Major infrastructure projects lag as AI adoption faces barriers

Recent studies indicate that despite the growth of digital tools, most large infrastructure projects continue to underperform, with artificial intelligence (AI) adoption lagging due to governance, cultural, and training barriers.

Reports from FIDIC and EY in 2024 reiterate a consistent trend: major infrastructure undertakings remain plagued by delays, high costs, and considerable risk, irrespective of technological progress. The Governance Institute of Australia has highlighted that the obstacles to AI use in the sector are less about technology shortcomings and more to do with slow adoption. According to its survey, boards most frequently list governance, an unprepared culture, and skills gaps as the leading factors holding back more extensive use of AI in megaprojects.

Muriel Demarcus, Chief Executive Officer of Marsham Edge, has focused her career on narrowing the divide between potential technological benefits and their real-world application in project delivery. With three decades of experience managing large-scale projects across Europe, the United Kingdom, and Australia, she has observed persistent patterns of inefficiency and missed targets.

Reflecting on her experiences, Demarcus recounted an example from her time advising a government agency overseeing a ten-year infrastructure programme. Originally, the plan was stress-tested against just a handful of scenarios. Implementation of AI-driven scheduling allowed them to model hundreds of possibilities, identifying opportunities that ultimately shortened the project timeline by two years.

In a separate case at a manufacturing plant in France, where frequent stoppages disrupted operations, analysis led by Demarcus revealed that only three variables significantly impacted output. Monitoring and acting on these reduced downtime by 70 per cent.

These case studies are indicative of findings across the industry. The Project Management Institute's recent report, "Shaping the Future of Project Management with AI (2024)", notes that while approximately 20% of project managers have integrated AI tools into their regular routines, most organisations lack comprehensive frameworks for using these solutions across complex programmes. The "State of AI Infrastructure 2024" study arrived at similar conclusions, identifying high demand for predictive and generative technologies but insufficient preparation for their large-scale deployment.

Singapore's approach

A pivotal moment in Demarcus's career occurred following a move from Sydney to Singapore. Describing the contrasting attitudes between the two cities, she said, "In Sydney, AI was discussed more with fear than curiosity. In Singapore, by contrast, she found open datasets, a strong research community, and a government pushing forward its National AI Strategy 2.0. That ecosystem, as the Asian Journal of Technology Innovation observed in 2024, has made Singapore a rare environment where applied AI can take root quickly."

She credits this environment for enabling her to establish Marsham Edge in Singapore, concentrating her efforts not on developing new applications, but on facilitating practical adoption. Her work aims to connect project teams more effectively with AI and smart technology tools, focusing on supporting the processes that keep major initiatives on schedule and within budget.

Bridging understanding

Demarcus's transition to founding a venture in her fifties, within a sector still dominated by men, is relatively unusual. Citing research that suggests women founders in technical fields often contribute resilience and a pragmatic bias, she said her goal is to help boards "understand the technology just enough to act, whether through direct consulting or bite-sized two-minute explainers."

Her point is not that AI is a silver bullet. It's that the excuses for projects failing in the same ways are wearing thin. The tools exist. The challenge now is finding the will and the right translation to use them.

Industry studies and first-hand project evidence both suggest that effective project delivery has more to gain from addressing organisational and cultural shortfalls than from waiting on the next major technological breakthrough. For Demarcus, the value lies in demystifying AI for boardrooms and ensuring that teams are equipped to capture the productivity and efficiency gains these tools can offer. Her experience highlights how cultural readiness and clear governance structures are critical to extracting tangible benefits from AI in major infrastructure projects.