Get smart: Why AI matters for the oil-and-gas industry by Scott Nyquist

"The new normal is continuous disruption,” says Lorenzo Simonelli, the CEO of Baker Hughes, an oil-services company. "We have to look at disruption as being our friend." Or maybe frenemy: Coping with disruption is not easy. Failure is most certainly an option. And now here comes artificial intelligence (AI), which could bring disruption to all kinds of industries and from the C-suite to the shop floor.

The McKinsey Global Institute (MGI) recently looked at the possibilities of AI for 19 industries, including oil and gas. There is a ton of detail in the report; for our purposes, though, I’ll start with MGI’s conclusion—that AI “stands out as a transformational technology of our digital age” that could create up to $5.8 trillion in value a year, just in those 19 industries.

That brought to mind two questions: Considering that the oil-and-gas sector is already drenched in data, does it really need AI? And second, why does it matter?

Yes, the oil-and-gas sector needs AI: Oil prices have recovered from their 2014-15 collapse, but it’s probably a safe bet that this is not the last time the industry will see dramatic price swings. So the companies that are most likely to prosper, through good times and bad, are those with the leanest operations and best management. By helping to identify and solve problems, AI can help in both ways.

My McKinsey colleague Richard Ward offers this example. Oil-and-gas companies can have thousands of suppliers and make tens of thousands of buying decisions. There is information about all of this, but no one, or even a team, can bring order out of this mass of detail in a reasonable amount of time. The use of AI, however, can absorb and analyze it all, helping to identify the best suppliers and thus save money within normal business cycles.

In general, AI techniques such as feed forward neural networks, recurrent neural networks, and convolutional neural networks can process real-time data in a way that is now not routinely done. Indeed, in 2016 McKinsey research suggested that “Within ten years, oil and gas companies could employ more PhD-level data scientists than geologists.” In just one example of how oil companies are exploring AI, ExxonMobil is working with the Massachusetts Institute of Technology to map the ocean floor. The abstract pursuit of knowledge is one reason; but so is the possibility of finding hydrocarbons. 

In a sense, it could be argued that oil and gas extraction is at least as much about bytes as bits: knowing where to drill, what technologies to use, and where the problems are. If that data is wrong, or mis-interpreted, the consequences could be huge. For example, given the colossal cost of major exploration projects, it’s obviously important to keep them working, and to spot small problems before they become big ones. AI is already proving effective at predictive maintenance; deep-learning technologies can analyze a combination of data—dimensional, audio, video—from a variety of sensors. MGI estimates that, compared to traditional analytics, AI could bring 79 percent greater value to the sector (exhibit); that comes out to $200 billion.

That figure may be conservative because AI technologies, and their effective deployment, is still relatively new. As MGI puts it, “as the technologies develop, the potential value that can be unlocked is likely to grow.” At the moment, exploration and sub-surface efforts use AI pretty effectively at improving targeting, locating new reserves, and optimizing drilling. Other operations, such as procurement and personnel, have some catching up to do. 

Source: McKinsey Global Institute

Why does it matter? There are at least three big implications to this. First, to some extent, AI will be able to do things that people do now. The result will likely be a continuation (and perhaps an acceleration) of the shift of personnel from offshore to onshore. And as with other forms of automation and digitization, there could be social issues associated with AI, with some jobs disappearing as a result. (Check out MGI’s recent work on the subject: “Jobs lost, jobs gained: Workforce transitions in a time of automation”).

Second, companies that learn to use AI well will have a significant competitive advantage. The reason is that AI is about making better decisions. For the oil-and-gas industry, these are where to drill, how much to bid for a lease, which bit to use, who to hire, will that hurricane disrupt us, and many others. The time to figure out AI is now.

And third, the development of AI is one more reason that oil and gas are not disappearing. We don’t hear much about “peak oil” any more; this was the idea that oil-and-gas production had peaked and could only go down. Peak oil has been predicted since the late 19th century, and got a robust lease on life when oil prices rose to more than $100 a barrel. But of course new supply came on line, much of it in the form of energy derived from shale, and prices went down. Peak oil always seems plausible but never seems to happen. In fact, in the most recent estimates, there are 1.7 trillion barrel oil reserves, or about 51 years of global production. That is up from 1.1 trillion barrels in 1996. Gas has seen a similar trajectory, with reserves of 186.6 trillion cubic meters, up from 123.5 tcm in 1996.

Sure, there is a case to be made that the world would be better off if it were less reliant on oil and gas. But their imminent disappearance is not one of them. The more interesting question is on the other side of the equation—peak demand The smart use of AI could drive down costs and increase efficiencies in the use of alternative energies. That could decrease demand for oil and gas faster than anyone thinks plausible at the moment.

To work, AI needs lots of trained people, data—and money: big transformations don’t come cheap. The oil-and-gas industry has all three. In addition, I believe it is increasingly seeing the opportunities. On that basis, I believe that the use of AI will help the industry continue to find and develop energy resources. Former Saudi oil minister Sheikh Zaki Yamani noted years ago that “The Stone Age did not end for lack of stone, and the Oil Age will end long before the world runs out of oil.” AI, as well as other technologies that will doubtless emerge, is just one more reason why this is the case.