October 2026
Smart tech, smarter people and stupid contracts
What FIDIC said about AI
What happens when new technology meets old contracting models? Fresh from the FIDIC Global Infrastructure Conference, Consult Australia CEO Jonathan Cartledge considers whether the industry is ready for the changes ahead.
Put some of the world’s sharpest engineering minds together in a room and it’s no surprise the conversation rapidly shifts to artificial intelligence.
At the FIDIC Global Infrastructure Conference in New Delhi last month the discussions went well beyond the business-as-usual ponderings about productivity.
For me, the most provocative ideas came from people thinking deeply about how AI will change the value of professional expertise, the way we develop it and the contracts we use to deliver infrastructure.
The price of knowing what you’re doing
AI will “change everything” and “perhaps that is the biggest challenge for our industry”, said Andrea Galli, Group CEO of Swiss-based ARX Group, currently working on Snowy 2.0.
If technology can compress days of professional work into hours, should clients pay less? Or should they pay for better decisions, risks avoided and outcomes achieved?
These questions are not new. We’ve long grappled with the tension between the value of professional expertise and procurement focused on hourly rates and lowest cost.
Check out the 1891 edition of the Institute of Architects of NSW’s Professional Practice and Charges for evidence: while the standard architectural commission was 5% of construction costs, the Institute noted that projects weighted to skilled labour over materials required a different approach to pricing.
Research published by Stanford University in August – dense enough it needs an AI to decode it – suggests that consulting firms are likely to embrace hybrid pricing models that combine traditional fees with charges linked to results or intellectual property. Some of the world’s largest consulting firms are already moving towards these models.
The enduring value of human judgement
Consult Australia’s 2026 Market Conditions Report found that 65% of firms identified digital, data and AI skills as their biggest emerging skills gap. But one of my favourite observations from FIDIC suggests some of the best AI skills may be the ones we already have.
Penny Kneebone, CEO of Tonkin + Taylor, recounted her experience with a young AI “super-user” who had noticed some of the most effective AI users were senior engineers.
Why? Because they already knew how to delegate. They could articulate a problem, provide a clear brief, interrogate the response and offer constructive feedback. Skills developed over many years of professional practice prove remarkably transferable to working with AI.
Penny also applauded healthy scepticism. “I have also been very encouraged by how people have challenged some of the responses that they have received from AI tools,” she told the FIDIC audience.
The next generation understands this too, it seems. New Australian research involving 48 engineering students across three universities found they recognised that AI would make routine tasks easier to automate, placing a premium on “communication, collaboration and accountability”. They also stressed the importance of knowing when not to use AI.
The obvious catch? Automating the apprenticeship makes it much harder to develop tomorrow’s experts.
When contracts become constraints
A conversation after FIDIC with the global CEO of a consulting business added another dimension to my thinking. He suggested that advances in AI might strengthen the case for collaborative contracting models, particularly alliancing.
The proposition is compelling: AI is considerably more powerful when it can draw on information across an entire project, rather than working within the boundaries of individual organisations. Yet many traditional contracting arrangements create precisely those boundaries.
Different delivery partners maintain separate information systems, protect their intellectual property and operate within commercial structures that can discourage open information sharing. Design and construct contracts, in particular, can create silos between clients, designers and contractors.
Could collaborative contracting models, with their emphasis on shared objectives, integrated teams and collective problem-solving, be better suited to unlocking AI’s potential?
Alliancing and other collaborative contracts do not guarantee better information sharing, any more than D&C contracts automatically prevent it. At FIDIC, Satish Suneja of New Delhi-based engineering firm Nippon Koei India, wondered whether standard contracts are keeping pace with AI-generated design outputs and digital twins.
Our fragmented understanding of AI goes well beyond individual projects. University of Melbourne researchers have found that estimates of AI adoption in Australian construction range from just 6% to 44%. We can’t even agree on how widely the technology is being used.
Smart tools, old habits
If AI’s greatest value comes from connecting information across a project, why are we still organising projects in ways that keep so much of it apart?
Bentley Systems’ Amit Shrivastava noted the barrier to AI adoption is often as much about the availability of clean, structured data as the technology itself. Two-thirds of infrastructure organisations are stymied by fragmented data and siloed systems, Bentley Systems has found.
There’s something absurd about developing sophisticated AI tools while the information they need remains scattered across incompatible systems.
At FIDIC, former Indian G20 Sherpa Amitabh Kant repeatedly emphasised the importance of preparing infrastructure projects before procurement: digitally mapping assets, understanding risks and resolving approvals.
The Australian Government is currently developing Australian Standards for AI to capture its opportunities, share the benefits and keep Australians safe. This is important.
But let’s start with the ‘low hanging fruit’ and the areas where governments already invest substantial public resources. Infrastructure should be high on that list.
This is why Consult Australia continues to advocate for a Digital by Default approach to infrastructure procurement. We welcome the leadership demonstrated by Infrastructure NSW through its Infrastructure Digitalisation and Data Policy, but need to see far greater ambition across Australian governments.
The challenge is one of timing. Procurement practices, information standards and organisational capabilities take years to change. The infrastructure we commission today will be with us for decades.
We all know we can’t build the infrastructure of the future with information practices that belong to the past. But changing those practices means challenging the contracts, conventions and commercial interests that keep them in place. And for all its intelligence, AI can’t fix that problem for us.