Arjen Boersma and Hessel Dikkers: ‘We are truly dependent on one another to keep our trains running

Let us start by having a look at each other's kitchens. What has one organized better than the other?
Dikkers: ‘I truly admire ProRail’s ability to manage extreme complexity. I am particularly impressed by how the colleagues at ProRail manage that, and I also appreciate their long-term vision. All the innovations needed for the future require meticulous planning, and ProRail excels at that. I believe they are leading the way in Europe in this regard.’
Boersma: ‘I am sometimes envious of the speed which Hessel is able to achieve with NS. He can relatively quick improvements through digitization that directly enhance the passenger experience. Look at the travel information in the app or on platforms. You can now see how long the train is, where it will stop, and where you can find a seat. These are functional, incremental innovations. At ProRail, we are almost always dealing with mission-critical systems. It is like performing open-heart surgery. We operate in such a complex landscape that even a small change can ripple through the entire system.’

Your IT landscapes intersect significantly, but they do not completely overlap. Where do the similarities lie and where are the differences?
Dikkers: ‘Naturally, NS has a slightly different scope than ProRail. We, for example, interact with customers, we manage stations and we have a division focused on train maintenance. Technology is increasingly being integrated into the trains themselves. These are all areas where we operate with our own technology. However, when it comes to logistics – everything related to train operations, planning, and passenger information – we collaborate very closely with ProRail. That is an important part of our work.’
Boersma: ‘At ProRail, I oversee one of the largest IT landscapes in the Netherlands. We manage one of the country’s largest fiber-optic networks, making us the third-largest telecom provider. Our IT extends to inside the trains. That is where NS and ProRail intersect. For example, with travel information. All the information that appears on the boards at the stations comes partly from our systems. If our systems fail, those boards remain blank. If I do not supply that information, the train service grinds to a halt. So yes, we are heavily reliant on one another.’
Dikkers: ‘For keeping our trains running, we are truly interdependent.’
Boersma: ‘Technically, we could develop IT independently of one another. But one way or another, it all comes together in the end. We both understand that without data sharing and collaboration, no trains will run. We have to support each other there. Also, to communicate when things go wrong. Where is the mistake? And how do you make sure that it is solved as quickly as possible? The impact of even a minor disruption can be massive. In many cases, we have a resolution time of 15 to 20 minutes. If it takes longer, the entire country might grind to a halt. Not only for a short time, but sometimes for the rest of the day. Also, in handling these types of disruptions, we are heavily reliant on each other's information.’

In my role at Celonis, we work extensively on supply chain integration, process optimization and - orchestration. The challenges often lie in obtaining data from across the supply chain. How do you organize that? Do you have project governance structures or communities of practice?
Dikkers: ‘On both sides, we have seasoned IT architects who have a deep understanding of the IT landscape, not only of their own organization but also of that of their counterparts’ at ProRail. These architects work closely together, and that collaboration is key. The foundation is in architecture and teamwork. Furthermore, we have teams in areas such as planning that work together closely, both figuratively and literally.
And what is also important: we try to learn from our mistakes. In April 2022, NS experienced a major disruption with significant impact. Restarting the applications afterward did not go smoothly, neither on the NS side nor on ProRail’s. We learned tremendously from that incident. Since then, we have mapped out all our chains explicitly and reestablished simple yet crucial processes, like exchanging up-to-date contact numbers. It might sound basic but knowing how to reach each other in a crisis is critical. Things get easier when you know the people you are working with. So, our teams regularly meet, not only in formal structures but also to work on establishing informal connections.’

Do you also have shared KPIs and goals?
Boersma: ‘Absolutely. We have agreements on the availability of shared systems and the data deliveries that support them. We have also set targets for track availability. These are all shared KPIs.’

And can you hold each other accountable for them?
Boersma: ‘In this sector, holding accountable does not make much sense. We share a common goal: ensuring smooth operations so that trains run on time. When things go wrong, pointing fingers or waving accountability metrics around will not solve anything. So, it is not about holding each other accountable. Of course, we address issues when something goes wrong. That is only natural in a partnership.’
Dikkers: ‘We are both of course accountable to the authority granting the concession, in our case the ministry. We have quality agreements with them about punctuality or seat availability, for example. The ministry directs us in this respect. And the same applies to ProRail. In part, these are also overlapping KPIs. The concession grantor can ‘hold us accountable’.
If something goes wrong, it is the chairmen of our boards who must explain the situation to Parliament. That is how accountability is structured.’

I often discuss digital twins with clients, meaning the digital representation of the physical chain. Are you doing anything with that concept?
Dikkers: ‘We do not have one large, shared dashboard where everything is consolidated. Both NS and ProRail have their own monitoring systems. At NS, for instance, we have a digital view of the physical state of switches, signals, trains and personnel. Again, relevant information is shared between us. The more detailed information is less relevant to the other organization. The technical condition of an NS train - how the braking system or the doors are doing - is not particularly interesting for ProRail. But to us, of course, it is.'
Boersma: ‘Our logistics systems are packed with technology. We track every train – freight and passenger –in real time. Sensors across the tracks monitor for example the weight of trains and whether the wheels are properly rounded. So, we have an abundance of process information. However, I am not convinced of a digital twin. It would result in an overwhelming amount of data, much of which would be of limited interest.’

What digital solutions work well for you? Are you using AI?
Dikkers: ‘We frequently use solvers, small software tools that assist operational staff, often powered by AI. For example, when there is a disruption, decisions need to be made immediately. If we are dealing with a defective train, several crucial decisions need to be made extremely quickly, decisions that often have a major impact on the timetable, both in the short term and in the long term. AI is very good at helping the operation make those decisions. If you make this choice, that is the consequence. If I deploy this train now, that is the impact on today’s passengers, that will be the impact on tomorrow’s timetable and that will be the impact on the planned maintenance. This allows us to make better-informed decisions. The rail system has become so complex that we truly need AI to understand the implications of certain choices. The use of AI is therefore crucial.’
Boersma: ‘Our train and traffic controllers also use AI to solve the daily logistical puzzle. This includes large-scale projects such as construction and maintenance. Every project has numerous variables, each affecting personnel and materials. We solve that puzzle for the here and now, but also for the future. This requires a considerable number of algorithms or simulations. Without AI, it would be far more complicated.’

Many CIOs I talk to struggle with AI. They experiment a lot, but it often does not deliver what they hoped for. CIO’s also to an extent struggle to identify the business context, they do not really see where AI can produce a significant difference. How do you decide where to invest with AI?Dikkers: ‘For us, the biggest impact is in logistics. We use AI also in other areas. For example, we use large language models for our customer service and for the work instructions of mechanics. Or we use AI to detect defects on trains with the help of cameras and image recognition, which saves a person checking the train with a flashlight. But the financial impact of that on our business is ultimately relatively limited. The real game-changer is logistics. That is where we deploy AI algorithmically. How can we deploy our trains and personnel optimally? That is where we have significant improvements thanks to AI. And that is at the core of our business. My advice to companies with a logistics component is to focus their AI investments there.’
Boersma: ‘Incidentally, I also notice that it is complicated to fully realize the benefits of AI. AI is not yet consistently delivering on its promises, but I am convinced it will get there. AI is often not smart enough yet to make the next connection. Or we still fail to feed it the right data. Plus, implementing AI impacts people’s work directly, which should not be underestimated, as this creates resistance. So, to be able to use and implement it, you need to bring your people with. A significant amount of time goes into explaining that AI is not a threat, but on the contrary, can provide support.’

What are the biggest challenges for your organizations in the coming period? What are the biggest tasks at hand?
Dikkers: ‘First, the modernization of our landscape. We are modernizing our systems considerably. We are busy migrating all our critical applications, including to the cloud. This allows us to make adjustments much faster. You can see this directly in our releases (the release of new versions of apps with, for example, improvements or new functions, ed.). We used to have a release three or four times a year, now we can do a release every day.
The second challenge is cost savings. Our company is struggling financially and our revenue model is under pressure. The advantage of being a CIO is that we can use digitization to help other parts of the company with cost savings.'
Boersma: ‘I think we face similar challenges: modernizing the landscape while being mindful of costs. We operate on taxpayers’ money, and must spend it as efficiently as possible. Another major challenge is workforce scarcity. Finding and training train controllers is increasingly difficult. We need to explore technological solutions to keep the network running with fewer personnel.’

Looking at your role as CIO, are you primarily ‘Chief IT Officer?’ Or do you also see a strategic role for yourself?
Boersma: ‘Both. I am ultimately responsible for IT—how it runs today and how to keep it running tomorrow. To do that well, you need a strategic perspective. What do we need to ensure that our trains keep running in the future? Without that perspective, you are nowhere. What challenges are coming our way? What technology is available? What is applicable? What is not possible, and what is? And in the meantime, everything needs to keep functioning. You are always wearing these two hats.’
Dikkers: ‘It is the only way to fulfill this role. In some organizations, the roles are split into an operational CIO and a policy-focused CIO, or something along those lines. But it is better to have consolidate it, so one person is fully responsible. Generating ambitious ideas is one thing, but only works if you are also accountable for making them happen. That is what makes our roles so interesting and diverse. One moment, you are discussing the logistics landscape of 2050 with the board of directors, and the next day you are solving an urgent operational issue. It is a constant interchange in levels of abstraction.’

This interview was published in Management Scope 02 2025.