Tenaga Nasional Berhad’s “Invisible Architect”: Building a Resilient, Data-Driven National Grid

Speaker A: The AIBP ASEAN B2B Growth Podcast is a series of fireside chats with business leaders in Southeast Asia focused on growth in the region. Topics discussed include business strategy, sales and marketing, enterprise technology and innovation. Hello and welcome to the ASEAN B2B Growth Podcast where we sit down with individuals responsible for driving growth within their organizations here in Southeast Asia. My name is Vanessa and I'll be your host for today. In this episode, we are peeling back the layers of a critical yet often unseen powerhouse of the Malaysian economy. We are exploring the invisible architect, the digital and data driven infrastructure that ensures the lights stay on for millions while the nation pushes towards a greener, smarter future. To help us understand how to build a, uh, truly resilient data driven national grid, we are joined by engineer, technologist Mohammad L. Jundi Abdullah, the lead of analytics at Tanaga National Berhad. Anjik Jundi, a pleasure to have you here with us. Perhaps to start off, could you give us a brief introduction of yourself, your role at tnb, and perhaps for our audience who may not be that familiar with the company, a brief overview, a brief introduction of the company as well.

Speaker B: Okay, uh, thank you, Vanessa. Um, hello everyone. My name is Jundi. I am the lead for analytics, uh, in Distribution Network Division at Tanagar National Berhad, uh, or tnb, uh, for the audience, Information for the audience. Uh, tnb, Tanagar National Burhard, especially the Distribution Network division, uh, is responsible, uh, to deliver electricity to our millions of customers across Malaysia. So for dn, the nbdn, uh, our focus is more on uh, reliability, uh, to deliver, uh, our supply in reliable manner to our customers, uh, in terms of asset performance and also to maintain our asset performance and also to maintain our operational efficiency, uh, to our customers. In terms of my role, uh, I focus on, uh, applying uh, advanced analytics and AI, uh to improve how tnb, uh, especially dn, uh manage our assets, uh, from predictive maintenance, for example, to support better decision making on the ground. And over the past few years I've been involved uh, in scaling, uh, up these capabilities, uh, nationwide, uh, in particular, uh, for our critical assets, for example, the underground cable. I think that's all for now.

Speaker A: Understand it's clear that your role sits right at the intersection of tradition and transformation. Also I want to um, maybe pivot a little bit now into how that transformation translates into business value for the company. Uh, for a national utility like TNB in Malaysia, growth isn't just looking at the revenue numbers, right? For yourself and Chundi, you mentioned that you are leading analytics for dn, uh, distribution network. Maybe could you also share how you define growth in your current role? Um, and you know you have had to navigate through the pandemic in the last five years as well. How did, did that shift, how did that evolution look like for you?

Speaker B: Okay, thank you, Vanessa. Uh, over the last uh, five years, I would say, uh, growth. Growth for tnb, especially for distribution network has really shifted not only to expand our assets, uh, but more in terms of uh, increase our optimization. Uh, previously growth would mean uh, building more assets, more coverage, more capital investment. But today I think it's more about maximizing our performance on what we already have. Yes, the pandemic was really a turning point for us uh, at that time. Starting that time, load patterns, the load patterns have become more unpredictable. We really couldn't rely on traditional planning anymore. Uh, that has forced us to accelerate uh, the use of data AI analytics. At the same time, uh, we are also seeing new uh, demand, uh drivers that come along, such as the increase of uh, data centers, uh, which require more dynamic forms of energy consumption, also require greater flexibility in terms of our network. So I would say now the growth not uh, really being defined uh as uh, uh growth or increase of our assets. It's more like uh, in terms of to have higher reliability, uh, better asset utilization and more importantly to have faster, more informed decision making.

Speaker A: Understand. And you mentioned a little bit about like data center. Right. Um, for our audience as well, there's a lot of conversations now in Malaysia around data center investments. How has that. How much of the growth strategy would you say is driven by this kind of external factors versus maybe your bosses, your board of directors pushing for digital transformation in your team?

Speaker B: Okay. In terms of growth strategy, um, I would say it's more like ah, external factors versus internal uh factors. It's more like combination of both. Externally in terms of external factors, we are seeing strong drivers. Like I mentioned before, the data center growth we are also seeing in terms of electrification, uh, for example EV charging infrastructure which are increasing and also from our consumers, changing demand patterns when it comes to the energy transition era. Increase of solar panels as well. So in terms of internal factors, uh, 10B is pushing uh for accelerating digital transformation, uh, to have a uh state where AI can help in terms of uh, asset management decision making. And I would say this combination of both, uh, is really important. External demand, uh, would create pressure for TNB to increase our capacity, but at the same time the internal uh factors or internal capability would determine how effectively we could respond to Those uh, external demands.

Speaker A: Understand? And what you mentioned is very true. Right? Balancing the external and internal factors as well as you approach digital transformation. Um, you know, last year when we had the conversation with you and your team, the spotlight was really on underground cables, uh, which won the 2025 AIBP Enterprise Innovation Awards for Malaysia. Could you maybe give us a brief overview about that particular project? And as you know your focus shifts towards overhead lines this year. What kind of unique data challenges more specific to your role in Cheek Joon D? What kind of challenges do these assets present compared to underground?

Speaker B: Okay, yeah, uh, yeah, last year we won the award for uh, underground cable project. Uh, I would say that at that time uh, that project was really important for tnb, especially our description network division because uh, underground cable is really our main pinpoint, uh the main contributing factor to our failures. But as you say, rightly say just now Vanessa, now we are only not uh, not only focusing on underground cable. Now we are moving ahead uh for our. For other asset type which is the next one with the overhead lines. So for your information, uh, underground cables and overhead lines are uh really different type of assets. They post. They have a really different uh data challenges. Uh, so as you know uh, underground cables uh, are buried underground. So so they are hidden. So we really rely on uh diagnostic data such as our insulation resistance, our partial discharge is uh more deep analytics problem to understand the internal degradation. Uh, uh, on the other hand for our overhead lines, uh, they are more exposed to the environment. The challenge is not really the lack of data but in terms of high uh variation of uh, external factors that can uh, uh contribute to this wahydland failures. So for example the factors like uh, weather conditions, uh, vegetation. What I mean here is from the three branches for example and also from third party interference. When I say third party is not really human being. Animals included, uh, animals can climb, monkeys can climb on our weight lines, bird can uh, cross our headline, so on. So uh, in a nutshell I would conclude underground cable we are more predicting failure from our internal signals. But for our headlines, uh, we shall interpret uh the complex uh external factors uh and also in real time. That I would say the main difference. Then one more if I would say uh, this is why for our weight lines we are leveraging uh more in terms of drone inspections, uh image analytics and also video analytics. Because the problem becomes more broader and I uh would say more dynamic.

Speaker A: So you basically need more people, whether people or technology on the ground for you to look after all these exposed overhead lines.

Speaker B: Right?

Speaker A: Understand and When I understand correctly, in some of our past conversations, one of your other key mandates as well, Njik Jundi, is using data analytics hotspots to predict and counter electricity theft from bitcoin mining. Right. How do you distinguish, I mean for yourself, how do you distinguish between a, uh, legitimate industrial search versus like illegal mining activities through data itself?

Speaker B: Okay, this is an interesting question. Uh, it's quite a technical one as well. Um, okay. To distinguish really between the industrial search and also illegal bitcoin mining, uh, it's not really, uh, we are not focusing on how much the energy consumption or how much energy is being used, but more on how the energy uh, consumption behavior, how it behaves. So we have to clearly, uh, differentiate between bitcoin mining, uh, energy consumption behavior and uh, the legitimate industrial search. For example, uh, for example, the bitcoin mining typically shows a very, I would say flat continuous load profile. Uh, it's a 247 operation with minimal uh, variation. Uh, whereas the legitimate loads for industry, it follows, uh, some cycles, some fixed operational cycles. So it's a real time based variation or variability. Uh, not only that, we are also looking at the power quality signals like harmonics, customer profile versus the actual usage, and also in terms of location and historical patterns. So the key idea is we are not just detecting how much energy is used or the energy consumption alone. We are also identifying uh, the energy behavior or the behavioral signatures of energy.

Speaker A: Understand. And in terms of this particular area where you look at. Right. Specifically around illegal bitcoin mining, are there still challenges right now that for TNB or even for some of your industry peers, do you think this is a common challenge that many of our Southeast Asia peers are also experiencing right now?

Speaker B: Yes, I would say yes. Yeah,

Speaker A: understand. So this is definitely something that is top of mind for many of the utility companies, especially in Southeast Asia. I think. Moving on to some of the metrics that you look at. Right. You know, earlier we spoke about like the underground cables, overhead lines. We spoke quite briefly about the bitcoin mining being a key issue, a key challenge many utility companies are facing as well. In your perspective currently in your role, how do you measure the success of this kind of initiatives? And also are there specific metrics that you look at when, let's say, um, your board of directors ask you about the status of these projects?

Speaker B: Okay. Uh, key success factor. I would say this question relies to the how much we measure, how can we measure success? Correct. Uh, and I would not limit to the project of uh, detecting crypto mining, legal, uh, theft alone as A whole. Um, traditionally for tnb, especially for distribution network, uh our key focus or our uh KPI would say would rely on the traditional side, uh what you call it system average uh interruption duration index, this id uh calculation. Uh but now we are evolving more towards uh value uh driven metrics or impact driven metrics. Uh not only in terms of how many minutes of outage we can avoid but also in terms of the uh, I would say for example prediction accuracy uh of our solution that leads to actual action and of course from this uh prediction uh or this uh better decision making to reflect how much cost savings that come along with it and also uh how ready for TMB to handle more dynamic and distributed demand patterns that we have now. So for my, especially for my team the most important uh one of the most key metrics is the how many uh operational decision making that we have actually changed because of the success of our analytics or AI solution. Because if the data doesn't really influence our decision making it really doesn't create value.

Speaker A: Understand? So it's really looking at the kind of value that the data insights can provide to the leaders at T and

Speaker B: D. Yes, got it.

Speaker A: And when it comes to like the different kind of projects that your bosses that your board of directors ask you to embark on, were you ever caught in a situation whereby there's like say too many projects and then you are not sure which ones to focus on and if you will, how do you usually find a balance between coping with all the projects?

Speaker B: Yeah. So uh, of course when it comes to the uh era of digitalization AI booming uh there are a lot of uh initiative or use cases that uh come or bloom uh which that is why in uh TNB we have this governance that looks into this uh, uh to ensure the delivery, the uh initiation and also the delivery of these uh projects. So when, when uh we could say in terms of how we determine whether we want to stop a project or not for us actually ties back to the value in terms of the practical value. For example if the project has uh shown uh no clear uh use case or operational use case. There's one example, maybe the project has high accuracy in terms of prediction but, but really no actionable outcome. Also if there is a data problem, data that cannot be sustained or scaled. And the most important thing is I would say whenever the initiative uh has lack of trust or lack of confidence from our end users, for me that's the biggest factor because uh, if the field team or the operational users doesn't trust or use the solution for Us the project has already failed.

Speaker A: M. Okay, okay. Understand. And when it comes to like balancing innovation against some of the risk. Right. One of the common conversations now that we are seeing across the region is how do you balance the speed of AI implementation, AI rollout versus certain governance guardrails, cybersecurity guardrails for the team at tnb. Is this also under the AI Governance Council? And how do you usually uh, evaluate or um, find a balance between the two? Do you work together with like the CyberSecurity teams, the AI governance teams as well?

Speaker B: Yes. Rightly said Vanessa. So in terms of governance, uh, uh, like I said just now, the most important thing to consider when you are evaluating a project, whether you will be uh, carried out or not, one thing, the most important thing is the value that comes from that. And uh, in that governance that we have, uh, we have uh, multiple uh, committee members that come from various backgrounds, technical it, uh, cybersecurity as well. So all of this aspect has been included and meticulously evaluate, uh, uh, evaluated for each project that comes to the uh, governance.

Speaker A: Understand. And for you personally, right. Anjeet Jundi, how do you feel or like when is your moment that you know that some of the projects, some of the initiatives to you is a success? Is there like a specific Eureka moment? I would say that you feel that you experience or when is it that you know that it's a success?

Speaker B: Okay, that's uh, an interesting question. So if you ask me, uh, personally is innovation or innovation is really a Eureka moment? For me, I would say no, no at all. Uh, for me, innovation, uh, is not really. It's not only about discovery. Uh, it's most important. The more important thing is about the adoption or the application. For me, innovation happens when for example, our solution would flag a risk, uh, or AI solution can uh, detect any uh, anomaly. Uh, and whenever someone on the ground changes their decision or action because of the AI or uh, solution, uh, insights, I think that is the moment of the uh, eureka for me. The real world impact, that's the innovation for me.

Speaker A: Understand? So really looking at like the business value, the impact, the outcomes that your AI insights can drive for the business.

Speaker B: Yes, correct.

Speaker A: Okay, got it. You know, we've spoken quite a lot about like the different machine, the models that you are looking at. But at the end of the day there are also people on the field interacting with these assets every day. Right. Um, you know, we mentioned before in some of our earlier conversations that you are currently brushing up AI for work management. Maybe can we dive A little bit deeper into that. You know, there's a lot of conversations now also around AI replacing the workforce, AI complementing the workforce. So what's your take on that and how is TNP approaching it?

Speaker B: Yeah, that is a really common question that we have, uh, especially in our industry, uh, are we going to be like a Terminator, like replace a human being? I think, uh, uh, I would say for this context, uh, when we say that, how does putting an AI really, uh, change the human involvement or how we say the change of human to asset relationship when AI is coming in? I think that is a big transformation. Previously, uh, our decision making, our decision really relied on, uh, heavily on our experience of uh, our technician, field technician or field engineers and our standard procedures. Now when AI is coming into the picture, AI can act as for example the real time assistant. Whenever this happens, this will lead to, for me, this will lead to more consistent execution. We can reduce, uh, variation. We can uh, standardize our action, of course. And also we have better, with AI, we would have better quality control. It also allows us to capture and standardize uh, knowledge across uh, all levels of our teams. So for us, um, the relationship between human AI, uh, would evolve from previously from human only from human to asset. Now human and AI working together on the asset. And as I would like to say my leader, my management, uh, uh, our chief Senior Network officer always said whenever AI coming to the place, uh, we would treat AI to enhance human intelligence. So uh, the term AI to elevate IA intelligence, augmentation of human. So that's what we are uh, currently moving towards to.

Speaker A: Thank you very much for sharing, Anjit Jundi. And I think I want to go back to where we started off. Right. We introduced you as like a technologist, as an engineer. And sometimes I hear that engineers like to say they often hear or sense a thought before they see it and like something like a sixth sense. And for yourself, you know, you spoke briefly about um, intelligent automation. We spoke about AI human. Um, but for you, in your data models, in the work around AI that you are dabbling in, are you trying to also digitalize that human intuition? And you know, in conversations also with like some of your regional utility players, they would say that the goal is not really for AI to replace, but rather is to complement the human element. Um, what's your take on that? And do you think that humans will really be replaced by AI?

Speaker B: No, um, I would say, uh, like I said before, uh, human AI should work together to increase our productivity. And also like I mentioned before, to have uh, better quality control, reduce variation, more consistent execution. And um, I would say, um, since in industry like power utility workloads and um, uh, work orders, for example, coming from various workloads are huge. Ah. Uh, our workers needs to uh, multiply their efforts, uh, daily, uh, to execute all those tasks. And uh, with AI, uh, for example, to have a better prediction that we have now, uh, we would really save or we have, uh, can have a targeted task or targeted executions and would really reduce unnecessary tasks or works that we previously would do that. So I think that where AI actually come into assistance, uh, in terms of uh, uh, human workloads, not really replacing them, but to assist them and to augment their intelligence in every task that uh, they are doing in the field.

Speaker A: Understand? Thank you Anjeet Jundi, for sharing and you know, for some of our audience who are listening as well, if they wanted to drive growth through technology, where do you suggest they start from? Should it be from like the data fundamentals? Uh, you know, we hear one of the biggest challenges in AI implementation is that the data they have is very messy. Right. Um, at the end of the day it's also like garbage in, garbage out. So how do you, like, what kind of advice would you give to leaders who are starting to embark on their AI journey and how big of a role does data play in it?

Speaker B: Yes, for me, start with people. Always start uh, with people. Because if uh, for data, yes, of course data without context can be misleading. But uh, the field teams, the users, the people that uh, they are the ones who really understand the real asset behavior and constraint. So for me personally, I would say my advice, uh, is to start with real, uh, operational problems, uh, translate them into data use cases. Then from there we would apply analytics or AI. So the best solutions come from the real pain points on the ground, from the people, from the users, uh, not always from the data sets. That would be my advice.

Speaker A: Thank you very much, Anjik Jundi. And it's very interesting to hear your perspectives on the Invisible Architect and for sharing how TNB is balancing high tech innovation with AI, with data analytics against the very physical realities of nation. Uh, you know, we spoke about the underground cables, overhead lines, we spoke very briefly about illegal bitcoin mining activities as well. Um, it's been great having you on the ASAN B2B growth podcast with us today. And as TNB scales and grows in the coming years, we certainly look forward more to some of the innovations that you are leading with the team and definitely keep you more on your toes, I guess, with a lot of the other new and emerging technologies coming in. Thank you for joining us today. And Chik Jundee, thank you.

Speaker B: Vanessa, thank you, AIB for having me on the podcast. Thank you very much.