Name: AI Transformation in Energy
Uploaded: 2025-02-05T16:30:03.386Z
Duration: 29 min 32 s
Description: AI Transformation in Energy

Transcript for "AI Transformation in Energy": Hello, everybody. Welcome to our AI transformation and energy webinar. I'm Miki Usiak, global marketing director for energy and resources at Microsoft. We have two fantastic speakers today. My colleague, Uwa Arriaveri, chief commercial officer of Microsoft's energy and resources business, and Steve Bowman, who leads enterprise AI at Chevron, who'll be sharing insights and experiences around leading AI practices. Over the next twenty five minutes or so, we're looking forward to into diving into the topic of AI in the energy industry in light of how quickly the technology is evolving and changing. In fact, the customer said recently, what we're doing with AI and Copilot today is different from what we were doing last week. If you have any questions along the way, please drop them into the q and a chat, and we'll do our best to address them now or offline after the webinar. Thanks again for joining us today. And with that, I will turn it over to Uwa. Nikki, thank you so much, for the welcome. And looking very forward, to this session, you know, we're seeing that energy operators are driving a lot of AI transformation. They're trying to create value by creating an intersection of their operations data, people centric data, customer data. And, this kind of integration, will create impact across, various spectrum, people centric use cases, operations, the path to net zero. We're even seeing, operators in the energy industry create new business models around sustainability and energy transition using AI transformation. So, we have a wonderful conversation today. We have Steve Bowman with us who's GM, of enterprise AI at Chevron. And, you know, we're very excited to have you here, Steve, and thanks so much for being here with us for this webinar. Yeah. Thank you very much, and, I'm really looking forward to the conversation today. Likewise. So, Steve, let's dive right in. Tell us a bit about, what you do, at Chevron and also some of the challenges and opportunities that you're seeing around AI. So we stood up our enterprise AI team in, June of twenty twenty three. We're a relatively small team and we really acted, in in the vein of a response team almost. The the technology was really evolving pretty rapidly, and we've been, you know, applying AI and machine learning, data science, advanced analytics within within the the company for many, many years. But we could see that that things really shifted, in in 2023, and we wanted to position ourselves so that we could take advantage of that in a, a really focused way. We understood that focus was really the key here. We didn't we didn't necessarily want a thousand points of light or really, like, small pockets of excellence. We wanted to really bring folks together and really drive toward a solution that or the solutions that deliver the most value for us. And that ties in really nicely with with our purpose as a company to provide affordable, reliable, ever cleaner energy and and enable human progress. And and that's really how we try to frame things. Well, that's wonderful. And and we've, we've been strategic partners for for many years now. You know, we've we've worked on, you know, sort of, data data state shifts to the cloud and also, helping to, explore ways to infuse, AI into some of your soft surface operations like the work we're doing around OSDU. Would you talk a bit about just how Chevron thinks about the business impact, through AI and AI transformation? Yeah. I I mean, the the number one item that I would say that really resonates with with folks in in all levels of the organization or across the enterprise is being able to connect with our data better. Yeah. We've we've done an awesome job as a firm. I don't think we're alone in terms of really accumulating a great deal of data, but, you know, the there can be a barrier in terms of people trying to access that. So going from data to insight to action where you're you're you're changing the future, that that that's a higher bar. And when you're talking about a company that has been around for a hundred and fifty years, we've operated in in almost 200 countries, over that over that long period of time. There's an enormous amount of information that's available for for for us to use, and we're pretty good at finding it. You you reactively or or or when we're trying to look backward and understand sort of what we could have missed or trying to explain what happened. But what what where I really see the the the point that resonates with people is when we start talking about how we can use AI to really change the future and change our relationship with our data. I'd love to double click, on that a bit, because there are, colleagues listening to this webinar who might say, look, we we just don't have clarity on business the kind of business impact we can get, from the use of AI. And for that reason, we won't invest. What advice might you have for companies like those who are just not sure of how to articulate that value and things like that. You know, what do you say to them? You know, we we take as table stakes that people wanna do their job a a bit better and a bit faster and a bit more efficiently. And and we look to solutions like Copilot or or the other kind of various various forms of of Copilot that y'all have been working on extensions to the you know, within the the Copilot umbrella as as really speaking speaking to that. So, but that's that's sort of only part of it. Whenever we talk about value, for for those kinds of models, we really try to lean in on how many users do we have. So we lean in on on the u on the user user accounts. We try to really lean in on the the total work hours that we're potentially saved. And then we really try to lean in on those value proof points. Right? So, like, really being able to articulate that it's not just about doing your job a bit quicker. It's that I did it quicker and I was more effective, and and these are the changes. This is how, use of these tools or these models impacted the future. Talking about business value can be can be a little little challenging, but it's it's something that that we end up talking about on a on a weekly basis with with some of our most senior leaders within the company because they see the potential that's that's in front of us here, and they always wanna confirm that we're working on, you know, very impactful items. And so that that's where we get away a little bit from sort of the the the kind of work and more efficiently items, but more really digging into, like, where are there insights that, that are that are just below the surface or, that that exist within these big datasets that we have that if we can connect all of the dots, we can really do something, something transformational. When we look at when we look at that, you know, kinda capturing value from from digital solutions can, can be a cottage industry unto itself. What we've chosen to do is we really try to work with who our end users are and really look at, well, what is the key business metric that your group is trying to impact? And then we have a sense of, well, within, like, a KPI tree, what are what are the handful of things that you're always really focusing on? We try to explicitly link the work that we do to that KPI or to that to that metric. And then, you know, we know when it's moving one way or the other, and we kinda have those proof points, we can we can then, maybe you can't perfectly answer the value conversation or the value question, but you can certainly go a long way. And the same thing goes with with really understanding who your users are. When when you have that, then you you know who to ask, I mean, at its most basic level, and you can kind of you can determine look. Like, what what is what it what is the value that you derive from this? You can almost start to think of a demand curve for the solution, and and that really goes a long way too in terms of articulating what the value is. Steve, thanks for for that. And and, because this ROI, question is a is is a big one that we get from a lot of, our peers in the industry. So thanks for sharing that. My sort of takeaway from that is, you know, one, sort of looking looking ahead and really trying to, understand the vision, what you're trying to accomplish as a as a business. And then two, and I love your commentary around that, just this concept of, you know, trying to figure out what the business metrics are and snapping to those business metrics. And then the people centric side, really your your users, and and making sure that the stakeholders, are aligned to all of that. So that actually leads me to, my next, question around people. And, you know, there's a big change management aspect of AI here. And, you know, people have to be enabled and empowered to adopt the use of AI. What are some of the lessons learned that, Chevron has around this topic? So we, we totally agree. I, I I I tell people, kinda when I when I since I've since I've been in this role, I learn something associated with three themes every week. First of all, I I've worked at Chevron my whole career. I've spent my my entire adult life here mostly in the drilling organization, and in engineering roles. But I've I've been here for twenty five years, and, I learned something new about the company. I also learned something new about technology as things are evolved going very rapidly. But the third thing I consistently learn is, is that people are resistant to change, and I learn new and exciting ways that, that that that that really consistently is a blocker. And and we really we really have to dig in deep and and have an under a real deep understanding of what the what the user experience really is. You know, we we we talk a lot as a team around, like, thinking about what are folks what are folks personal change curves look like, where do we have folks who are maybe leaning forward, where do we have folks who are laggards about this particular aspect of of of their role? They're probably not laggards about everything, you know, but in this particular case, this is just, you know, it's nothing's gonna get them over the line quite like time. So we we understand that. We also have a sense of the different personas that we're working with. I can give you a couple of examples where when we worked with maybe some of our shale and tight business units and and, where maybe a little bit more rapid cycle time. You know, it's it's more of a a linear thinking kind of, kind of environment there and a quicker cycle time. You know, that that tends to be sort of a little bit different mindset than, say, when we do the work we've done for with our Earth scientists in the exploration world. You know, there there you're really kinda relying on you're talking to people who really value their interpretive skill. And so you can't use the same behavior change plan for for for both of those people. You gotta really understand with who you're talking to there. And when I get into who you're talking to, that's really identifying who are your key leaders there. So we know that we we deploy technology through people. So we, we really strive to to drive toward leader led change. So who are the who are the folks who we're willing to work with in terms of building up a model, and making it very powerful? But then who who are the key who are the key people? May maybe individual contributors and maybe team leads or superintendents. But who are the key people who when you win that person over, you know, you you you get a lot of cache from from that group of folks. They can put their stamp of approval on it. And then who is directing work? Because people just naturally are resistant to change and and and we, we we certainly understand that. We see that all the time. But when when we think about, like, well, who's directing work? That that gets us at the spot of understanding, well, what does this workflow really look like? And how do we embed these tools within the workflow? So we know that it shows up like help, and, and we know that people are getting value from it. So that's really working at the individual user level. The other item I'd say is you you really have to engage your your higher level folks who have profit and loss responsibility or folks who are leading business units or, different functional groups. And and we did that, at the end of last year, or the end of twenty twenty three, I should say, where where we were very focused in in bringing those folks together, making them first aware that, that you've heard a lot about AI in the news, but the future is now. This is not this is not something that's, that's over the hill that these capabilities exist today. And we wanted them to start thinking about AI and machine learning and really advanced analytics as, as a new tool in their business tool tool chest. So you're you're trying you you have a you're performing at a certain level. You have a goal that you're trying to hit. You don't necessarily have a clear path on how you're gonna get there. Well, begin to think of AI as, as a potential bridge into trying to get from your to to from state. Now one thing that we have observed relative to this is that when we when we engage with senior leaders, their their view is they're they're thinking of AI in really an innovative kind of way or they're thinking about, kind of inventing the car or inventing the the airplane. Right? But when you talk to individual users, like, their their mindset goes like, look. I got a job that I gotta do today. Like, I need a faster horse. And you gotta really understand sort of the the you have to you have to kinda work almost both angles there because when you start giving folks, folks the the the faster horse, then it really begets more sophisticated use cases from the on the path to trying to really deliver that big innovative moonshot that, that senior leaders envision. So there's no one answer. You know, we, like I I think I said before, we deploy technology through people and really understanding that that that user experience, that people that that people experience being human centric. That's to me what what when we talk about human centric AI, human centric AI, that's really what I'm thinking. Steve, I I love that framework around sort of engaging from both ends. Right? You know, having it leader led and then sort of coming in from the the operating level and and people who'll be using it. Just I it's just such a big topic, so I'd love to double click into that a little more, especially around the leaders, you know, and and how, you know, you, have been able to, pull in leaders and and convince them that this is important. You know, there are leaders on this call right now listening to this webinar and sort of saying, hey. You know, I'm not ready for this. I'm not ready to sponsor a GenAI or even a BroadAI initiative. You know, what are some words of wisdom maybe to help, those leaders sort of think through, you know, how to wrap their heads around, AI in general and and value from it? Just just curious how you how you would think about that. I I've you know, we really leaned in on what were examples from from other industries. Right? So, like, kinda helping people sorta connect. You know? There there's there's there's a lot that really seemed to resonate with, for with with our senior leaders with the idea of, well, how how is, how is AI used in in in pharma and in particular with prospecting for new new treatments and new therapies or really trying to understand, understand the complexities of a of a of an illness particularly like a, you know, something that's quite rare. Right? That that really seemed to grab folks. The other thing was really looking forward into, like, well, how how does this impact agriculture? There's a you know, I I know that that really kind of, really grabs the the a lot of our a lot of our more engineering kind of, focused leaders. They for some reason, the farming examples really, really grab them. You know? So, so that that was one thing in terms of trying to make it relate. To. I I think kind of the you know, when we when we think about trying to make it relate to, a key thing there was convincing folks that, look, this is not a a solution looking for a problem that we, are the you know, we don't hold a hammer and the whole world looks like nails. We really wanted them to understand, well, look, this is what the capabilities are today, but the crucial thing was the focus. You know, we one thing we saw pretty early on was the the proliferation of pilots. You know? Like, kinda you we we could we could do a pilot. People saw some value from it. Then you you get in this trap of, like, yeah. But can I do this? Or, yeah. But can I do that? Now now we're in a spot where we're kinda doing our tricks. Right? And we're like, hey. Like, we're not gonna get value from this. We really have to focus. Big thing with convincing those leaders to go along was by by saying no to other things. Right? So being really prescriptive about what we're gonna work on and making visible, like, look. This is what we're not. Now as a team, one of the one of the dynamics that we've we've always tried to set up is, we we try to be a partner of choice, within within the organization. It's part of kind of the Chevron way, which goes back a long time. And when we internalize that, a lot of folks come to us with an idea of, like, hey. Can AI help help me with this particular business problem? You know, oftentimes, you don't need AI to solve that. Right? And so this is that's where we we try to connect them to, look, if you do some some some data engineering here, some pretty straightforward data analytics, or you connect with kinda some existing machine learning models that we have, or just connect with some other people within the company, you're gonna get value. And and we really see that as critical in terms of we're we're turning them not necessarily away, but we're connecting them to a solution. And we know that they're gonna come back with something something even more impactful as as they get value from the advice that we were able to provide. That that's extremely helpful, Steve. And and maybe just one more double click into the operating level. You know, so we've seen some success, with, just companies coming in and finding champions at that level. So if people who are, well known to to help, drive and scale different new technologies. So would you talk a bit about maybe some lessons learned that have helped you to identify the right people at the operating level, to help with AI adoption and just even technology adoption in general? Just curious how Chevron does it. Well, you know, we talked a lot about kinda that leader buy in. A couple couple other items is really, when when we've opened up AI training kind of within within the enterprise, we've seen significant pull. People wanna wanna learn more, whether it's whether it's participating in the training in terms of how do how do I learn how to just just, you know, kind of, tap into the wisdom of prompting, that's available kinda across the entire enterprise with with Copilot. How do I do more with my m three sixty five Copilot? Significant adoption there, significant, significant pull for training. When we've done kind of what we call, in, generative AI intensives, where we brought cross functional groups of people together in groups of, like, kind of 15 or 20, and we ran about 10 of those over the last year. We saw a lot of folks that wanted to be a part of it, but we we looked at those folks as a beachhead within different organizations around around the company being able to talk about, well, what are the capabilities that are available? And then now they have a network that they can call upon as they're trying to think of different solutions that are available or as as people come to them with new problems. These are really largely functional folks. These were not AI machine learning engineers, not necessarily IT, professionals. And then, our our our other level here at DoubleClick and Annie more is we we created, an AI change champions network. We we didn't know kinda what the demand was for that, but we we have, we have over 2,000 people that volunteered to to join that. And and they really, we we really saw a tremendous pull from those populations of folks. They wanna be a part of this. They wanna try to lead change. And so tying into all of those existing networks, tying into that, that latent passion around the company, really really that that makes our job, significantly more straightforward because now we know who to talk to really at the coalface if we need some additional context. Or if we're trying to understand sort of, what what an ask is, if we're trying to drive something, if we need somebody there to to be physically present when we're we're we're here in either Houston or San Ramon and, the the user is somewhere else around the world. You know, trying to trying to trying to, again, think about the people who are the end users and connect with them in different ways. Steve, thanks for those practical, ideas and and very executionable. And actually, this conversation of adoption has, driven a a flow of questions. And I wanted to add just one more, build on question. I'm sort of combining Ian's and and, Lorena's. But but but it but they asked really, it's around the same question. Are there, any business units or or verticals within Chevron that you found, were faster adopters and sort of early adopters? And whatever you can share, obviously, within within the confines of confidentiality. But just I think there's some curiosity around, you know, are there particular units that ended up being, early adopters? Yeah. We've we've done quite a bit with our Shale and Tight business units, in particular with our Mid Continent business unit. And, there's there's a few kind of technical reasons for that. There's a few kind of behavior change type reasons for that. One is we a lot of the folks on my team, we we previously worked in that business unit. I worked in that business unit. So we had good relationships within within that sort of that that larger sort of community within the company, and we really had a pretty key understanding of some of the challenges that, that that folks face. Whether it was trying to really understand the the huge amount of non operated joint venture data that we get, the huge amount of unstructured data that, that that sits in various systems of record, or some of the operational problems that that present themselves where, you know, you you have you have two two different assets, maybe a drilling rig and a frac spread that, you know, they rapidly find themselves in some sort of, temporal or spatial conflict and then the ramifications for that. Right? So, so we really saw some some strong early early pull there. Some of it just based on on the bit that business in particular, but then some of it based on our our personal relationships. It's a very short time business. Right? So you have the ability to get in a very data rich, part of the business as well. So so you had a lot of things, a lot of a lot of tailwinds working working in our advantage there. I I like that idea of of because it's kind of a short cycle, you can very quickly see value from, you know, projects. That that and that for sure, plus the relationships, access and accelerant to, to impact. So so I think we we're we're we're coming down to I think we'll probably have enough time for for two questions. I'm gonna take one from the q and a and then and then finish off with, with a just a a question around governance and responsible AI. But but Manu asks, you know, if energy companies value certainty, how can, Microsoft AI offerings, help here? And I I would love to zoom out maybe a bit and just talk about AI broadly and and and a question to you, Steve, because it's similar to some of the questions we talked about earlier, but maybe another angle to it. Right? Because we're we're in an industry that is relatively conservative for for the right reasons, health and safety, compliance, the scope, and volume and quantum of the investment. So we've gotta we gotta keep the operation safe, run the business safely. So how how do we help an industry like that that values that certainty and then we sort of come wanna bring it in and snap in this concept of AI. How should we think about that? Just what's your what's your industry point of view? Yeah. So so like I said, I I I come from, I come from the drilling function with within the company. I come from an from an engineering background. And, and I I really, I I totally, I very much like the question because to me, especially when we have items that have potential operational impact or really, we're really good as a company in under at doing risk assessment. We're trying to really understand what really really what could happen and and understanding that it's, it is a challenging environment to operate in. And having a keen sense of, well, look, where where do we need a human in the loop? Where are we gonna be very intentional about that? And also, where, what what are the terms of use associated with with with some of these models? Right? We want people and ideally, you you want people really, boxed in a bit in terms of how how are they gonna use these items where they can go into the task at hand. Maybe it's an unfamiliar task or maybe it's something that they they they do very rarely in that particular area, but it's done pretty routinely in another part of the industry, where those folks have the most information, the most important information available to them during the planning phase. Because that's the the the ideal time to really impact what that that future state looks like. And it's a time when, you can add knowledge, you can you can really grow that organizational capability, you can really impact what that next step is. That's that's, that's personal to me. I grew up in an oilfield town. I've spent most of my most of my life kind of work working in in operationally oriented jobs. And really, I I think about what does it mean to make that person's life better, when when they're walking out on location with wrenches or or getting ready getting ready to open a valve. How can how can we equip that that person with more knowledge so that they can do that job better and, and in a more prepared way and minimize the the risk to to to injury, than ever before. And that's something that I I personally really am very invested in. Steve, thanks for for sharing that. And and, final sort of wrap up question for you here, because all this is exciting. Right? AI and leveraging the organization, but we've gotta do it responsibly. So for example, at Microsoft, we we build our, Gen AI tools with with responsible AI principles. Curious how Chevron thinks about governance and and creating and using and executing AI in a in a in a safe way and responsible way. It's a core part of what what we do. When we stood up our team, we intentionally had representation from HR law and cybersecurity so that we could make sure that we were always thinking ahead in terms of what the policy landscape, but also understanding, what true resiliency meant. We, we joined the Responsible AI Institute very early on. We conducted an assessment. We inventory all of our solutions, so that we always have a handle on where, where where AI is in use. We're we're always working with our, folks around the enterprise in terms of doing a lot of vendor evaluations as well. So, the first step is really knowing and and really being able to answer that question. But it's, it's core to what we do. It's core to our training as well and really trying to train people on what responsible AI means, at its, at its most tactical. Steve, wanna thank you so much for, partnering with us in this webinar. I wanna thank, all the participants for joining us and for all the fantastic questions. We couldn't even get to all of them. I mean, just so many rich, fantastic questions coming through. Thanks for the engagement, and, have a fantastic day, everybody. Thank you all very much. Thank you.