Every sector currently wrestling with how to deploy AI responsibly faces a version of the same foundational question: before you can use your data intelligently, do you have your data under control?
In financial services, healthcare, and retail, that question has been wrestled with for years and often driven by regulatory pressure, competitive necessity, and hard-won lessons from early automation. But the $13trn global construction industry, representing a massive segment of the world economy is among the slowest to digitalise.
Our latest research, surveying 600 CIOs across the US, UK, Europe, Australia, and the Middle East, puts the scale of the concern into sharp focus. 96% of construction CIOs report anxiety about data ownership and control across their technology stacks. In the UK, that figure rises to 99%, the highest of any market surveyed, and a striking contrast with sectors that have had a decade or more to build data governance muscle.
2D and 3D design platforms, coordination software, field applications and cloud environments are procured to solve a specific problem. In many respects, those investments have delivered. With Revizto, some of our customers have saved as much as $15 million on a single project through early coordination issue resolution, and report rework cost reductions of between 2 and 5%.
Yet, as the number of tools grows, so does the complexity with each operating in its own data silo, each governed by vendor terms that weren’t interrogated closely at the time of signing. The result is what any technology leader will recognise immediately: tool sprawl, fragmented data estates, and a creeping realisation that critical project information doesn’t actually belong to the organisation generating it.
Across a number of other sectors such as manufacturing, logistics, and professional services where there has been a high level of automation and adoption of AI, the response to this problem has been consolidation and governance, often painful, but always necessary when the stakes are so high. The construction sector is now following the same path, with 39% of global CIOs planning to consolidatetheir tech stacks over the next 12 to 18 months. But it is doing so under pressure from a new direction: the arrival of AI as a genuine operational capability, not just a boardroom aspiration.
This is where construction becomes a genuinely important case study for the wider technology community.
The pressure to deploy AI is intense and familiar across industries. What differs in construction is the physical and financial consequence of getting it wrong.
There is a reason why construction is one of the most risk adverse industries. Multi-billion dollar and highly complex projects that span years. Decisions made at the design stage propagate forward through procurement, fabrication, budgets and field execution. Our research found that 92% of construction teams experience cost overruns of 6% or more, representing billions of pounds in lost value annually. In an environment where errors compound at that scale and with unpredictable payoffs yet to be realised by AI, deploying it on poorly governed, fragmented data doesn’t accelerate delivery, it accelerates failure across timelines measured in years rather than sprints.
The AI readiness gap the data reveals is instructive. Across the markets we surveyed, 24% of respondents cited regulatory uncertainty as the biggest barrier preventing firms from gaining value from AI. And in the US, 21% cited lack of tech integrations as the largest obstacle – above the global average of 17% and the highest of any English speaking market. The tools are there, but the connections between them aren’t.
These barriers add up to a stark disconnect between ambition and outcomes. Despite executive and board-level pressure to explore or mandate AI strategies, only 10% of respondents globally said they were already seeing value from AI with no barriers remaining. The cost challenge compounds this further: as Big Tech struggles to meet insatiable demand, pricing is rising, and firms have yet to see meaningful returns at scale, those bills are becoming harder to justify.
The industries extracting real value from AI-driven decision-making such as predictive maintenance in energy, fraud detection in financial services or demand forecasting in retail have one thing in common: they built the data foundations first. They knew what they had, where it lived, and under what terms they could use it. The AI capability came later, and was built on top of something solid.
Construction is being asked to compress that journey. And the firms that are doing it well are not the ones moving fastest on AI adoption. As AI capabilities mature, the underlying data may no longer need to be fully cleaned, labelled, or centralised. But understanding where that data sits, in what format and who can access it will remain critical. Often much of the data sits within flat PDFs and drawing so extracting the data from these will be critical before any AI can be laid on top.
The companies that are making deliberate decisions now about data portability, vendor terms, and platform consolidation, treating those choices as strategic infrastructure decisions rather than procurement details will benefit from the efficiency gains that AI can and should deliver.
For technology professionals watching from adjacent industries, the construction sector offers a live experiment in what happens when AI adoption pressure arrives before data governance is in place. The firms increasing their tool sprawl in already fragmented environments are not gaining capability, they are adding complexity and cost while reducing visibility. The ones consolidating, standardising, and asserting control over their data estates are positioning themselves to actually use AI when it matters.
The question worth asking in any industry right now is not which AI tool to buy. It is whether the underlying data infrastructure, governance, accessibility and ownership are solid enough to build on.
