Can AI help keep major infrastructure projects alive?

One of the biggest challenges in major infrastructure delivery is continuity.

Major projects span years, and over that time, people move, sponsors change, and priorities evolve. A project may retain its documents, but lose track of their meaning.

This matters because infrastructure decisions are cumulative.

A strategic objective becomes a business case assumption. An assumption becomes a design requirement. A requirement becomes a technical compromise. A compromise becomes a change.

Years later, it may be difficult to reconstruct why the decision was made, what evidence supported it, who accepted the risk and whether the original public value case still holds.

AI provides an opportunity to rethink how major projects preserve continuity by helping sponsors and project teams retrieve, connect, and test the evidence that informs those decisions.

The scale of the challenge

In 2024-2025, the Government Major Projects Portfolio (GMPP) included 213 of the UK’s largest, highest-risk projects.

Infrastructure and construction made up the largest share of that portfolio. It included 68 projects worth £433 billion, with each lasting nearly 11 years on average.

Throughout that time, the challenge is keeping the reasoning behind decisions visible as the project moves from business case to design, delivery and operation.

Yet major projects can lose that continuity.

Why? Because the infrastructure delivery system is structurally prone to fragmentation.

The Competition and Markets Authority’s latest civil engineering market study highlighted several contributing factors:

• a divided public sector landscape;
• funding and pipeline uncertainty;
• capability and skills gaps;
• short term, fragmented procurement approaches; and
• complex planning and regulatory barriers.

Learning from past projects

Heathrow Terminal 5

Heathrow Airport shows how difficult it can be to bring a major project into operation.

Problems can arise when passenger services, baggage systems, IT, airline operations and staff readiness aren’t aligned at opening. Terminal 5 showed us that.

Terminal 2 later adopted a different approach, with testing and phased airline moves helping to manage risk and smooth the transition.

The lesson is that building the asset is only one part of successful delivery. Success also depends on operational readiness and how well the asset is brought into service.

As the government considers future Heathrow expansion, that lesson remains relevant.

Many people involved in earlier terminal projects may have moved on, but the lessons shouldn’t. The challenge is ensuring they’re available when future teams are making similar decisions.

Crossrail

Crossrail offers another useful example, as it demonstrates that major projects are systems made up of thousands of connected decisions, requirements, approvals and operational commitments.

One of the project’s clearest lessons was the importance of managing how these systems are integrated from early design through to installation, testing and handover.

However, the value of Crossrail’s learning legacy goes far beyond that. It includes lessons on planning approvals, environmental commitments, asset information, staff training, and operational readiness.

In other words, Crossrail isn’t only a record of what was built. It’s a record of what project teams learned along the way, which other projects must find a way to use.

HS2 and East West Rail are two current examples where this learning could be relevant.

The opportunity: continuity intelligence

Continuity intelligence offers a way to keep critical evidence, decisions and lessons connected over time.

This approach would use AI to help preserve, connect and interrogate the evidence, assumptions, decisions, risks, requirements, interfaces, and benefits that shape a major project over time.

It should not be understood as a dashboard, file-sharing platform or a chatbot, and it would not replace project controls.

Instead, this should be viewed as an intelligence layer that uses verified project data to help sponsors and project teams trace key decisions and proposed changes back to business case objectives, benefits and commitments.

AI could identify assumptions that have changed since project approval.

It could show the potential longer-term implications of a cost-saving, or a scheduling decision.

It could flag possible conflicts between project requirements, regulations, risks, environmental commitments, assurance evidence and so much more.

Crucially, judgement and accountability would remain with people.

AI would not make the decisions. It would help the project remember, challenge and explain itself.

This is technically achievable with today’s technology. The opportunity is already within reach.

What is needed now is a joint effort to organise trusted project knowledge, connect it across programmes, apply an intelligence layer and put it all to work at the point of decision.

Keeping priorities aligned throughout the whole project

If the UK is serious about delivering its infrastructure pipeline, major projects need a disciplined way of keeping objectives, decisions, evidence and value connected throughout the whole lifecycle.

With the right safeguards and human oversight, AI-enabled continuity could well be the answer.

*The ICE welcomes guests to share their views about infrastructure policy issues on the Infrastructure Blog. These views are the views of the individual.

If you're interested in writing for the Infrastructure Blog, please email [email protected]. The ICE reserves the right not to publish articles that have been submitted.