peshkov -

Shell CTO Yuri Sebregts talks about using AI to amplify the human impact of its workforce

Digital transformation is a multi-decade process for Shell, with artificial intelligence presenting the oil and gas giant with new opportunities to ramp up the productivity of its workforce, its CTO explains

This article can also be found in the Premium Editorial Download: Computer Weekly: Drilling into AI at Shell

Digitisation is fast becoming a process that every enterprise throughout the world needs to embark on, as technology becomes increasingly embedded into how the economy at large functions.

While some companies are under more immediate and urgent pressure to digitally transform their entire operations, for the technology team at oil and gas giant Shell, it is an ongoing, multi-decade process.

“digital transformation for Shell started 50 years ago, and it has been a constant evolution,” says Yuri Sebregts, Shell CTO and executive vice-president for technology. “We have had people working here who are at the edge of computational sciences, in the broader sense, for decades.”

Proof of that is the company’s newly announced artificial intelligence (AI) and machine learning push, which will see Shell primarily lean on these technologies to aid the predictive maintenance of the equipment used to run its upstream and downstream operations.

The initiative will see Shell combine the capabilities of Microsoft Azure with C3 IoT’s platform-as-a-service (PaaS) offering to create the underlying infrastructure for its AI effort, as the firm looks to build on its earlier work in this field.

“We have already done a number of cool projects at a smaller scale, with a single production unit in a single chemical or refining facility or doing predictive analytics on a couple of hundred pieces of equipment in a development-type environment,” says Sebregts.

“We are at a point now, though, where we really want to scale this to tens of thousands and hundreds of thousands [of sites], and more than a million pieces of individual equipment, and that scale-up requires you to go from more of a development environment to a platform that allows you to do that.”

“Digital transformation for Shell started 50 years ago, and it has been a constant evolution”

Yuri Sebregts, Shell

As Sebregts points out, getting the timing right on predictive maintenance tasks can be a big source of cost savings for a company the size of Shell, which is a major reason why it opted to press ahead on its scale-up efforts with the help of Microsoft and C3 IoT.

“It’s a really solid platform on the one hand, and it provides us with all the data security and rigour, which means we have assurance that it works, “ he says.

The setup also provides enough flexibility for Shell’s teams to bring their own domain expertise to bear in terms of how the platform operates, and it is extensible in other ways too, says Sebregts.

“It is not a ‘black box’ solution. It’s a platform where we can build our differentiating capabilities into that environment and it’s a solution that allows us to continuously pull in the latest open source developments so that you stay at the forefront and you don’t get locked into a system that is outdated after a time.”

Exploring the use cases for AI

The decision to draw on the Microsoft Azure public cloud to deliver the project came to light days before the start of the software giant’s Ignite developer and customer conference in Orlando, Florida, with Microsoft CEO Satya Nadella hailing Shell’s efforts to wholeheartedly embrace AI across multiple parts of its business.

As well as the predictive maintenance project, the company has also created a service called Machine Vision using Azure-based deep learning technologies that combines CCTV footage with internet of things (IoT) devices to alert employees at its service stations to potential safety hazards occurring on the forecourt in real time, such as someone lighting a cigarette or driving erratically close to a petrol pump.

There is also potential for this technology to be applied in a stocktaking context in Shell’s warehouses and petrol stations, says Sebregts, so that staff can intervene and replenish suppliers, as and when needed.

Beyond its retail sites, robotics is already commonly used to install equipment in offshore environments where it would be hazardous and impossible to send humans, and Sebregts also sees potential for AI to enhance how that work is carried out in future.

“Our inspection robots regularly go underwater to see if all the equipment is where it is supposed to be and that there are no leaks or whatever, and that is an increasingly autonomous process,” he says.

“We believe in technology as a key differentiator of our business”
Yuri Sebregts, Shell

“The interpretation of the video footage generated can be automated more and more, with pure machine learning algorithms autonomously interpreting the camera footage and immediately alert when it sees something that is not supposed to be. In the past, that was not possible.”

Sebregts adds: “Really, an industry like ours has an almost infinite number of applications you can think about and apply this technology to. The challenge is choosing the ones you will start with and get to full scale, rather than losing yourself in 10,000 different projects, but not scale any of them.”

Shell is a keen consumer of Microsoft cloud technologies across its business in other ways, too, with its 84,000-strong workforce making active use of the Office 365 online productivity suite, while its development teams are flagged as reference customers for the Microsoft Azure DevTest Labs service.

But that is not to say it does not use cloud technologies from other providers for other use cases, says Sebregts, but there are a couple of reasons why – on the AI side – it has gone with Microsoft.

“For us, it must be a platform solution, and it must be a big enterprise-grade solution, in terms of the quality, and it must be open in a way that means it works with other service providers and it embraces that concept. That works well for us,” he says.

“We like the ability of Microsoft to combine AI in the cloud and at the edge. It really makes that easy for us. So there is a series of good things why we make that choice.”

Using AI to amplify human efforts

While debate continues to rage across the technology industry about the impact AI and automation will have on the global workforce in years to come, Sebregts says that Shell sees AI as a means of amplifying the work humans can do, rather than taking it away from them completely.

“Like other technology breakthroughs in the past, AI drives up productivity tremendously,” he says. “So, to me, it doesn’t replace humans.

“We have a certain amount of human talent and capability in our organisation, and access to a certain amount of natural resource and a certain amount of capital, which is all traditional economics. So how do you create the most value from that? AI is your new tool and next step up to unlock that productivity.

“It is not so much about replacing people, as it is about giving them the tools to do more, to have bigger impacts and make new things possible.”

For many companies with big AI ambitions, sourcing staff with the right skills to support their plans can be problematic, but this is an area that Sebregts says Shell’s heritage works in its favour.

“We’ve always had a relatively large technical workforce, around maths and IT, and that gave us a head start as we’ve always been a very tech-heavy company,” he says.

“The most iconic application in our industry is how you interpret seismic signals to picture what is in the sub-surface, and in the ground and that has always been quite an advanced information technology.

“Also, if you compare what Shell has spent on R&D over the last couple of decades with other major players, you see we’ve always been at the high end of the band because we believe in technology as a key differentiator of our business.”

Read more about Microsoft and AI

However, what has become apparent as more and more enterprises move to incorporate AI into their technology strategies is that competition for the best talent in this field has started to rise markedly.

“What we find to be compelling for really innovative, talented people is really cool problems to work on,” says Sebregts. “We can often offer the most complex problems to solve, and that’s just fun for people who like tech.

“They like to work on the really complicated problems. They hate doing boring things. So from a strong base, we’ve actually been able to grow our size and capability in this area, and we continue to grow and attract people to that part of the company as we continue to expand the number of use cases for AI.”

Looking ahead, Sebregts says technology will have a critical role to play in how Shell addresses one of the energy industry’s biggest challenges to date – how to meet the growing demand for power across the globe in a sustainable way.

“This is the challenge for the world as we segway from today’s energy system to an energy system that has much lower and zero carbon emissions, while energy demand continues to rise rapidly, because the world’s population is growing and people aspire to a better quality of life,” he says. “Digital unlocks new opportunities to enable that transition.

“It will be a digitally enabled transformation, so I see that development bringing lots of new and unforeseen innovation as an exciting field to be working in.”

Read more on Infrastructure-as-a-Service (IaaS)

Data Center
Data Management