failure. Name a responsible person for each interface.
A network availability question is not the same as an architecture question. Distinguish internet connectivity, local Wi-Fi, vendor cloud dependency and validated offline or on-premises operation.
03 / Make exceptions part of the workflow
Determine what happens if a pallet is skewed, the destination is occupied, localization fails, the battery is low or an operator calls the wrong mission. Define alerts, safe stopping, authorized recovery and how unfinished tasks return to the queue.
Exception handling is not a minor service topic. It affects throughput, operational acceptance and user confidence.
04 / Plan the whole ownership chain
Assign ownership for cleaning and marking routes, first-level assistance, preventive maintenance, updates, spare parts and escalation. Build an acceptance plan that covers relevant factory and site tests, training and handover evidence.
Define how the organization will evaluate benefit at three, six and twelve months. Compare output against the original business baseline and record why results vary.
05 / Use readiness as a decision gate
A responsible feasibility review may end with a pilot, a process redesign or a decision to postpone automation. Each is better than committing to equipment while ownership, interfaces or acceptance remain ambiguous.
The purpose of asking difficult questions early is not to delay a project. It is to prevent a poorly understood process from being automated at greater cost.
An AMR is not only a vehicle moving from A to B. It changes the logistics process around it.