2025 will go down as the year construction finally went all-in on AI. Every conference, every podcast and I’m sure every boardroom conversation seemed to have the same message: “AI is going to change everything.”
And it will. But many teams learned a brutal lesson along the way: AI only works when the data behind it is strong. The American’s have a great saying “Garbage in, garbage out” and that is never been more true than when applied to AI.
Construction projects generate mountains of paperwork but very little of it becomes usable, trusted data. In 2025, as AI tools spread, that weakness became impossible to ignore. Teams learned the lesson the hard way. When the data is weak, the results are weak too.
Still, this was not a failure. It was a wake up call.
For years, construction has under-invested in high-quality, accessible data. We’ve been document-rich but data-poor. In 2025, that finally came home to roost. Teams rushed to adopt AI, only to discover that their foundations weren’t strong enough to support it.
The good news? All is not lost.
As we storm into 2026, construction has a huge opportunity to learn from the last year and do things better. Not by chasing every shiny new tool, but by fixing the fundamentals.
Here are the top three risks to watch out for, and the cures that will actually set you up for success.
Risk #1: Crappy Contracts
Too many contracts still treat data as an afterthought. There’s no clear requirement for how data should be captured, structured, or maintained. So what happens?
Data isn’t:
- Captured properly.
- Shared consistently.
- Usable when you actually need it.
That means all your investment in technology, AI included, gets massively diluted.
The cure:
We have to demand better. Contract structures need to set clear expectations around data: quality, format, timing, ownership, and sharing. Data should be treated like any other critical project deliverable, not a “nice to have.”
If data isn’t written into the rules of the game, it will always lose out to time pressure and cost cutting. This is something industry groups like Infrastructure Client Group (ICG) are already starting to look at. If we don’tbake data expectations into contracts, it will keep getting treated like an optional extra, and all our tech investments go straight down the drain
Risk #2: Avoiding the shiny new thing!
We’ve all seen it. A flashy demo. A slick interface. Big promises about AI doing everything automatically.
So teams invest in exciting technology… without checking whether they can actually support it, integrate it, or feed it good data. Or worse, they plug shiny new tools into the same old broken data flows and hope for magic.
Spoiler: magic does not happen.
You just get expensive disappointment.
The cure:
Focus on quality data and usability, not flashiness. The best tools are the ones your teams can actually use, that fit into real workflows, and that make data better, not just prettier.
AI doesn’t need theatre. It needs clean, structured, accessible information. If a tool can’t help you improve your data, it will never deliver on its AI promise.
Risk #3: Sinful Silos
Data trapped in unintegrated systems is one of construction’s oldest problems. And let’s be honest, most roads still lead to a spreadsheet.
Now, I love a spreadsheet. Truly. It is “one tool to rule them all.” It can do almost anything.
But it can also do everything… badly.
When data lives in disconnected systems (or worse, in someone’s desktop file) it becomes outdated, duplicated or simply forgotten. AI can’t help you if it can’t see your data, trust your data, or connect your data.
The cure:
Find the right tools for the job and make sure they talk to each other.
That might mean a central data lake. It might mean direct API connections between your core systems. Either way, your data has to move freely, safely, and reliably between the tools your teams depend on.
If you get that right, something powerful happens: as AI evolves, you can keep unlocking new value from the same data, instead of starting from scratch every time a new tool appears.
So Where Does Qflow Fit?
At Qflow, our focus is simple: get you the data and make it high quality, accessible, and valuable. From there we deliver insights and actions that act as a point of intervention when we spot something that doesnt look right.
We focus on materials and waste data, because that’s where huge amounts of money, carbon, and effort are being lost today. By getting accurate data into the hands of cost, quality, and carbon teams, we help them:
- Reduce rework
- Control costs
- Cut carbon
- Build right first time
But we’re only one piece of the puzzle. There are incredible technologies out there focused on safety, staffing, procurement, design, logistics, and more. You can solve most of your problems today.
The trick is not finding “the perfect tool.”
It’s finding the right tools, and making sure reliable data flows between them.
That’s how you future-proof yourself for AI. Not by betting on one platform to do everything, but by building a strong data foundation that any good AI can stand on.
Looking Ahead to 2026
2025 showed us what happens when you rush into AI without fixing the basics. 2026 is our chance to do it properly.
If I could leave you with three priorities for the year ahead, they’d be these:
- Treat data as a core project asset, write it into contracts and expectations.
- Choose tools that improve data quality, not just ones that look impressive.
- Break down silos so your data can move, connect, and grow in value over time.
Do that, and AI stops being hype. It becomes what it was always meant to be: a way to help construction teams build better, waste less, cut carbon, and deliver more value for everyone.
Want to find out more?
Wishing you all luck in your data and AI journey for 2026.
If you have any questions or you’re just not sure where to start, connect with the team here.
Check out our State of Data Quality in Construction report here.
