Future digital twin & AI
MIDDLE EAST 2025
30TH OCTOBER 2025 | ABU DHABI
TAKEAWAYS FROM #FDTME2025
AI implemented right offers big returns
You don’t need to invest £100m to make $100m return on investment, Nadun Muthukumarana, CEO, of AI solutions provider V2T told the event.
“AI is perhaps the only technology in history where you can realistically expect better than tenfold returns. Recent data shows AI leaders are achieving 10.3-times ROI per dollar invested whilst average implementations see just 3.7-times. That gap? It’s widening every quarter,” he said.
Arvin Singh, Partner, Head of Digital Advisory, ME & SE Asia, Wood agreed.
“We are already seeing the benefits of AI, however the return on investment is going to be far greater in five-years’ time, because AI is going to be so readily available at a price point where we can roll it out almost without thinking about it,” he said.
AI is unlocking the full potential of digital twins
AI applications are turbocharged the development of digital twins, as Erika Parn, Research Assistant Professor at NYU Abu Dhabi, told delegates.
“Think of digital twins as an analogy of the human brain-body connection. As a Digital Twin matures over time, just like a human, it learns to speak and expands its vocabulary. AI has truly unlocked this potential. Before, we viewed them as static dashboards; now, AI is giving digital twins a voice and a language with which to communicate with their operators, transforming them into tools for action,” she said.
Speakers said production optimisation was the area where AI and digital twins were adding most value in the energy industry, but that their utility is spreading fast.
“Using predictive analytics powered by AI to improve asset reliability and reduce downtime in operations is leading to a quantum shift in how we strategise, plan and execute maintenance,” Wood’s Arvin Singh said.
AI needs to be implemented with care and responsibly
AI is proving to be an invaluable tool, but it needs to be introduced cautiously to avoid investment being wasted, some speakers said.
“We need to be disciplined in assessing how much value digital twins are going to provide for our businesses – it’s easier to scale up than to scale down,” one manager from a leading Middle Eastern energy producer told the conference.
Others highlighted key elements for successful change, such as a robust database foundation, worldwide trust and governance – and making sure the workforce has the required skills to implement the changes associated with AI.
Agentic AI is here to support human decisions, not replace them
Øystein Hole, Chief Product Officer at digital twin developer Kongsberg Digital, told the conference, the goal of introducing AI is to provide prescriptive options for critical decisions, ensuring that humans remain in control while benefiting from AI insights.
“Where agentic AI can really shine is in providing industrial intelligence on which we can base our own decisions. It can take the data and analytics representing operational scenarios to compare them and identify which scenarios provide the optimal outcomes, and which ones need improvement. That’s the step that gets us to the prescriptive future we envision for complex engineering workflows,” he said.
But we are not at the stage where agentic AI can be fully autonomous. Experts said
humans still needed to be kept in the loop to verify output and train agents to ensure responses are correct and grounded in the relevant information to reduce hallucinations and errors.
AI has a key role to play in extending asset life
Extending asset life is now a hot topic for the industry, as it seeks to maximise profitability through the energy transition.
AI-driven technology can help extend asset life, particularly enhancing the maintenance aspect, driving equipment and assets as far as they can possibly go,” Clayton Ludik, Vice President Digital, Worley Consulting said.
Digital twins come in many flavours – interoperability is crucial
The industry is switching from vendor-specific to interoperable solutions, speakers said.
“Interoperability is critical. There are around 250 classifications of digital twin today and many inputs are required to feed a digital twin to develop AI-leveraged models. Those components need to be speaking to each other and need to be open source and open architecture,” one expert from a digital twin solutions company told the conference.
Don’t wait to line up the historical data before digital twin implementation
Don’t make a lack of historical data a barrier to launching your AI-driven digital twin project.
“If you’ve got historical data, then use it, but don’t let it stop you getting your $100m project off the ground. You can still use ‘effervescent’ data — your operational data,” Nadun Muthukumarana of V2T said.
Train the workforce, involve the workforce
They say it is not the camera, it is the photographer that is crucial for a good photo. In the same way, you need smart people to drive AI – but it may take time to train engineers with the right skills and mindset.
Worley Consulting’s Ludik told delegates: “Training and upskilling are critically important. They tie in very closely to strong change leadership, making sure that AI adoption runs from top to bottom and bottom to top of an organisation. Users need to be involved in the design process – it is critical that their voices on the processes and the pain points they are trying to alleviate by using AI are heard. That buy-in is central to success.”
A facility manager from a leading global energy producer added: “You need to understand the needs of people that AI can address – investment will be wasted if you don’t bring people with you on the journey”.
A new approach to regulation and standards is needed
The use of AI in the energy industry will require new regulations governing its use to be put in place. Wood’s Arvin Singh provided an example:
“AI could mean risk exposures can be assessed in real time, effectively moving away from the need for a facility to benchmark against an industry standard. That means regulatory and standard bodies will need to work with new AI regulations that have yet to be put in place,” he said.
Leaders are looking for faster digital twin development
The days when developers were given several months or even years to trial AI are over.
“Today, most organizations are asking for rapid time to value. We don’t do projects that last longer than 12 weeks anymore. So, what if every digital twin project you do can no longer take any longer than that. Could you do it?” Nadun Muthukumarana of V2T asked delegates.
Get ready – AI-driven tools are going to be everywhere
“AI is not just to enhance the traditional tools, it is here to help humanity achieve much more, especially through autonomous AI agents – and we need to be ready,” Jean-Paul Vencatachellum, Director – AI Solution Engineering, at G42-owned Inception AI told the conference.
“To get a sense of the speed at which this is evolving, the Covid virus took 13 months to infect 100 million people. ChatGPT took just two months to reach the same number of people,” he said.