Automated Transcript

Deep: Hi there. I'm Deep Dhillon. Welcome to your AI injection, the podcast where we discuss state of the art techniques and artificial intelligence with a focus on how these capabilities are used to transform organizations, making them more efficient, impactful, and successful.

Hi there I'm Deep Dhillon, your host. And today on our show, we'll be talking to Carolina Torres, executive director of energy industry transformation at Cognite. We're gonna touch on all things, industrial data ops, including the great energy transition underway and sustainability as a data problem.

Carolina, get us started. Tell us a little bit about how you wound up in the energy industry and at cogni and what is this transformation that you're talking about or that you're, that I, that I'm reading.

Carolina: Sure. Um, I started out life as a geologist actually, and  okay. So, uh, I, I entered into the oil and gas industry as a geologist. I spent like 10 years working for Arco and then the rest, another 20 years working for BP, um, at, I only really worked as a geoscientist for the first. 10 or 15 years. And then after that I went into more of a strategy role, organizational change and technology. I got very, very interested yeah. In how we can use technology to make better decisions.

Deep: It sounds like you were, you, you had a really interesting list of locations around the planet that you've been is that, did I read that right?

Carolina: Yeah. I, I mean, I don't know if it's that way anymore in the oil and gas industry, but when I joined in the early nineties, there was a lot of moving around and, and going to a lot of interesting places in the world to work. So I, I got to see quite a lot of different geographies and different cultures. And I, I, I love that about my career.

Deep: Yeah, I think you were, if I recall, right? Like south Amer like pretty much all over the planet, it, it, it looked like

Carolina: so America, Asia, um, spent quite a bit of time in Myanmar, which was super interesting. And then, uh, in the UK for quite a bit as well.

Deep: Awesome. So like, so what is this transformation that's going on this energy transformation? Are we talking about the transformation to clean energy? Is that basically what we're talking about? Or am I,

Carolina: you know what, I, I actually think that there's multiple things that are kind of coming together to a head and some of them are techno technology and some of them are actually social and, and there's interwoven. So let me unpack that cause it's a complexity.

Deep: Yeah, please do. 

Carolina: so we have digital digitalization. People want to do a lot of automation and digitizing of their existing historical processes. Okay. We have the energy transition, which is really driven by, you know, climate change and, you know, there's actually a big push from both consumers and the financial sector. Um, to, to kind of do things in a really different way and to operate our planet in a more sustainable way. And then there's a whole bunch of, you know, what are called fourth industrial revolution technologies like, like AI automation, machine learning, um, you know, computer vision. There's a whole bunch of stuff that's like coming to fruition. That is. Can be used to enable, um, those, those transitions, both the digital digital transformation and the energy transition. Um, so I think all of those things are coming into play and the social aspect of it, which I find super interesting as well is that, um, it's gonna require a different way of working. It's gonna require a significantly more. Collaboration, uh, than, than competition. So in the past business, um, and industry has all been about competing with other, other folks to win, to win the business. Yeah. To, you know, make the most widgets to sell the most widgets. And when you start thinking about trying to operate more sustainably, you get to like, for example, making a change from a supply chain, which is basically a waste stream from one side to the, you know, to the other from, you know, pulling something out of the ground, using it, making stuff out of it, selling it to people and having them throw it away. um, to something which could be more like a circular economy or a circular, uh, not a supply chain, but a supply cycle. Um, and that's gonna require a different level of collaboration, a different level of transparency and sharing of information and having like a common source of truth. That, you know, we haven't really been able to do before, but we have the technology to do it. And that technology is, is available in the public sector. You know, like what Google does, for example,

Deep: can you, can you maybe give us an example of what you mean by, like, what would be this an example of a circular supply chain?

Carolina: So, um, there's a manufacturing company in Norway called Rucker who has basically. You know about, um, in sustainability you have scope one, two and three emissions. Right. Have you heard about that?

Deep: Nope. So tell us all about it. Yeah, I don't, I don't know anything.

Carolina: So scope scope. One is, you know, emissions that your business. Generates. Okay. So two, um, could be sort of side things that are not related to your business. Like say you are, um, a manufacturer who makes drill bits. Um, so in the manufacturing of drill bits, You are creating emissions that scope one scope two. It could be like the supply chain, the, the material that comes into your shop, right? Like the gas that was used to, to transport the metal or something. Exactly the transportation, the electricity and all that kind of stuff. And then scope three is the emissions that your customer, uh, immense, well, during the usage of that okay. Of your product. Of your product. Yeah. Uh, yeah. So if you're an oil and gas company, you have a huge scope, three footprint because you are selling petroleum to consumers who are then that's coming out the tailpipe and into the yeah. Consuming it and yeah. Um, but you know, how in the world are we gonna keep track of all of that? When, you know, my scope to my scope, one is somebody else's scope three, et cetera. It's like, there's, there's no way to actually track that unless we actually share information in a different way than we've, you know, done. And the data transparency has to be much better. So this customer of our, of Cognite's called RRB blocker created an algorithm to calculate. You know, basically to automatically produce their CO2, uh, reporting at the plant level at the machine level or even at the product level. So they could, you know, take into account their suppliers, emissions, add their own emissions for the machine tooling and what they did with that product, that bit they made. And then they can actually hand it over to, um, the oil company that buys that bit, um, with a certificate that says, this is the footprint of. Piece of equipment, this product right now at the moment that we're handing it over to you. And then after that it becomes yours and you can do whatever. And that's so that the purchaser of the, of the drill bit, if you will, can like take that whole chain of information about the product, I guess, in that case, that would be scope two, uh, take that whole chain of stuff and then incorporate it into their own reporting that goes onto the next step. Is that. Yeah, but what, what ended up happening, which was really interesting is that then they started renegotiating, like, okay, what happens if instead of me, I'm the oil company now, instead of me buying that bit from the manufacturer, what happens if I lease it? Oh, okay. And then when I'm done with it and it's all worn out, I give it back to the manufacturer and they can melt it back down and reuse the metal. Mm-hmm .  um, you know, and then how then that doesn't, that's not really, how, how does the scope one, two and three work in that scenario, but it there's actually an incentive in that situation to actually reuse something as opposed to just dispose of it and kind of get rid of that linear waste stream type supply chain scenario, and start create creating something a little bit more circular, which then reduces everybody's. Footprint and, um, you know, gets people closer to the net zero goals.

Deep: I see. So it's like, you've got the, maybe the emissions from the production of the drill bit, but you also have the emissions from the production of the new drill bit that didn't necessarily have to get built. Right. But, and, but you're using you, don't reset if you don't reuse it, but if you do reuse it, then all of a sudden, you know, you just save that. And so. For the sake of our, um, of our, of our listeners. Why are we doing this? Like what's new and why is everyone accounting for all of these emissions? And like what's driving it down to the business level. Is it like executive choice? Are investors demanding it? Is it coming down from UN mandates?

Carolina: That ripple in like, uh, what's happening all of the above and it's all very interlinked. So yes, there are new regulations in place. Those new regulations are there because of consumer sentiment and the science that supports the fact that climate change, um, And, and global warming is happening and it's because of the emissions and, and the waste and everything else that we, as humans are creating. And so there's a big drive and I think the financial sector has moved into it. The S E C just recently, I think it was three weeks ago, announced some new, um, reporting requirements, uh, around, uh, emissions reporting, sustainability reporting, um, where customer. Or, uh, companies that have made net zero claims have to now in their reporting and their S E C filing show progress that they're making towards that. And it can't just be like, wave my arm and paint everything green. It has to be like, These are the steps that I'm actually taking. I am now recycling my bits. I am now, you know, tracking the emissions of my, my, uh, products, um, and, and handing them that information over to my customers so they can do their reporting, et cetera. So I think, uh, You know, there, there's just a lot of impetus in the world for that.

Deep: Yeah. I mean, it clearly a huge chunk of us realize that we need to, you know, keep the global temperature rise below 1.5 degrees, and there's a lot of motivation to sort of force change here. What's the, what are some of the differences in the energy industry? And maybe walk us through a little bit more of like, what's the role AI is starting to play there.

Carolina: So I think that, um, the oil and gas. You know, the hydrocarbon energy industry has been quite behind in their digital transformation. I think, you know, the financial sector banks, marketing, um, tech has moved quite a lot faster into that space and being able to manage and operate their data, even huge big data, you know? AWS, uh, Amazon and things like that have been able to manage their data in a way that they have been able to use machine learning and AI to enhance their business, to operate more efficiently, to, you know, even to manage their sustainability information in their data. Mm-hmm, , you know, real time they can track like who's buying what, when, how much, et cetera, and gain some insights by using machine learning and AI to do that. The oil and gas industry also. Significant amounts of data, big data, but it has really, really struggled to do that digital transformation and to actually take that data and be able to use machine learning and AI to gain further insights and make better decisions and optimize their business and automate their business. They've struggled with that. And a lot of that has to do with the nature of the data. It's extremely siloed. So, uh, if you go to an offshore installation anywhere, there are hundreds, if not thousands of moving parts, and many of those are come from different vendors, the different machines, you know, the pump system and the, whatever they all come from, different vendors. Each of those vendors. Keeps and manages the data associated with those, their particular machines in their system, in their data model. And so if you wanna try to aggregate and amalgamate all of that and get some sort of a view of the entire facility or even components of that, that, that facility, so that you can get some real time monitoring and be able to make some real time optimization decisions. Either towards production enhancement or towards, you know, emissions reduction or whatever it is. Companies really struggle. So both in the digital transformation and in the energy sustainability, the energy industry has really struggled and it's because it's a data problem.

Deep: Yeah. I mean, that, that makes a lot of sense. Like we see that in the manufacturing sector too.  Right. You've got, I, I sort of break it into two buckets. One bucket is. Gathering the data. So this would be like the whole quote, big data movement where we're trying to get everything into data warehouses. So you've got, you know, companies have, are sort of at various points on the spectrum. Some companies like pretty much every single thing generated is like being warehoused. And then you can have machine learning, AI teams going in and trying to figure out how to predict what from this.

Carolina: Yeah. But I mean, warehousing isn't enough because you have to putting it in one place and that's, it's just companies have made, it's like, oh, well we have a data. Oh, you know, we've got a Azure and we've got all our data in this data lake. Yes. And they still can't. There's no relationship mapping. There's no semantic layer. There's no ontology. There's no way to actually connect the different pieces of data in a relational way that then allows people to pull the data and the relationships around that data. Into a model that they can then do AI industry.

Deep: Yeah. I, I agree a hundred percent. I, I completely agree with you, but I just wanna like go kind of up, tighten the bow here, if you will, on, on your, your characterization of the energy industry, not quite yet having their data sort of, even in that mode. So like I'm, I mean, I guess I'm wondering, or maybe it's, it's a question for you. Is this not maybe a result of the lack of internet ready sensors on some level like you've cuz because if you think about like a software company, it's fairly straightforward to get all of your data into a warehouse because every everything you generate is written by oftentimes like a good chunk of it by folks in house who create the data and just need to put it via API somewhere else. But if you talk about something like an like an oil company, you know, they've got machinery parts, That maybe had sensors on them pre um, you know, where they're all wifi enabled and there's, you know, wifi sitting maybe out there and then boom, it's like immediately onto the inter internet. Therefore you can centralize it easily. Therefore you can get it, whether it's a warehouse or something. With, you know, more structured representations, but is, do you think that's part of it?

Carolina: It definitely is. And we have a name for it actually at Cognite, we call that the OT IT divide. So operational tech versus, you know, informational tech it's, you know, is there, there's a, there's a big. Gap. Um, and the operational tech, I mean, a lot of folks think that they have to upgrade all of their facilities. And many of these oil and gas facilities are 20, 30 years old. But, um, there are, there's a lot of aftermarket sensors that can be added and they're not very expensive. So we we've got a lot of customers that have been systematically doing that. Um, but there's also, you can use AI for doing.  um, sort of combined physics and AI models that can help with that. So for example, one of the products that we, uh, help enable is something called the virtual flow meter. So if you think about an oil and gas, well oil or gas, well, You can put a meter on the well retrofit one and that's what can be quite expensive on a multiple Wells, or you can create a virtual flow meter using the data that you do get from that. Well, like the pressure and the, you know, the, the volume of flow and several things. And you can write a hybrid AI model that will enable you to create a virtual flow meter. Um, so that's one of the things that, and what, what, and what do you, what is your goal with. Meter, um, well, it allows you to track the amount of flow production out of that. Well, which then on the sustainability side, if we're going to that, um, you can also, you know, track the emissions, um, by using some calculations to go from production, you know, barrels produced to emissions. Um, but you can also optimize, you can create some, you know, neural net or other, you know, optimization. To do some forecasting or something. Yeah. Or, you know, basically you each well will have a meter that tells you how much it's producing. It's going through multiple systems. It's going through a pressure handling system and a, maybe a water handling system where it's separating the oil and the water out. And then, you know, it's going through multiple systems on the facility and you can optimize those to maximize production. Turn down this, well, turn up that one, you know, this one's really gassy. Let's, you know, you can basically continuously optimize that system as a whole, and you might have 10 Wells that are linked up to that one system. And so you wanna continuously optimize and that's, that's all machine learning and AI. Systems that can do that once you've got them each metered and you have the equipment and understand what the equipment throughput is, then you can optimize that system.

Deep: You're listening to your AI injection, brought to you by xyonix.com. That's xyonix.com. Check out our website for more content, or if you need help injecting AI into your organization.

Deep: How are they done? Maybe pre AI. And, and, and where are we with respect to the transformation from that pre-World to the current?

Carolina: So pre AI, you had operators, human operators that got to, you know, that were there all the time and got to kind of know the Wells, like personally, like, like, you know, the humans do, it's like, oh, that one sounds funny and that one's doing this and that one is slugging. They get to the point where they understand and they know the system really well, and they're optimizing it manually as a human would, you know, they're like trying to take these 10 Wells and try to figure out what's, you know, and it's by trial and error, it's empirical, it's, you know, a human process.

Deep: Right. So I don't know enough about an oil. Well, to know what the flow rate changes. Look like, but sure.

Carolina: Well Wells can behave differently. Some of them slug because of the. Dynamic geometry within the wellboard itself. Mm-hmm  um, you know, so they might be sluggish. They might like burp up like that. You know, others are just flowing like steady state, you know, like a nice train , you know, it's, um, you know, they're gonna behave differently depending on. You know, is one of 'em close to the oil water contact in which case it's bringing up oil, but it's also bringing up water, which needs to be filtered out. Um, or it might be near the, it might be, have more gas in it than, than the others, that kind of thing. So there's, there's a lot of variables in the system that need to be managed and. You know, some of these operators that have been out there for 20 years operating the Wells, they get to know, you know, in this very instinctual way and they can be really good operators, but they can't really beat a machine that can handle, you know, 50 different variables. Or an algorithm that can handle 50 different variables in like real time, kind of be creating suggestions on how you can operate that facility and give us a sense of the knobs and dials they're twidling like, are they like,

Deep: What are they. Maybe just a few of them?

Carolina: They're called chokes every well has a choke, uh, there's choke settings. And you can think about it like a clock maybe, and there's, you know, like 12 settings and one's wide open. And the other one is, you know, like different degrees of shut. So you can, but when you choke it back, you're also creating some back pressure, which sometimes helps it produce better. I mean, we're getting into like super detailed, like how, well, I dunno if you wanted to go there, but it is interesting.

Deep: No, I, I. Kind of go with whatever I'm trying, cuz I'm trying to understand like create a picture in my head of what's going on and I don't know whenever else I'm gonna get a chance to tuck somebody who knows going out oil. What else? So walk us through a little bit, like on the machine side. So you're trying to automate this with the system. Are you learning from the humans or are there, are you basically able to kind of model the physics of this problem and or is it a kind of a combination maybe where the, like you said, like the humans are twiddling some knobs checking out flow and based on that they can kind of adaptively hone in on optimal settings. Like, how do those systems tend to work?

Carolina: The automation, couple of things. So the, the company that I work for Cognite is primarily a data operations system. Mm-hmm  so what, what we have is a machine learning and AI. Software, uh, that brings in data from multiple different sources. So each of the individual Wells plus the water handling system, plus the pumps, uh, the gas compressor system, you know, all the different systems, both the, the Wells underground plus the stuff that's on top, the, the facilities. Yeah. And we take sensor data in, we take production data from all the Wells. We take temperature, pressure, you know, all the different data from all the different systems. Bring it into the platform create, uh, you know, linkages through entity matching and really, you know, creating basically a, like a knowledge graph of the entire system, like a digital twin, if you will.

Deep: Okay.

Carolina: And then from there on top of that, either we, or sometimes our customers. Citizen data scientists and other folks that do this themselves. So our data platform is an open API system. We create those linkages, we create APIs and then either we can write programs on top of that, or they can write programs on top of that, you know?

Deep: Gotcha.

Carolina: So. Basically we, we develop that platform or that system, and then you can write smart stuff to do insights or automation on top of it. Um, when we do it, we have quite a lot of domain experts that work for our company that understand, you know, facilities, engineering, Wells, engineering, how these things work. Production engineering. And so, you know, in addition to our data scientists, we have domain experts that help write these algorithms and they understand both the physics and the AI parts of how all these pieces work together and how, how optimization can happen.

Deep: Um, so going back to maybe the. Sustainability part, what do you do at on a day to day basis, maybe at Cognitee that has something to do with sustainability? Like what's the, what, what does that mean?

Carolina: So I, I think up till now, sustainability, um, and, and sort of this net zero goals. That, um, folks, companies have been, uh, you know, pledging, I guess, to be, you know, net zero by 2030 or 2040, um, relatively new it's in the last couple of years that all that's come about. Prior to that, there were requirements around environmental reporting. So companies had to generate, you know, for the S E C or for the European commission or whoever their reporting body was some sort of a report, um, to stockholders. If they were public publicly traded company about their environmental. Performance in other words. Yes. You know, did you have any spills, you know, how much did you admit? Um, yeah, a, a lot of different kinds of things like that. More recently with the advent of, you know, pressure from the financial, uh, sector around ESG kind of reporting and investors are now really looking at different companies and their performance in a much, much deeper way. Um, and they're looking at, you know, environmental and social and governance and how that happens in the company and making investment decisions around that. It's getting to the point, like for newer startups, it's very hard to get funding if you don't have, you know, very, very clearly stated net zero goals and a path to get there. And you're demonstrating that you are reducing your environmental impact. For example,

Deep: Are we talking about energy companies like our energy companies, trying to get them at zero too? How does an oil company get to net zero? Well  are, are they building carbon capture systems or something to some of them, some of them, uh, you know, there's all different strategies that companies have.

Carolina: So some of them are building, uh, are, are really like mostly American companies are putting all of their eggs into that. Carbon capture and sequestration basket. They're, you know, I think that they're a bit behind the ball compared to European and other, all the other companies, uh, com countries basically, including Canada, are looking at also, um, alternatives, alternative energy. So if you look at shell or BP or Ecuador, or, you know, some of the Canadian companies, they are trying to reduce the footprint of their traditional hydrocarbon based business. At the same time they're transit grow their renewables. They're transitioning to other either hydrogen or, you know, wind or solar they're buying EV charging systems. They're creating new businesses that have to do with, um, partnering with cities, for example, to create, um, holistic, circular energy solutions. We'll supply you the energy and also the carbon tion of all the energy for your city. So that makes sense. There's a lot of different strategies. The companies are deploying to do this, but I would say the traditional American companies tend to be like, oh, we're just gonna keep producing the same thing we are, but then we're gonna stick it in the ground.

Deep: Like, what's your take on that? Does the energy usage pencil on the CCS systems like. I mean, now you need a clean source of energy to power the CCS system. That's gonna extract the carbon from the atmosphere to put it back in the ground. And yeah, we don't wanna use oil for that.

Carolina: No, I think they do. I mean, I think they're betting on that oil is gonna be around forever. They're still gonna need tons of oil and then they'll just come up with you. They'll just gonna pump it back into the ground. But so then, so then presumably they're, they're gambling on a significant increase in energy efficiency on the CCS front. They're they're banking on gaining some efficiencies on the production side and then basically, you know, whatever they produce, whatever footprint they have from there, they will net it to zero by putting it all in the ground or by supplying the capability to others. So it's not just, we're just gonna put our own. Waste into the ground, but we'll also, you know, supply that for manufacturers. For example, we will, you know, in do the injection and the monitoring of, uh, that carbon for, for customers as a business, which will then that's a, that's a negative number on their balance sheet.

Deep: Yeah. I mean, it sounds reasonable if you, um, pull a barrel of oil out of the ground and you sell it to somebody who puts it into the air, if you can pull a barrel of oil back outta the air and shove it back into the ground and do that at a cost that keeps it profitable for you. Sure. But I guess I wonder if that. Pencils.

Carolina: Yeah. I mean, it's gonna require it's not gonna be profitable unless the government creates that regulatory environment to make it profitable. In other words, if they, if, if the government makes it so that, you know, every ton of carbon equivalent that you inject is worth, you know, X dollars like that market is doesn't exist and it needs to be created. And it's gonna have to come from some sort of a taxation scheme or something like that, or a price scheme where, you know, every carbon, every ton of carbon that anybody produces is gonna cost them. You know, X dollars unless they get rid of it somehow. Okay. So it's, it's complicated. It's not just the, technology's not there, but also the regulatory environment and the tax, you know, the tax environment, isn't there, there's a lot of social things that need to line up in order for that to work. And it's very difficult when, you know, you have American companies that produce, you know, hydrocarbon out of fields in Angola. So how does that. You know, there's some sort of an agreement that needs to happen between the us and Angola about what the price of that oil is and the hydrocarbon, sorry, the, the CO2 equivalent value of that and what it is. So it's, it's very complicated. I mean, all that I personally don't really like, to be honest, I I'll just give you my personal opinion. I don't like the whole carbon hydrocarbon, um, the CC sequestration concept at all. I mean, I think carbon capture is great. There's, you know, some technology that's very, very incipient, very new, where they're taking carbon out of the air and making, you know, product out of it, like basically bricks and things like that. Mm-hmm , or you can, you know, use it to pave roads and things like that. It's not energy efficient right now, but it seems like it would be a lot easier to regulate a value for that, where you actually use it for something. As opposed to just disposing of it. To me, the whole sequestration part of that equation is just more of the same linear waste stream thinking it's not transformative enough, perhaps you're not sure whether AI can really transform your business.

Deep: Maybe you don't know what it means to inject AI into your business. Maybe you need some help actually building. Check us out at xyonix.com. That's xyonix.com. Maybe we can help.

Deep: Yeah, I don't know. I mean, like right now there seems to be. A market for high quality carbon capture and proof of injection into the ground. Like Microsoft's paying big bucks for this and there. And there's a bunch of companies that are willing to like run contests, put a lot of money on the table. The challenge seems to be nobody's doing it well yet, but there's a lot of startups being funded at pretty significant rates. So going back to the AI question, How does the, how do folks in the, you know, this industrial energy sector, how do they think about artificial intelligence? Do they think about it as just like another tool? How do, how do they think about it?

Carolina: I think, let me step back a bit. There's every range, um, of client that we have that is on the digital maturity scale. Okay. So there's some that are quite immature and they're just trying to figure out like how they can improve their bottom line, how they can optimize their production, but also like answer. They're uh, stockholders requirements for reducing emissions. Um, and they're, they're just like barely just trying to figure that out all the way to the other extreme. We have some clients that have very, very robust. Citizen developer programs where they're retraining all of their domain experts. They're drilling engineers, they're petroleum engineers, they're geologists, you know, they're financial people. They're training them all to in data science and data science thinking. And they're basically saying basically we're gonna change our entire business. We're gonna digitally transform our business so that we're gonna have. Uh, some, some sort of, um, you know, enhanced decision making across our entire value chain for how we do business. And we're gonna take, you know, all the historical knowledge and information that we've got buried in our data files. Uh, and we're gonna bring info out of that and make it possible for people to query it and analyze it and, you know, perform like analytics on it and, and gain insight.

Deep: And on the other side of the fence where you see these companies that are sort of struggling to just even capture the data, like what's the difference between these two companies? Is it size? Is it like management commitment? Like.

Carolina: No, it's not, I mean, it varies medium size companies seem to actually be the most advanced, I would say in the sense that they're the most agile and, and I'm talking about medium size companies like HES, for example, mm-hmm, , that's what I would consider a medium size company. A big company would be like Exxon or BP or so Uhhuh. Um, but the media size companies seem to be able to be a bit more agile and make these kinds of decisions and train up their people and, you know, establish the things that they make decisions quite a lot quicker. The big guys are wanting to do it all themselves. So those guys are talking to Microsoft and Microsoft is telling 'em yeah, you have all this stuff in Azure and it's in your data lake and you can just use Microsoft product to mm-hmm , you know, basically make a digital twin or get your data, you know, contextualized, you know, create that, that relationship knowledge graph around your data, that then enables you to do AI and machine learning at. And so those guys are quite slow because they're big. They have a lot of different assets all over the world. Those assets are very different. The one from the other, in terms of how they're organized and plumed, and they actually don't have the, they don't have the people who the know how to do it themselves, but they're wanting to do it themselves. So in some ways they're quite behind.

Deep: That's such an interesting take, right? It's like you have money and resources. You have large pools of employees. So you're committed to like bringing this stuff in house, even if it takes longer. But because of that, And you're and your non willingness to kind of go wherever the expertise already is you end up being slower. I, what I would've guessed, uh, would be a little bit different. Like, what I would've guessed is that they would go on shopping SPS for startups and just pull the talent in that way.

Carolina: No is, I mean, some of them are going for big shopping. They're paying huge premiums on really smart people that they're poaching from Google or Uber, or, you know, other, other industries that know it. But then these folks get into the industrial world and they get completely bogged down cuz they just don't understand the OT. I think one of the things that you, you probably need to bear in mind on, in the oil and gas business too, is that they are used to, they more than any, they're very secretive, their DNA and their history and their culture is, you know, around exploration and keeping everything very, very close to their chest. They don't, you know, The tech business and even now, and now sort of the marketing in retail is all about being open and they're like, oh yeah, we're writing open source stuff and everything's open and everybody can see our stuff. And here copy me, you know, it's fine. Uhhuh  they have this like totally different culture. Whereas the oil and gas and probably manufacturing is a bit the same and definitely power and utilities is like that. It's like all you. Competitive advantage is me keeping everything right here, you know, and very well DOD. Nobody can see sector is like that too, right? Yeah. Yeah. And the oil and gas company is, you know, so that's why this whole DIY thing they're not going out and, you know, buying tech and they're not. You know, leveraging those kinds of things at all. They're all just trying to figure it out themselves and be, you know, they think that that's gonna give 'em a competitive advantage.

Deep: And then what about like a new energy company? Like a, you know, like Tesla, I mean, do they, they come at it from a completely different angle. They don't necessarily care about sort of legacy energy sources and they're sort of tech from the core up. Like, what is your take on. You know, maybe not just Tesla, but like some of the newer players.

Carolina: I mean, most of them are playing more in the renewable space, right. It's gonna be wind and solar. It's gonna be EV charging. It's gonna be battery technology. Um, and, and things like that. And I, I do think that that's, that's gonna keep growing and it's already growing at a rate that nobody foresaw, you know, if you look at. The, you know, the shape of the graph on, you know, solar and wind it's massive. And even this year, they bumped up the number of, um, you know, wind, uh, bid rounds for wind leases by, I don't know, they doubled or tripled them this year. And the like, I don't know, you'll know the stats way better than me, but I keep hearing things like the costs of building a new wind or solar plant are significantly lower than the cost of building a new coal plant or a new. Yeah, and, and quite a lot quicker too. If you think about the oil and gas sector from discovery till you actually have a facility, um, and what, what we, what used to be called first oil, like first drop goes into the pipeline on average, it's 10 to 15 years and costs billions of dollars, whereas a wind farm, you know, maybe three to five years and you know, then you're making money. So I think the, the hydrocarbon based energy sector is going to have to compete at a totally different level. And I, and I think that's driving the behavior around, you know, digitization and sustainability as well. You know, they, they need to be able to make decisions a lot quicker and a lot and, and have get at their information a lot quicker to be able to make those decisions quicker and better.

Deep: Why would they not just bet on some of the new technologies? Like why not say, okay. Yes. You know, Exxon I've done oil and gas forever. Uh, I also have a very large bank account. Like why don't I go shopping for new solar companies, new wind companies, new energy storage companies. Are you seeing that? Or, or

Carolina: I'm not seeing that in Exxon or Chevron or American companies, but all of the European majors are there already and have been doing it already for a while. So if you look at BP or shell or Elanor, they're all investing. They've all bought startups in wind and solar. They're all, they've all bought EV charging companies and, you know, they're do and battery, they're investing in tech, all of that.

Deep: It just feels obvious to me like, cuz I, I don't know why you would tie yourself to a type of energy. If you're an energy company, it feels to me like diversity is a natural thing, but do you know why they're not?

Carolina: So I think there's a couple of different things. So Europe is, uh, in a world to hurt cuz they're very dependent on Russia and they have not wanted to be independent on Russia. So the government, the European union has incentivized companies to do alternative. And there's a lot of good tax incentives for doing that and have been around for quite a long time. So, you know, the north sea and, and most of the European oil is dwindling quite a lot and has been for the past, you know, 20 years. Um, so they've had a lot of drive to invest in alternatives in the us. It's really different. It's the opposite of that. We have tons and tons of product. It's all over the place. Um, there's a very, very friendly regulatory environment, um, for, for hydrocarbon production and, um, you know, the government really hasn't stepped up to try to, you know, incentivize in different ways. You know, I think, you know, Biden administration is trying to. Do that, and I don't know if they're gonna pull it off. So I see I'm trying to, I kind of wanna stitch the story together cuz it's, and maybe you can fill in the gaps for me.

Deep: So this has been a fascinating conversation by far for me. I, I, I feel like I'm entering a world that I really know very, very little about. So I feel like this is maybe the first of many conversations that I need to have. But you're painting a picture, uh, for me of a world where clearly the climate is being affected. We've got European entities moving towards, you know, wind and solar as energy companies, less so here in the states. But we're also talking about artificial intelligence and the role of it in increasing data driven efficiencies. Bring it all together for us. Like what's, what's the best case scenario. If we go out 10 years with respect to artificial intelligence and energy.

Carolina: So I think that, um, I think the us is gonna catch up, uh, regulations wise. Uh, eventually, uh, we, we're gonna start to put in some, some regulations in place that will incentivize different behaviors. That's it's already happening. The consumers are. Climb for it. Um, the financial sector is climb for it. Um, and that, and that pressure's gonna continue to rise and, and, you know, lead to different regulations. And I think that also the, the energy industry is becoming more mature. I mean even two years ago, I think there was a lot more digital immaturity and lack of understanding. People were still thinking that sustainability was about reporting. What happened last year, as opposed to optimizing what's happening right now. And maybe even making some forecasting and future. Decisions based on the environmental impact of those decisions, like including your environmental impact in your cost, risk benefit, risk analysis, or your, you know, your, your business strategy. So I think all of that is already happening and it's, it's gonna continue to happen. I I'm actually very positive. I think that people are, are getting with the program. They're understanding it. They're, they're moving in the right direction. And they're gonna, in order to do that, they're gonna have to invest in, you know, something different in the way that they're managing their data. And I think they're waking up to that. Like I said, most of the medium companies are already there. The bigger ones are still trying to DIY, but eventually they'll get there and the little guys will get there too. I mean, Data operations is gonna have to change. They can't manage the data the way they have been in the past in these separate little silos, they're gonna have to be able to figure out how to merge it, contextualize it, and automate the generation of this data model that enables AI and machine learning and automation. And they they'll get there. They'll get there because if they don't get there, they won't survive. And all of that paints a picture of efficiency. Whether or not you're pulling oil and gas outta the ground, or you are running windmills or solar farms. All of it paints a picture of increased reporting of your goal towards net zero. And all of it lets you drive immediate optimizations that help you. Push further towards net zero.

Deep: It sounds like everybody, even the American companies are trying to go net zero. Is that correct?

Carolina: Yeah. I mean, Exxon has made claim has made, you know, net zero goals and Chevron. Uh, everybody has all major companies have. They've all said now, whether it's net zero to 2030 at 2030, or 2050, they all have those goals stated. It's not just about efficiency though. They also need to think about, you know, the supply chain issue and, you know, coming up with different circular solutions and, you know, either carbon sequestration or carbon capture in some sort. uh, you know, productization of that, you know, it's the multiple things are gonna have to happen. And I think American companies eventually are gonna need to diversify into renewables and other things too, which they're lagging, but they will get there.

Deep: I mean, it seems like just simply by the fact that them stating they want to hit net zero and you couple that with the increased decrease in cost of wind and solar and renewables and all the innovation happening there, not to mention. All the automobile fleets and, you know, are moving to electric. It just feels like there's no choice. I don't see how you don't go there.

Carolina: Absolutely. And you know, you asked the question, so how does this all relate to AI? Well, We're not gonna have a single energy source solution. Like we've had up till now. Everything's been hydrocarbon. It's very simple to manage. Now we're gonna have multiple sources. And those source sources are not steady. They're intermittent, solar and wind and things like that are, you know, at the mercy of the elements to some degree. And so you're going to need to do grid optimization in order to, and, you know, use storage. In order to be able to stabilize those grids. So that's you sort of a whole new flowering field around AI and managing multiple sources of data, you know, multiple sources of energy coming into a grid. And then how do you keep that grid really stable?

 Deep: Yeah. There's also like a lot of distributed pieces there too. Right? So now you've got the, the whole automobile fleet sitting on electric. You've got solar panels and wind, uh, turbines going all the way down to homes and neighborhoods. And, you know, you can imagine a world where, you know, you you've got extra charge in your Tesla or your, you know, or your Rivian, you park in a garage. The garage is competing based on the fact that it's got solar panels up above. So you can get, you know, carbon free electricity to reduce your car while you're in the office or, you know, or going wherever.

Carolina: I mean, it feels like in that more, there's some really cool like projects and studies that people are doing like California and Southern California. They did a study where I think it was PG and E. Um, took a bunch of, uh, you mentioned Tesla earlier. There's enough consumers that have Tesla home batteries. Yeah. In Southern California, that they were able to do a test to say, okay, what happens if we, you know, give these consumers, you know, some break on their per kilowat hour charge. If they allow us to use their storage, their, you know, distributed battery storage, 25% of it. If we need it in order to stabilize our grid. And that was a very successful pilot and it's been done in several other places.

Deep: Oh, that's interesting. I had not heard about that.

Carolina: That was that's really, there's another company that's looking at doing that with cars, like hybrid cars that are plugged into network. Yeah. Let us store the store, the power in your car. We pull it out later. Interesting. Yeah. I mean like that stuff becomes feasible. Be between like the compute capabilities, the intelligence we can deploy at the algorithmic level. Just like we can manage a much more complex energy system that could be more robust even.

Deep: Well, thanks so much. I feel like we've touched on a lot of topics that are a lot of fun. If, uh, is there anything that we didn't talk about that you would like to talk about?

Carolina: No, this was a great conversation. I really enjoyed it.

Deep: That's all for this episode of your AI injection as always. Thank you so much for tuning in. If you've enjoyed this episode and wanna know more about AI in the environment, you can check out a recent article of ours by Googling Xs environment, or by going to xanax.com/articles. Also, please feel free to tell your friends about us. Give us review and check out our past episodes at podcast.xyonix.com. That's podcast dot xyonix.com. That's all for this episode, I'm Deep Dhillon, your host saying check back soon for your next AI injection. In the meantime, if you need help injecting AI into your business, reach out to us@xanax.com. That's xyonix.com. Whether it's text, audio, video, or other business data, we help all kinds of organizations like yours automatically find and operationalize transformative insights.