AI will not deliver meaningful gains on construction projects until firms fix the disconnect between office and site, says Steve Smith of Bluebeam.
The industry’s AI uptake isn’t a technology problem. It’s a workflow problem. That may sound odd in a market full of talk about automation, efficiency and transformation. But the data suggests a real barrier to progress is more basic. Bluebeam’s research shows that 84% of AEC firms plan to increase investment in technology this year, yet only 11% describe themselves as fully digital across all project phases. Usage is strong in the design phase but falls away the closer you get to the site.
The real obstacle is not a lack of software, or even the software itself, but rather that too many projects still rely on old habits. Many still use paper within handovers, often combining printed documents with digital tools. The result is disconnected teams, duplicated versions of information and missed insights.
Too often, drawings are updated in one place, marked up in another, issued by email, downloaded locally, printed on site and, before you know it, there’s a newer version. This leads to rework, increased costs and project delays.
A recent Get it Right Initiative research report says avoidable errors cost the UK construction industry an estimated £10bn-£25bn per annum. The root cause is often the same: information did not move through the project in a controlled and trusted way.
This is why the industry should be careful about treating AI as a shortcut to better performance. AI can help summarise, classify, search, predict and flag risks, but it cannot rescue a workflow built on inconsistent naming, patchy version control and undocumented decisions. These failures are often discussed as technology gaps, but in reality, they are process and behaviour problems.
Digital process is not just a head office exercise
“Progress will not come from layering intelligence on top of broken habits. It will come from fixing disconnects between office and site.”
Firms need to stop treating digital process as a head office exercise. The businesses making the biggest strides in technology are connecting the dots between tools, workflows and outcomes. New tools are only as good as the people who use them.
Part of this comes down to training. Our data shows that 65% of firms spend less than 10% of their technology budget on training, which is why adoption remains uneven. If teams do not understand when to use a system, why it matters and what good looks like, the old habits will survive.
Firms also need to design workflows that reflect how projects are actually delivered. If the formal process is slower or less practical than an alternative, teams will use the quickest and easiest route. For digital processes to stick, they have to be simple, consistent and clearly better than the alternatives.
It also means being realistic about where AI can add value now. For most firms, the most useful application will come after core workflows are connected, not before. AI can only work with the information that is given. If that information is incomplete or inconsistent, the output will be limited.
The key question is not whether AI can produce something impressive in a demonstration. It’s whether the tool will reach every person who needs it, in the right version, at the right time. If the answer is no, that is where attention should go first. Progress will not come from layering intelligence on top of broken habits. It will come from fixing disconnects between office and site. The next step should not begin with AI: it should begin with the basics.
