This is Leaders in Finance, a podcast where we find out more about the people behind a successful career. We speak with the leaders of today and tomorrow to discuss their motivations, their organizations and their personal lives. Why? Because the financial sector could use a little more honest conversation. We’d like to thank our partners for their ongoing support. They are EY, MeDirect, RiskQuest, Kayak and Roland Berger. Your host is Jeroen Broekema. [Jeroen]
Welcome to a new episode of Leaders in Finance. Today, an extra episode recorded in the run-up to the FLAIRS event on the 27th of September. The event theme is Banking from the Past to the Future and we’ll specifically zoom in on sustainability and data and hopefully the use of AI if I was the one to design. Super important topics in itself and we will try to see how they are connected. I’m sure we will learn more today with my guests here at this table, which are people from very different institutions, all banks, but they’re also represented at the FLAIRS event. I will introduce them to you.
So first, Dennis de Reus, who is at ABN AMRO and he is the Head of Innovation Technology. Welcome, Dennis.
[Dennis]Great to be here. [Jeroen]
Wonderful. Next, Adrian Collien at ASN Bank, Operations and Transition Manager there. Welcome. [Adrian]
Thanks. Thanks for having me. [Jeroen]
Wonderful. And Dafina Nonkulovska at Rabobank, CFO, Lead Sustainability. Welcome. [Dafina]
Thanks, Jeroen. [Jeroen]
Wonderful. And last but not least, she is at the Organizing Committee at FLAIRS, Fleur Gieskes. Welcome. [Fleur]
Thank you. Thank you for having us here, Jeroen. [Jeroen]
And first, before we dive into the topic, I’d love to chat a little bit with you, Fleur, to learn more about FLAIRS, because I’m mentioning FLAIRS in the introduction and maybe listeners have no idea what FLAIRS is. So tell me more. [Fleur]
I’ll definitely tell you something about FLAIRS. So FLAIRS actually started in 2006. That’s quite a while ago already. It’s currently organized by the five banks. So Rabobank, ABN AMRO, Van Lanschot Kempen, ING and the Volksbank. The main goal was really actually to connect the younger bankers to the topics that were really going on at that moment within the banking sector. So that’s why we organized FLAIRS. During the event, you have like two workshop rounds, two panel discussions, plenty of time for networking and some great keynote speakers.So pretty cool. [Jeroen]
Do you have one or two names for us to share that will show up that day? [Fleur]
I will actually share three names. Barbara Baarsma will be there, Mathieu Willems and Robert Swaak. [Jeroen]
Wow, good names. And how many people approximately are joining this event? [Fleur]
400 people. [Jeroen]
That’s a lot. Wow. Yeah, that’s great.
And so as I said in my introduction, is that correct? Is that going to be the main title, the main theme of the day?
[Fleur]Yeah. Banking from the past to the future. So really showing the difference from banking in the past and where we are heading in the future. Focusing on like sustainable financing, digital transformation, financial crime detection and customer centricity is really important during the day. [Jeroen]
Good. Thanks for that introduction. That’s great. I mean, hopefully I’m going to be there as well. We now know about FLAIRS, we go into the topic of sustainability and data and AI. But I’d love to know more about you guys, because it’s interesting to know a bit more about your background. Maybe I can start with you, Adrian. Could you tell me more about yourself and what you actually do within ASN Bank? [Adrian]
Yeah, I’m happy to introduce myself. So as you say, Adrian, I have a German accent. That’s because I’m from Germany, from Munich, the south. I made the decision to move to the Netherlands for my bachelor studies. And fast forward 15 years later, I’m still here. Maybe on a personal note, I recently moved to Utrecht, which is exciting. But I’m now happy to discover the place, the city center, which is beautiful. And I’m also biking outside of the area. And I’m playing tennis and making art as well. I have a creative side. And as you say, I work at ASN Bank. Maybe good to know ASN is part of the Volksbank. So good to mention and Volksbank being participating organization of FLAIRS. And at ASN, as you say, I’m operations and transition manager, project finance department. So what we do is we provide financing, mostly loans to different types of capital projects, to entrepreneurs, to companies that want to make a sustainable impact, for example, in the fields of the circular economy and energy transition, biodiversity, and many other exciting new fields in sustainability that we want to be a part of. So very excited to be here to discuss sustainable banking. It’s obviously a topic very close to my heart. I’m really looking forward to have this discussion. And of course, technology will have a big impact. So it can help us in the transition. But I’m also happy to share a couple of concerns or a couple of other viewpoints that are worth considering when it comes to the use of the technology responsibly. [Jeroen]
That’s great. Let me start with a compliment. We had a short conversation before the podcast and your Dutch is wonderful. It’s just amazing. I mean, you’re hardly able to hear that you’re originally not from here. So that’s amazing. Do you miss Germany? [Adrian]
I do miss it. And I’m actually happy to come back once in a while. So I try to go there three times a year, approximately. It doesn’t have to be my hometown, Munich. Sometimes I just go to Hamburg. [Jeroen]
You’re closer now than Rotterdam. [Adrian]
Yeah, I’m closer now. Yeah, there’s an intercity part. [Jeroen]
It saves you a bit of time. Wonderful. Thanks for this great introduction. Dennis, over to you. Could you introduce yourself and tell me more about what you actually do at ABN AMRO? [Dennis]
Yeah. So I basically do two things at ABN. I have the engineering team for all our new propositions. And so we build apps and websites and backend and platforms for new things that we launch. And that’s, for example, we have a proposition called Payday where we help gig workers where they can say like, hey, I’ve done my gig, can I already get paid? And they have instant payments. So we have like 15 of those initiatives. And then the other half of my team is the AI team. And over the last, let’s say, I’ve had the team for the last five years, but over the last two years, it’s really accelerated. And so we’ve rolled out our own version of ChatGPT for the bank, AI support in the contact center, and a number of other things. So that’s basically what I do. Then how I got here is the longer story. I’m actually a chemical engineer that kind of ended up in technology and banking with an in-between step at McKinsey. But as a consultant, I did not just do PowerPoints, but actually had a lot of software implementation projects as well. So that’s kind of the link how I’ve managed to end up where I am. [Jeroen]
And have you learned writing code yourself as a chemical engineer, or do you have people that are really good at that? [Dennis]
Not as a chemical engineer, but I did kind of when I was 14, I started coding. So that’s just before the dot-com bubble, dial-up internet, these type of times. And I’ve always coded a little bit, but never professionally. But I do code. And I think sometimes the teams appreciate it. And sometimes they’re like, it would be easier if this guy didn’t know how to code. [Jeroen]
Makes sense. Thanks. Wonderful introduction. Last but not least, Dafina, could you do the same thing and introduce yourself? Tell me more about your role. [Dafina]
Thanks, Jeroen. It’s interesting because I’m also German, but by naturalized German. I’m originally from Macedonia. And I have been in the Netherlands for the last four years. I consider myself very international as I have a Portuguese husband and a Spanish cat. So I consider being quite European in that sense. And it’s always difficult to belong in one place and knowing where you are and what you take away. I think that I’m taking the best of every place where I am. Yeah, I live also close to Utrecht. In my role, I’m the CFO lead for sustainability and I consider my role as being the bridge between the CFO domain and sustainability. In my role, we are trying to basically help Rabobank transform and embed sustainability in the strategic cycle. Trying to embed sustainability into all the processes within the CFO domain from external reporting, which is basically the preparation of the annual report according to the CSRD, or up to now, we were always working within the impact report, all the different types of sustainable information. But also to embed that into the thinking of the bank on how can we make better decisions for a future fit businesses or future fit bank. And in that sense, we are trying to incorporate the balanced perspective of risk return and impact into the decision making process. I love my role. And I think with my role, we have been trying to make a change with the organization quite successfully. And in my dream, I think you mentioned, Jeroen, when we were talking, is it a CFO or CVO? I wish in future it’s a CVO. [Jeroen]
It was a CVO. [Dafina]
It’s a chief value officer. I believe that financial professionals can provide value for the organization, not only from financial perspective, but also in a balanced perspective with integrating impact into all the processes that they do. And I wish that CVO becomes a reality in that sense. And I always think of financial professionals can be the ambassadors for the planet. Yeah, kind of trying to motivate in that direction. [Jeroen]
Thank you. That’s a great introduction as well. I was just wondering from an international perspective, your personal one, but then to the business one, are you doing this for the Dutch Rabobank business or globally? [Dafina]
No, I’m doing it globally. So I do this for the whole bank, including DLL and the international business and the Dutch business. Obviously, my Dutch is not as good as Adrian’s. I think even with my German background, it takes a while, but my team and the colleagues that I work are, yeah, it’s an international bank, so English is well embedded. But there are stark differences, so when we talk about Netherlands and the Dutch businesses, sustainability is prime focus. However, when we talk about countries like US or Australia, or Brazil where we are actually operating. There’s still steps that needs to be taken by the governments and the policies there so that we can have further steps on sustainable future. [Jeroen]
So let me start with talking about importance, right, about these two major words we have here, sustainable finance and data slash tech slash AI. Dennis, you mentioned your team grew rapidly. Does it mean that for the leadership of the bank, AI tech data is really one of the top priorities or is that too much? [Dennis]
We really see a big potential, especially now in the area of gen AI. And I think the example for me that illustrates that is that last year when the open AI APIs, so the way that we could programmatically talk to open AI, came out, we were able to launch that in our contact center about four weeks. And that’s kind of an unusually short timeline for traditional IT.
And then it turns out that especially when you talk about, you know, data and modeling, these type of things, and then with generative AI, I’ve been able to do that very, very quickly, and then scale it very quickly as well. And we still see the technology evolve very quickly as well. And so our broader perspective is that, yes, it’s going to be a very impactful development for the financial sector.And I think that’s also supported by what we see externally. So if you look at the consultants, you see these reports where it’s like, okay, so these industries are going to be most affected by gen AI. And it’s basically sorted by how much text they handle. And banks happen to handle a lot of text or a lot of kind of written data at least. And so I think for us, but also for insurance, and a number of other activities, this is going to be a meaningful change. [Jeroen]
So you use the word potential, right? So you’re already using it, but the part that is potential is much bigger than what you already do, right? I assume. Is that right? [Dennis]
Yeah, exactly. So I think we do have it, of course, live in a number of areas. And you can see the quality of the feedback being great. Like people love working with this type of tool. I’ll pick again on the call center as example. So what we do is we record the conversations and then automatically summarize them afterwards. And that means that people don’t have to log during or after the call. And so it saves them time, which is, I mean, that’s great from a business case perspective. But if you go and talk to the people in the contact center, they’re saying like, look, usually I’m on the phone and I have to log and help the client and manage the systems. So I’m juggling these three balls all the time. Now I don’t have to log anymore. So I’m only juggling two balls. And that means I can spend more time with the client. And so qualitatively, we think actually we are able to offer people a better experience. And since we’re somewhat faster, shorter waiting times and these type of things. So they were very happy. But it’s also just the tip of the iceberg, right? This has been around only for a year and a half. So we expect a lot more when it comes from writing code, writing content, personalizing content, and these type of things. [Jeroen]
Adrian, I assume that ASN Bank is well known for its sustainability, right? So sustainability has always been important, right? Or are there things changing now with all the tech developments? [Adrian]
No, of course, sustainability will always be the core of our mission at ASN Bank. And we see technology mostly as an enabler to actually make that impact. So when you look at the impact that we are trying to make, we put it in three categories, climate change mitigation, and adaptation, by the way, as well. Second, biodiversity loss, preventing it or restoring biodiversity. And third, taking care of human rights and the social aspects of sustainability. And whenever AI can help us actually make that impact, then we embrace that technology, of course. And also at Volksbank, we have an AI center that is working on lots of different pilots whereby we use both internal and external information in an ethical way, adhering to privacy, anti-discrimination. So we are embracing the technology to actually make that impact at ASN. [Jeroen]
Right. And it’s great you mentioned the word Volksbank, because that’s a nice segway into what I wanted to ask you around the importance question. Because do you see that ASN has a lot of impact on the other, I think there are four brands, right? Four major brands on the other three brands. So Regiobank, SNS, and BLG Wonen. Do you see that? Does the whole bank become greener or is it still the island ASN? [Adrian]
Well, the way you can see it is that the Volksbank has an impact mission, an impact strategy, and each of the different brands, as we call them, has a different focus areas. But ASN is the pacesetter when it comes to sustainability, not only internally, but also externally, by the way. We have been one of the first to, yeah, take this issue very seriously. And now more and more of the financial sector moves with us, which we are very happy about. But also internally, of course, the same thing happens that the expertise center sustainability, which is located within ASN Bank, now has the scope of the whole of Volksbank. So actually, they set the minimum standards that all of the brands adhere to. So yeah, we use a lot of that insights across the brands as well. [Jeroen]
Pacesetter, both internally and externally. That’s what I take away. I’ve asked the two gentlemen here around one more question on the AI side and one more on sustainability side.
And Dafina, I want to ask you both at Rabobank, how important are these two topics for Rabobank?
[Dafina]Yeah, absolutely. Hot topics, I would say. It’s interesting because if I can mold this all together a bit on the recent experiment that we have been doing within the Rabobank, is that we have actually looked into, especially our research team, sustainability analytics team, have looked into how can we actually use AI to actually source data from our clients for potential use cases in terms of external reporting or assessing of our clients. You know, traditionally, bank would approach, it will gather data directly from the client, if there are big corporates, or if they would go through the credit bureaus, if it’s a mass of data that SME businesses and so on. And what we discover is that when we talk about sustainability, it’s very difficult for relationship managers to know what they need to gather and how precise they need to be and what all the potential information that they need to gather.
So it becomes quite difficult to collect relevant or let’s say correct information or accurate information from a client. And our team, sustainability analytics team, they run a script by asking questions to the general information, so everything that is publicly available. And they actually were able to source quality, good quality data, four out of five. So that is already a great step to think about data sourcing with regards to sustainability using GenAI. So I think they are both very important. And if we can bring better use cases, I think it will make it our life especially as understanding the impact that our clients have on nature and climate will help us actually tremendously.
[Jeroen]No, that’s good. That’s a wonderful answer. And I will get back to how you guys are using or you know, want to use AI tech data for sustainability purposes and to see if you have examples, use cases, you guys call it. [Voice-over]
You’re listening to Leaders in Finance with Jeroen Broekema. [Jeroen]
When I was preparing for this podcast, the first thing that came to mind, and actually before this podcast, we already talked about a little bit off the record, I love to do it on the record as well. When I hear these two words, sustainable finance and AI, the first thing that comes to mind is energy, right? You hear everyone talking about AI costs so much energy, and which is really bad for the environment. So any thoughts on that from anyone here? [Dennis]
So this is my cue. [Jeroen]
You go for it, Dennis. [Dennis]
So I think on a macro level, what we foresee, and I don’t know, we as ABN, but what you see in general and what the media writes about is that people are investing tremendously in compute capacity to build ever larger and ever more capable models. And I think if you extrapolate that line, I’m not sure if we have the actual numbers on the table, but I think it’s something like doubling the energy consumption of data centers over the next so many years. And data centers represent some percentage, I think it’s a small percentage, but we’re talking about a very large amount of energy if you look at the global energy pool. And so adding a percent on top of that actually is a lot, a lot of energy. So on the macro level, AI is supposed to kind of really impact the amount of electricity or power in general that we use in our data centers. And so that’s a concern. And there’s a few things happening I think that will dampen that movement. So the AI companies are very careful about releasing impact data, but we can read a little bit into their pricing. Like clearly they’re making a profit and if they drop their pricing by a factor of 10, then underlying that they are apparently using less energy to some degree to provide these services. And there we see, for example, from the initial chat GPT, which was a GPT three and a half model to now four mini, where they’re saying, look, the performance of this model is better and it’s priced at half the price of the original model. So that’s an indication that we’re seeing a trend where we can get better performance at a lower price and by proxy, I think, a lower impact score per chat that you’re having as well. And so we’re having a number of interactions that people are having versus the consumption per interaction. And I think that that second number is going down, even though the quality of their answers is going up. So there’s some dampening mechanisms in there, I think, that are promising in terms of what’s happening. [Jeroen]
Yeah, but if you use your distinction between macro level and more like, let’s say, your own bank level, is it fair to say that for you, for your bank, it’s not an issue at all yet? [Dennis]
If we start using a thousand times more AI, then at some point, it will become an issue, of course. But if you look at today, so our largest consumption is summarizing all of the calls in our contact center and the availability of ABN AMRO GPT for all of our employees. And if you sum all of that up, then so we’re doing the broader in-depth assessment at the bank. But if I sum all of it up right now and I look at what we’re paying per month, then I can kind of estimate the maximum amount of energy that underlies that. And that’s actually a very small number. And so if a few people decide not to go to the car, to the office, I’m saving way more CO2 than all of it that we’re using for these use cases. [Jeroen]
Really interesting, because it’s kind of counterintuitive, right? I mean, I think there’s probably a lot of your colleagues actually think that you’re already using a lot of energy, which is not the case then. [Dennis]
Yeah, well, today it’s not. And this is by proxy, but yeah, this could change. Like it’s very easy with IT to say like, oh, you know, I’m going to do all of these additional checks and suddenly have 10 times the consumption. And then those 10 cars become 100 cars. And if we do that times 10 again, then it doesn’t fit in the parking garage anymore. And so it can become significant if you don’t watch what they’re doing. And so I think it’s definitely something that we take into our assessment framework.
Like, hey, are we making the right choices here? If I look at it today, it’s still a very small number.
[Jeroen]It’s very interesting. I mean, I love the fact, it’s just very fact-based at the moment. Any other thoughts here in the room? [Dafina]
Maybe if I could add here, I think it always also depends on the energy that is generated from which type of source it is generated. [Dennis]
So let me go there. We’re using the models in France. France has predominantly nuclear energy. So the CO2 equivalent impact of a kilowatt hour in France is like 10 times less than it would be in, let’s say, a coal-heavy power generation area. [Dafina]
Exactly. So I think in Netherlands also predominantly we are using renewable energies. I think in Europe we are moving towards the usage of more renewable energy. However, if we take this in China, where predominantly the energy is produced using coal and in Russia, in gas, then you are having huge CO2 emissions. And these are the countries, actually, that the emissions are growing still significantly within the energy. And yeah, knowing where the data centers are, it’s quite important where you’re making the Gen AI and making sure that you’re still contributing to those transformation in those countries. [Jeroen]
But somehow you measure within your bank, do you know, do you measure how much you use for the AIs you’re using? [Dafina]
Yeah, I think there are measurements like one hour of Teams call, if I would ask you, what do you think it is emitting in CO2? [Jeroen]
I hope you’re not asking me, I have no idea. [Dafina]
Yeah, well, these are the new measurements that we actually have to get accustomed to, like, what is the emissions? [Jeroen]
So you know the answer, actually. [Dafina]
Yeah, it’s one kilogram of CO2. [Jeroen]
One kilogram, an hour of Teams call. [Dafina]
Yes, if it is in the most polluting sources, and almost zero if it is from renewable energy. [Jeroen]
Yeah. [Dafina]
So that’s the difference. [Dennis]
I used to have this older laptop that wouldn’t handle Teams very well, and like, I could heat my house with that thing if I was doing a Teams call. Now they gave me this really nice MacBook, and it has like this 20 hour battery life, and it doesn’t even break a sweat when I’m doing a Teams call. So like, I have like a strong suspicion that even the hardware you use for the call is going to be like a measurably impacting what’s happening. And actually, there you can see the reach of Microsoft. If they make Teams 10% more efficient, that’s going to be a huge dent just in power consumption on these type of things. [Jeroen]
Great example. Let’s not talk about all the minerals that were used to get your laptop there. [Adrian]
Yeah, I have maybe one more addition to make indeed. The intensity of the use of AI, I think the call center example that you mentioned, that’s something, yeah, that’s based on the number of calls per day, right? But there are also a couple of other activities within banks where you can actually make a decision of how often you run a certain algorithm.
So for example, if you use it for market research purposes, we could use it in MySpace and project finance or loan finance where you would want to know, you know, what are the trends? You can ask AI to screen all of the information that’s available on the market and provide you a report. Once in a while, these are the newest technologies, sustainable technology that we see in the market. These are upcoming customers and this is some growth rates and some economic information. And you could use that to then actually adjust your strategy. So you can run this report maybe once every day or once every minute, but it doesn’t add much value. So you can, based on your own capacity to actually consume all of the information, adjust the intensity of the algorithm you’re running.
[Dennis]So I think in a similar fashion, like the smaller models are, you know, a factor of 10 or a factor of 20 cheaper than the largest models, and they still tend to get very good quality outputs. And so picking a smaller model allows you to do, and so again, I’m proxying cost and the actual underlying energy impact here, but that means that if I can use a smaller model, I can do 10 times as much and still have fewer of less CO2 impact, for example, than using the larger model.
And so I think it’s easy for a developer to say like, oh, I’m just going to go for the highest quality, but so you’re going to get a 10% better answer at 10 times the impact. And so there’s always this trade off as well. And I think it’s in a similar way of thinking is like, am I using this thing in a kind of, ‘oh, I’m just going to run this every minute with the best model I can get.
It’s not that expensive anyhow,’ or is this more of like, hey, at scale, this is actually going to matter. And that way I should start thinking about using smaller models if they work well enough. And I should start thinking about, ‘is this okay on a daily, weekly or monthly basis.’
And so yeah, in the same for us, we use one of the things that we’re looking at in terms of impact at the ABN AMRO or GPT side is, so you get the conversation that you’re having, but we also have guardrails that are looking at that conversation saying, hey, is this appropriate usage? And that’s also GPT. So for every call that you make, we make another two or three for compliance.
Well, we get to the dilemmas as well. Let’s switch gears to how you use, could you give me any example that comes to mind around how you use data and AI for sustainability purposes? Whoever likes to talk. [Dennis]
I think Dafina gave the great example of we’re doing the same. [Dennis]
So what we have is, you get a lot of unstructured data from your clients. And that happens in the ESG domain. [Jeroen]
Unstructured data, you mean, for example, annual accounts or text statements. [Dennis]
PDFs, text statements, emails, et cetera. And for all the sustainability reporting, that’s even worse than for the other reporting because it’s still kind of an emerging area in that sense. And so to be able to take that and turn it into something that we can put in an Excel sheet, so to say, that’s a very powerful thing when it comes to reporting from a regulatory perspective, but also in general understanding kind of our scope to three emissions and these things. So I think that’s one of the interesting areas that we’re exploring just as well. [Jeroen]
You’re speaking quite fast, right? I’m just making sure that I understand it. So you get all these data from customers, you need data to report, and then you use AI to get that data you need out of all that data. Is that right? [Dennis]
So let me simplify. I get from my client a report on their sustainability behavior. But it’s going to be a written report with an introduction from the CEO and a bunch of other things. But what I want is get a sense of, okay, so what specific activities is this client doing? What is the linked CO2 emissions if it’s on a CO2 level, but also let’s say human rights and other activities. Ideally, if I do this across, let’s say a thousand companies, I can put this in a table where I have the same fields for all of them. The thing is there’s not really a standard in which they will go and write all of it like this is the table we use. For financial data, at least we have kind ofIFRS type of agreements on how you do this. And AI is really good at both extracting insights from the unstructured information you get. So just a written long form text, and then also classifying them to like, hey, I would like to have it in this format. And that’s what we’re exploring. [Jeroen]
You’re exploring it or using it already? [Dennis]
Exploring it. [Dafina]
Yeah. So we’re also exploring it yet. This is exactly what we were trying to achieve to see if we would go on the general, yeah, basically on the general web and we go through and all potential, then we could actually extract information. If I would ask you, do you know how balance sheet looks like? You would always imagine it has assets and liabilities and equity, right? If you’re a financial professional, you would know. But if I would ask you, so do you know how climate report looks like from a client? Well, I think if you have seen one, you would see it’s quite messy.
It’s quite a lot of information, a lot of graphs, a lot of different types of information. So that’s the unstructured part. And that’s what we were talking about, having this text that you need to transform into a table, into an information that you can use for further information. We have not yet used it for our external reporting, but what we would like to actually explore if we can use it through the process of having the conversation with the clients. So imagine if you have the data you have extracted from a client and you could compare it and say, what are your activities, where you need finance to improve on your scope three emissions? Or how can you, how can I help you improve your circularity or what kind of practice is useful?
[Jeroen]Could you just tell me scope three, for people that don’t know? [Dafina]
Those are indirect emissions that the clients make, that the clients have from operations within the value chain. So for example, usually we finance clients. So we are not directly responsible for the, or we don’t have direct emissions from the clients, but because we are financing the client, we are influencing and we are taking on board their emissions as well. What additional type, employees are traveling by planes, the emissions from their travel, from their business travel is also scope three emissions. So that’s how you can imagine scope three emissions. So having a conversation with a client is quite important for us as banks to basically help the clients transition to better practices, to better operations that are with lower carbon footprint, or that they are helping in reusing of materials.
So circularity or better water usage, better resource usage. And so those type of activities, if we can help them, will be a benefit for both parties. But sometimes we need to go through understanding those metrics, those information, and how wonderful it would be if we can have it structured and we can immediately approach a client already proactively in those situations.
[Jeroen]So we have seen CSRD coming in. I’m sure you’ve been busy with that. So first of all, things are going well in the implementation. I think January 2025 is the first time where you need to report, right? [Dafina]
Yes, financial institutions are quite busy right now. I think we have, we are, as Rabobank, we are well on the way. And I think the work around financed emissions have tremendously helped us over the last three years. I’m very proud to be the one leading the group that works on financed emissions with a very, very long list of colleagues that are helping out in this process. And yes, CSRD is a challenge in itself, but at the same time, it’s an opportunity for the financial institutions to understand where they are in the process of integrating sustainability in their own operations. And at the same time, helps the Europe in future, and this is a European regulation, helps Europe reach their goals with regards to sustainability. [Jeroen]
Is the major challenge, at least that’s what I hear when I talk to people, a data challenge, or is that not fair to say? [Dafina]
Oh, absolutely. It is a data challenge. [Jeroen]
And why is that? [Dafina]
I deal with this every day, I would say. We as financial institutions, we want to be precise, accurate. We want to be consistent. We want to be complete. And it’s nice when we talk about euros and dollars, everybody knows what does that mean. But when we talk about one kilogram CO2 emissions, or one kilogram of methane, what does that mean? So how much of my operations means this amount of CO2 emissions? And not everybody has that picture, not everybody reports that information. And it’s quite a new information.
So not knowing, not having complete information from our clients, and also being uncomfortable with this information is a stretch for the financial professionals. So I always say, well, right now we, at least in Rabobank, we have for majority of our wholesale businesses around 65% information directly coming from the clients, which is tremendous amount, at least until last year, and I’m hoping this year we will have even more.
[Jeroen]So what kind of information? [Dafina]
Clients report their own scope one and two information, right? [Jeroen]
And then 65% of them are now delivering that to you? [Dafina]
65% of our exposure towards them. It’s 30% of our clients. And I’m hoping that it will be, yeah, but the bigger clients understand, and I think CSRD will help us have more and more information available for us to actually have meaningful conversation with our clients.
But then how do we value or how do we measure the farmers in Australia or the farmers in Brazil? And not only on their scope one and two and three, but also their water usage, their energy. So there is a lot of additional information that we need from our clients in order to have a good perspective.
[Jeroen]Yeah, and I can imagine in environmental issues, it’s relatively easy and super hard, but emissions at least is a number where, you know, Adrian mentioned the more the social side of things, but even harder to measure. [Dafina]
Yes, we need to talk about workers in the value chain, just transition for inclusion in the workforce. So those type of information will not always be very easily measurable. And that is the confrontation for financial professionals, right? Having not this being uncomfortable with estimation or being uncomfortable with a qualitative disclosure, it’s something that we need to bridge and get more comfortable. And at the same time, I say, if we wait for the data to be complete and accurate, then we are in 2050. And then our net zero ambitions are gone. And we are two degrees… [Jeroen]
Yeah, having all the data doesn’t mean you’re taking decisions not to finance something, right? And you guys have been very clear on what you do and what you don’t do, right? So actually, do you need a lot of technology to get all data or is it just not lending to those businesses, for example? [Adrian]
Yeah, it is valuable to use technology still, I would say, to make better decisions and sustainability-related issues. But I think the main position that I would take in this is that you will have to have the human taking final decision. So in the end, what AI can do really well is prepare a decision.
It can actually provide you all kinds of information, take something which is unstructured, put it in a structured format, give you all the arguments that you might need. But as a professional, as a sustainable banker, you actually have to look at all of that information and decide what is most significant. In the end, it’s a judgment call, I believe. You need to make that judgment. What is important in terms of your own values, in terms of the values of the bank, in terms of the values that the bank projects towards stakeholders? What am I actually focusing on in a specific situation?
And I think that that remains something that you cannot really delegate to a machine or that you cannot really automate. You will have to look at the data yourself and make that judgment. Is this specific decision relating to sustainability actually aligned with my values?
[Jeroen]Don’t you think it also has to do with the fact that it’s ultimately a forward-looking decision that you need to make versus analyzing data that is, by definition, from the past, even if it’s a minute ago? [Adrian]
So what all of these regulations do is they try to make things more measurable on a macro level. So at the macro level, you hope to see the right movement of the market towards more sustainable actions overall, more sustainable decisions overall. But the thing about sustainability, is that often it happens at the micro level. So for example, if you try to decide and make a good decision, whether or not building a solar farm on agricultural land is a good decision. The CO2 box you will tick immediately, right? It will avoid CO2 emissions. So that’s fine. But you have to think about social issues indeed. Where have the solar panels been produced? And is this according to our social standards? You have to look at the specific location of the site. How does this solar park compete with agricultural land? So when you add these additional layers, the social part, the biodiversity part, you actually get a more interesting decision. But more importantly, you really need to use your judgment. And the computer will provide you with some of the information, but it will not make the judgment call for you. You shouldn’t let the computer. [Jeroen]
So it’s, yeah, it’s a mean, not the actual goal. You use it to prepare decision-making. You want to keep the human in the loop. Kind of my summary of what you’re saying. [Dafina]
And if I can add, for CSRD, the biggest challenge is that you need to provide forward-looking statements. So you can use data till the extent that you can for the past and maybe look at the past and think where I have been in the previous year, but where I want to be, it might be different outcome. And it is quite uncomfortable at the board level to have those discussions at this moment of time, not knowing and not having sufficient amount of data from the past as well. We have been faced with estimating our scope three for the last two, three years. It’s not like 10-year model of credit behavior of our clients is just three years. And most of them are estimates, averages from sector information. And that is quite uncomfortable to know. So where would we be, where do we want to be? How far can we be? And what does it mean if we don’t reach it? [Jeroen]
Well, and as you said, there’s not much time, right? [Dafina]
Exactly. [Jeroen]
So it makes it hard for you as well, I guess. [Dafina]
Exactly. Yeah. I think for every bank and for every company… [Jeroen]
If you want to go fast, you work for a really big bank. I mean, you know, not necessarily Rabobank, but I think it’s for every large organization is hard to change and to change fast. [Dafina]
Exactly. And also at the same time being, being too cautious playing the safe game or too ambitious which might actually, if not achieved, come back with a knocks on the door from the Friends of Earth or Greenpeace. So how to balance exactly what you can do and what you cannot do at this point of time for the future. [Jeroen]
Ultimately, it’s leadership. That’s why it’s called leaders in finance. Dennis, let’s see if we can find some disagreement here, because maybe not necessarily focused on sustainable finance, but more generally speaking on the use of AI and to Adrian’s point, do you think it’s always a human in a loop or can you see a lot of things that can actually be decided by AI? [Dennis]
I take a bit of the contrary stance there. [Jeroen]
That’s what I’m kind of looking for. [Dennis]
I think for material decisions, we’ll often have human in the loop. And I think in the end, you get to kind of three potential outcomes. There is decisions that are completely immaterial at the scale that we made. And so, and you might argue it’s a decision or not. So one of the things with AI, since I’m summarizing calls anyhow, I can also determine what people are calling about and if maybe there’s something wrong on the website that we can fix. And I can have AI extract that at scale and make a decision that, hey, we need to go and fix this page on the homepage. Does that need human in the loop to send that out to the e-commerce team? Not really, at least my opinion. And then there’s some stuff like, ‘hey, you know, this client could be involved in money laundering and we should consider, you know, some very stern conversations or potentially an exit.’ Does that require a human in the loop? I would say yes. And I think it’s going to require a human in the loop until the end of days. And then there’s a whole piece in the middle where the examples from insurance, let’s say I have some AI that determines whether I should pay out your insurance claim or not. I can choose to only do it in one of two situations. I can basically say, look, if the model says I should pay, I will just pay. And if the model says I should not pay, then I’ll have human review. And so I could basically tell people that regardless of the fact that I’m making 80% of my decisions autonomously, you still have the best outcome. And so I can have, partial human in the loop and everybody wins. So I think there’s a nuance on how material a decision is and kind of, you know, up and downsides for both in this case, the bank or the insurance company and the client, and they use that to determine this. And I think that is a bit of looking at, you know, today and today plus a few years or plus five years. I wonder kind of depending on how steep the AI development curve remains, I will look at this in five years. And because at some point we’re going to have the discussion, what if structurally the AI is able to take better decisions, like significantly better decisions than the humans. [Jeroen]
And also whether you’re able to reproduce the decision, because if you throw in all the data today and you throw it in tomorrow, it may be a different outcome. It’s not always easy to reproduce a decision using AI, right? [Dennis]
So there are some things where you’d say, like, I really should be able to explain this. And I think there’s other, I think what you’re learning is the ethical side of it as well. Like, so I think at some point, even if we have near perfect AI and the decision making from the AI, even in what it would come to very high risk things like credit proposals and these stuff. Imagine you get to a world 10 years from now where the AI is always giving you a better answer than the human. We might still want to choose on some of these areas to not have AI make those decisions because we say like, look, from a, from an ethical perspective, we think it’s important that there is, you know, a human system that looks at it as well. But I think in the short term, I kind of stand where I stand. It’s like immaterial stuff, maybe it’s okay. And then there is a whole bunch of decisions where you could say like, I can structure it in such a way that everybody is better off in the system, even if some of the decisions are made autonomously. And then there is always going to be a bucket of decisions where we say, we need to have somebody that we could look at and say, like, how did you come to this decision? [Voice-over]
You’re listening to Leaders in Finance with Jeroen Broekema. [Jeroen]
I’m slowly going towards the end of this episode. I think time is always flying, especially here, because I’m really enjoying this topic. It’s super interesting learning a lot of things. But maybe we can go around the table and talk about what you ultimately think is your biggest concern and your biggest opportunity you see when you combine these two major themes of this episode and also the FLAIRS event coming up around sustainability and data slash AI slash tech. I mean, who wants to take the mic, then please go. [Dennis]
I’ll start with a concern, but it’s not necessarily for banking. I’m just back from my vacation and we went to Denmark and we had amazing weather. And I had this conversation with someone who was like, oh, it’s great that we didn’t go to Greece because it’s like 40 degrees. And it has been for the last few years. And then suddenly it hit me. It was like, okay, so everybody’s saying well this whole climate change thing, we don’t really see it. But then, you know, vacation homes in Greece have basically become unsellable. And people are looking for cooler countries to go on vacation because it’s it’s become unpleasant to go to the places I would go when I was a kid. And so it is really happening. And I don’t think that as a society, we’re on track at all to do something about this. So there’s this big concern that even if we have all the reporting in place, this is all just nibbles at the much bigger problem. And I don’t see anyone really tackling at scale, this bigger problem. And maybe it’s because it needs to be solved by, you know, tens of thousands of small initiatives. But it worries me. [Jeroen]
Yeah, well put, I don’t think anyone disagrees here at this table. I’m pretty sure that you guys don’t and I don’t for sure. And yeah, Denmark being the Mediterranean, it’s crazy. [Dafina]
Yeah, or the Netherlands. I was in like mid-August to the Noordwijk aan Zee, just exactly like being in warm water, no wind, perfect day on the beach. My biggest concern is that we have this responsibility right now. And we are doing this. And we’re always thinking from the past and from our own experience and where we are right now. But I have two kids and my son is now 10. But when he was seven, with his friends, he wrote a book. What he wrote was a book about how to save the planet. So he, with seven, is worried about the planet. I was talking about ecology at that time when I was maybe 18. If I take that in perspective, our future generations are worried. And the future generations, we give a lot of responsibility of things that will come to them. And they are the ones who actually currently should hold us responsible. This is the generation of leaders that we have in all the organizations that need to make steps forward. And it’s a concern that whether we will balance this short-term view of the next four years or the next eight years until my retirement versus what the future generations would need. So that’s one thing that I’m thinking about. And the other things I’m always thinking about, we talk a lot about regulation and reporting, I think that we are, as financial institutions, we sit in the middle of the economy. And if we can always think that environmental issues are not just new issues, but they are financial issues. These are going to have economic costs for the banks, for the clients. We really need to think ahead and think better for how to make our organization’s future fit. So this is the time to have that leadership on the table and not think, ah, this is just sustainability. So how to really embed it into that perspective. I loved what you were talking about, balancing between the price and the efficiencies and impact that can have. So how to balance all those perspectives. That’s where the opportunity is. And I think that everybody, if it starts to think that, not just how much it costs, but also, yeah, what is the impact out of it? What is the all the causes that it has: where our food comes from, who made it, who picked it up? How much pesticides was used?Is it healthy? Is it enough? You know, so that kind of thinking, if we can embed in most of our decisions, this will be something that I’m thinking that’s an opportunity to give us the next steps. [Adrian]
Yeah, I very much agree with everything that has been said, Rafina and Dennis about opportunities and concerns. Maybe one to add: I think the great opportunity of AI is that you can actually use the time that you have or the time and the thinking capacity that you have as humans, and the values that you add to that and the human dimension, the feelings, you know, the emotions that you have as professionals in the end as well. You can spend that time on value-added activities while automating things that a computer can do. So you can actually spend time building relationships with clients. In my specific case, you can spend time with clients working on sustainable business cases, brainstorming how you can structure financing in a specific way so you can make that innovative project possible. So that’s my main point actually, that you can outsource a lot of work that can now be automated and spend the time on the things that really add the value. [Jeroen]
That could have been a great end of this episode, but I have one more thing I want to discuss and also to link back to my short chat with Fleur earlier this episode on the FLAIRS event. Because as you know, this is for young bankers, this event, and you are in different stages of your career. I mean, Adrian, you’re more closer to the young bankers than probably you are, I don’t want to insult you, but I think it’s quite factual. [Dafina]
And myself- We cannot join FLAIRS, that’s all right. [Jeroen]
No, okay, that’s a very factual. But my question really is related to sustainability or AI or both, do you have any tips for the young banker, whatever bank he is or she is at the moment, do you have any thoughts or tips you have for him or her? [Dennis]
There’s this really big opportunity happening in terms of using AI to automate very significant parts of our work and so that allows us to spend time in other places. And I think if you look at the younger generation come in, they’re much more digitally native than the older generation that’s been around in banks for 20 plus years. And so they have the opportunity to really use that to the best ability. And they’re much more natively grown up with that than people that are basically saying, oh, I need to do a chat GPT course to see how I get most out of that. And so that’s where you could- If you’re a young banker, that’s where you can lean into, right? Use the advantage you have of growing up with this type of technology and use that for your work. And so there’s always this joke, like you won’t be replaced by AI, you will be replaced by somebody using AI. So be that guy, right? Be the person using AI. [Jeroen]
Well put, great answer. [Dafina]
Yeah, and if I could say, I would say always use your voice and hold us accountable, I would say. Make sure that you know your purpose, your value, what your drive is and use it with your own voice. Don’t look up at authorities and hierarchies. Make sure that you are a leader or yourself or try to be a leader in whatever you do. Yeah, maybe it’s too bold in that sense, but I have seen small examples of future boards being introduced, the voice of the future generation being part of decision-making in the financial institutions. Scale that up, talk about it more and yeah, hold us accountable. [Jeroen]
Wonderful, great one as well. [Adrian]
Yeah, I would say use the technology, but stay authentic and true to yourself. Think about what your values are indeed and how you think about it personally. [Jeroen]
Wow, great statement. Very, very clear. I wanna thank you. I was planning on doing a podcast for 40, maybe ultimately maximum 50 minutes, but I didn’t, which is a really good sign in a way because I really enjoyed it. It’s very interesting. There’s a lot of things in this podcast that may inspire someone or makes people think differently. So thanks a lot. I also think, I hope at least, I’m looking at Fleur now, that it’s a nice start or run-up episode towards the FLAIRS event on the 27th of September, which is probably already sold out, but I don’t know. If not, then try to get a ticket and go there. Fleur, thanks for introducing FLAIRS. Dafina, thanks for being here. And Dennis and Adrian also, thanks a lot for taking the time to speak to Leaders in Finance, to speak to me. And to make sure that it’s not a thank you just by saying thank you I’ve also brought a small present to thank you after we have finished this podcast. So once again, thanks for your time. [Adrian]
Thanks. [Dennis]
Thank you, everyone. [Dafina]
Thanks. [Voice-over]
You’ve been listening to Leaders in Finance. We hope you’ve enjoyed the episode and would love to hear from you. What’s on your mind?
Who would you like to hear next? Tell us in an Apple or Google review, via email or our social media channels. We’d greatly appreciate it.
Finally, we’d like to thank our partners for their ongoing support. They are EY, Zanders, MeDirect, Kayak and Roland Berger. Thank you for listening.