Global infrastructure spending reaches trillions annually, yet billions are lost to errors, delays, and poor data coordination. Experts from EY-Parthenon argue that agentic AI can cut processes from weeks to hours and reduce costly rework. For Poland, facing an unprecedented wave of investments, the stakes are especially high.
Global infrastructure productivity has stagnated or even declined, while investment needs grow rapidly. According to the Construction Industry Institute, direct and indirect costs of errors reach 10–15% of project value – hundreds of billions of dollars wasted each year. The problem lies in fragmented data, siloed systems, and slow decision-making among multiple stakeholders.
A new report by EY-Parthenon, titled „The intelligence layer”, proposes agentic AI as a solution. The technology connects isolated processes – planning, construction, and operation – in real time, providing actionable data and enabling faster, more accurate decisions.
How agentic AI transforms infrastructure
Agentic AI uses autonomous software agents that collect, analyze, and act on data. It can detect design conflicts before construction begins, forecast risks, and automate routine tasks like data reconciliation or material ordering. Processes that currently take weeks can be compressed to hours.
The report outlines four levels of autonomy: from simple notifications, through recommendations and human-approved actions, to full autonomy within defined boundaries. High-risk decisions, such as structural changes affecting bridge safety, remain under human control. Repetitive low-risk tasks can be fully automated, freeing experts for higher-value work.
A crucial moment for Poland
Poland is entering a period of massive infrastructure investment, with state spending on roads, railways, energy, and digital networks rising sharply. The investment gap remains large, making efficiency critical. – Poland stoi dziś przed bezprecedensową falą inwestycji infrastrukturalnych. Przy takiej skali nawet niewielka nieefektywność przekłada się na bardzo wysokie dodatkowe koszty – powiedział Maciej Ziomek, partner w EY-Parthenon Polska.
Ziomek emphasized that Poland lacks a layer linking information from different sources and translating it into actions. Agentic AI can fill that gap without replacing existing systems. He stressed the need for faster digitalization and integrated data approaches. Trust in AI-driven decisions is essential, especially in the public sector, which handles most Polish investments. Without it, the technology’s potential may remain untapped.
Źródło: WNP.PL, Fot. Andrew Angelov / Shutterstock
