How artificial intelligence can unlock a new future for infrastructure

Despite its potential to substantially improve decision-making, increase productivity and enhance outcomes across the asset lifecycle, the current adoption of AI in the sector is still relatively slow, with mixed levels of investment by industry and technology companies. Our survey responses from the FIDIC GLF members indicate that some organisations invest conservatively, up to 2% of their revenue, while others take bolder steps, up to 10%. This is in line with other industry reports on AI investment.

Figure 1: Yearly AI investment as a percentage against the turnover of associations

The growing startup and venture ecosystem has primarily driven the adoption of AI in the infrastructure industry. Our survey of FIDIC GLF members indicates that machine learning (ML), computer vision (CV) and natural language processing (NLP) are among the industry's most widely used AI technologies. And, where investment is being made, the focus is primarily on early project stages, particularly in the design phase. This is understandable, as this is where AI can significantly improve efficiency, accuracy and innovation.

Figure 2: Ranking of project-level challenges associations perceive could be supported by AI

While the steps that the industry is taking are a good start, they are merely tinkering around the edges. AI use cases in the infrastructure market are also still largely discrete point solutions, with AI integration remaining fragmented and limited. To make meaningful change in the industry, investment in AI and AI adoption need to be far greater. To do this, the industry needs to overcome key barriers (such as bridging the knowledge gap in the infrastructure industry and historical reliance on traditional practices and established methodologies), which will only be accomplished with a distinct change in mindset and cross-stakeholder collaboration.

The infrastructure sector is inherently complex and interconnected across the asset lifecycle, but it is constrained to a linear approach as processes and technologies have forced the sector into silos. What if AI could unlock opportunities to work differently, more productively and efficiently across the lifecycle stages?

This is exactly the future that AI can enable – it can be a driver for productivity and change. AI technologies can process vast datasets, uncover hidden patterns, predict potential issues, and optimise resources across interconnected project phases. By breaking down barriers and bridging gaps across stakeholder groups and project phases, AI enables collaboration across the lifecycle, facilitating better decision-making, outcomes and more sustainable infrastructure. The adoption of AI has many strengths and opportunities, however, it is still important to recognise that human-AI collaboration in infrastructure delivery is imperative.

While organic AI solutions are emerging, accelerating growth and maturity needs industry-wide strategic and targeted interventions from across the infrastructure ecosystem - from asset managers and owners to contractors and software vendors. Each group has diverse interests, concerns and influences that are critical to balance their complex interactions and needs.

To support this approach, we have defined five guiding principles and ambitions that create opportunities for greater adoption of AI. The ambitions aim to facilitate further automation, reduce duplication, strengthen controls and enable evidence-based decision-making.

1. Determine the purpose – Increase the certainty that infrastructure will meet its intended purpose before funding commitments by taking a proactive industry approach with incentives and enablers initiated and implemented by asset owners and governments. Prioritise end-user needs and societal benefits in all planning and decision-making processes. Ensure assets are developed with resilience, responsiveness and sustainability as key priorities.
2. Plan for end-to-end delivery – Focus on enhancing the quality and reliability of project planning and preparation. To achieve this, develop comprehensive digital strategies that integrate AI technologies throughout the project lifecycle. Invest more time in thorough up-front planning to ensure projects are deliverable and affordable before making commitments. Implement AI-driven solutions from initial planning through operation and maintenance phases, emphasising data continuity and interoperability.
3. Confirm the operating model – Set up projects and programs with governance and contract arrangements that support the ambitions of wider portfolios as a strategic priority. Integrate AI tools to create more flexibility and address data ownership and liability concerns. Continuously reassess and adapt traditional delivery and operating models to meet evolving infrastructure needs. Establish flexible contractual frameworks that encourage innovation while managing risks associated with new technologies.
4. Integrate ways of working – Actively ensure the right systems and teams are in place to operate more efficiently. Build capacity and capabilities for AI implementation and invest in training and upskilling programs to ensure the workforce can work alongside AI technologies. Foster partnerships between industry stakeholders and educational institutions to develop relevant curricula. Cultivate a culture of collaboration and continuous learning, emphasising cross-disciplinary skills and AI literacy across all levels of the organisation to enhance project outcomes.
5. Operate responsive assets – Ensure operational considerations are at the forefront of decision-making during the planning, design, delivery, handover and initial occupation of built assets. Leverage AI for optimised asset performance and maintenance by implementing AI-driven predictive maintenance systems and digital twins to enhance asset performance, reduce downtime and extend asset lifecycles. Implement data-driven solutions to analyse asset operational performance against assumptions and targets. Facilitate continuous improvement by incorporating feedback loops from operational data into future project planning and design processes

Find more examples of how AI can be applied across the framework in the report.

For the successful integration of these models, industry professionals need to adapt their mindset, skill set and toolset:

• Mindset: Cultivate an open and collaborative environment. This involves raising awareness, fostering discussions, and addressing legal and ethical considerations to build confidence and encourage sharing of data and insights.
• Skill set: Develop the competencies required to leverage AI in infrastructure. This includes providing training, certifications, and workshops that equip professionals with the knowledge and expertise to integrate AI technologies effectively and keep the infrastructure industry attractive to future generations.
• Toolset: Offer the right tools and frameworks for practical implementation. This involves developing standardised data collection methods, secure data-sharing protocols, and AI-specific contract provisions that support innovation and collaboration.

In collaboration with FIDIC, we have also shaped several actions that they will explore to further support the industry’s efforts to adopt AI. These are listed in the report.

The bottom line

The infrastructure industry stands at a critical juncture. By embracing AI and fostering collaboration across the entire ecosystem, it can unlock AI's full potential to address the immense challenges it faces. This requires all actors and in particular eight key stakeholder groups in infrastructure – asset managers, owners, consultants, contractors, educational institutions, government, and technology and capital providers. It's time to usher in a new era of innovation and efficiency in infrastructure development and management, ensuring well-functioning infrastructure continues to be the backbone of society, supporting how people work and live in an increasingly complex world.

Summary

AI has the potential to transform the infrastructure sector but unlocking the value across an asset lifecycle requires more investment and adoption of AI, and a change in mindset.