The Connector Podcast - DFS Digital Finance Summit -Inside Belgium’s Business Data: Compliance, R...

Welcome to the Connector Podcast, an ongoing conversation connecting fintechs, banks, and regulators worldwide. Join CEO and founder Koen Vanderhoydonk as you learn more about the latest available trends and solutions in the markets. Welcome to a new podcast from DFS, and today I have a very special guest with me, Toon van Acht. Toon van Acht wearing multiple hats. So explain us, Toon, who are you and what are your hats?

Well, I think I got invited here to speak about Daddle BE, which is my fintech company since 2012. But on the side of that, I also have the pleasure to be the chairman of the board of Fintech Belgium, our grassroots nonprofit that unites all the Belgian fintechs. And early this year, I also had the pleasure to I would say join up with great entrepreneurs to address the Peppol e-invoicing challenge that many companies still have to overcome. We're almost halfway in Belgium now with that new legal obligation. We're going to be talking compliance in Data.be afterwards, but with e-invoice.be, as you can see, I like good domain names. So we're going to switch to Data.be for the rest of this. But don't hesitate to also look at the e-invoice one.

Okay, great. So tell us a bit more about Data.be.

Yeah, absolutely. So Data.be basically fixes the who am I dealing with problem in business that is not only limited to banks, insurers, and fintechs, but is of course much wider related to knowing who your counterpart is. And so what we typically do, a user can look up a company, so typically a business entity. That's what we focus on, and typically a Belgian business entity. And we will actually structure and put together all the different Belgian official authentic sources. So this information coming from what we call the State Gazette, België Staatsblad, Moniteur Belge in Dutch and French, or the National Bank, which is publishing all the Belgian annual accounts. And so all this information is compiled together to know who can legally represent which company and in which state they are as a counterparty. So it's in one structured view from onboarding over credit decisions to counterpart checks to reduce risk and to be more compliant with all the regulation that those in the financial industry already have to adhere to. And more and more companies in the near future will have to adhere to.

Yeah. So we're talking about the famous KYB, know your business.

The B one. Yeah. So the know your customer is, I would say, the acronym at large. But as we focus on the business part of it, it is indeed KYB.

I'm kind of glad that you already mentioned regulations in here. And I know you're only focusing on the Belgian market, but can you tell us a little bit more about regulation versus the data that you actually provide?

Yeah. So if you look at regulation, there is a number of things that companies have to do to be more transparent. One of them is the famous UBO, the Ultimate Beneficiary Owner declarations that you have to make as a company owner. to know who is the ultimate owner behind a legal entity.

Mm-hmm.

This is one of the things to raise transparency in the economy as a whole. But of course, behind that is not just about transparency for the sake of transparency. It is to know who your counterparties really are and to avoid that you have scams or people doing tax dodging and other things that shouldn't be happening out there. So it's both used for governments, mainly for that, purpose, but also to know, like, okay, if you go for the very high scary stuff like anti-terror, looking like how is legal money being funneled to the economic system to fight against that, the UBO, ultimate beneficial owner registrations are an important element. And they're also, they have to be checked by the banks. And it's a pesky thing as a company owner, which I'm also, because you have to keep sending that several times a year to all your to all your counterparts that request it. And if you have several banks, then indeed it becomes several times a year that you might have to send stuff, the same stuff over again. And so that's only one of them. The most important one you have to see is under what we generally call compliance, different rules. And those rules are much stricter if you're active in the financial services industries where those counterparts really have to both check and challenge the fact that they have the right representative of a legal entity in front of them dealing, asking for a credit, opening a bank account, and then recertifying them every— they have to regularly check. And in the past, it was a very, I would say, retroactive and passive reactive. Now it's more proactive. So they would check every— depending on the policy and the country, you would have to check every 2, 3 years that your file was still up to date. Now it's becoming much more proactive. Our company also has a tool which sends out alerts on a customer base to say like, hey, there's a change here, there's a new official business address, or there's a new director, or hey, there's new financial numbers, you should maybe check them out because the company is in a worse state than it was when you onboarded them.

So you actually—

those are— yeah, go ahead.

No, go ahead.

So that's the basics of what we do is to make that process more seamless while we try to strike the right balance between financial regulation and the access to official company information regarding transparency and balancing that with privacy.

So you went from reactive to proactive to perpetual.

And I think what your name here indeed is That's the cycle that has happened over the past years. That in the past, it was clearly something that was done once every, depending on the policy or on the risk profile of the counterpart. That was already the advanced part, the ones that had a risk profiling, and then we're asking more questions to the ones and more regular questions to the ones with a higher risk profile. Today, we see that amongst all the parties that take this seriously, they do it perpetually and they want to be in sync. With the official data and sync it with their own databases. But it also involved, because more and more it is not just of getting the data from the official authentic government sources. And Belgium, in that respect, is a very open country. Many of those sources, they're open, but they're scattered and they're unstructured. And so that's the value we bring to make sure that we unite them and we structure them to make the manual checks less manual and the automation a lot faster and more seamless. But nevertheless, within that process, the banks and insurers often have to ask the counterpart to send more details, send more information, upload it, and especially they have to declare. So it's a self-declarative statement to say this is the current truth. And more and more, we are giving a score on the delta between what a company says is true and what the official registers claim is true.

True.

And you'd be surprised how often there is a delta. This is not— most of the cases, it is not fraud or it's not even badly intended. It is because the process to publish this information is quite cumbersome. And so you might appoint a new director who goes to the bank and says, okay, I'm the new director, I need access to the accounts. And then the bank will say, well, sorry, you're not published in the Crossroads database. So no, we can't do that. And then this person will say, yeah, but here's a decision of the shareholders. And they will most of the time say, okay, but we need to wait for that information to be in the official authentic government sources to grant you this access. Because we, for this one, we cannot just trust yourself decorative statements. We want to see that it's liable against third parties, such as a bank, that they have to adhere to that decision. And this publicly checked, able to be checked and challenged. So more and more, the whole process is not just about having data access through APIs or a web tool. It is more about the process where you're going to check a delta between what you have on file, what the customer is confirming towards you, and to make sure there's no delta between those 2. And when there is one, there's a reach out. And that's the part where, I mean, maybe we're going very fast in where this industry is going. But of course, this process of compliance is being looked at at all financial institutions to be more and more automated. And so you have AI agents that will typically try to reach out if a delta is identified and tell the customer, watch out, we've seen a delta here. You have published something, you have not notified us. Please put this in order. Please confirm. And typically those declarations, they have to be done with strong customer authentication or in some countries already through a wallet where you will upload the document and share it with your bank and insurer and fintech.

You mentioned other countries because in the world of KYB, there's a lot of difference in terms of continents, regions, and countries. So what is your view on Europe?

It's a bit of a mixed bag. Certain countries, for example, Italy, but also to a lesser extent, the Netherlands, you have to pay an official database. So in the Netherlands, it's called the KVK. If you want to get official extracts about companies, you have to make a payment to get a report. And then private companies like ours who do the structuring and who enrich that data and who will actually validate and normalize it, They also exist in those countries, but of course they have a much higher cost themselves to also procure this report and then to resell it to their customers. Some of them do what I would refer to as a recycling trick where they have one, if they have to buy it today for one bank and they can establish that there was no updates between yesterday and today, they can actually send yesterday's report that they bought for another bank at for today's request without having to pay the report again. So that's one thing you see in certain countries where you still have, like, I would say, a paid semi-government agency that is selling the authentic business data in PDFs. And those prices can be marked. So on the one hand, you have registers in Europe and countries where it's closed, but the general tendency is for companies to open this up and to get to this transparency, but also to make it very very clear for everybody. Like, if you're doing business, this information needs to be up to date and everybody can check it. So the more people can check it, the more people can also report a discrepancy between what's being relayed in the official sources and what they know to be the truth. So that's the other way it gets changed. The first way is you as a company owner, director, will make sure that the information is up to date, either by publishing it yourself, or if you're a bigger company, through a legal counsel, or if you're an even your company, you might have typically a Big 4 accounting consulting company that might do this sort of stuff for you to make sure that your international group information is published everywhere according to the local legislation. So the trend clearly is towards more openness and more public distribution. But I think the UBO database was a great example, and there was a 2022 An ambition? Yeah, I know there was a Court of Justice ruling through which the public access was restricted. It was very broad before in Belgium. It was public. Anybody could search for anything. Now it's pushed back to a role and a need-to-know basis that you need to show a legitimate interest before you actually will get access to the UPO database. And when you do checks in there, it's all being logged and audited. And you must be able to say, okay, I am A banker and I have a credit file. And in the application of the credit file, I informed my customer that we need to check their UBO. And so I'm now checking the UBO against the declaration they made towards me of their filing. So clearly, that was something that was in Belgium open and it has now been closed. So that's a little bit of a counter movement happening for some sources. But I think for the general business information, the trend for me is that it's going to be more and more open data, the high-value dataset Even under other European legislation, which is the open data type of frameworks where this is being pushed and promoted to make sure that this information, as it has a high value for the whole economy, to make sure it's open and free for the users to reuse it.

So that's clearly a journey to be held in Europe, in Belgium as well. I'm sure that by this, there's a lot of work also coming your way. We promised the audience also to talk a little bit about fintech. So let's maybe spend the last few minutes on getting a better understanding how you look at the fintech industry as we stand today out of your role of being in Fintech Belgium.

Yeah, well, I'm not sharing any secret in saying that the big theme— and we just had the Digital Finance Summit— clearly was agent e-commerce and more broader, the use of AI agents within your back office, but also in customer-facing interfaces. And so the big question is like, how are we going to make sure that technology providing agentic commerce is going to remain accountable? What about liability? What about the technical standards that are today emerging but are clearly not at all clear yet and not well defined? It's, I won't call it a little bit of Wild West because it's a negative term for it, but everybody's trying to figure out, okay, so now we are still pretty, most people are still in the discovery phase. Like you're using an LLM to like, okay, I'm going on holidays, make me propositions, and then it will come with, you can stay there and you can travel by those means with those transport providers. And then Some people already have browsers where they've linked their credit card information to it. And so you can say, okay, shall I book it for you? That's the next step. That's what agent e-commerce is about. Your agent will start negotiating with agents on the other side to make a transaction. But there's evident drawbacks, like what if it goes wrong? What if you were not happy with the outcome of the purchase made by your agent? Who's going to pay you back? And so everybody is looking a little bit like, okay, how do we solve this in a way that the agent does what you intended to do, but also that the one providing the service and who made also a certain cost even before providing it and delivering it is sure that not everything gets sent back or canceled upon arrival, meaning that you would have empty hotel rooms because people say, it was not me who booked, but I don't like this while I'm standing in front of the reception desk, so I'm going to cancel my night. Meaning the hotel not being able to sell the room at such late notice. And so I think that's just one very simple example, but you can figure out how complex this is and how our current system is not really built to deal with that.

So I guess the buzzword for the future is guardrails.

Guardrails is indeed one of the big things everybody is working on. But as this is a complex chain, from the information discovery, like how will your product be found? And also, let's be honest, if you look at financial products, the way they're being found by customers, how they're subscribed to, if you're going to ask your agent to say, okay, I want to— this is my risk profile, this is my, I would say, timeline on which I will need these savings, make me like a best-in-class mix of assets allocation of typical products I can get which are in my interest and not always in the provider's interest where you're currently maybe having those savings, that is really going to change the market. And so the big question is how quickly will end users use those, the advice they get through the different commercial chatbots? Will they use that advice to change how they allocate their own savings or credit sourcing? And so then the first question is like, how do those LLMs find them? How independent are they? Because it's not hard to imagine that just as in the old day with, and I'm just gonna name the incumbent Google, where you had a whole industry of doing SEO, you now already have the GTO for the generative text to make sure, okay, you want sure your company can be easily found by an LLM and that you're high up there as a trustworthy provider for this product. So there will be this race of making sure it's in everybody's interest. But at the same time, these LLMs and the people providing services under have another interest to make sure that they end up high in the negotiation process with your agent and to get your purchase, even if you're not doing it yourself manually after having done maybe a Google search like you would have done in the old days.

Yeah, exciting times ahead and maybe Taking us one step back, back to Data.be as we are closing, close to our closing of this podcast. If people want to talk to you about KYB and the product Data.be, how do they contact you, Toon?

Well, easy, through our website. You can first of all start even the discussion on our website, but you just reach out and we follow up depending if it's a standard. I mean, We are, we've been since 2012, we're a SaaS company. So it's clearly like, if you do not understand our product on the website, we've done something wrong. So please reach out because we want to improve how we do that. But it's self-service to say so you can onboard yourself. Of course, for banks who are selecting a new provider or a data stream provider to integrate with their own, because that's another trend we've seen happening. We've built KYB tools for all Belgian major banks over the past years. They have nicely paid us for that. But we also see a trend where more and more now they start to integrate this because it's a core process for a bank. It became a core process. And so they are insourcing more and more of that. And I would say the low-end data feeds, they also, I mean, the larger banks, they are capable now to integrate those themselves. So that's definitely something that happened over the past few years and becomes more and more about the outliers, more and more about edge cases, more and more about the famous 80/20 Pareto rules. Like in the past, it would be relatively easy to maybe automate 80% of your process. And that's one of the very big takeaways when you start automating also with agents that you have to keep a human in the loop, both for legal reasons, but also because in practice you cannot fully automate. all your processes. And so the hard part is not that you're going to have a new automated process. The hard part is it's always going to be a hybrid process with human oversight, with edge cases that cannot be fully automatically treated. And so to do that, banks today have looked at their backends and are looking at ways to split the workflows of the use cases they can do fully automatically and the ones where they have to manually intervene or assess, like, okay, this is something where we do need to step out ourselves as humans. and reach out or make an interpretation to see what is the correct way to handle a transaction.

Well, Toon, thank you very much for your contribution here at the podcast. Thank you also to the audience. And please stay tuned because we're going to reveal more amazing talks coming from DFS on this podcast channel soon. Thank you very much.

Thanks a lot, Koen. See you soon. Bye-bye.

Bye. Thanks for listening to another episode of the Connector Podcast. To connect and keep up to date with all the latest, head over to www.jointheconnector.com or hit subscribe via your podcast streaming platform.