Rail freight optimisation conversations tend to follow a predictable arc. Someone presents an impressive use case. Everyone agrees that AI and automation offer enormous potential. The session ends with cautious encouragement to move faster. 

The Optimisation panel at railXchange 2026 took a different turn. 

With Bernd Fritzges (CEO, MICE DESK), Fabian Stöffler (Co-Founder & CEO, Menlo79), Burkhard Bräkling (Principal, Transport & Digital Consulting), and Michael Lehmen (Head of Operational Staff, TX Logistik) on stage, what emerged was something rarer and more useful than consensus: a genuine disagreement about whether the industry’s real problem is technical, organisational, or something harder to name. 

The view from inside operations

Michael Lehmen represents a perspective that is often missing from innovation panels — the person who has to make the trains run the next morning, regardless of what the strategy presentation said. 

As Head of Operational Staff at TX Logistik, one of Europe’s largest railway undertakings, Lehmen’s daily reality involves complexity that most optimisation tools have not fully reckoned with. Operational disruptions don’t arrive in isolation. A delay at one node cascades. A staff shortage in one location creates knock-on effects across the network. The plan that was valid at 06:00 is already wrong by 07:30. 

His observation — that the people dimension consistently gets underestimated in rail’s digitalisation ambitions — cut against the grain of a day that had largely framed the problem as technical. Works councils, he pointed out, are often consulted too late in the process, when the design is already fixed and resistance is the only available response. The organisations that do this well involve them early — not as a compliance step, but as a genuine input into what the system needs to do. That changes what gets built, and it changes whether people use it. 

This is not a new observation. What was striking was who was saying it. Not a change management consultant, but someone running operations at scale, who has watched well-designed tools fail at the point of contact with the workforce that was supposed to use them. 

What “just start” looks like when it meets procurement reality

Stöffler’s provocation from his impulse — just start, learn iteratively — held up reasonably well in the panel, but with important qualifications from Bräkling. 

Burkhard Bräkling brought something to the conversation that the day had largely avoided: the structural friction that sits between a good idea and a working deployment in a large rail organisation. Strategy consulting in the transport sector for 35 years gives a particular view of the gap between what is technically available and what organisations actually implement, and Bräkling is direct about where that gap most often lives. 

The problem is rarely the technology. It is rarely even the will to change. It tends to live in the procurement and vendor relationship architecture — the way large operators are locked into existing system providers, the switching costs that make even clearly superior alternatives hard to adopt, the multi-year contract cycles that create a structural lag between what the market offers and what a given operator can actually deploy. 

Stöffler’s answer — start small, build iteratively, avoid vendor lock-in through an API-first architecture — is theoretically correct. Bräkling’s point is that the organisational appetite for starting small is often lower than it sounds, because starting small still means navigating the same procurement processes, the same IT governance, the same stakeholder alignment requirements that large-scale implementations face. The path of least resistance is frequently to keep doing what already works, even when it doesn’t work well. 

This tension — between the available tools and the organisational capacity to deploy them — is the most honest description of where rail freight’s optimisation gap actually lives. 

The excuse inventory

Fritzges, arriving with an outsider’s perspective from the hospitality industry, had the least patience for the structural arguments. His read of the situation was blunter: the excuses that rail offers for moving slowly are structurally identical to the ones that hotels offered five years ago, before the industry was forced to automate by competitive pressure and labour cost dynamics. 

Complex regulation. Legacy systems. Many stakeholders. High coordination overhead. These are not wrong descriptions of the rail environment. But they are also, in Fritzges’ experience, the standard inventory of reasons why every industry believes its particular version of the problem is uniquely hard to solve. 

The hotels were also right that their environment was complex. They were also right that the human element was critical. The technology that transformed their operational processes didn’t solve those problems by making them simpler. It solved them by making the handoffs between steps so fast and so reliable that the coordination overhead shrank from hours to minutes — and then the old excuses didn’t apply anymore. 

His challenge to the room was not that the structural arguments are wrong. It was that the structural arguments are being used to justify inaction rather than to design better solutions. The question is not whether the environment is complex. It is whether the response to that complexity is to wait for it to simplify, or to build systems that function well despite it. 

Where the panel actually landed

Three things emerged from the friction that were more useful than a consensus might have been. 

The first was an honest acknowledgment that rail freight’s optimisation challenge is not primarily a technology problem. The tools exist. The case studies exist. The gap is organisational — in procurement architecture, in change management, in the gap between what strategy says and what operations can absorb. Anyone claiming the problem is purely technical is selling something. 

The second was a clearer picture of what “just start” actually requires. Stöffler’s iterative approach is correct in principle, but it needs an organisational container: someone with budget authority who owns the outcome, a technology architecture that doesn’t create new lock-in, and a workforce engagement strategy that treats the people who will use the tools as co-designers rather than recipients. The boxXpress results — 98% schedule coverage, near-zero manual correction, meaningful productivity gain — happened because all three of those conditions were present. 

The third was the most uncomfortable: the awareness that the delay itself has a cost that doesn’t show up in any cost centre. Every year that a rail operation runs its scheduling on Excel and its payroll on manual processes is a year that competitors in adjacent transport modes are widening their efficiency gap. The disruption risk is not that a competitor will suddenly deploy perfect AI and win all the customers. It is the quieter, slower accumulation of a competitiveness deficit that is hard to measure until it is impossible to close. 

Bräkling, whose career spans the full arc from pre-digital transport to the current AI moment, put it plainly: the organisations that will define what rail freight looks like in ten years are the ones that are making decisions now, under uncertainty, with imperfect information. Waiting for certainty is itself a decision — and not a neutral one. 

 

The Optimisation panel at railXchange 2026 featured Bernd Fritzges (CEO, MICE DESK), Fabian Stöffler (Co-Founder & CEO, Menlo79), Burkhard Bräkling (Principal, Transport & Digital Consulting), and Michael Lehmen (Head of Operational Staff, TX Logistik).