ASHBURN, VA / ACCESS Newswire / September 9, 2026 / Baker Hughes announced a deal in late January that didn’t get much attention outside trade press, but it captured something bigger happening across the oil patch. Expand Energy, the largest natural gas producer in North America, agreed to roll out an automated production system called Leucipa across thousands of wells in the Marcellus, Utica and Haynesville shale plays. Buried in the press release was a smaller detail: Expand Energy also agreed to pilot “Lucy,” an AI assistant that analyzes production data in real time and talks back to field engineers through a conversational interface.

That pilot is one small data point in a much larger shift. For the past two years, oil and gas companies have leaned on generative AI mostly for summarizing reports and answering questions. In 2026, the technology took a step further into software that doesn’t just respond, but acts, adjusting equipment and flagging problems without waiting for a person to sign off. The industry has settled on a name for it: agentic AI. And this year, it stopped being a conference buzzword and started showing up in actual field operations.

The gap between chatbot and coworker

The distinction is important to those who wish to capitalize on this trend. Much of that distinction comes down to systems integration, stitching together decades of mismatched field equipment into a single decision layer, work increasingly handled by specialized oil and gas IT services providers such as DXC. A chatbot that summarizes a well log saves an engineer a few minutes. Constant monitoring of artificial lift performance, prediction of equipment failure before it occurs, and the ability to automatically adjust operating parameters change the cost of operating a field day to day. The latter is what Baker Hughes’ Leucipa platform does, embedding physics-based failure prediction, digital twins, and scenario optimization into everyday decisions. Building something that could integrate with decades of mismatched hardware from different vendors was not easy. The challenge of integration, as Baker Hughes chief digital officer James Brady described it, is “you deal with equipment from various vintages from many different vendors.”

Adoption is still in its infancy, but is rapidly gaining momentum. As of late 2025, only about 13% of oil and gas companies have deployed agentic AI, with approximately 49% aiming to do so sometime in 2026, which would nearly quadruple the number of companies using agentic AI in just one year if those targets are met. One constant slowdown cited by analysts is security, governance and intellectual-property issues, especially within operational technology systems that were not originally designed to accommodate autonomous software.

The service companies got there first

The most visible activity isn’t coming from the oil majors – it’s coming from the companies that sell them technology. SLB moved first, launching an agentic assistant called Tela in November 2025. Its chief technology officer, Demos Pafitis, has framed the competitive stakes around data and expertise, saying the winners will be the companies with “the best data, the deepest domain expertise.” By June 2026, SLB had built a digital marketplace around Tela offering roughly 200 products – software, AI agents, foundation models – from SLB and more than 30 partner firms. It also struck a deal with ADNOC to deploy an AI-driven production optimization platform across the Abu Dhabi producer’s fields, drawing on millions of real-time data points from thousands of wells. In March, SLB paired the software push with infrastructure, partnering with Nvidia to build what it calls an “AI Factory for Energy.”

Baker Hughes’ Leucipa ecosystem has grown along a similar path. A mid-2025 collaboration with Repsol added a generative AI assistant to the platform; by January 2026, the Expand Energy deal extended it across one of the largest gas-well footprints in the country. In both cases, the AI isn’t replacing engineers; it’s shrinking the time between a well throwing an anomaly and someone, human or software, doing something about it.

The majors are chasing margins – and power demand

ExxonMobil and Chevron have taken a slightly different path, folding AI into operations while also eyeing a new business entirely. In Guyana, ExxonMobil has expanded its use of deep learning and high-performance computing to analyze seismic data and identify prospects that were harder to evaluate before, work that’s continued alongside new drilling activity in the region this year. Internally, the push is tied to a specific number: CEO Darren Woods has described AI producing a “double effect” of higher revenue and lower costs, part of a plan to reach $15 billion in operating cost savings by 2027.

Chevron’s version of the AI story has taken an unusual turn. In June, the company announced a natural gas power deal with Microsoft – called Project Kilby – to supply electricity for AI data centers, with facilities expected online by 2028. Chevron New Energies president Jeff Gustavson said interest picked up immediately: “After making the announcement about Kilby, everyone wants to talk to us.” It’s a reminder that “AI in oil and gas” now runs two directions – companies using agentic systems to tighten their own operations, while some simultaneously become power suppliers to the AI industry itself.

What the math says

The financial case rests on a handful of widely cited estimates. Rystad Energy projects digital initiatives, agentic AI chief among them, could save the sector more than $320 billion between 2026 and 2030, concentrated in drilling, predictive maintenance, reservoir management and logistics. Boston Consulting Group puts it in profit terms: oil and gas firms that fully integrate AI agents into operations could see incremental profit equal to 30% to 70% of earnings before taxes over five years.

Budgets are starting to catch up to the enthusiasm. Deloitte’s 2026 oil and gas outlook found AI and generative AI currently make up less than 20% of total IT spending among U.S. producers, with that share projected to pass 50% by 2029. Part of the urgency is geological: Baker Hughes has pointed to internal research showing that by 2030, roughly 80% of oil and gas production will come from existing, declining reservoirs rather than new discoveries, putting a premium on software that can squeeze more out of aging assets around the clock.

Not every pilot survives contact with reality

All this will not necessarily bring a return in time. While there has been significant investment in digital technologies, the study revealed that many oil and gas companies have found it difficult to move beyond the pilot phase, with legacy IT and governance structures proving to be an obstacle to the autonomous systems required for full deployment, in line with a wider industry estimate that nearly 70% of digital transformation projects stall before they are fully implemented. Centralized governance with distributed execution, known as a “hub and spoke” model, is the structure that EY says more companies are turning to avoid that fate.

The other live question is security. The convergence of these systems into direct control of equipment is becoming more common, and, as such, the boundary between IT and operational technology (OT) is becoming fuzzy – and it’s becoming one of the more serious emerging threats to energy infrastructure. It is the same independence that makes agentic AI valuable that makes it worth protecting carefully.

At least for the time being, the better indicator of whether or not a company has a demo that’s worth watching is whether it can give you deployment numbers and avoided costs to support the claims.

FAQ

What exactly is “agentic AI,” and how is it different from the chatbots oil and gas companies already use?

Traditional generative AI tools answer questions or summarize data – think of a chatbot that reads a well log and explains it. Agentic AI goes further: it can monitor operations continuously, make a decision within set boundaries, and act on it, such as adjusting a lift system or flagging a failure before it happens, without a person triggering each step.

Which companies are furthest along in deploying it?

SLB and Baker Hughes are currently the most visible, having built commercial platforms – Tela and Leucipa, respectively – with paying customers including ADNOC, Expand Energy and Repsol. ExxonMobil and Chevron are applying AI within their own operations, particularly in seismic analysis and cost reduction, while also expanding into supplying power to AI data centers.

How much money is actually on the table?

Rystad Energy estimates digital initiatives, including agentic AI, could save the oil and gas industry more than $320 billion between 2026 and 2030. Boston Consulting Group estimates companies that fully adopt AI agents could see incremental profit worth 30% to 70% of earnings before taxes over five years, though these are projections, not guaranteed outcomes.

Why is this happening now rather than five years ago?

A few forces are converging: declining output from mature reservoirs (Baker Hughes estimates 80% of production will come from existing fields by 2030), persistent labor shortages in the field, and AI models finally becoming capable enough to handle the messy, multi-vendor data environments typical of oilfield operations.

What could slow adoption down?

Governance and security remain the biggest obstacles. Roughly 70% of digital transformation projects in the sector have historically stalled before full deployment, often due to legacy IT systems and unclear decision rights. As AI agents get closer to directly controlling equipment, the merging of IT and operational technology networks is also raising new cybersecurity concerns that companies are still working through.

About DXC Technology

DXC Technology is a global technology services company that helps organizations modernize IT environments, strengthen digital operations, and manage complex technology infrastructure. Headquartered in Ashburn, Virginia, DXC Technology provides technology services and solutions designed to support enterprise transformation, data management, cybersecurity, cloud adoption, and operational efficiency across industries. The company works with organizations to integrate modern technologies with existing systems and build scalable digital capabilities.

Company Name: DXC Technology
Contact Person: DXC Team
Email: social@dxc.com
Address: Ashburn, Virginia, United States
Website: https://dxc.com/

SOURCE: DXC Technology

View the original press release on ACCESS Newswire

Media gallery

About The Author
DXC Technology