Olivia Kabell, Associate Editor, World Oil

Holding consistent with many conference discussions in 2024, the Women’s Global Leadership Conference also tackled AI implementation. In fact, the first day of the Nov. 19-20 conference did one better, featuring multiple panels on the topic and discussing how female leaders can make strides with this new technology.

Optimism in spades. One such presentation, hosted by Cortney Arenstein, Legal & Data Analyst at Marathon Oil, was titled “Embrace AI.” Though Arenstein ranged topically from practical application to resources for education, there was a consistent theme of optimism. “There’s no barriers to entry,” she emphasized, highlighting some of the increasingly accessible ways to learn skillsets critical for this new technology.  “Seventy-five percent of knowledge workers are using AI, so far,” she noted, adding that many hiring managers are less willing to hire workers without AI skills and more willing to hire those with less experience if they do have those same skills.

Arenstein also touched on a key factor of implementation, which is keeping ahead of the ever-evolving landscape of AI technologies. The fast-paced changes in regulation, technological advancements and legal and security concerns are critical to decision-makers when it comes to enhancing existing workflows and practices with AI. Though the presentation raised many potential gains and opportunities the technology presents, difficulties remain.

The human element. One major concern that another panel brought up is the matter of getting the larger workforce onboard with AI—not just the decision-makers. In a panel titled, “Technology and innovation: AI, automation and digitalization,” moderator Molly Determan, president of the Energy Workforce & Technology Council, raised such a question: “How do we make sure employees don’t get left behind?”

According to Lucy Blount, Senior Wells Engineer in Digital & Automation at bp, the answer is communication. In an era of rapid digitalization, “AI feels like one more thing” of many being thrown at the energy workforce, leading to what she calls “digital fatigue,” or the dwindling drive to implement new digital technologies.  Providing the resources to learn, plus a simplified approach, helps, Blount added. Milos Milosevic, Head of Integrated Digital Well Construction at Halliburton, agreed: “It’s not about replacing the people; it’s about getting them the tools.”

Blount noted that ultimately, AI implementation is about a cultural and mindset shift.  For everyday energy field workers, it’s “not their job to focus on implementing this; their job is to focus on safety and safe operations.” For decision-makers, this creates a critical juncture, where AI needs to be a top-down strategy to be successful, according to Blount. Karin Guardia Phillippi, Deep Water Strategy and Performance Lead at Shell, said that for more reluctant adoptees, “friction points” are where companies should focus their efforts. Contrarians, she pointed out, are critical to early adoption research, smoothing the path for future technology implementation. She added that companies should embrace the “good” they offer and leverage it to the company’s advantage.

The data issue. Another issue that Milosevic raised is that AI is not just changing processes that companies already do, but it is also driving the need for standardization. “A lot of our processes [involve] people making the connections,” he said, and for AI to do the same, those connections will have to involve standardized pieces of data. Marie Sutton, Chief of Staff, Global Feedstocks and Energy, at LyondellBasell, was quick to echo this point: “The underlying issue is with data,” she emphasized, stating that without good, consistent data, AI is only capable of very specialized tasks.

With proper standardization, however, AI could free up valuable and much-needed time on the part of specialized experts in the energy field. As Blount pointed out, many experts spend much of their time finding information, but AI presents a path “to have something else aggregate and find the information for safer operations.” The dream scenario, in her mind, is to have the entire lifecycle, end-to-end, run by AI, though Milosevic pointed out it’s unlikely that onsite personnel will ever truly disappear.

The panel also touched briefly on IT and cybersecurity concerns, given how intertwined AI must be with vast reserves of data to be effective. Though the discussion was limited, Phillippi noted that a sandbox, or a small, safe environment to test new AI use cases is a good place to start. From there, she said, executive support can escalate a demonstrative use case to wider-scale implementation. As oft-repeated in the panel, a top-down approach is critical, and a strong CIO, according to Sutton, can take a one-off use case into a company-wide tool.