Construction doesn’t have a data problem. It has a connection problem

Sara Ibraheem
Sara Ibraheem AI Research and Transformation Lead, Wates Group
Sara Ibraheem
Home Insights Construction doesn’t have a data problem. It has a connection problem

There is no shortage of data in construction.

Across our industry, we generate huge amounts of information throughout the lifecycle of a building: from early design and planning through procurement and construction to operation. We have invested in digital tools, BIM and increasingly sophisticated systems to capture it.

Yet we still struggle to turn that information into knowledge that can be carried from one project to the next. I believe this is because the industry has been focusing on solving the wrong problem.

Construction’s challenge isn’t simply data quality, or a lack of technology – it’s fragmentation.

So how did we get here? Earlier this year, I published the Theory of Reconvergence in Construction, a research framework exploring how construction knowledge has become fragmented over time and what it might take to reconnect it.

Two waves of fragmentation

There are two waves of fragmentation. The first happened gradually as the traditional master builder model separated into increasingly specialised professional disciplines. That specialisation brought the benefit of enormous expertise, but knowledge also became distributed between multiple roles and organisations.

The second wave came with digitisation. Rather than reconnecting those disciplines, many of the digital platforms and data systems introduced from the 1990s onwards reproduced the same silos in digital form.

The industry digitised a fragmented system without fundamentally changing its architecture.

That matters because construction is not short of lessons. The biggest challenge is to connect those lessons systematically. A project finishes. People move on. Teams change. Systems change. Valuable knowledge is captured in different formats and different places, but the connection between what happened, why it happened and what we should do differently next time is easily lost.

I describe this as an open learning circuit: early strategic intent and later operational evidence are not architecturally connected, making systematic feedback difficult to sustain.

More data isn’t necessarily the answer

I recently joined a roundtable hosted by our partners at the Centre for Construction Best Practice (CCBP). The discussion focused on data quality and will form part of its next industry research project, with contributions from representatives including contractors, consultancy and academia.

What was striking throughout the session was how consistently the conversation returned to the same underlying challenge – the industry is already capturing significant amounts of information, but collecting more of it does not automatically create better decisions.

Standardising what we collect is valuable, but standardisation alone cannot solve the problem if information remains disconnected or is not subsequently used.

What we need to ask is a more fundamental question: How do we turn the information generated by individual projects into knowledge that becomes more valuable every time we build?

And then there’s AI

That question becomes even more pressing as construction accelerates its adoption of artificial intelligence. There is huge potential here. But AI cannot compensate for weak foundations.

Its value depends on the information and context available to it. If our underlying knowledge remains fragmented across disciplines, systems and individual projects, simply applying increasingly powerful AI over the top will not resolve that fragmentation.

At the CCBP roundtable, this led to a discussion about the importance of benchmarking AI outputs before trusting them, and the need to address underlying fragmentation rather than allowing AI to inherit it.

So before asking what AI can do with construction data, perhaps we need to ask whether we have created the right knowledge architecture for AI to work with.

From fragmented data to connected intelligence

This is where I believe connected intelligence offers a different way forward.

My Theory of Reconvergence proposes a future in which specialised expertise does not disappear. Instead, knowledge from different disciplines and systems is connected through a shared causal layer, allowing what we learn on one project to inform what happens on the next.

That distinction is important. The goal shouldn’t be one enormous system containing everything, or removing the specialist tools people need to do their jobs. It should be creating connections between knowledge so we can understand relationships, causes and outcomes across the built environment. And we need to think beyond individual projects.

During the roundtable, I raised the opportunity to consider connected intelligence at framework and sector scale. That conversation is still at an early stage, but it points towards a much bigger question for construction: how could learning compound across portfolios of projects rather than repeatedly beginning again?

Building an industry that learns

Construction has spent decades becoming better at capturing information. The next challenge is becoming better at remembering, connecting and learning from it. That requires us to look beneath individual technologies and question the structure that sits underneath them.

If we can create an environment in which lessons persist, knowledge crosses organisational and project boundaries and evidence from what we build today shapes what we build tomorrow, the potential goes far beyond better data. It gives us the foundations for an industry that can genuinely learn.

And that may be one of the most important foundations we can put in place for AI.

Sara Ibraheem AI Lead at Wates

Sara Ibraheem is AI Research and Transformation Lead at Wates Group and author of the Theory of Reconvergence in Construction, a conceptual framework examining the fragmentation and future reconvergence of construction knowledge.

Read Sara’s Theory of Reconvergence in Construction here.