preloader
Insights
02/07/25

Low-Code is running out of road

Wrecked Gray Vehicle Unsplash
Low-code was once a clever shortcut. But it’s no longer the future. Generative AI is changing the rules faster than most businesses realise. For those managing essential IT, this shift isn’t academic. It’s operational risk.

For over a decade, low-code platforms like Mendix and OutSystems have promised a faster way to build software. Less coding. More visual flows. Easier to manage with smaller teams. That promise made sense—especially in a tight labour market and under budget pressure. But the landscape has shifted. And fast.

Generative AI tools can now do what low-code set out to do and more. They generate working software in real code, based on natural language. No platform logic required. No vendor contracts. No visual interface. Just a clear instruction, and working output.

Ask for a customer portal with a dashboard and Salesforce integration. You get it in native Python, Java, or React. It’s fast. It’s flexible. And critically: it’s yours to maintain.

The issue isn’t AI hype. The issue is low-code.

Let’s be clear. Generative AI hasn’t transformed every business yet. But it has exposed the cracks in the low-code model. What was once a modern alternative is starting to look like yesterday’s workaround:

  • Closed platforms
  • High licensing costs
  • Poor scalability
  • Weak integration with DevOps and AI tools

Ironically, low-code is turning into the same type of legacy it was meant to replace.

What this means for leadership

For CIOs, CTOs and CEOs, this isn’t a technical debate. It’s a question of risk and relevance. If you’re investing in low-code for critical systems today, you may be building dependencies you can’t afford tomorrow. The software will work—until it doesn’t scale, can’t be extended, or locks you into pricing you didn’t plan for. And when AI-native tools start outperforming it in cost, speed and flexibility, you’ll be the one holding the bill.

There’s a way forward

At Utilus, we work with organisations to future-proof their essential IT. That means being honest about what’s truly scalable, maintainable and smart in the long term. Our advice is simple: If you’re about to modernise core systems, test AI-first approaches before committing to low-code. Try a small proof of concept. Build something with AI-generated code. See how fast and flexible it really is. You don’t need to bet the farm. But you do need to know where the road is going.

In conclusion

Low-code helped bridge a gap. That bridge is no longer needed. The future of software isn’t visual—it’s generative. And it’s already here. At Utilus, we don’t do hype. We help organisations do the work. Clear, scalable, and built for what’s next.

Sonja Meijerink


More articles

Insights
11/06/25

Build or Buy? Why real IT delivery needs a deliberate mix

In this blog, we break down when to build internal IT teams, when to bring in external expertise, and why the best delivery models combine both. Clarity, continuity and control — that’s what ma...

Insights
16/05/25

Reflections from CEO / CIO in Tech: The human side of technology

At CEO / CIO in Tech, Sonja Meijerink and her Utilus colleagues saw how IT leadership is shifting. Not towards more tech — but towards more understanding between people.