Energy

We met with Olav, Team Lead for Operations and Maintenance AI to learn more about how Equinor is using AI and Machine Learning to enable better decision making for the future.

Olav, you are very passionate about AI and Machine Learning. What is it that fascinates you about this topic?


"Yes, it was one of the reasons I wanted to work in Equinor. The organisation has an incredible amount of data, gathered over many decades and through tens of thousands of employees and many millions of operations. This, in the combination with the need for the energy industry to shift direction, is very interesting. AI and Machine Learning represent a largely untapped potential for a data-heavy company like Equinor. The possibilities are endless, and the development of the technology happens at an extremely high speed. That’s where the thrill lies in it, for me.”

How do you see the role of AI within operations and maintenance in the energy industry?



“I’m convinced it will play an increasingly important role. Our industry is entirely dependent on finding new ways of working to secure both our own success, but also to secure safe and effective operations for the future. To be able to this, we need to make sure we’re doing the right things, at the right time. That’s where the role of artificial intelligence comes into play. If we manage to use the immense amounts of historic data that we have, we can train AI models, to make more accurate decisions, faster.”

More specifically, what types of decisions can AI help with?



“For Equinor, there are many processes that involve complex decision-making related to both maintenance and operations. This goes for everything from maintenance on wind turbines and oil riggs, to planning of supply chains, optimising production and predicting environmental conditions and impact. This data is abundant, but not always easily accessible. By liberating it and using it to train AI, we will be able to make much smarter decisions when it comes to when and how to do different types of maintenance, pauses in operations, production planning, and so on. It has the potential to free resources – both people, money and time – to do better work on the things that AI can’t solve.” 

Can you see any risks with the use of AI?

”There are some risks, for sure. When it comes to working with it in our Equinor context, we deal with complex operations and handle products and processes that require a high level of safety, so we can never compromise on risk mitigation. We never launch any type of technology without thorough testing and evaluation. With the European AI Act in place, we are of course adhering to that and are so far focusing on the low-risk categories. This means we use AI solely for decision support – not for automated operations. We also maintain full transparency around what artificial intelligence is used for, and not. This is absolutely essential.”



Where is the human factor still relevant in this?



“Oh, almost everywhere! Even though the development of AI is happening very fast, there are things that, so far, humans do better. I think the idea should be to use AI and Machine Learning to enhance and complement human skills – not to replace them. Like in the example I’ve mentioned with Equinor – we need to analyse large amounts of data to help people make better decisions. AI is excellent at that. But people are still better at seeing the big picture, to connect the dots and, of course, to relate to other people.”

What is the role of design in developing these types of decision-making tools?



”The systems we’re dealing with are extremely complex. Having designers and design competency on the team is really helpful – designers are great at bridging the gap between abstract and complex systems, and the people who are going to use them. You help us make sure we’re all seeing the same picture and have the same end goal. For example, we have people from EGGS working together with us right now to ensure that the tools we’re building are properly integrated into the work reality of the people who will be using them. Everything from dialogue with developers, stakeholder management, testing, prototyping to the overall direction of the project, is being supported through their help. But for that to be possible, design has to enter already at the start. It can’t be an add-on at the end.” 

What are some of the common mistakes when companies want to implement AI?



”I think for more traditional companies, that are used to running projects through a waterfall approach, there is one evident challenge – you can’t innovate like that. If you want to be innovative, for example through the use of AI or other technology, you need to work more agile. Another important thing is to manage expectations very early in the project. Involve stakeholders early and start with a feasibility study – what is doable? Test the idea. There should be a low threshold to try things, but the threshold to throw what’s not working out the window needs to be just as low.”

Equinor – on an AI journey to better decision making