Opinion | Standardization is dead

"The construction industry is fragmented into hundreds of contractors, software houses and consultants with their own tools and data formats. Each project forms a temporary ecosystem that disintegrates after completion. The innovations pile up but do not reinforce each other. As a result, digitization rarely leads to acceleration," considers Jan Willem van de Groep, program maker, future thinker and publicist.

The impetus for this article was a conversation with an AI model about standardization. I asked if standardization was outdated in times of artificial intelligence, APIs and self-learning systems. The model explained that standards are still needed to capture meaning and prevent misunderstandings, until we found out together that that reasoning undermines itself. If an AI can explain the difference between a dwelling, a residence object and a building with a residential function, it also understands the context and can apply the correct meaning. Classical standardization is then no longer a prerequisite, but a relic of an era when machines could not interpret what humans meant.

The real challenge is not technical but mental. The limitation of AI is not in its semantics, but in our trust in it. Until we recognize that systems can reason meaningfully, humans will remain the biggest brake on digitization.

From standard to code

Since the 1960s, standardization has been the backbone of construction. It made mass production possible, but also uniformity. When digitization took hold in the 1990s, that logic shifted to information: CAD, BIM and data structures. Everything revolved around compatibility and control. That suits stable chains, but not the dynamics of area development, where every collaboration is temporary and every context different.

Today, adaptive systems are taking over the work of fixed standards. Software monitors tolerances in real time, robotic arms adapt to the material, and precast components are generated from parametric models. The standard becomes a code of conduct. The same thing is happening in the world of data: APIs and semantic networks make fixed formats obsolete; AI understands meaning and dynamically translates it. Standardization is shifting from uniformity to intelligibility.

From size to meaning

Where dimensions used to be recorded, we now record relationships of meaning. A wall no longer has to be 120 millimeters thick because the standard says so, but because the system knows that this thickness meets comfort, fire and structural requirements. Terms such as dwelling, residential object or residential function need not be identical as long as their meaning is shared and software understands that relationship. Standardization shifts from form to function, from capture to understanding.

The paradox of fragmentation

And yet the practice remains unruly. The construction industry is fragmented into hundreds of software houses, consultants and contractors with their own tools and data formats. Each project forms a temporary ecosystem that disintegrates after completion. Innovations pile up but do not reinforce each other. As a result, digitization rarely leads to acceleration.

The reflex is to call for new standards, but that is a repetition. The solution lies not in new agreements on data exchange, but in harnessing the intelligence that can already do so. AI recognizes connections without people first having to record them. What is needed is not a new platform, but a shared learning environment in which that intelligence develops and functions reliably.

From Standard Management to Intelligence Management

The key lies with parties with scale and continuity: the larger construction companies and public clients. They don't need to create new software, but can start training algorithms that learn from existing projects. That requires three things:

1. Feed data. Construction companies own vast amounts of data from design, construction and management. That information is now fragmented, but it forms the breeding ground for self-learning systems. By making data accessible, not necessarily public but usable, AI can recognize patterns that increase quality and speed.
2. Organize learning cycles. Instead of standardization committees, the industry can work with learning cycles: projects that act as practice spaces in which AI models are tested and improved. Each project provides data and feedback to the next. In this way, the industry builds knowledge rather than losing it after each project.
3. Transparency and oversight. As intelligence evolves, oversight is needed of how systems learn and decide. Clients and knowledge institutions must monitor that the learning process remains explainable and socially responsible.

When these three elements come together, the smart layer that the industry now talks about as if it were a product will naturally emerge. It does not need to be built; it needs the space to learn.

From control to understanding

The challenge lies not in what software we use, but how we teach it to work together. AI enables us to understand rather than capture meaning, but requires that we nurture, validate and guide it. This is not a technical project, but a cultural shift: from control to trust, from specifying to training, from agreeing to learning to understand.

Those who dare to take this step not only build faster and smarter, but develop a new kind of craftsmanship: a craft in which knowledge, data and experience merge into a collective intelligence that continuously learns.

Standardization began as an attempt at control, grew into a foundation of efficiency and now ends as a principle of understanding. We no longer need identical parts, but well-connected meanings in our data and in our buildings. The future calls not for more rules, but for systems that understand what we mean when we say build with reason.

This article was written by Jan Willem van de Groep, program maker, future thinker and publicist. He is known, among other things, for the government program Building Balance.

Respond

Would you like to respond to this story? Or would you also like to share something on here? Then send an email to Construction & Installation editor-in-chief Thomas van Belzen at Thomas.van.belzen@jaarbeurs.nl

Digibouw  

On Nov. 19 and 20, the two-day event on digitization in construction DigiBouw 2025 will take place at the Jaarbeurs Center. You can be there. View the entire program here or register for free here.