The tipping point from low-code to AI

How the way we develop is changing, and what that means for the choices you make today

Over the past few years, low-code has played a clear role in how we build software. Platforms such as Mendix have made it possible to develop faster, with more structure and less dependence on traditional code. At JAM-IT too, we worked with it intensively for years.

But at the same time we see something else happening. Something that does not feel gradual, but fundamental. In a short space of time, AI has gone from a handy aid to a fixed part of the development process. And that creates an interesting moment. Not because one directly replaces the other, but because the choice is no longer a given.

From experiment to daily practice

Not so long ago, AI was mainly seen as something to experiment with. Interesting, promising, but not yet something you would base your development process on. That picture has changed fast.

Over the course of 2025 it became clear that AI is mature enough to use structurally. Not as a bolt-on, but as part of how you build. Within JAM-IT that is now the reality. AI is woven into the daily work.

We notice what that delivers straight away:

• Ideas can be tested faster
• The step from concept to working solution is smaller
• Development cycles get shorter

But perhaps more importantly: the work itself is changing. Where a lot of time used to go into writing code, it is now shifting towards steering, checking and refining what gets generated. The developer's role does not disappear, but it shifts. Less manual work, more direction.

Two ways of solving the same problem

At heart, low-code and AI are trying to achieve the same thing: developing software faster. They just do it in fundamentally different ways.

Low-code brings structure by abstracting away the technology. You work within the frame of a platform, with visual models and standard components. That gives you something to hold on to, and predictability.

AI takes a different route. It speeds up working with the code itself. Generating, adapting, combining. Less constraint from a platform, more freedom in how you put something together. And that is exactly where the difference arises.

Where low-code makes some of the choices for you, AI puts that responsibility back with the team. More freedom also means more thinking. More steering. More safeguarding.

The trade-off becomes explicit

The choice between low-code and AI-driven development is rarely black and white. In practice it always comes down to context.

There are plenty of situations in which low-code is still a logical choice. Certainly when an organisation already works with a platform such as Mendix and has built up knowledge and processes around it. Then it offers structure, speed and predictability.

At the same time, we see that trade-off changing. Traditional low-code platforms often come with fixed constraints. Think of licences, dependencies and the way you build solutions. That need not be a problem, as long as it fits the situation.

But AI shifts the playing field. It makes it possible to deliver custom work faster, without those same constraints. Particularly for organisations that do not yet have an existing low-code infrastructure, that trade-off becomes more relevant. Not because one option is better, but because the differences are growing.

What changes in practice?

For clients, the difference is often not only in the technology, but in pace and focus.

The speed at which ideas are turned into working solutions is higher. Iterations follow each other more quickly. Experimenting costs less time.

At the same time, the centre of gravity of the work shifts. Less building, more assessing. Less creating from scratch, more validating what emerges. Because however fast AI can produce something, you still need to look at it critically:

• Is it functionally correct?
• Is it secure?
• Does it fit the context of the organisation?

AI speeds up the process, but it does not take over that responsibility.

Not a replacement, but a shift

The rise of AI does not mean low-code is disappearing. What we do see is that its role is changing.

Where low-code was for a long time the default choice for creating speed, it has now become one of several options. An option that still fits perfectly in some situations, but is no longer automatically the first choice.

For organisations, that means looking ahead matters more. Not only at what works today, but at what will be needed tomorrow:

• How flexible does your solution have to be?
• What knowledge do you want to build up in house?
• How dependent do you want to be on a specific platform?

Those are questions that will weigh more and more heavily.

Our role in that development

Within JAM-IT we no longer work from a single way of building.

We combine low-code, traditional development and AI, depending on what a problem needs. Sometimes that is Mendix. Sometimes it is not. More and more often, AI plays a central role in it. Not because it is "new", but because in many cases it has simply become the most logical choice.

What does not change in all this is how we look at software development. It is not the technology that takes centre stage, but the solution. What fits the organisation, what works in the long term, and what stays manageable.

The tools and the way we build change. But the responsibility for quality, choices and direction stays exactly where it belongs: with the people who make it.