Can AI help humans build sustainably?

Contractors on London’s newest Underground route, the Elizabeth line, were guessing what to build.

 

Not estimating. Not improvising. Guessing.

 

Construction ran ahead of design on the capital’s £19 billion Crossrail project. Workers installed systems, ripped them out, reinstalled them. They called it “the hokey cokey”. Management teams with competing incentives clashed across sites. Information from the ground got passed up the chain and disappeared into meetings.

 

For two years, there was no effective visibility.

 

The project opened four years late and £3 billion over budget. But here’s what matters: Crossrail isn’t an outlier. It’s the norm.

 

Another major LNG project, a critical $10+ billion initiative for US energy security, faced a major setback in May 2024 when its lead contractor, having achieved nearly 80 per cent completion on the first train, laid off thousands of workers and filed for bankruptcy. This outcome demonstrates that even a 100-year-old company, with all the incumbent systems and tools, could not successfully manage the project. Construction of trains two and three remains unfinished.

 

 

The construction industry has poured $50 billion into planning technologies between 2020 and 2022 alone, from BIM software and AI analytics to project management platforms.

 

But here’s what actually happens on projects.

 

Teams use AutoCAD or 3D for design, SAP for procurement, project scheduling, documentation systems and more. Each system speaks its own language. Data doesn’t flow between them.

 

The result is severe fragmentation. Information gets trapped in silos. Teams can’t share, sync or act on critical data when they need it.

 

 

The technical fragmentation is bad enough. The organisational structure makes it worse

 

Construction projects operate like a relay race, where runners keep changing mid-sprint. Designers hand off to engineers. Engineers hand off to contractors. Contractors hand off to subcontractors.

 

Every handoff loses information. Context disappears. Critical decisions get made in one silo, but dependent teams stay in the dark until the last minute.

Most planning systems serve only the small functional team using them. Other teams that rely on that information get left behind.

 

 

In a typical mega-project, the truth changes daily. Between 20 and 100 data points shift every 24 hours: design updates; supply chain disruptions; weather delays; labour availability; regulatory changes; equipment failures; stakeholder decisions.

 

A human project manager tracking these variables is like someone trying to play chess on 50 boards simultaneously while the rules change mid-game.

 

We’ve hit a ceiling. Project complexity has outpaced human processing capacity.

 

 

AI systems are already demonstrating what happens when you remove that ceiling. Sequence optimisation that took two weeks now takes one day. Safety observation workflows that required 30 days for 5,000 workers now complete in 24 hours.

 

These aren’t marginal improvements. They’re different categories of capability. Teknobuilt, a pioneer in the built environment technology, has been delivering such capabilities for major project owners and engineering contractors for a few years now.

 

Teknobuilt was brought in to assist one of the largest energy projects in the U.S. after its first phase encountered severe difficulties with the contractor and other project systems that did not deliver. Early performance reviews of Teknobuilt’s technology have shown improvements exceeding 30 per cent across complex energy infrastructure and data centres. Furthermore, Teknobuilt is also helping one of the world’s largest data center hyperscaler reduce delays and accelerate their project timelines.

 

Platforms such as Teknobuilt’s PACE OS consolidate fragmented data and systems into intelligent containers – or Digital Construction Blocks. The platform contextualises information across design, supply chain and execution in real time. It provides a single window to the truth that updates as conditions change, and a shortest route for intelligence to the decision makers to orchestrate with the power of a digital master builder.

 

The result is projects that can guarantee costs, schedules and carbon budgets.

 

 

Here’s where the maths gets uncomfortable for traditionally managed projects.

 

Insurance and funding gaps for infrastructure and energy projects with high execution risk are already massive. When AI systems can demonstrate guaranteed outcomes, traditional approaches become comparatively uninsurable.

 

Why would an insurer back a project with a 90 per cent chance of significant overruns when AI-managed alternatives can prove controlled delivery?

 

The industry knows this: 65 per cent of construction professionals now rate AI’s role in project management favourably. Only 13.5 per cent remain skeptical.

The shift is already happening.

 

 

If you’re planning a $500 million infrastructure build starting in five years, your project management team will look radically different.

 

Roles in planning, co-ordinating and supply chain tracking will require 50 per cent fewer humans. AI agents will handle the synthesis of daily variables, sequence optimisation and risk identification.

 

Humans will focus on final approvals, complex scope interpretations and removing bottlenecks. The physical build still requires foremen. Complex process design still needs human expertise.

 

But the co-ordination layer, the information synthesis, the real-time optimisation? That’s AI territory now.

 

The digital ledger of execution and rapid learning can put back the project manager now as a master builder with clarity and confidence that never existed in this industry before.

 

 

If we look at mega projects breaking ground today without this technology, it’s like seeing Crossrail repeating. The warring incentives. The information black holes. The contractors guessing. The billions in waste.

 

Historically, there’s a 90 per cent chance these projects significantly blow the budget and get painfully delayed. AI doesn’t eliminate all project risk. But it eliminates the category of risk caused by information fragmentation and human cognitive limits.

 

That’s most of the risk.

 

The question isn’t whether AI will help manage your next mega project. The question is whether you’ll still be competitive if it doesn’t.