In this conversation, we spoke with Max van Riel, co-founder and CEO of Struck, about reframing construction compliance as a data and execution problem and how AI can move regulation from static documents to continuous validation.
Q: Your career shows a consistent pattern of removing friction in heavily regulated systems, from mobile payments at Payconiq to building regulation with Struck. At what moment did you realize that construction compliance was fundamentally a data and execution problem rather than a legal one, and how did your FinTech experience directly shape Struck’s first design principles?
A: For me, what has been a pattern is to apply new technology in an existing Business. At ING DIRECT we fundamentally changed how people interacted for their daily banking online and saving them money on banking costs. At Payconiq, we used API technology, new regulations on European payments together with the possibilities of mobile to process payments. At Eagle Eye Networks, we used Cloud technology for traditional on premise camera surveillance. The insight on construction came because I saw many solutions pop up to create smarter buildings but few solutions on how to build smarter while we are in a time when there is a housing crisis everywhere. Digging deeper, you quickly realize how rules and regs are a roadblock, not an enabler and that AI could change that paradigm.
Q: In the Netherlands, the Omgevingswet is often described as a legal bottleneck. From a systems perspective, do you see today’s document-driven regulatory workflows as a form of accumulated “regulatory technical debt,” and why has the construction industry remained dependent on manual PDF interpretation long after other sectors digitized their core logic?
A: I think there are a couple of answers here. The value chain in construction is extensive and requires the involvement of many. Standardization has proven to be difficult. Even if you look at how 3D models are built up within one design software, it varies widely. On the one hand, this creates a difficulty for us as well in the field of Model design checks with AI. On the other hand, you can argue that as a lot of this data is so unstructured, it is well suited for AI to unravel, hence the name of our product; Unravel.
Q: Struck positions compliance not as an afterthought, but as a real-time feedback loop during design. Why is shifting compliance upstream so structurally important, and what breaks in the construction system when legal certainty only arrives after designs are already locked in?
A: Early on in the projects, a lot of capital needs to be set aside which is at risk. The goal obviously is to get the risks out of projects as soon as possible. The costs of compliance itself are increasing as there are more and more rules and regs coming our way which turns the project into a maze. We believe that assisted compliance will reduce the costs but moreover increase the time projects can be turned around. As all requirements can be easily tested against the regulations, the gaps can be addressed quickly.
Q: Your Revit integration moves compliance from text interpretation into spatial reasoning. How does Struck translate non-structured legal language into computable geometric constraints, and what does this reveal about the future relationship between BIM systems and regulatory enforcement?
A: The expectation of the general public is extremely high in this topic, when can I upload my model and AI checks 200 things. We will get there and we have our first products to show for it but there are nuances. Can I read from a drawing that a room is 3 by 3 if that has been indicated by an architect. Absolutely, but have we then verified that that room is 3 by 3. No, that is where the challenge is. And where do you measure that, dots in a PDF that signify a distance or checking the line in a 3D model.
Q: Traditionally, compliance expertise lived in the heads of a small number of senior consultants. By encoding this logic into software, Struck effectively redistributes regulatory knowledge. Where do you believe human judgment remains essential, and where should it deliberately be removed from the loop?
A: This is an interesting topic. I think democratizing regulatory interpretation in a comprehensible way. For example if in the Netherlands you want to look up if you can build a shed in your garden you have to look for words like “adjacent buildings in a Achtererfgebied”. In addition, we have seen examples where people from the building control departments can have more meaningful and at par conversations with lawyers because the language becomes more unified. In the end, there are subjective arguments to make as well for interpretation of rules and regs but we believe AI will be able to distill the most neutral correct application of the rules.
Q: Struck serves both private actors such as architects and developers, and public actors such as municipalities. When both sides rely on the same system, are we moving toward a form of algorithmic consensus in permitting, and how do you prevent transparency from turning into governance risk?
A: We have more the perspective that we are able to unify actors in the value chain and stop the prejudices that generally exist where private sector claims municipalities are slow and municipalities argue that the private sector does not have complete files or miss information in the permit process. In our view, the system should support alignment and transparency, but final decisions and accountability must always remain with human authorities.
Q: As Struck expands beyond the Netherlands into markets like Germany, Spain, and the UK, how do you balance building a general compliance reasoning engine with the need for deeply localized legal interpretation, and what determines whether a new jurisdiction is technically viable?
A: We have already crossed the Dutch borders by adding building codes from various European Countries. In Denmark, 100 students have already tested our solution without us making many changes to our model and we already came out superior compared to other initiatives. Of course, work needs to be done but one of our advantages is that we build the infrastructure from an AI first infrastructure instead of AI as a bolt on to an existing library of documents.
Q: Building regulation directly affects public safety, making compliance software a high-stakes system. Under the EU AI Act, would you classify Struck as a high-risk AI application, and how do you ensure auditability and accountability when users act on AI-generated compliance conclusions?
A: First of all, we always present the source of our answer and we sometimes say that we don’t know the answer. This is actually highly valued by our users as they require accuracy. In addition, we do pass answers by various models to ensure AI accuracy in the answer. Of course, we adhere to Data privacy and GDPR Compliance standards as we are nearly ISO27001 certified.
Q: Investors often frame Struck as an enabler of Europe’s green transition. From your perspective, does compliance automation accelerate sustainability primarily by reducing approval friction, or by freeing cognitive bandwidth so architects can focus on energy performance rather than regulatory survival?
A: I actually think this is not a choice but a must. The fact that we are now already getting carried away with the difficulty of basic building regulations and zoning laws, makes it harder to prioritize sustainability whereas we know that more and more emphasis will be put on this in the coming years.
Q: Coming from fast-moving technology sectors into construction, which assumptions about software scalability or adoption did reality force you to abandon, particularly when working with conservative municipal institutions?
A: Funny enough I have not encountered conservatism at municipalities in the Netherlands. Yes of course the decision cycle could be faster but in general municipalities have embraced us quite quickly. Probably because the pain is clear there as even last week’s news came out that in 10 years, ⅓ of the municipality workers will retire leaving an enormous gap in knowledge and workload. Compared to other industries, this product is great because you don’t have a two sided proposition like merchants and clients to make it work. However the challenge still is, how to change people’s behaviour and processes to adopt an AI solution knowing that we are only at the beginning of its potential.
Q: Looking toward 2035, if building regulations become fully machine-readable and continuously validated, do you believe the concept of “building permits” will still exist, or will construction move toward a CI/CD-like model of continuous compliance, with platforms like Struck acting as the execution layer of the physical world?
A: I think that is a nice analogy. My cofounder Nikhil always has the comparison that in Software there is unit testing all the time. We aim to bring that concept for the construction and real estate industry as well.
