Chevron CFO Sets AI Test: Does It Improve Performance?

Chevron CFO Eimear P. Bonner has a simple filter for AI projects. If the technology does not improve performance or remove a bottleneck, it does not get funded.

Bonner, who served as Chevron's first female chief technology officer before taking the finance chief role in 2024, told Fortune that CFOs are "uniquely positioned to take a view on whether AI will create value or not." The question is not whether AI can do a job, she said, but "how can AI help improve performance or overcome what's limiting performance?"

That framing matters for a company spending billions on exploration and production. Chevron runs about 15 enterprise AI workflows and use cases. One of the most consequential is ApEX, a proprietary tool designed to improve how the company discovers oil and gas resources. If ApEX reduces dry-hole risk or speeds up reservoir characterization, the savings show up directly in the P&L.

Chevron uses Microsoft Copilot and Anthropic's Claude internally. The finance organization, which employs roughly 3,500 people globally, has deployed AI in investor relations and audit functions. Bonner's goal is to train every finance employee on effective AI usage. She framed the technology as a partner: "Think of AI as your partner to glean more insights."

The emphasis on training comes at a moment when most companies are not ready. A Skillsoft survey of 2,000 global professionals found that only one in four employees feel prepared to use AI at work, even though 86% are already using the tools. The same survey identified a 53-point readiness gap between what leaders think and what employees report. Chevron's top-down push – overseen by CEO Mike Wirth and corporate officers Jeff Gustavson and Ryder Booth – positions it ahead of the curve.

Bonner herself uses AI for data synthesis and as a sounding board. She is exploring building a personal AI agent to handle routine tasks like preparing for speaking engagements.

The company also runs an AI-focused program targeting high-impact "moonshot" projects, particularly in exploration and reservoir recoveries, often with technology partners. For a company whose core business depends on finding and producing oil at competitive cost, even a small improvement in recovery rates scales to hundreds of millions of dollars.

On the revenue side, Bonner pointed to the intersection of energy demand and growth: powering data centers and partnering with hyperscalers. Last year, Chevron announced a partnership with GE Vernova and Engine No. 1 to develop large-scale power projects for U.S. data centers. That deal ties Chevron's natural gas production directly to AI's fastest-growing physical demand.

"Technology is the air we breathe in a company like Chevron," Bonner said. It underpins how the company extracts oil and gas and delivers products where they are needed.

For traders watching Chevron, the AI strategy is a lens on capital discipline. Bonner's performance filter means AI projects must earn their keep. The company has already scaled a small "skunkworks" team and is building AI capabilities across the workforce. If ApEX and other tools meaningfully lower finding and development costs, the margin improvement could outpace what the market expects from a pure commodity-price cycle.

Chevron's Alpha Score of 51/100 carries a Mixed label in the Energy sector. The score reflects a company that is balancing traditional cash flow generation with efficiency bets that take time to show up in quarterly numbers. A skeptic would say the AI program is still small – 15 use cases across a global operator. A supporter would note that Bonner, a former CTO, has the technical credibility to separate hype from real leverage.

The next concrete marker is not a specific catalyst date. It is the pace at which the 15 workflows scale and whether ApEX improves discovery rates in the company's upcoming drilling campaigns. Bonner's own test applies to every line item. If it does not improve performance, it does not stay on the budget.