AMRs: smart and flexible, but it doesn’t happen by itself
Autonomous mobile robots are widely hailed as a revolution in logistics and manufacturing. They move independently, avoid obstacles, and are both flexible and scalable. No wonder more and more companies are investing in them. But how guaranteed is that success, really?
In this blog we take you step by step through the world of AMRs. We explain what they are, what you can do with them, what the prerequisites are, and what they cost and how they pay off. Above all, we answer the question: how do you get AMRs truly working in your operation?
Because one thing is certain: it doesn’t happen by itself.
AMRs are smart
Put an AMR in a warehouse or production environment and it will navigate around racks, machines, people and obstacles on its own. These robots don’t just perform tasks — they “think” along the way: they sense, decide and act on their own to complete their task. Something no traditional machine or vehicle can do.
Broadly speaking, logistics AMRs come in three types, in order of increasing intelligence.
- Transport AMRs move goods from A to B — for example between production lines and storage, or from inbound to storage. They can carry or tow loads and come in various sizes and capacities.
- Pick-assist AMRs support operators by dynamically navigating to pick locations, reducing walking distances. Well-known examples are the systems that bring shelving units to pick stations. There are also AMRs that retrieve totes from racking, making full use of vertical space. Another application is dynamic zone picking: pickers stay within their zones while AMRs move roll cages or pallets between them.
- Hybrid or specialised AMRs: with built-in lifting mechanisms, collaborative robots, or models that work across multiple floors. This group also includes AMRs that cooperate in swarms.
The market is evolving rapidly. Suppliers are investing heavily in software intelligence, enabling AMRs to collaborate better — with each other and with people: swarm intelligence, in which robots coordinate their routes with one another without central control; AI-driven planning based on real-time information about congestion, stock or downtime; standard interfaces with WMS and MES that shorten the time from purchase to productive use; and collaboration with cobots that perform pick or assembly tasks autonomously.
AMRs are flexible and scalable
AGVs and AMRs are often mentioned in the same breath, but they differ fundamentally. An AGV follows fixed routes, guided by physical markers such as magnetic strips or QR codes on the floor. That makes the system predictable, but also inflexible: any route change requires a change in the infrastructure.
An AMR navigates autonomously using sensors, cameras and laser technology to scan its environment and detect obstacles, and adjusts its route without human intervention. This makes AMRs a better fit for dynamic environments where processes or layouts change regularly.
A major advantage is scalability: when demand for capacity increases, it is relatively easy to add an extra robot to the fleet. Another advantage is redundancy — if one robot fails, it usually causes only a minor disruption to the whole.
AMRs are hot
They get plenty of attention at trade shows, in industry publications and on social media. They are strategically attractive: they help to tackle labour shortages, boost productivity and make operations more agile. And companies that use them love to share their success stories.
In short: AMRs are smart and flexible, and they look easy to implement. But success doesn’t just happen. Our clients, too, ask us about the feasibility of AMRs in their operations. Sometimes we advised not to go ahead (yet). In other cases, the answer resulted in a successfully implemented solution.
AMRs require the right environment
Give them space
AMRs are more manoeuvrable than traditional vehicles, but they are not a panacea for a poorly designed work environment. They literally need space to navigate, turn, load and unload safely — especially at intersections, aisles, dock areas and around production lines. Allow for extra space for pick and drop locations, and provide enough buffer positions so that processes can keep running when many AMRs arrive or leave at the same time.
The interaction between people and robots also deserves attention. AMRs do account for pedestrians, but every stop costs speed. So where possible, separate pedestrians from AMR traffic. Think carefully about where to place charging stations too: strategically placed chargers let robots top up their batteries without disrupting the operation.
Finally, AMRs often demand stricter control than a fully manual operation — and that can require additional space. An example from practice: where AMRs stage roll cages at outbound docks, the staging space is planned meticulously in advance. A fixed spot is reserved for each roll cage, in route sequence, and every position must be reachable. Only when a route has been fully loaded and the dock is clear can the next one be staged. In a manual operation, an employee can push containers closer together or move them temporarily. With AMRs, that flexibility is not there.
Make your environment AMR-proof
A stable, flat floor without thresholds, ridges or damage is essential for smooth driving and positioning accuracy. Dust, moisture and temperature fluctuations affect sensors and battery life. Lighting and reflections matter too: shiny floors, windows or mirrors can confuse some systems. Cold stores and freezers require special models.
A real-life example: we once declared an AMR business case unfeasible, even though the ROI was positive. Why? The production process frequently left puddles of water on the floor. On a wet, slippery floor, the AMRs could not brake safely, so safety could not be guaranteed.
Make your IT infrastructure AMR-proof
Without a solid digital backbone, an AMR fleet will quickly grind to a halt. What is needed: a stable wireless network for continuous communication, integration with WMS, MES or ERP so that AMRs automatically receive tasks and report back their status, and smooth real-time data exchange for route planning and live monitoring.
In practice, we have seen AMR systems literally come to a standstill because messages did not arrive in full: the host systems no longer knew what the AMRs were doing, and the AMRs no longer knew where they were supposed to go. Cybersecurity is another key consideration — AMRs are connected devices within the IT landscape, and therefore potential points of entry.
AMRs require a well-founded business case
Investing in AMRs goes far beyond the vehicles alone: charging stations, IT infrastructure and software licences, adjustments to the physical environment, integration with existing systems, and training and support during go-live. Investments quickly add up to tens of thousands of euros per robot; for large-scale implementations, to several hundred thousand euros or more.
On top of that come recurring annual costs: maintenance and service contracts, software updates and licences, replacement of wear-and-tear components such as batteries or wheels, fleet management and monitoring, and training for new employees. When used properly, these costs are manageable — but underestimating them often leads to frustration and disappointing results.
Against this, there can be significant savings — provided the AMRs are well embedded: fewer manual transport movements and less forklift use, less damage to products and infrastructure, more reliable internal logistics, better capacity utilisation and less waiting time for operators, and improved safety and lower physical strain. Environments with a lot of manual handling, high error sensitivity or fluctuating workloads usually have the highest potential.
Keep in mind, however, that it can take some time before these savings can actually be cashed in. Implementation lead time plays a role, but so do start-up challenges, which may delay the expected gains. This is exactly why a well-founded business case is so important.
In practice, we see payback periods ranging from 18 to 36 months, depending on the number of shifts, frequency of use and level of integration.
How do you make it work?
A solid business case is only the first step. Now comes the real work. Implementing AMRs means more than a technical implementation: it often requires a significant change in ways of working, roles and collaboration. Involve key people from production, logistics, IT and technical services early in the design phase. Ensure clear ownership: who is responsible for performance, maintenance and control? And redesign your processes — AMRs move bottlenecks, and if you don’t adjust the process, you only move the problem elsewhere.
Design starts with the process
Whenever you replace a manual process with AMRs, the process changes — always. A warehouse operator can interpret and improvise; an AMR cannot. That is why you design in this order: process, control, organisation, information systems.
Map out the new process in detail, and also look at the preceding and subsequent processes — these often change too. If you deploy AMRs to pick from storage cabinets, the inbound flow and putaway also need redesigning.
Every task in the control layer must be crystal clear and executable, also when disruptions occur. A robot cannot interpret or delay an assignment. Think of decoupling points, so that a failure in one subprocess doesn’t bring the entire system to a halt.
In the organisation, define who issues tasks, who may intervene and what information is needed to do so. Should a team lead be able to intervene in the system, or does a control room have a better overview?
Only after that do you design the information systems, including their interfaces. That sequence matters: a solid IT design can support and accelerate processes, but it can never correct a poorly designed process.
Integrate with the system landscape
AMRs usually run under a WCS with a fleet manager that assigns trips and controls the robots. That WCS operates under a WMS — or an MES in production — which orchestrates the whole: stock management, priorities, planning and coordination with supply and transport. Above that, the ERP directs the supply chain.
The challenge lies in the boundary between WMS and WCS. Should the WCS know about stock, or is it enough for it to know which handling unit is being moved? How many open tasks should the WCS hold — enough to optimise routes, but not so many that the WMS can no longer prioritise in real time? Who determines the pick sequence across multiple locations? A good principle is to treat the WCS as a black box that optimises as autonomously as possible, while remaining integrated into the larger system.
Keep control in your own hands
Even an automated warehouse needs human intelligence. Set up a control room or dashboard that keeps the systems aligned and lets you monitor, adjust and intervene when disruptions occur.
AMRs often work side by side with employees who pick orders or receive goods. The pace of the system must follow the pace of the employees — not the other way around. When a product or task deviates from what is expected, or an unexpected peak occurs, it is people who provide the solution. That requires orchestration — and operators and team leads who understand the AMRs and their role in the process.
Make AMRs part of daily operations
An AMR system only works if operators and team leads understand how it works and what role they play in it. Without that knowledge, confusion follows: AMRs come to unnecessary standstills, operators get frustrated and people revert to the old way of working. Provide training, make the status of the AMRs visible in the operation, and define who takes which action in case of breakdowns, delays or rush orders.
A cobot AMR should follow the employee, not the other way around. When the person stays in the lead and the AMR follows, the employee feels supported. When the AMR takes the lead, frustration arises.
Maintenance and securing know-how
Preventive maintenance and monitoring make the difference between a smoothly running system and costly downtime. Many suppliers offer service contracts with remote monitoring, software updates and support. But the basics must be in place internally too: appoint a technical point of contact who can quickly solve first-line issues, keep spare parts in stock, and secure knowledge by documenting working methods, troubleshooting procedures and maintenance instructions and making them part of regular work meetings.
Conclusion: it doesn’t happen by itself
Autonomous mobile robots often seem to be a ready-made solution: they navigate on their own, perform tasks without breaks, and interact with systems as if it all just happens automatically. But appearances are deceptive. In reality, their success stands or falls with how well they are embedded in the organisation.
The technology is smart, but without clear processes, engaged people and proper control, the promised benefits won’t materialise. The assumption that you can just add a few robots to solve your logistics challenge rarely proves true in practice. Do you really want to benefit? Then don’t start with the technology. Start with the question: how do I make sure my organisation is ready for it?
