Despite investments in recent years, the global infrastructure sector faces a significant funding gap of $64 trillion. Governments worldwide and the private sector will need to invest nearly $140 trillion by 2050 to cover this deficit, keep pace with ever-changing global conditions, and meet the needs of people around the world. A significant increase in productivity is required.
Infrastructure productivity is limited not by a lack of technology, but by the fragmentation of systems, data, and stakeholders, which reduces coordination efficiency.
The report, titled The intelligence layer: how agentic AI can connect the infrastructure industry, Analyze how governments and organizations can integrate agentic artificial intelligence into project, program, and portfolio structures to improve coordination between segmented yet highly interdependent data sources. While current execution models require manual coordination and periodic reporting, agentic AI brings a continuous model that analyzes data in real-time, fluidly directing it between systems. Agents identify and connect information at scale, while humans apply subject matter expertise, use their judgment, and retain accountability.
According to the report, investments in digital tools for improving efficiency and optimizing workflows have not generated cumulative effects at the project, program, or portfolio level, as productivity gains have remained localized. Data is fragmented across isolated systems, limiting visibility and hindering coordination.
Industry standards indicate that rework alone can account for up to 15% of total project costs, resulting in losses of hundreds of billions of dollars. Agent-based artificial intelligence, when viewed as an information layer, addresses these issues by acting as an interconnecting layer situated above and between existing systems, providing clear, actionable real-time insights and allowing infrastructure professionals to shift their focus away from time-consuming administrative tasks. This shift increases the complexity of the decision-making process and enables the delivery of large-scale critical infrastructure with the same workforce.
Aurelia Costache, Partner, Consulting, AI Lead, EY Romania: „Romania's ambitions in infrastructure require a change in how we transform investments into results. Agentic artificial intelligence helps correlate data and decisions across projects, so organizations can advance faster, reduce rework, and ensure solid governance. The real benefit of this technology is not just automation, but a clearer understanding that supports faster delivery without eliminating human control.”
A governance framework based on five critical areas is needed.
For agentic artificial intelligence to develop, agents must be capable of reasoning across organizational boundaries and between multiple information sources that were initially designed to remain separate. This creates the need for an appropriate governance framework. The report proposes a governance framework based on five critical domains: accountability and legal liability, transparency and explainability, data sovereignty and commercial privacy, oversight and human competence, and systemic resilience.
