Future digital twin & AI
amsterdam 2025
3-4th 2025 | , TEXAS
in collaboration with:
TAKEAWAYS FROM #FDTAM2025
AI-driven digital twins are now firmly embedded in the energy sector
Digital twins and AI-driven processes are part of everyday life for many in the oil and gas industry.
As Shane MacArdle, CEO of digital twin developer Kongsberg International, told the conference: “Digital twins are already business-critical in many sectors. If you switch one off, the asset will still run – but it won’t be as competitive.”
Nadun Muthukumarana, CEO of AI solutions provider V2T, said digital twins were starting to add serious value to the bottom line.
“Next generation digital twins are now core to $100m-plus value delivery,” he said
Think big, start small, scale fast
AI applications are changing the industry, but they are expensive and so need to scale up fast, panellists said.
Sofia Silvestre, AI Delivery Manager at Galp told delegates that organisations need to adopt a value-driven approach to achieve business impact using AI
“Organisations need to adopt a value-driven approach – start small with targeted wins but scale fast, create an AI roadmap based on business value, and adopt cross-functional collaboration,” she said.
Expectations of success have been raised
The need to scale up to obtain full value from digital solutions is reflected in a more aggressive attitude towards AI implementation.
“Winning companies are starting to demand that digital initiatives pay off in 12 months,” V2T’s Muthukumarana said.
Tarun Dhall, Lead, Digital Offerings, Compressions Business, Siemens Energy told the conference: “Don’t wait for everything to line up across the company. Find that one use case for which you have the data and move to develop agentic AI applications for that”
Successful Al implementation needs careful planning
Holistic thinking is vital when implementing AI.
“Putting AI on top of a process like ketchup will not solve all your problems – you need to consider AI as an enabler of business transformation,” Achilles Georgiu, Head of Data, AI & Innovation, MOL told the conference.
A tailored approach to the technology needed to be adopted, according to many speakers. Different assets and businesses have different realities and expectations. One size does not fit all for digital twins, so a modular approach is key to success.
AI is not perfect, but the benefits outweigh the shortcomings
Misleading output from AI continues to be a major problem for winning trust in the technology, so rapid identification of “hallucinating” AI should be a top priority. But that should not deter companies from adopting AI, panellists said.
“The question is not whether we can trust AI, but whether we can afford not to. The output from AI may have its shortcomings, but these are likely to be outweighed by the benefits in the long run,” said Achilles Georgiu, Head of Data, AI & Innovation, at energy firm MOL.
Christian Markussen, Global Practice Lead for Digital Twins and Technology Qualification, at DNV said there was a delicate balance to be struck.
“If users have low trust in an AI-driven system, they won’t use it and that will kill your business model. But too much trust is also a concern, since it can lead to complacency which again can result in unsafe or costly actions. So a good understanding of a system’s capabilities and limitations, and how it interacts with the user, is vital,” he said.
Gaining trust in AI means involving the whole workforce
“You must put people at the centre of everything you do,” according to Galp’s Sofia Silvestre.
That was a sentiment shared by Louise Ibbs, Shell’s Digital Lead, Digitalisation Pernis, who told delegates: “The human element is often overlooked. Humans create the data and consume the data. Computers do not make the decisions – we are the interpreters and gatekeepers. We get very excited about technology, but we also need to get excited about people interacting with that technology.”
Panellists said it was vital for the leadership to bring the workforce along every step of the way when adopting an AI.
“Obtaining trust is very hard but losing it is easy. It’s not all about the technology, it’s about the people,” MOL’s Achilles Georgiu said.
And persuading people to adopt AI means making it user-friendly, according to Kongsberg’s Shane MacArdle.
“We dedicate as much effort to the user experience of AI as we do to the technology itself. You need to make it simple to use, because widespread adoption is crucial to achieving a return on investment at scale,” he said.
Collaboration is vital
The industry needs to collaborate, both within and between organisations, to speed up successful use of AI-driven applications.
“AI capabilities are only going to continue to grow, with asset owners owning and maintaining the semantic data in an interoperable format. Allowing multiple contributors will accelerate the pace of change,” Michael Hotaling, Technology Scouting, Innovations & Ventures Executive Advisor at ExxonMobil, said.
Gerrit Tjallema, GM, Transformation & Excellence at Shell told the conference that collaboration was crucial to make the sweeping changes necessary for effective AI implementation.
“We need to be honest about what we do and don’t know about AI and go on the journey of change together within the organisation,” he said.
Panellists said asset performance management was embracing AI, but many teams were still doing this in siloes, which increased the chance that AI-driven projects would fail to be adopted. These need to be broken down.
Data quality, interoperability and standardisation need more work…
“A digital twin is only as smart as the data that feeds it. Prioritize high-quality, trusted data before scaling your twin’s capabilities,” Christian Farra, Director & Practice Leader, Data Supply Chain, at Informatica told the conference.
That effort also needed to be applied to data standardization.
“The most expensive part of the digital twin is not the technology, it is the data. If you standardise that data, it’s going to get much cheaper to obtain from the supply chain, simpler to manage and access in the source applications and easier integrate in the digital twin,” Anders Thostrup, Principal Engineer, Data Standards at Shell, said.
…But progress is being made
Panellists agreed that data standardisation across large, decentralised organisations remained challenging and that companies needed to put in structures to address this issue, but they said progress was being made.
“We’ve been doing a lot of work on the relevant standards over the last 15 years, and they are now getting to a point where they really can add value in putting a digital twin together, both when specifying data requirements to the external supply chain and when seeking to manage consistency of data across disciplines and application landscapes that feed the digital twin internally,” Shell’s Thostrup said.
ExxonMobil’s Michael Hotaling said companies needed to adopt a flexible approach to data.
“The industry has changed to a point where data interoperability has become the norm expectation. Users should not be wedded to one hardware or software supplier,” he said.
Clear data ownership is vital for successful collaboration
Panellists said clear data ownership and governance was crucial to get the most out of digital twins, especially to enable information sharing around commercially sensitive information.
Cooperation between the industry and academia, and between companies, creates a great dynamic to accelerate AI advances, but sharing proprietary information remains challenging, they said.