How artificial intelligence is shaking up the oil and gas industry

“The net result of all that is we produce thousands [more] barrels a day than we otherwise would be able to do because we can intelligently manage the integrity of that operation," says Eyton.

At a time when AI is promising new possibilities to streamline, expand and advance activities across all frontiers, few sectors stand to benefit from its capabilities as much as oil and gas.

The International Energy Agency estimates digital technologies like AI could boost production volumes in oil and gas by 5pc and reduce operational costs by 10-20pc.

The industry’s adoption of artificial intelligence couldn’t come at a more crucial time. Last week, BP published the 2019 edition of its Energy Outlook, which explores the forces and uncertainties that will shape global energy markets by 2040.

In the report, BP outlines the dual challenge of increasing energy demands and the rapid transition to a lower carbon energy future in what it describes as an “evolving transition” scenario: a doubling in global GDP by 2040 driven by “fast-growing developing economies” and newer sources of energy as the world looks to clean up its act.

For a company like BP, the challenge is somewhat existential, as it must orient its business to enter a new world of energy in favour of clean fuel.

But to support this desired future built on renewable energy, all scenarios assessed by BP see requirements for “significant levels of investment in new oil” in the next 20 years at least. Figures from the IEA point to an expected growth in demand in 2019 of 1.4 million barrels per day amid a global fall in supply to 99.7 million barrels per day in January.

To confidently take on the challenges ahead, oil giants must find ways of applying frontier technologies like artificial intelligence to their operations across the board in a bid to boost oil and gas production.

Take upstream activities, for instance. Exploration of new oil and gas sites, field development and production require a careful understanding of a location’s geology and physics, as well as historical information that can give insight into a place’s past.

According to a report by PricewaterhouseCoopers (PwC) on oil and gas trends for 2018-2019, global upstream capital expenditure is forecast to rise 6pc year-on-year in the medium term after dropping almost 45pc between 2014 and 2016 following the crash in oil price five years ago, with exploration being “on the rise again for the first time since the global recession”.

Though BP doesn’t necessarily see itself as being a primary developer of the technologies that can bring greater efficiency to upstream activities, it does see the need to be a rapid adopter through a partnership-first approach.

Last month, BP invested $5m in start-up Belmont Technology, which has developed an artificial intelligence platform named Sandy to help it map out its subsurface assets and make better-informed decisions based on data processed through machine learning.

“Instead of using the hunch of a geologist or the guess of some trends, they are going to apply algorithms against all the data that BP is sitting on to help them figure out where the best place to explore is, getting rid of human bias or internal politics or regions of the world that may or may not be favourable,” says Mike Zamis, chief product officer at Sphera, a provider of risk management software.

Zamis, who describes oil and gas as “dirty and dangerous” businesses requiring careful management, believes the sectors are in serious need of artificial intelligence as the transition to solar and wind-powered energy systems will be a slow one.

Various activities such as those involving pipeline infrastructure and transportation in the midstream as well as refining of crude oil in the downstream portion of the business can see boosts through the use of artificial intelligence. 

“It'll be difficult to supply the world with all that energy and not use oil and gas in the near-term,” he says.

Sphera sees the role of artificial intelligence as one that “scrubs intuition out”. By using data processed by machine learning, operational risk at a refinery, for example, can be minimised, information can be fed to engineers about whether an equipment is impaired before a problem ensues and companies like BP or Shell can clean up their image by avoiding environmental disasters.

Oil and gas giants have sets of data in multiple places which previously have been difficult to access but having artificial intelligence systems aggregate it all will allow greater operational efficiency.

“The operational reality of a plant is changing hourly based on maintenance and other activities.... it's virtually impossible for human beings to make sense of all this on the fly,” says Zamit.

After years of pessimism following the collapse in crude oil prices in 2014 paving the way for a new found optimism rooted in technology, the industry is ready for change.

Of course, much of the new embrace of artificial intelligence is driven by competition. “The competitor set is not just the conventional competitors in the oil and gas industry”, but now includes “the Microsoft’s, the Amazon’s, the Didi [Chuxing’s], the Alibaba’s of this world," says Eyton.

“These are the people who are increasingly entering the space and finding ways of competing in what was conventionally an energy world, so you have to compete with them and these are formidable people.”