How Is Shell Driving Its AI Future?

Well, hello, and welcome, everybody. It's absolutely brilliant to be back with you. I think it's fair to say that Amjad and I last gave a presentation at MathWorks Expo in 2016. So treat this as a bit of an update of the Shell journey six years on. And hopefully, you'll see some degree of progress since then.

My name is Dan. I'm the Vice President for Computational Science and Digital Innovation. And Amjad now leads one of our largest data science groups in the organization. And we're here today to give you a bit of an update of how Shell is accelerating the process of energy transition, leveraging digital innovation, AI, and data science.

Before I go on, of course, it would be remiss of me not to say, please don't invest in Shell stock on the basis of what I'm about to say. And of course, everything that-- all the opinions that I share are going to be my own. Maybe jumping straight to it, last year Shell announced a new strategy.

And I think it's important always, when we talk about the progress we're making in digital innovation and AI, to put that in the context of our overall corporate strategy. Shell has set out some very bold ambitions for the sort of company that we want to be in the future.

We've said that we want to be net zero by 2050 or sooner, in step with society. And what that means is it covers off our Scope 1, 2, and 3 emissions. In other words, all of the emissions that we're directly responsible through our operations. As well as all of the emissions associated with our products.

Now, as many of you will imagine, that is a very sizable and lofty ambition. And it's a challenge in an energy system that today is very heavily dependent on hydrocarbon based fuels. Similarly, we've also set some interim targets. We've said by 2030, we want to reduce our Scope 1 and 2 emissions-- those are the emissions that we are directly responsible for through our operations-- by around 50%.

And once again, that's a very challenging task. So net zero is very, very important. And I think what's fair to say is that if we want to achieve those targets, the energy system of the future needs to look very different to the one that we operate today. And we believe that digital technology is going to play a critical role in getting us there. And we'll unpack that more as we go through this presentation.

But it's not just about net zero. It's also about how we do it. All of us benefit from the energy sources that we have today. The prosperity that we enjoy in the Western world is largely driven by having access to energy amongst other things.

And so energy is very important. It's an important part of our lives. It's an important part of wealth and prosperity generation. And we want to make sure we keep that in mind as we go through energy transition. We also want to think of those who are currently in energy poverty. You don't have access to the same privileges that we have in the west.

And so powering lives is another tenet of our strategy, as is respecting nature, where we want to ensure that we are able to, as we develop a new energy system, do it in an environmentally sustainable and responsible way. And of course, we need to remain profitable.

We're a for-profit corporation, and we need to ensure that what we do continues to drive the shareholder value that our shareholders expect. So this is a challenging task. And I think at the core of this is a sense of fundamental change. And I find that very exciting. I believe that we're going to see massive shifts in the next decade in the energy system that exists today.

And I believe that actually one of the ways that we can get there quicker is by leveraging digital technology and AI and data science to the fullest extent possible, to ensure that we're able to rapidly transform the entirety of the energy sector, and also, to manage a future energy system. And maybe just to unpack that a little bit, at Shell we've been really focused on driving digital and AI into our business as a core enabler for the energy transition that we want to go through.

We see that in three core ways. The first is that we recognize that digital technology has a huge opportunity, a huge potential to transform our core business, the business we operate today, making it more profitable, but also making it more energy efficient and reducing the CO2 emissions that come from our existing assets.

And we believe that this is one of the key levers that we need to pull in order to meet some of the ambitions that we've laid out as an organization. But we also need to do more than that. We need to use digital technology to transform the energy system, and design a new energy system, which is going to have a lower CO2 footprint. So leveraging digital and simulation in the design process of the new energy system is fundamental to reducing the CO2 footprint.

But finally, also, as we see an emergent energy system, which is more diverse, more distributed, and more complex, we recognize that optimization, data science, big data is going to be fundamental to allowing us to operate an energy system of that nature.

And so we're investing too in how do we operate something like that. How do we operate it remotely. How do we ensure that it's digitally enabled. So right across Shell's business, you see digital transformation happening at an industrial scale. And we're making great progress.

To put that in perspective, we have about 300,000 people already using Shell applications for EV charging, as we grow our EV network. We're developing over 100 different AI applications across our organization, which we're deploying to production at the moment in various states of maturity.

We have about $1.9 trillion rows of data aggregated across everything from our energy and chemicals parts, to our wind farms, to our upstream facilities, to emerging new assets like hydrogen electrolysis. And we're bringing all that together into a common data plane, that allows us to bring all the sensor data together for the purposes of data science machine learning.

We're monitoring north of 10,000 pieces of equipment using AI today, as part of our proactive technical monitoring processes, all run from remote digital centers, where AI is embedded into the day-to-day work processes. We've got over 80 million digitally connected customers who visit our retail stations and also participate in part in our energy retailing, where they purchase electricity from Shell.

We also have north of 2.4 million people in the UK, who are leveraging our AI powered loyalty program, Go Plus. And we have over 350 professional data scientists and computational scientists and AI engineers who are helping us to develop solutions right across our organization, and of course, a much broader number of citizen data scientists who are using tools like MATLAB and Azure ML to develop their own models for their own parts of the business, which in turn can then be scaled up to solve enterprise problems.

And as part of that, we're extending the awareness of AI through what we call the Shell.ai network, which is a community of interested parties, everything from managers to practitioners who want to understand the journey that we're on in using digital technology to transform the energy sector.

And that represents now about 5,000 people. These stats are all hard to get your head around. But what I want to paint a picture of, with this, is the scale and the pace of transformation that we're undertaking, and the level of ambition that we have to truly transform the energy system using digital technology and AI.

And so it's a great privilege. Amjad has led the journey of leveraging MATLAB and MathWorks technology within Shell at scale as part of this journey. And it's been an exciting journey. We've come a long way. And I wanted Amjad just to come and talk a little bit about the progress that we're making with MathWorks technology. So Amjad, over to you.

Thanks, Dan. And it is a real privilege to be invited back to speak to the MathWorks community, and also co-present with my mentor, Dan Jeavons. I'd like to say that the last eight years have been a privilege for me on this journey. When I started out, when we came into this space working for Shell IT, we were challenged with what was going to be our digital strategy. How were we going to leverage all the great things that were happening in the marketplace with cloud, big data, APIs?

And the first thing we actually did was look around, do a 360. What was happening in Shell? What was happening in the external marketplace? Who were the key suppliers and vendors we were working with? And from where I came from, my previous job, I knew that lots of developments were happening with MATLAB tools, the packages and stuff. So that was a good place to start.

And eight years ago, we sat in the same room with MathWorks executives, and basically explained what we wanted to do, and whether MathWorks could help us on that journey. The first thing we looked at were some of their products and tool sets, and how we could kind of use them. So we embarked on a strategy of doing lots of experimentation testing, proof of technologies, proof of concepts, take some use cases, see if we could kind of deploy in some of the native cloud technologies.

And the success of those early collaborations working with the great people at MathWorks, the consultants, the developers led us to actually cement and purchase what I think is today still a great product, the MATLAB Production Server. That was going to be our deployment engine for deploying algorithms that we were developing in-house with the help of MathWorks engineers, to show us how to deploy, but actually teach us how to do it ourselves.

And on the back of that, we built several environments, which we typically call development and acceptance and production, and we deployed our first real production application in 2016. It was what we call the quest solution, a surface monitoring laser AI solution, managing to look out for exceedance and leaks across-- above a, well, actually Shell's first carbon capture and storage facility, for a facility in Canada.

We were very excited by this. It was launched even officially by our CEO, Ben van Beurden at the time. I remember it. But we then left that ticking along, collecting data. The lasers were scanned, and the data would be collected into our systems, and would be able to access that through our normal tools.

But that solution was a 24-7 solution, and it's been running up to this point in time. In fact, we're going through a refresh on that. We're just replacing the lasers, but it's a great solution to date. And also, carbon capture for Shell remains one of the key enablers to meet our carbon ambition.

So what happened after that Well, the relationship grew. More use cases, more testing of great MathWorks products. And because of that, we were cementing this relationship for a COE being created.

We looked at where MATLAB was being used across Shell. I think it's safe to say that MATLAB has been in use in Shell for best part of 25, 30 years. And what we found was that it was all a bit discrete, disparate, and we needed to consolidate. So we entered into an enterprise license agreement. Basically, we want to understand what tools and tech we were using.

There was a lot of redundancy, because people had come and gone in the organization. So I think it wasn't giving us an accurate picture. But it also allowed us and Shell to create COEs, joint COEs that would work seamlessly together. And that proved a great worth, because that meant the next set of challenges that we had, working DevOps, building a DevOps framework to accelerate, operationalize projects while we were working on digital solutions.

Shell helped enormously in our Azure environment to create and test DevOps solutions. And we embarked on that with a key project, what we call capture connect. But basically, we build and manufacture catalysts. We sell it around the globe to different chemicals and refining companies. And we need to monitor the health, the activation, and health of that catalyst over time.

These catalysts can last for three to four years, but you have to keep intervening and seeing whether there's still enough life in that catalyst, or whether you need to change it out. So we digitized all of that. We moved away from the manumatic processes. And we use DevOps to do that, because we continually need to develop and redeploy, test, work with the business and the customer. So I was very proud of being involved in one of the first cases in Shell where we have DevOps working for us.

So then, where did we go next? Well, MathWorks keep producing these new solutions and products. I'd like to think that we influenced some of that in time, and some of the things that we thought were going to be the big trends. But what came next was like the MATLAB web app server.

I and others believe that the true power of digitalization is enabling people to do some of this stuff themselves, and being able to build small app stores. So MATLAB and MathWorks worked with Shell to test and improve on some of their web app server products.

And for that, the consequence for us was we were going to deploy-- redeploy, actually, another MATLAB application, what we call MADA. MADA is a modern analog data analysis tool used to kind of analyze the subsurface in geological analogs to work out where the best place to do further exploration or drilling.

And that application went live in 2020. And then moving on, we're in this new realm. Dan has been a great advocate of working collaboratively with all our key partners and stakeholders. C3 is a-- C3 AI is as a major platform that we're invested in. And we wanted to invite all our key partners and vendors to come and join. I'm not going to steal his thunder. I'll let him explain more about OAI. But that is something that was launched, and I'm delighted to say MathWorks are on board.

Moving on to the future, I know there's a great product coming downstream, which we invested to go live in Shell later this year, MATLAB Online Server. We believe that the DIY element and the DIY community will grow. It will grow across all enterprises. And I think the MATLAB Online Server for a certain type of user would be a very vital and valuable tool.

And then, much beyond, we're in this energy transition. There are things that we've not even touched upon, things like Simulink and Simscape, and these are tools that we know got great value. We've had one or two people who use them. But what we haven't really done is really powered these tools into what we want to do next. Modeling, simulating, physical objects, marrying them with physics-based equations and stuff, and also optimizing and doing networks. That's the future. That's where I hope the relationship will continue to be strong and grow upwards. But to you, Dan.

Super. A huge thank you to the whole MathWorks team. This has been a true collaboration, and we've been working hand in glove with MathWorks throughout the process. And we've seen real maturity in the technology as it's advanced. And we're excited about what the future holds.

And maybe, to build on that, what we see is the potential now from the place that we've got to for digital to become really part of the way in which we do everything across all of our businesses. So that rather than this is becoming an optional extra, it really becomes embedded into the core of what we do.

And I think a lot of that is going to be around the confluence of science and data-driven modeling. And Amjad alluded to it, of where this is going. What we see is that everything that Shell does is physical, from running remotes, production operations, through to solar parks and wind farms, and even into retail stations, which increasingly are becoming very complex as we add LNG and charging stations to an existing fuel retail station.

And so we see these miniature assets emerging right across our business. And as we see that, of course, we see the potential to bring together the science that Shell has always had great strength in with the data, and bring that into, if you like, common digital solutions, which transform the way in which we operate our business.

And of course, this has been a strength of MathWorks for many years, with the Simulink, and also MATLAB products. And we see the potential to bring that together. We see the potential in some of the projects were already running. Amjad talked about capture connectors. We're using a combination of data driven and chemometrics techniques to monitor the catalysis processes.

We're also looking to extend that into high resolution mass spectroscopy. And also, we're trying to develop into other areas, like for example, optimization of processes, where we look to create new ways of generating synthetic gas to liquids technology, where we start to integrate, if you like, the design process for the engineering, with the actual data science to allow us to optimize the configuration of new LNG trains that can produce both traditional liquefied natural gas, but can also create synthetic natural gas from CO2 intake. And so these types of technologies that we're developing within Shell are all dependent heavily on our ability to integrate both the chemistry and the physics with the data science, as I've mentioned.

We also see the importance, as Amjad alluded to, of enabling this through partnerships and integration. So no one solution is going to solve it all. We recognize the benefits of cloud, of data platforms, of modeling technology in all its forms, as well as a digital application deployment at scale.

And so we're trying to bring that together to create an integrated digital ecosystem that allows us to form a common backbone across all of our plants. And that's what the OAI is really about. We're trying to make sure that we deploy technologies in an integrated manner that allow us to have a common data view of our business from day to day, but also allows us to develop consistent modeling methodologies on top of that, which are integrated with each other and sit on top of those common data foundations.

And so we're delighted that MathWorks has joined the Open AI Energy Initiative, which is our intent to try and make these platforms interoperable and scalable, so that we can deploy them consistently across our business.

So I'll say one final thing before I close, which is we've talked a lot about the journey. We've talked about the progress. We've talked about the vision. We've talked about the importance of the convergence of science and data driven methods as we look to the future.

I think the final thing to dwell on is also about culture. It's not just about the technology. It's also about how we enable the community and the business to take advantage of the technology to transform the way in which we do things day to day.

And so for us, it's all about trying to find ways in which we can facilitate change. If you like, the whole AI program, the digital transformation journey, is as much, if not more, a change journey than it is a technology journey. And we do that in a variety of ways, where we try and stay open. We have open innovation processes where we collaborate with startups and small companies, as well as universities.

We're open in terms of we're working hard on open source standards like OSDU, for example, and Open Footprint. We're also starting to develop communities where we work closely with external partners, through things like the OAI to develop tight integration, as I mentioned.

And of course, internally, we're building out networks like the Shell.ai community, to be able to help the business to understand what this transformation means, the potential of the technology, and how it can impact the day to day. We also run things like hackathons and boot camps to upskill our staff, to enable them to take advantage of this new technology, and to understand the standard integration patterns that we're running within Shell.

I think what I want to leave you with is a sense of a true transformation. We recognize the scale of the challenge that we have as Shell. We recognize that the energy system has to change in order to meet the net zero ambitions. And we recognize that digital is a key lever to get there.

But we also recognize the need to run the transformation in an integrated, holistic way that recognizes the need for change. And to do that, we need strong partnership. We need companies that are going to come with us on the journey, who are committed to the same values that we have, who are willing to roll up their sleeves and get stuck in.

And I think through all of the products and the journey that Amjad's outlined today, hopefully you get a picture of that level of collaboration with MathWorks, the way in which we've been really hand-in-glove helping to shape each other's journeys as we look to achieve something meaningful and impactful for society.

And so again, just a huge thank you to the MathWorks team for their journey so far, or the journey so far, and all their support and collaboration that we've had. But I think also, as we look to the future, we recognize the need to take things to the next level.

To achieve our ambitions, we're going to have to move quickly. 2030 is not very far away. And so we recognize that there's a need to double down, to accelerate, to move quickly, and to leverage this technology to the fullest extent, and to integrate to the fullest extent, to achieve the ultimate aim, which is the digitalization of our assets, and the decarbonization that goes with that.

So again, thank you for having us. I hope you found it interesting and insightful. And I look forward to continuing the dialogue with many of you.