- 5 hours ago
AI is driving huge energy demand - but it could also help cut it. AVEVA CEO Caspar Herzberg tells Cynthia Ng how industries are using data to run smarter operations.
Category
🗞
NewsTranscript
00:00Hello, you are tuned in to Awani Review with me Cynthia Ng.
00:15Today we are taking a closer look at a company that uses data to keep everyday systems running.
00:24From keeping the lights on to making sure the cereal that you eat tastes the same every time.
00:31Joining me on the show today is Kasper Herzberg.
00:34He is the CEO of Aviva which is one of the world's leading industrial software companies.
00:40Thank you so much Kasper for taking the time to speak with me.
00:43My pleasure, thank you for having me.
00:45And welcome back to Kuala Lumpur.
00:47Fantastic to be here.
00:49All right, so Aviva is a company that works quietly behind the scenes.
00:55Not many people know of the name, so let's start with that.
00:59What do you actually do and why should people care about the company?
01:04Well, you said it yourself very, very nicely.
01:08We are an industrial software company.
01:11What that means is that we design, we operate and we optimize with our software industrial assets.
01:18So you have to imagine this is anything from a nuclear power station, to a refinery, to a milk bottling plant, to a shipyard that builds a ship, a cruise ship or a massive container ship.
01:33All of these what we call industrial assets are designed with Aviva software and are then operated and optimized, meaning made better, with Aviva software.
01:45So you mentioned the yogurt or the milk that you eat.
01:50This is a bottling plant.
01:51It would run on Aviva software.
01:53The quality of the yogurt would be assured by Aviva software.
02:00The water that you consume, that you take, that you take a shower with in the morning or that you drink, it would come from one of the big utilities that assures the quality of that water, that manages the pumps that pump this water with Aviva software in an efficient and clean way.
02:20So this is what we do.
02:21And it's fun.
02:22It's true.
02:23Not everyone knows us, especially in the consumers.
02:27But it's a good feeling to work in a company where you know that much of the world relies on the fact that the software works well.
02:35So you are essentially everywhere.
02:37It's just that you're not usually mentioned by consumers, but you run the operations, or already the systems, the software behind it.
02:45Okay.
02:46So just taking a look at Aviva's history, the company has been around for more than half a century.
02:51Yet, if you take a look at some of the biggest technological advancements, we're looking at data, analytics, AI.
02:57This happened, I think, largely in the last decade or so, right?
03:02Yes.
03:03So in your opinion, why did the pace speed up and how has that changed the way you run your business or rather your clients?
03:11How are they running business differently now?
03:13Well, I think the important thing with technology is that it is a tool.
03:17It is a tool to achieve something that you want.
03:21So our customers use technology, Aviva technology, other people's technology, in order to achieve outcomes.
03:27And the more you know what you want to achieve, the better you will be at using and at identifying which technologies to use.
03:34Now, to the question of the advancements, I would say there are two in recent time.
03:41One is clearly the internet, the ability to move data from one place to another at scale, right?
03:48And how that led to the cloud and the ability to see data in the cloud to work with it.
03:57I think this is one big change in the last 20 years or so, 15 years or so,
04:03that has led to changes and to a lot of opportunity in the world of industry on how to make better use of the data,
04:11how to analyze it, how to use AI, you mentioned it, in order to predict which machines will break down one year from now,
04:21one and a half years from now, and what to do about it now instead of waiting until it surprises you when,
04:26as happened in Heathrow, for example, last year, a substation in your electrical distribution system blows up.
04:34These things you can predict, successfully predict, and this is probably the key driver of making use of AI in the industrial world.
04:47It is the ability to predict and then to act on these predictions.
04:51So, in your role, where do you see some of the areas where you think companies are perhaps under-investing at this point?
05:01So, clearly, all across the world, in some places more than others, utilities, specifically electrical utilities,
05:14have under-invested in digitization and have under-invested in the ability to model and act on predictions in their networks.
05:23This is a global issue. It's more pronounced in parts of the world. It is probably a lot more advanced, interestingly enough, in Asia.
05:34I mean, the electrification of China, for example, is one example where a lot of technology was put into the system
05:41to manage the flow of energy. But generally, when compared, for example, to the oil and gas industry,
05:50that are early adopters at digitization and what you can do with digitization,
05:54there are differences in both how the industry uses digital technology
06:02and in how the industry, as a result of that, is ready for the use of industrial AI.
06:10Can I pick your point about when you say you observe that some parts of the world are investing more heavily in this sort of technology and Asia.
06:21Why is Asia more advanced, in a sense?
06:25I think that because Asia is very competitive and Asian nations are very competitive and are able,
06:37they compete with each other and they compete with the rest of the globe.
06:41They're also quite good at long-term strategic planning and digitization lends itself to strategic planning.
06:49So you have Malaysia, you have Singapore, you have Korea, you have countries that have invested nationally
06:57in both the backbone, the digital backbone, and in prioritizing industries that work in digitization.
07:06Semiconductors is one example, right?
07:08So as a result of that, you know, they leapfrog and now they're quite ahead.
07:14You know, the frog that's quite ahead of other frogs, if you like.
07:18And I do see that in parts of industry and the Americas as well.
07:24Where do you see areas that are putting most strain on operations right now based on the clients that you're seeing?
07:33So, I mean, one thing to say is clearly the advent of AI and the big data centers that serve AI,
07:43in combination with the data centers that serve everyone, right?
07:47That's putting a lot of strain on both, on the network, on the...
07:52So it's a dilemma. So you need more energy for AI at the same time?
07:56Yeah. At the same time, you can use AI to significantly reduce the energy that you use.
08:03So the best energy is the energy you don't use.
08:06I mean, this is a fact, right? This is the cleanest energy.
08:09So there is a lot that AI, industrial AI, predictive analytics, for example,
08:17can do to significantly reduce the energy consumption, for example, of pumps.
08:23Pumps use a lot of energy, of chemical plants, of...
08:28So there's a lot of optimization that you can still do.
08:31And my argument would be that what you save in industry, it can power what you need in, let's say, in data centers, right?
08:41But of course, it means you have to use AI for constructive and useful things, you know?
08:51Maybe less cat videos and more optimizing industrial processes.
08:56That's where I'm going to get at.
08:57So where are the areas where you see most impact or measurable impact?
09:02And where do you see where AI is more promise than a reality?
09:06So very clearly in industry, when you combine predictive analytics,
09:11using algorithms to take time series data, so this is the data that measures the temperature of a machine or a pump,
09:20the vibration of it, and you analyze that, you can predict when a turbine, for example, will go down.
09:26When you are able to predict that correctly, you will save tens of millions, potentially hundreds of millions of dollars by doing preventive or predictive maintenance.
09:41So this is clearly an area where AI has massive benefit, right?
09:48If you then combine that with large language models and you allow an operator like a new engineer or an engineer to make sense of,
09:58in the average plant you have 100, 150 manuals, you know, no one reads them.
10:03And AI can summarize the manuals, can give advice, and the engineer then can decide, based on what he or she sees, what the correct action is.
10:16This is a very useful way of using AI.
10:19This works pretty much in any type of rotating equipment.
10:23Wind farms, another big example, right?
10:25A lot of times wind farms stop, especially when they're in the sea.
10:29It takes a long time to fix them.
10:31If you can predict when they start to have a struggle and the operator can schedule the maintenance,
10:38you make a wind farm operator significantly more profitable, right?
10:42Which is good for the planet, good for everyone.
10:45There are clearly other areas where you wonder sometimes, when you look at your Instagram feed,
10:50if there's the best use of AI, but I leave that to others too.
10:54I think now companies are facing pressure.
10:56It does sound like they are told, you know, to cut energy costs, cut the use of electricity,
11:03but at the same time they are also sucking up energy because of the demand for AI.
11:08So is this even realistic?
11:11And when will such a predictive technology, the ones that you mentioned,
11:15be able to offset the consumption of energy, the cost of consumption of energy?
11:20I think more widespread adoption across industry, I think will definitely pay for what we want to do in terms of investing more into AI.
11:32I think also at the same time, I mean, the energy transition is happening.
11:36A lot of new generation sources are coming.
11:39You know, there's a nuclear renaissance at the same time happening, meaning people are reinvesting in nuclear power as well.
11:48So I think it's absolutely possible to put enough generation together to be more efficient, more smart in how you distribute energy,
11:56and in making the processes on the industrial side much more energy efficient to allow, you know, a world that has the energy it needs.
12:06Absolutely.
12:07So what's interesting is that you see some companies are on board and are doing well,
12:13and some are seemingly still firefighting, I would say.
12:17So what separates the two? What separates companies that invest well in technology and don't?
12:24Well, we are in Malaysia, so let me use Petronas as an example.
12:27Petronas uses predictive analytics in upstream. They do extremely well with it.
12:32They save a lot of money by predicting when assets have issues and taking the necessary action.
12:39This is clearly a company that has, from the top down, understood the power that digital brings,
12:47and is also very clear about what they want to do with it.
12:50I mean, they use it to optimize industrial processes to save energy and to predict when machines go down.
12:57The challenge sometimes in, not just in industry, but generally in business, is that new technologies come.
13:05It's very fashionable to experiment with them. It's not really clear what they are meant to be used for.
13:13Large implementations happen, and then, you know, leaders change and the adoption in the business isn't big,
13:20and people move on, and then, you know, it's the equivalent of digital ruins that you have in your company.
13:31This does happen, absolutely, right? But again, I go back to what we said at the very beginning of our conversation.
13:38Digital, AI, all of these, these are just tools. You need to know what you want to do.
13:44And when you know what you want to do as a leader, as a business leader, as a business,
13:47then you find the right technologies that support that, not the other way around.
13:51So tech matters, but leadership matters just as much.
13:55And common sense.
13:56Okay, so let's zoom up for a moment, looking at where the world is heading.
14:01We are in the middle of a tech race or rivalry, predominantly the US and China.
14:09We're seeing restrictions on chip imports, on critical software, not just minerals, critical softwares as well.
14:16From where you sit, do you see that clients are being forced to rethink where to invest and how they invest,
14:26because of the geopolitical tensions that we're seeing?
14:29Yeah. I mean, there's no doubt that we are at the, not at the beginning, but in the middle,
14:36of a complete reshufflement of the global economic order, right?
14:43I mean, what has worked the past 50, 60 years will no longer work.
14:48And we probably go more to a world that is similar to the 19th century,
14:52when you have a lot of competing nation states and they try to leverage everything at their power in order to, you know, do well.
15:00And that means that supply chains are going to be continuously reconfigured as you react to these changes, right?
15:08So the businesses that are able to react, not just in the enterprise, but across their supply chain,
15:16where you buy, how you buy, who you sell to, in the most flexible and dynamic way to something that will change,
15:24well, sometimes changes every week or every day, but definitely changes over months.
15:29Those businesses that do well there, they will outperform those that don't.
15:34And one of the key ingredients of doing well there is to understand where, what data you have, what that data is,
15:42not just your sales data or who you sell to, where usually a lot of the focus is,
15:47but your production data, your sourcing data, your industrial data, right?
15:52And if you're able to share that effectively across your supply chain, then you're able to become much, much, much more nimble,
15:58much faster in, you know, changing processes, buying somewhere else, building a new plant somewhere else.
16:06I think this is the response, the smart response to this changing world.
16:11That sounds like there is more investment.
16:13So do you see that companies are investing maybe twice as much just to manage that geopolitical risk or duplicate systems even?
16:23Yeah.
16:24I see that smart companies, smart supply chains, invest in understanding the data that they have.
16:34This is not necessarily building new systems or necessitates large investments.
16:40This is in breaking down the silos that sit even within a company where processes, where people, departments,
16:48don't share data with each other.
16:50And they share even less of that data with other companies.
16:53When you're able to break down these silos, access this data and make informed decisions at the leadership level
17:00and delegate that then, then you will outperform.
17:03But that's not necessarily big investments.
17:06That's a change in thinking and how you think about most data is trapped.
17:12You have to imagine that the value is trapped within an organization.
17:16And you want to, you want to free that data.
17:19Okay.
17:20It makes sense of the data.
17:21Okay.
17:22I want to bring this closer to home.
17:23So Malaysia is wants to move up the value chain.
17:26Yeah.
17:27But if you look at adoption of industrial intelligence, it is still quite uneven.
17:31I think local companies, I think the smaller ones especially are facing, you know, budget constraints, talent constraints as well.
17:40So in your view, where are the local companies getting right?
17:44You mentioned Petronas.
17:45But if you were to look at perhaps then industries or maybe the smaller supply chain providers, where are they getting right?
17:55And where are they falling behind in respect to Malaysia?
17:58So, I mean, what my impression is, is a slightly different.
18:06So my impression is that there is a massive amount of digital talent in Malaysia.
18:11I want to start with the talent.
18:12We have a very big office here.
18:14We have 200 plus people.
18:16Some of the best people that serve customers in Malaysia and in Asia in general are here.
18:22And digital adoption understanding of what is possible, of what is also possible in the future is very high.
18:29And my understanding is that's a fairly widespread in Malaysia.
18:34When I see at the adoption, now I personally interact more with bigger companies.
18:40But when I see, for example, the digital adoption in water.
18:45I mentioned earlier clean water.
18:47I mean, some of the most impressive systems are here in Malaysia serving everyone.
18:53Now you can say these are state owned companies.
18:55They have more money.
18:56True.
18:57But nevertheless, there is a level of sophistication here that you don't see everywhere.
19:05Now, when it comes to smaller companies, a lot of small companies are quite sophisticated here.
19:11They use, for example, from us, our SCADA products.
19:16They do look on how they can access that same data that I talked about that is trapped, how to access that, how to make smart decisions in their processes.
19:26So I would say that one can always do more.
19:31But the overall level of digital maturity, if that's a word, is fairly high in Malaysia across all sectors.
19:40But if I can crop a bit further, where do you think we are falling behind or maybe under-investing?
19:46Well, again, it comes to what do you want to do as a nation.
19:55If you want your small and medium business sector to prosper and digitize aggressively,
20:03then you need national strategies that support specifically that sector or a sub-sector.
20:08You need to identify where exactly you want to do that.
20:13This has worked well in other countries and it's maybe something to double down on here as well.
20:18Okay. So I want to touch on talent.
20:21When people talk about automation or AI, people tend to worry about losing their jobs.
20:26But from where you sit, are our roles disappearing or are they changing?
20:34You know, I think about this a lot.
20:37And I'm increasingly coming to the conclusion that at least in the world of industry,
20:44there are actually not enough people for the jobs.
20:47Not the other way around.
20:49So I'm not disputing that there are some repetitive tasks, not necessarily jobs that will disappear because of AI.
20:57But I don't see massive reductions in the workforce, at least in the world of industry that we serve as a result of AI.
21:07I don't see it. I want to be very clear.
21:08So when you say there are not enough people for certain parts of the ecosystem, where are they?
21:13What are the roles?
21:14So chemical industry, just to give you one example, most heavy industries, most process industries struggle to find the talent,
21:21struggle to find the people that want to work there, enough people.
21:28So actually, using AI as decision support for the next generation that comes into the job,
21:36capturing what all the engineers know and make that available to new employees, to graduates,
21:46I think that's a huge value add of AI. And I think that's going to be very, very needed
21:51because the biggest problem we have in industry is not, is not AI.
21:56It's that there is a whole generation of experienced people that is retiring.
22:00This is the problem in Aveva. We're an industrial software company.
22:03So we have, you know, a lot of very experienced people.
22:06There's a problem in industry in general.
22:08And the biggest benefit AI is going to bring is to act as an advisor to the next generation.
22:15I don't see mass displacement of people by AI.
22:21I can't speak for other industries like insurance or I can't speak for that.
22:27But in what we cover, I don't see it.
22:30Okay. In this world of AI and automation, where do you see that?
22:34This is a bigger picture sort of question.
22:36Where do you see that human plays a more important role, not less?
22:41Because when you talk about technology and automation, we tend to reduce human to, okay, repetitive tasks, automation, machine.
22:47But where do you see the role of human that plays even a more important role than before?
22:52Well, I mean, we talked earlier about how to predict correctly what is going to happen in your supply chain,
22:59how to reconfigure accordingly, how to make the right decisions, you know, also when they are unpopular or not necessarily, you know, when it's the first time.
23:13This is something that AI will empower humans to make those decisions so that you are not so surprised by industrial events, by systems not working, by sudden changes in production quality, etc.
23:29So I think that people will make a lot more forward looking decisions in the industrial world as a result of the advice they get.
23:45And I think this is a benefit AI brings.
23:47And I think that this is something that will require talent because someone who is not a talent will not make those decisions.
23:56Right. So I think it's going to be very important to finding the people that want to make decisions, that are motivated, that care about the job they do and that they work in industry.
24:06Okay. Now, let me end with this. Now, if you were advising a CEO today, which I know you happens a lot of time, but maybe someone...
24:14A lot of CEOs advise me.
24:16I'm sure they come to you for advice.
24:19And it's overwhelming. What do you invest in? How much do I invest in? What is my ROI?
24:25What do you think are some of the decisions that they cannot delay any longer, that they need to invest now to future-proof their business?
24:35I think you have to be very clear. I go back to that, to be very clear on what you want to do as a business and very clear in a very complex world.
24:48Yeah, in a world that is changing and in a world where the supply chains are changing, where you get distracted every morning by the news and you think, oh, my God, what is this going to mean for us today?
24:59What do you want to achieve in this world? Next year, the year after, in five years?
25:08And take this long-term view of where you want the business to be and then reconfigure everything else accordingly.
25:15It sounds really easy. It's really hard to do. I think this is the thing.
25:20Okay. All right. Thank you so much, Kasper, for your time. I appreciate your insights very much.
25:24My pleasure. Thank you. Great to be here.
Comments