The energy industry is buying AI, but not scaling it. Here is why.

Across Falkor's sessions and customer discussions at ONS, one message surfaced again and again: technology is not the hard part.

So what is? In our experience, it’s creating the foundations that allow people, data, and expertise to work together at scale. Whether the discussion was about industrial AI, collaboration between operators and suppliers, or the future of subsea developments, the organizations making the most progress are focusing on the systems that allow those technologies to deliver real value.

Building the foundation vs. plugging in an AI model

One of the strongest discussions of the week came during Falkor CEO Shane McArdle's session with Fernanda Lopez Larsen, EVP Corporate Development at Yara International.

Their conversation focused on a familiar problem for industrial companies: bringing information together in a way that people can actually use. Decades of engineering drawings, maintenance procedures, operating history, and frontline expertise already exist – the question is how to make that knowledge accessible and useful in day-to-day work.

As Shane put it: "You don't put intelligence on top of chaos."

The phrase resonated throughout the week because it captures a reality that many organizations are discovering as they move from AI experimentation to large-scale deployment. Success depends on getting the fundamentals right long before the first model goes into production.

Yara's experience reinforced that point. While technology played a critical role, many of the most important lessons came from involving operators, understanding how people actually work, and ensuring that digital solutions fit naturally into existing processes. As Fernanda noted, the goal is not to "move fast and break things." In industrial environments, the challenge is to move fast and build things.

The importance of strong foundations was a recurring theme in Falkor's session with Vår Energi, From Data Silos to Shared Value.

The discussion explored a familiar industry challenge. Inside teams, collaboration often works well – it is at the point of handover that problems arise between operators and suppliers, projects and operations, or different engineering disciplines.

According to Oddgeir Dalane, SVP Project Development and Shane, those barriers are rooted in fragmented data, inconsistent context, and organizational habits that develop over time.

The session highlighted how contextualized information and digital twins can help bridge those gaps, allowing engineers to spend less time locating information and more time applying expertise and judgment. By focusing on a single asset and aligning operations, maintenance, and IT around a common scope, Vår Energi was able to go live with a digital twin in six months. The approach also enabled the team to screen up to 1,000 inspection findings in three hours, compared to 3-4 findings manually per day.

The message was practical rather than theoretical: start with a focused use case, create value early, and improve the data along the way.

As Oddgeir concluded, organizations should "act when you can, not when you have to."