HOUSTON (Reuters) - (This March 13 story has been corrected after Devon Energy ( DVN ) amended data to say that it saw a 15% improvement in drilling efficiency, not a 25% improvement in productive life of its oil and gas wells, in paragraph 12, and to remove garble from headline)

Artificial intelligence is speeding up oil and gas drilling and prompting companies to take a second look at places they had viewed as too difficult or expensive to develop, executives detailed during the CERAWeek conference in Houston. 

AI took center stage in many sessions at the world's largest energy gathering. Oil producers are seeking ways to remain profitable in an environment of plummeting oil prices and worries that U.S. President Donald Trump's tariffs could slow global energy demand. 

UK oil major BP is using AI to steer drill bits and predict potential problems in wells before they happen, said Ann Davies, BP's senior vice president of wells.

"We are able to drill more wells per year and have a better capital allocation," she said.

BP announced last month it would boost annual spending on oil and gas production as part of a major strategy shift to improve investor confidence. 

AI has helped U.S. oil producer Devon Energy ( DVN ) drill in areas where it was unfeasible before, said chief technology officer Trey Lowe in an interview.

For example, the company can gather information about a fault in a formation, then drill on the other side to avoid it, he said. 

Chevron ( CVX ) is using AI-powered drones that fly over its shale operations in Texas and Colorado to remotely monitor potential problems like emissions leaks and alert field workers. 

In three months of testing drones through a partnership with autonomous drone company Percepto, Chevron ( CVX ) reduced the amount of time that production was shut in for repairs or maintenance, said Russell Robinson, a deputy program manager of facilities and operations at Chevron ( CVX ), in an interview on the sidelines of the conference.

The drones helped workers spend less time criss-crossing the shale field performing routine inspections, he said. 

"We've continued to have more assets that are running at a longer time, so this is all around just producing more oil or gas," he said, adding Chevron ( CVX ) is evaluating whether to expand use of drones to monitor its refineries. 

Devon Energy ( DVN ) has machine learning models monitoring each of its oil rigs across the U.S., Lowe said. The company has seen a 15% improvement in drilling efficiency. 

AI is also speeding up offshore drilling. BP is evaluating vast amounts of seismic data in the Gulf of Mexico in just eight to 12 weeks with the help of AI, versus six to 12 months previously. This helps geoscientists determine where to drill a well and predict difficulties, a spokesperson said. 

While the oil and gas industry has used AI for years, recent advances like large-language models are revolutionizing the sector, said Chicheng Xu, founder of OpenPetro AI, a company building AI tools for the energy industry, and a former petrophysicist at Aramco. 

For example, building three-dimensional visualizations of features deep beneath the ocean floor would be time-consuming for humans, he said. 

"AI can dig through the data and find the features you want to see and visualize it to you. That's the real difference," Xu said.

Cutting time and costs means gaining a competitive advantage. 

"Companies that don't deploy it (AI) will get left behind at this point," said Devon's Lowe. 

(Additional reporting by Arathy Somasekhar in Houston; Editing by Simon Webb and David Gregorio)