Artificial intelligence (AI) is taking the world by storm—and the energy industry is taking advantage.

Complex machine learning (ML), data management and AI are rapidly evolving to enable autonomy and efficiency. The energy sector is looking for best practices to put the technology into effect.

With AI set to consume 3% to 4% of all global energy by 2030, companies such as Quantum Capital Group are beginning to invest heavily in smart technologies and data platforms.

“The effect of these technologies is really tremendous, and experts agree that AI will have a fundamental impact on how we understand the world around us as well as how we interact with the world around us,” Sebastian Gass, CTO of Quantum Capital Group, said at Hart Energy’s SUPER DUG Conference & Expo in Fort Worth, Texas. “We believe that the companies that have a digital DNA and have a beginning in becoming data-driven will outperform the rest of the industry.”

Yet with all the benefits AI can provide to the oilfield, Gass said the energy industry is lagging behind due to size: size of the market, size of datasets and size and scale of different platforms.

Gass said data-driven companies will outperform the industry because of their impact on decision making. Older methods of data management, such as Excel, despite being widely used, only offer a limited contextual understanding of a problem.

AI data management systems are able to process and handle more information, taking more variables into consideration and offer a broader understanding of the issue an operator is facing, said Gass.

“20 or 25 years ago, when databases came about and when Excel spreadsheets came about, petrotechnical applications tried to structure the subsurface and help with production optimization,” Gass said. “Then this digital transformation happened where we were able to actually integrate the data… now with AI, your ability to generate data to drive insight and the data that you’re managing within your companies is exploding.”

The expanded data contextualization brought on by AI also helps to mitigate biases that might appear in simple Excel spreadsheets. Having representative data that can support decisions is one of the keys to making a high quality decision, said Gass.

Instead of looking at a decision as a point estimate, AI enables decision makers to query a dataset over and over to ensure the data is trustworthy and unbiased. The process results in a more “repeatable, algorithmic decision maker,” said Gass, letting them “see things that others don't see,” he added.

Success rate: 11%

Still, the industry faces challenges.

The market for generative AI for energy is around $800 million, said Gass. The oil and gas sector is the highest growth market for AI and GenAI—21% larger than any other industry. A high growth market can dilute limited resources across an unlimited number of unique buyer needs, making it difficult to effectively target specific segments.

Gass also listed the size of datasets as an issue when it comes to implementation in the oilfield, citing the immense amount of data—a zettabyte, which is a million petabytes.

While the massive amount of data could be read as the industry having the resources to tackle its challenges – transitioning to cleaner energy sources, measuring methane emissions, assessing climate risks, creating virtual power plants and understanding the impacts of climate change— success ratios are much lower than desired.

“The success of traditional data science is 11%... and for GenAI, it's even lower than that at 3.5%. So very, very low success ratios. If you look at the technology that most of the oil and gas companies deploy, only a few of them are in the cloud, only a few of them are actually able to query the data that they have,” Gass said.

Only 5% to 10% of data held by companies actually be queried and used to drive algorithmic decision making.

While AI implementation isn’t seamless, there are ways to smooth out some of these issues, the first of which is by having a strong team.

“We see a lot of people just hire a data scientist, but this data scientist needs data. And if they don’t have a data engineer, how are they going to find insights…how are they going to develop if they don’t have [a] software engineering skillset or software engineers with them?” said Gass. “It is complex and it does require a comprehensive approach.”

Any company looking to incorporate AI data into its workflow also needs to have a solid architecture and technology basis for AI, said Gass. Quantum has partnered with companies such as Microsoft Corp. and Databricks Inc. to create the cloud-based stack that they use. Partnering with third party vendors also enabled them to create a solid foundation for data.

Gass also believes that to further the push for AI in the oilfield, companies need to announce and make public the successes they see as a result of AI and ML.

“You need somebody on the petroleum engineering side that wants to increase EUR, that wants to have the drilling department drill better, lower cost wells,” Gass said. “An operations department saying ‘our maintenance costs will go down by X’ or ‘there will be zero preventative maintenance anymore’… I mean all of those are goals that can actually drive you towards business result.”

Despite not being the easiest transition, the energy industry looks to catch up in terms of usage. As Gass puts it, AI has the ability “to build a competitive moat between digital haves and digital have nots.”