Patrick
Welcome back to the Barclays Brief.
AI is everywhere. It's changing how we bank, how we shop, how we learn, and even how we manage our health. But when it comes to the energy sector, the power behind everything we do, it feels a lot less tangible to me.
Joining me in the studio here today in London is Lydia Rainforth, head of European energy research in our Equity Research division. Lydia, thank you so much for joining me today.
Lydia:
Patrick. Thank you. And I've been a big fan of the Barclays Brief since it started. So, I’m actually really thrilled that you asked me to come on here.
Patrick
Well, that's very kind of you to say.
And, well, in that case, you know, the format, it's ten minutes. Concise answers, one subject. So, let's get into it. Clearly, AI has been shaking up lots of industries, and I listed some of them just a moment ago, but the energy sector just feels less obvious to me. What does AI really mean for the energy sector and why should we really care?
Lydia
So, this idea of what impact can you have on industrial AI has been a really challenging thing to think about, but it is making such a huge difference.
And I’ve been , looking at this sector for 25 years now, and this is one of those transformational opportunities that I think there is. This idea of how do you deploy AI and particularly agentic AI across the energy space? And when I think about that, I probably should explain what I mean. Generative AI is the stuff that we all use every day.
We can create images, it can create notes for you. Agentic AI is more a system that independently sets goals that can make plans and take actions, all without really much human interaction. And the impact that this can have on the industrial space and energy is remarkable. And I don't think There's been enough time spent on this.
Patrick
Yeah. And you mentioned the note that you published it’s called ‘Agentic AI, the $80 billion game changer for energy’. I was rereading it again on the tube this morning, and you talked about this $80 billion number in terms of productivity gains by 2030. Now, that's not a long way away, and that's a big number. So where do you see the biggest wins for the sector?
Lydia
Yeah $80 billion I think we probably underestimated overall I think we went with the cautious end of what is possible. So, the scale of this is significant. And when you think about the oil and gas sector, this covers everything from finding the oil and gas to getting it out of the ground, through to refining and making into useful products, through to the marketing it and selling it to you and I at the pump. And most of that opportunity is going to sit in the upstream.
But if I just give you a little example of something that I like is in the marketing sector. So this is less than 5% of your, sales, really. But if you imagine that you're driving up to the petrol pump, the gas station, and it will recognize your number plates, and it will also recognise who is in the car with you, and it will then go through, well, what would be the typical profile and what would someone buy that is of this profile?
And so for me, I've got a beautiful golden retriever, and if he's in the car with me, it may say, yeah, we'll offer her sausages and I will click yes, we'll buy sausages for the dog. And so the idea is that you try and boost your sales that way. Now some of this technology and the ideas have been around for a while, but to be able to go number plate recognition, people recognition and then run through what those offers are to get them on to the screen. That's a big step change to actually get this done. And if you can double those sales, that suddenly starts adding up quite materially, but it is a small part of it and there's a lot more to do on the other sectors.
Patrick
It's a small part of it. I mean, I'm intrigued as to what they would offer me with my three kids in the car.
But let's talk about the bigger part of it, because yeah, I think you said about 80% of the $80 billion is going to come from upstream productivity gains. Can you unpack that a bit more? Because some of the stuff that you were writing about is fascinating, what they're going to be able to do in terms of drilling and locating oil?
Lydia
Yeah, and it depends on how geeky we want to get?
Patrick
As geeky as you want.
Lydia
Brilliant. So if I go into this and actually let’s start with the most difficult part of this: actually finding oil. At the moment, the success rate for finding oil is probably about 1 in 10, when you go out and drill.’
So if I think about the detection of oil and gas, you do it through ‘seismic’ and effectively, this is sending sound waves down to the seabed. And it comes back up and it generates an image for you.
You've got AI which will fill in the image if there's any gaps in it, that's really easy to do. Then you've got the next part, which is interpreting the image. And what you're effectively trying to do is find what looks like an oil and gas pocket. That is several thousand metres below the sea level. So can have 3000m of water and that can be down to another 2000m to go and find this.
And geologists spend months looking for this stuff and looking for what's the right pattern, where should you drill. They are now able to run through several thousand models within a couple of days. So you're taking the time down significantly, and you're now able to drill into what you think is a prospect to within 12 inches.
That is something where actually previously, if you were with 100m, you think you'd be doing well. So there's a lot around that where your exploration success rate has suddenly gone up. So it used to be 1 in 10. Now you actually can be down to 1 in 3. That's a material step change and that's just on finding the oil.
If I then fast forward to actually how you can produce it, you can actually improve your safety. You can take people off the rig floor, and I can increase the maintenance, increase theproportion of time that your field is operational.
Patrick
How does the maintenance work? Because presumably a lot of the materials, both on the rigs and in the sea, you know, rust and, need to be replaced regularly. How does AI help facilitate, improvements in productivity there?
Lydia
So, for example, what you would do is send out a drone which would be automatically sent out to go around the vessel or the refinery and try and detect rust areas, or is there a point of failure. And so it'll recognize what is rust, what could be a significant issue.
And then what you'll get to is that a computer program will say this is a problem, then it will send out a robot to go and fix it. And so you're actually keeping the uptime going. So can I get 2 or 3% more production per year out of the facilities, which obviously reduces costs, to actually get this done?
So that's an incredible way of actually improving it. But it's all done without human intervention. The other bit is let’s say a part needs replacing. You recognise what part it is that needs replacing. It sends the signals back to the onshore, business that then gets printed in 3D, and then a helicopter takes it, out to the field.
So there's a lot of things that you can do, but there is a lot of efficiency that can be gained by using these computer programs.
Patrick
I mean, it sounds completely transformational for the sector. Have you ever seen anything like it?
Lydia
I haven't, and we spent 12 months looking at this, as a process. It's, it is transformational. And I know I've used that way before, but I do want to use again, because it is so important. There's been a lot of time where we want to talk to the heads of the businesses the heads of private companies that are actually having access to the data. How do management teams think about structuring things?
How do you get adoption of this? So this is one of the biggest game faces that we have. We think we've seen incremental improvements in technology before, but nothing as game changing as this.
Patrick
And what are the biggest challenges that the management teams face? Because, you know, you and I work in the sell side, Research business, and we're using AI a lot more this week than we were last week than we were, the week before.
But, you know, we're going through a process of learning how to utilise AI for productivity gains and improving products. What's happening from a top down perspective, as the management teams are trying to implement changes to these massive global companies?
Lydia
This is the hard part because the technology exists. But as we know that from our own experience, actually getting people to use it, can be harder.
And I think there are three things for us. One is about a mindset and a vision that's set by the top level. The second is the ability to scale things. And the final point is really the biggest factor of all is the data.
Do you have good data and how do you make sure that everybody has access to it? Because anything in the energy industry or industrial AI space: it's got to be safe. We cannot afford to have accidents. So where we ended up with this is we set up a new framework.
It's called the ‘Barclays AI Readiness Framework’. And what we do is try and assess where companies are on this. How much data do they have? Is the data structured well? Does the management team have the vision to be able to do it? And ultimately are they set up to start deploying these and gain those benefits from it.
Patrick
Interesting. Andso if you fast forward to 2030, what does an AI powered energy company look like, you know, across the whole piece? And do they all bunched together and look fairly similar,or are you going to have winners and laggards and a big gap between the two?
Lydia
Yeah. So scale is going to matter here. The more data you have, the better quality data and the ability to allow everybody in the organisation to be able to access that and create use cases is really going to matter for it. When I think about what is that AI enabled company look like, it's one that has production that's probably 3 to 5% higher than it is the success rate on exploration is much better.
And you’ve probably got a 40 to 50% improvement in free cash flow from where we are today. That's a big step change from where we've been. And that gives us a lot of options. Say we're safer, and we've got better profitability.
There will be differences between the companies.
Patrick
A game changer indeed. Lydia, thank you so much for joining us here today.
Lydia
Thanks Patrick, I’ve really enjoyed that.
Patrick
What struck me most in this conversation is the sheer scale of change. AI isn't just tweaking the energy sector, it has the potential to completely rewrite the playbook. From drilling and maintenance to marketing and logistics. It's everywhere. But it's not just about algorithms.
It's about people. It's about culture. It's about skills, and it's also about data. And the companies that embrace this early could pull a long way ahead, creating a whole new set of winners and laggards. So find out more about which those companies it might be, clients can read Lydia's reports on Barclays Live, and there's a link in the show notes.
And finally, don't forget to leave us a review and hit subscribe if you want to be notified when the next episode of the Barclays Brief comes out.
