Contribution by Maaike Stoops – Head of R&D at LINKIT
The ambition to modernise is there. 94% of Dutch IT decision-makers call it a top priority, according to Kyndryl and Edelman DXI. But anyone who starts a digitalisation project soon runs into the same problem: the new system gets built in the image of the old one. How do you decide what to keep, and just as importantly, what to let go?
In this article we answer three questions:
- Why digitalisation projects copy problems instead of solving them
- How to decide what does and doesn’t belong in the first version
- When AI is a logical step, and when it isn’t
Modernising without starting over
When a process is digitalised, the starting point is almost always the existing system. What steps are there now? What exceptions exist? What works right now?
Those questions are understandable. But as a result, the existing process often becomes the basis for the new system. Including all the exceptions, extra steps and rules that have built up over the years. In Excel files that have become the source of truth. In mailboxes that function as a workflow.
In practice, we see three reasons for this:
The first is capacity. Many teams are already full with daily work, with questions from the business, with keeping what’s there running. The room to really zoom out and rethink a process from the ground up simply isn’t there. According to research by Ipsos I&O, commissioned by LINKIT, 24% of Dutch organisations experience limited opportunities for innovation due to IT staff shortages (Ipsos I&O, 2026). Those under pressure often take the short route: capturing what exists now and building on it.
The second is buy-in. Deciding what doesn’t return in the new system means making choices about whose way of working changes. Not everyone is waiting for that. Completeness feels safer. But when you carry everything over, you carry over all the inefficiencies too. The result is that a long list of functional requirements still emerges that don’t necessarily contribute to the best solution in the long term. In the short term, you only keep more people ‘satisfied’ with it.
And the third is expertise. Many organisations lack the knowledge to properly map what the process really needs. Not the technical knowledge, but the analytical kind: who are the users, where is the friction, and what does a first version need to deliver real value straight away? Without that analysis, rebuilding is the easiest option, and the discovery phase that makes all the difference gets skipped.
There’s a good chance this is playing out in your organisation too. According to research by Kyndryl and Edelman DXI, 94% of Dutch IT decision-makers call modernisation a top priority, while at the same time 54% of business-critical IT infrastructure is currently outdated or will be soon. The ambition is there. There’s just no room to really start over.
Modernising starts with choosing
When determining the first version, it’s all about focus. Not all parts of a process contribute equally to daily work or to the organisation’s objectives. So: which steps in this process actually deliver value: for the user, for the employee, for the organisation?
A solution that works better than the current situation doesn’t have to do everything the old system could. As long as it does the things that matter day to day, it delivers value straight away. The rest comes later. Based on what users turn out to need in practice, and not on what was once requested.
Three questions help to determine where that focus lies:
Where does it go wrong now?
Which step causes errors, waiting time or escalations? Employees answering status questions, information kept in different places, and exceptions solved manually.
What happens most often?
Which actions do most people perform every day? A system that handles the most common situations well solves the biggest part of the daily load in practice. What rarely happens can come later or stay manual.
What fits the direction of the organisation?
Some functions in the current system exist because they were once useful, not because they’re still relevant now. Those don’t need to come along. The first version is not the final version: it’s the version that already makes the process better today than it is now.
The questions above help to determine the right focus. This lets you prioritise and decide what matters most. You can then make that value measurable for the organisation/users. It’s important to actively involve key users in weighing up the priorities and the value. That way you test whether the value you choose matches what they experience in practice, and you can properly substantiate the choices you’ve made. This encourages good adoption later in the process.
With a sharp eye on priority and value, you can:
- Work on the right things with limited capacity
- Tackle politics within your organisation. It’s clear why one thing is more important than another.
Cases: modernising with a focus on what adds lasting value
Lidl Sweden: every store visible in real time
At Lidl Sweden, checks on organic products ran via paper lists. As the assortment grew, so did the chance of errors. The choice was not to rebuild that paper process digitally, but to start with what an employee on the floor needs every day. The Eco-products App in Mendix does exactly that and gives head office real-time insight per store.
Read how Lidl approached this →
Hollard: always the right customer information within reach
At Hollard, customer data was spread across multiple systems. Call-centre agents had to switch and search during every conversation. That cost them valuable time. The solution was not to replace all the systems, but to start with what agents need during every conversation. The Customer Launchpad in OutSystems brings all relevant information together on a single screen, without anything having to change in the underlying systems.
See how Hollard approached this →
4SURE: claims processing that grows with the law
At 4SURE, the claims platform LINKIT built together with Acumen, the temptation to cover everything was greatest. Claims processes contain dozens of decision rules and exceptions that have built up over the years. The choice was to start with the actions that occur every day and take the most time. By modelling decision rules flexibly in Mendix, the system grows along as legislation changes. Without the core having to be rebuilt.
Read more about how 4SURE approached this →
When AI is a logical next step and when it isn’t
With every digitalisation project comes the question: what can we do here with AI? It’s a good question, but also one that is sometimes asked at the wrong moment.
AI adds value when the information it needs is available and reliable. If status questions still run via email and you solve exceptions manually outside the system, there’s no usable data to act on. Those who digitalise the process first lay the foundation on which AI can then do something. This information is then captured in a structured and consistent way.
Think of automatic decision-making in claims processing, smart prioritisation in workload distribution, or flagging based on anomalies in a compliance check. These are applications that work when the underlying process is right. Not when AI is deployed to make a broken process run faster.
So the question is not only what you can do with AI, but when, and what needs to be in order first. With enterprise low-code platforms Mendix and OutSystems you can make AI part of your processes: controlled and scalable, not as a standalone experiment. But that step only works if the foundation, the process itself, is already in place.
Curious where the focus lies in your organisation?
What do you keep and what do you let go? Schedule a session with our experts about how to bring focus to your modernisation journey.
