Oil and gas firm BP is embracing automation and machine learning to run what it calls 'intelligent operations' and maximise business value in its supply and trading division, which accounts for 3,000 employees across 28 countries.
In an interview with Computerworld UK, Ayman Assaf, CIO for compliance, regulatory, risk and finance at BP Supply & Trading said: “We are moving away from technology being a cost-centre to actually a revenue-generator, and this is where we come into the new technology like automation.
“The energy business is being disrupted with digitisation, but for us, trading is an absolute. It depends on the digital technology side. We see ourselves as a technology company because there’s so much dependency we have on technology here,” Assaf added.
BP, which has its roots in the Anglo-Persian Oil Company more than 100 years ago, trades in power, chemicals and finance, with the physical trading – supply and logistics – one of its core services.
Oil and gas operations are typically divided into three sectors, upstream, midstream, and downstream. Generally speaking, upstream is the discovery and then drilling of oil and gas, midstream covers trading and supply, while downstream refers to the refining of crude oil, processing and purifying raw natural gas, and then marketing the products derived from it, for example lubricants, diesel oil, or petrol.
For the midstream, where Assaf's unit is focused, they are using automation to “look for opportunities where we try to maximise in the value of access in terms of where we buy from and where we sell”.
BP has already dabbled in automation through the adoption of new technologies like robotic process automation (RPA), across its business. “Automation is allowing us to consolidate data on the trading floor, using robotic process automation to mimic repeated processes,” Assaf said.
“This shifts the role of our analysts, increasingly freeing up their time to focus on higher value tasks. So rather than collecting data sets, they can spend time interpreting and interrogating the meaning of that data." BP is using technology from vendors like Automation Anywhere and NICE to drive this.
“We are seeing that automation is not only about streamlining a process to achieve more efficiency, but it is also allowing us to reduce process risks and to increase reliability across our operations,” Assaf said.
“With the introduction of AI and ML applications, along with our data lake, we believe there are huge business growth opportunities,” he added.
The rest of BP is currently in the middle of an IT modernisation, taking the lead from the supply and trade business.
"We traditionally, because we work in the trading space, have always taken the lead and in fact, they used us as a print to modernise the entire organisation," Assaf said.
BP is also taking this automation approach into how it runs IT by shifting to an Agile and devops culture.
BP has been working with a number of partners - Splunk, AppDynamics and Jenkins, to name just three - to get the best technologies across its business, and is also exploring open source, having recently entered a partnership with Red Hat to start working with microservices.
“We can practically go live every day because we can select particular functionality," Assaf said. "In some cases, we’ve been going live every two weeks with some new functionality because that continuous automation process across the whole conveyor belt that we built from idea all the way to production.”
This automated approach to production contrasts with the legacy of the over 100-year-old company. Assaf added that the business is complementing this by keeping its eye on the technology landscape, picking the best of breed technology.
“What we do now is different. We get ideas and we see who’s got the best technology and the technology providers like Amazon, Google, Microsoft and IBM and a lot of startups have technologies, but they don’t have the ideas of what to do with them,” he said.
BP currently has a number of proof of concept projects in the works and plans to bring some of the applications to market later this year. Watch this space.