As adoption of AI in construction continues to accelerate, the uptake varies greatly. Some organisations are experimenting with advanced use cases like predictive maintenance and contract analysis, while others are still grappling with where it fits into their operations. But as time goes on, chief technology officer at Pagabo Group Dave Stott warns the industry risks missing the point if it continues to treat AI as a standalone capability, rather than a structural part of how modern construction businesses operate
What’s becoming ever clearer is that the challenge isn’t access to AI technology – it’s now in-built to many of our ‘doing business’ platforms. For the construction sector, a recent survey from the Association for Project Management revealed that the use of AI in project management nearly doubled in just two years. Adoption is well on the way, but the challenge remains in how to approach implementation.
Too often, AI is framed as something to be deployed to solve isolated problems: automating reports, meeting notes or improving scheduling, for example. While these use cases undoubtedly deliver time saving and therefore value, they reinforce a narrow view of what AI can do.
Is AI just a tool or is it something more?
Answering that question straight away – one of the biggest wrong turns a business can take at the start of its journey is to treat AI as a single-purpose tool.
In reality, it’s far closer in nature to the internet. It’s an enabling layer that, when properly embedded, quietly enhances productivity across everything – from the administrative uses already mentioned through to technical and design functions and project delivery.
This distinction matters. When AI is treated as a tool, it tends to become siloed and organisations only really scratch the surface of its potential. Sometimes this comes with dedicated teams, ringfenced budgets and sometimes even chief AI officers. On the surface, this may seem like a logical step but in practice, it keeps AI isolated from the very processes it can improve.
Drawing on the parallel to the internet, it would seem laughable today to appoint a chief internet officer. The internet is not a department; it’s infrastructure that allows the business to function, just like electricity. The same should ultimately be true of AI, taking it out of silo and into business structure.
Construction is already a fragmented industry – especially when it comes to data – with multiple stakeholders, tight margins and complex operations and project lifecycles. Introducing AI as another standalone function only adds to that complexity and serves to silo information and data further.
That’s why the focus needs to be on introducing AI as a structure, embedding into the fabric of operations. This shift in perspective is critical. Without it, AI will remain confined to pockets of innovation, rather than flourishing and delivering transformation at scale.
From ‘shiny project’ to core business capability
A further challenge is that AI is often treated as a shiny new toy or innovation project. Pilot programmes are launched, proof-of-concepts delivered and early results are often promising.
But AI is not a one-off implementation. It’s something that compounds value over time as data improves, processes enhance, models learn and use cases expand. So, having AI specialists within a business can be powerful, but they should be working to nurture this growth business wide, not manage something in a vacuum.
There’s a need to move beyond experimentation and embed AI into the core of their business structure. That means it flows into the everyday operations, aligning with strategic objectives and working to support decision-making at every level.
Importantly, it’s not there to replace humans – especially in a sector like construction that is so heavily built on relationships. What it should be doing is creating the efficiencies that allow people to focus their time into activity with a higher value, tying back to business strategies.
But effective processes have a part to play. It’s a well-known saying from Bill Gates that applying automation to an efficient operation will magnify the efficiency, but the opposite is also true. Established processes have been developed with humans in mind, which now need to shift towards something that’s now designed with ‘machines’ in mind. The capabilities are similar, but different, so embedding AI into existing workflows will probably require the workflows to be redesigned.
Getting it right: Data sharing and top-down leadership
Achieving this is not without its challenges. There’s the issue of data – a particular challenge for the construction sector. Firms often sit on vast amounts of information but it’s frequently siloed or in varying formats or states of completion. Without a solid foundation, even the most sophisticated AI tools will be unable to deliver meaningful insights, which in turn feeds the conversation on whether we can trust AI.
Addressing this requires investment not just in technology but in data governance, standardisation and a big change in thinking across the supply chain to achieve greater levels of data collaboration.
But if there was this strong, industry-wide data warehouse then there are huge efficiencies to be found. For example, across the Pagabo ecosystem we have a wealth of data on projects procured through the frameworks we manage for contracting authorities. That means that when a new scheme comes through procurement, we’re able to examine comparable historic projects to identify where improvements can be made – particularly when it comes to public sector spend, impact on local economies and outcomes for communities.
This may mean understanding where a high proportion of this project type hit delays at a certain point and examining cause trends – as well as identifying what was done differently on schemes that didn’t see similar delays.
That is a very basic example but imagine how much more momentum could be unlocked in project delivery and social value generation by learning from the whole industry, not just single organisation experience.
A cultural shift is also needed. Teams need to be able to trust AI-driven outputs and are confident in using them in day-to-day decision making. Here it’s less about technical expertise and more about usability, training and clear alignment with how people need to do their jobs.
Finally, leadership plays a crucial role. Moving AI from experiment to infrastructure needs a long-term view – one prioritising integration over quick wins and recognising transformation will take hard work and effort to reap the rewards.
Despite the challenges, the benefits to be uncovered are huge. AI has the potential to improve not just how much a business can do quickly but completely transform how it does business. And for an industry facing ongoing pressures – many outside of its own control – these gains aren’t just desirable, they are essential to futureproofing.
It’s easy to ask how we will know when we’ve got AI implementation right in our business or sector. The simple answer is that we’ll know when we’re no longer talking about it. A goal should be for AI to become “unremarkable” – not because it lacks any sophistication or value but because this reflects how truly transformative technologies behave over time. For example, we don’t stop to think about the complex engineering behind the water that flows from our taps or compliment a sat nav for getting us to our destination, we simply expect it to work.
The same applies here and when AI is implemented right, we’ll simply be able to rely on it operating in the background, enabling better outcomes without requiring constant attention.
The post Stop treating AI in construction as a bolt-on: Businesses must embed it to the core to unlock outcomes appeared first on Planning, Building & Construction Today.
