Where the morning sessions at FIDIC's Global Infrastructure Conference 2026 asked why the industry needs to collaborate and how it finances that collaboration, the third plenary turned to the tools reshaping how infrastructure is planned, delivered and managed – and to a panel that was careful to separate what artificial intelligence and digital twins are genuinely delivering today from what remains, for now, more promise than proof.
Moderating the session, Manish Kothari, CEO of Sheladia Associates and FIDIC vice president, opened by framing digital transformation as the “how” that follows the mornings “why”. He told delegates the digital twin market alone is growing at more than 30% a year and is expected to be worth several hundred billion dollars within less than a decade, while AI-driven design tools can now generate and test thousands of options in the time it once took an engineer to sketch a handful by hand. The panel for this session brought together technology providers, delivery firms, and a contracts specialist whose job is to think about whether legal frameworks are keeping pace.
Where the technology is genuinely mature – and where it isn’t
Amit Shrivastava, senior director of solution engineering at Bentley Systems, addressed the much-discussed adoption gap, arguing that the bottleneck is as much a data problem as a technology one – most organisations, he suggested, do not yet have the clean, structured lifecycle data these tools actually need. He drew a distinction between infrastructure AI and generative AI more broadly, telling the panel that the two are fundamentally different in what they require and what they are safe to be trusted with. Interpretation of AI-generated information was highly dependent on the context in which that information was generated in the first place and this highlighted the need for organisational and professional knowledge, he said.
Devansh Kakkar, senior technical lead for public sector and enterprise accounts at Autodesk, spoke to how AI-powered platforms are transforming projects and assets into what he characterised as a stronger digital backbone, and set out the strategic considerations he believes infrastructure professionals need to bear in mind as they continue their own digital transformation journeys. He spoke about the continuing need for “human validation”. “AI does nor relace people, it enables us to do more productive things,” he said. Its role in optioneering hundreds of potential choices was one example of that.
Andrea Galli, group CEO of ARX Group, gave a frank assessment of the changes that AI would bring to the industry. “It will change everything and perhaps that is the biggest challenge for our industry and one which we have to help them with,” he said. As part of that pace of change, he also raised the challenge of keeping people’s knowledge current when design methods can look outdated almost as soon as they are adopted and reflected on how big a challenge it is simply to harness and share the additional knowledge now available to firms.
Adoption at the organisational level
Hari Somalraju, CEO of Systra India, addressed how his organisation is balancing new technology with human judgement, cutting through what he described as a great deal of “noise” around AI to focus on what is actually real on Systra’s projects today. He spoke to design automation specifically, addressing how AI-assisted outputs are validated and how their value is quantified in terms clients can actually see, and to the data security considerations shaping how AI is being brought into practice.
Penny Kneebone, CEO of Tonkin & Taylor, offered a New Zealand perspective alongside Somalraju’s Indian one, addressing what it actually takes to get a large engineering organisation to change how it works, rather than simply buy new software. Kneebone highlighted the paramount importance of people in every engineering and business process and said she had been impressed by how staff have adapted to the growing use of AI and digital technologies. “I have also been very encouraged by how people have challenged some of the responses that they have received from AI tools, which is good so see,” she said.
Accountability, disputes and responsibility
Satish Suneja, deputy director at Nippon Koei and chair of FIDIC Contract Experts, raised the pointed question of who is actually accountable when an AI-assisted design decision goes wrong, and whether FIDIC’s standard forms of contract are keeping pace with AI-generated design outputs, digital twins as contractual deliverables, and questions of data ownership. He believed that there was a real danger of more disputes in the future as a result of these issues and this raised the importance of dispute avoidance quite starkly. Ultimately though, responsibility still rests with people, not the technology itself.
Closing the session, Kothari reflected on the massive advances in technology and the tools available to the industry and said that this would inevitably change and influence the future. However, as Kothari said: “The best way to predict the future is to create it.”
