AI’s Role in Energy Markets


AI’s Role in Energy Markets


We have a very large task ahead of us.

To create a safe and equitable future, we need to figure out how to feed a growing planet and stimulate economic growth, whilst reducing emissions and managing resource use sustainably. And this has to be done in the face of a changing global climate.

Enter artificial intelligence.

Energy needs to change – it needs to fulfil the “energy trilemma” of being cheap, reliable and clean. At the moment (on a global scale), it’s quite cheap, moderately reliable, and mostly not clean. Adopting AI will make the energy trilemma a lot easier to achieve. Over the course of 4 notes, I’d like to explain how.

Artificial intelligence is an umbrella term encapsulating many different verticals – the most well-known today being machine learning (ML), computer vision and natural language processing (NLP). You’ll have come across these technologies multiple times today – if you have Alexa, you’ve used NLP; if you’ve walked outside (particularly in a city) your face will have been processed by a computer vision aided back-end of a CCTV system; and machine learning optimises your internet searches and ad preferences daily.

All AI today is “narrow AI” – much better than us at one very specific task, such as processing millions of data points and spotting patterns the human eye can’t see. The AI the media wishes to scare you with is AGI – artificial general intelligence, and we’re a long way off this all-thinking, all-doing AI.

It’s hard to ignore artificial intelligence at the moment. A lot of companies seem to be moving towards offering AI solutions (in my opinion, most have AI solutions looking for problems) and early adopters (Google, Amazon) are moving ever further ahead of the rest of the field. However, there are some technologies out there that can genuinely change the way most companies and sectors do business, and it’s no stretch to say that AI will fundamentally change human society.

Big changes will be seen in all sectors and corporate functions due to AI (sales & marketing is leading the way at the moment), and energy is particularly ready for this change due to the nature of its vast, global supply chains (ripe for optimisation), and its reach into people’s homes, offices, and factories.

For a good example of this complexity, think of the amount of processes involved in the oil market. From finding a potential oil play in Texas, to assessing geodata, to extracting the crude, to transporting the oil to a refinery in Houston for processing, to transporting the various products to end-users such as a car in France or a factory in Mexico, each stage offers potential for optimisation – and I’ve vastly simplified these stages for brevity.

This can be very intimidating for a company looking to adopt these technologies. Where do you start? What problems are you trying to solve with AI? And from whom can you learn?

I’d like to inspire the energy sector to answer these questions and embrace these new technologies wholeheartedly. I believe that in doing so, the sector will become considerably more efficient than in its current form, with obvious positive benefits for our planet.


AI’s Role in Energy Markets: Oil & Gas


Following on from my first, introductory piece on AI in energy, I’m going to start adding some more colour to the space. This second note focuses on AI in the oil & gas industry, and some recent interesting developments in this area. This AI in energy series is written in my personal capacity.

Energy is a vast, $6 trillion industry, with a correspondingly big supply chain. As a sector which has historically had deep pockets and a talented workforce, energy has nearly always been quick to adopt new technology (and woe betide those who don’t adapt fast enough), and adopting AI is no exception.

Upstream oil & gas has some excellent case studies for adoption of AI. Prospective oil fields essentially depend on a good understanding of geology and hydrological data. Using AI to spot patterns in this data and draw new conclusions from existing data, will enable the humans in the loop to make a better decision, and free engineers up to focus on more important, strategic thinking (such as scenario building on peak oil demand). Some of the world’s biggest energy companies have had success adopting AI – Eni, Total, Gazprom and BP present some good case studies – and though we’re very much in the early stages of this space, there’s some exciting lessons to be learned.


Analysing millions of geodata

The adoption of AI in oil & gas is particularly pertinent in the era of raised oil prices ($60-$80/bbl) and peaking oil demand – ensuring that an oil & gas company’s decisions are as well-informed as possible means less risk and more taking advantage of current high prices. Energy companies can worry less about sinking CAPEX into a project that yields far less than expected in an uncertain future for your products. They want more oil to sell, now.

With this in mind, Eni has partnered with Stone Ridge Technology, using their ECHELON petroleum reservoir simulation software. Eni’s HPC4 supercomputer (the world’s most powerful commercial supercomputer) processes 100,000 reservoir models in under 16 hours, which would take human engineers weeks or even months to do.

In a similar vein, Total partnered with Google Cloud to apply AI solutions to subsurface data analysis during oil and gas exploration and production. Total’s geoscientists will work alongside Google’s cloud machine learning experts, with the aim of interpreting subsurface images far better than the human eye using computer vision, and automate analysis of technical documents using natural language processing.

BP recently invested in BeyondLimits to aid the supermajor in “operational insight” and “business optimisation” – read making sense of the vast amounts of data an oil well produces (let alone making sense of the data a supermajor itself produces).

Outside of Western Europe, Gazprom Neft and Yandex signed an MoU recently to collaborate on the “analysis of large volumes of geological and technical information”, opening up Gazprom to more accurate information on future well profitability and sharper project risk analysis.

Judging by the enthusiasm with which the industry has adopted AI in its upstream business units, there’s clear value in using machine learning to improve decision making on your most valuable assets. We can expect to see a continuation of big oil adopting AI for this purpose.



Oil & gas uses a variety of autonomous and manned machines in the deepwater and onshore exploration and production of core products. Total’s ARGOS challenge aimed to address the difficulties that robots have working in extreme conditions, and the safety risks that these challenges raise. Opening up the challenge to external companies and inventors enabled Total to access external expertise, coming up with a solution to the safety problem for a fraction of the cost and time of doing the research internally.

All of the above are process-and-data-heavy tasks which humans are not able to perform as well as machines. They can also replace the dirty, dangerous and difficult nature of much of oil extraction. This enables the human in the loop to be freed up to focus on higher value tasks, such as making more informed decisions about assets that can cost multiple tens of billions of dollars.



As an energy company, your biggest asset is your customer base. Away from the upstream, there are some good examples of companies implementing AI solutions in their downstream, customer-facing operations.

AI, specifically a branch called Natural Language Processing, allows any company to act like they have thousands of support staff to help their customer base. By learning from each interaction with a client, a chatbot can answer client questions quickly and efficiently in most cases. Difficult questions can be sent to a human when necessary, dramatically reducing the amount of time your expensive staff are answering mostly low-value questions.

Shell has a couple of B2B chatbots for customers of its lubricants services. Their LubeChat bot serves this niche particularly well, as a lot of the questions customers ask will be very similar (pricing, delivery options) and easily answerable with a good chatbot.

There are so many options for a chatbot or NLP solution to give your clients and staff a better experience that we won’t go into too much detail here. Suffice to say, the world is talking, so making it easy for people to talk with your business has enormous benefits.

AI can also provide opportunities to engage with your customers in ways that weren’t possible a few years ago. To improve point-of-purchase experience, ExxonMobil has partnered with IBM to offer customers tailored promotions when buying fuel or a carwash through their Speedpass+ app. The app uses IBM Watson’s Campaign Automation on IBM Cloud, offering ExxonMobil’s customers targeted offers and loyalty programs through their device. This encourages greater customer loyalty, with the added benefit of more revenue per customer. BP has a similar service with their BPme app.

We’ll see a lot more AI investment from the oil & gas industry in the next few years. Faced with uncertain future demand for their products (electric vehicles, renewable energy uptake and heightened competition being three key areas of uncertainty), the future, more efficient supermajor will look very different from today’s sprawling conglomerates.



George Hackford
Commercial Director at CognitionX – AI, blockchain and emerging tech


Leave a Comment

Your email address will not be published. Required fields are marked *