In this episode of the Connector Podcast recorded after the DFS Digital Finance Summit, host Koen van der Hooijdonk interviews Karel Lodewijk Lefere, CEO of NABU (nabutech.ai), about the company’s approach to modernising KYC and compliance for the regulated financial sector. Karel explains that traditional onboarding is a manual, data‑heavy process that hampers both fintechs and their customers. NABU’s mission is to shift from "onboarding" to "no‑boarding" by leveraging technology, user‑centred UX, and AI to automate data collection, verification, and ongoing monitoring. The conversation covers the regulatory pressures specific to payments (AML, CTF, UBO registers, sanctions), the need for multi‑country API orchestration, and the distinction between structured data (public registers) and unstructured data (social media, street‑view images). AI is used to analyse unstructured sources, flag inconsistencies, and provide confidence scores. Karel also discusses how NABU balances compliance with a seamless user experience, the broader market focus beyond payments (private equity, B2B banking), and the company’s fintech‑for‑good ethos of reducing friction to make financial services more accessible. The episode ends with contact details and an invitation for “shadow sessions” where NABU reviews client processes under NDA.
Chapters
Welcome to the Connector Podcast, an ongoing conversation connecting fintechs, banks, and regulators worldwide. Join CEO and founder Koen van der Hooijdonk as you learn more about the latest available trends and solutions in the markets.
Welcome to another podcast here at DFS, not live, but just recorded just a few days after DFS. And today I've got with me Karel Lodewijk. Karel Lodewijk, who are you and for which company do you work?
Hi, Quinn. Good morning. Nice to be in this podcast. So I'm Karel Lodewijk Lefere. I am the CEO of Nabu. And Nabu is a company specialized in supporting KYC and compliance processes for the financial industry and the regulated industry as a whole.
Oh, that's a very good start because that means we can really deep dive a little bit what you guys do more in specific. So Could you explain a little bit deeper what exactly you guys do?
Yeah, absolutely. So within the regulated industry and the compliance industry, there's a lot of, as it says, regulation on how you should evaluate and monitor your customers. If you are, especially within the B2B segment, if you are a payment provider, a bank, or an insurance company, you have to check all the customers before they become your customer because you will be in a sort of a money flow. So from the AML perspective, which is anti- money laundering, but also counterterrorism financing. Especially from those 2 perspectives, you have to check and really be sure that you know who your customer is. So, the know your customer, KYC. Within that space, up to now, it was a very, very manual and cumbersome process for, I mean, for 99% of all the regulated industry. We as NABU, we have the mission to really support those customers, or so those fintechs, in becoming much more digital and especially providing value to both themselves by being faster, cheaper, and so on, but also value to their customers by making this a much less hassle or cumbersome process to onboard.
You made a big statement on your website because there you say from onboarding to no-boarding. And that really struck my eye. So what does that mean exactly?
So, I mean, which is commonly known as onboarding in the industry, means onboarding a new customer. So if you want to become a customer of a payments company or a B2B customer from a bank, you have to onboard to them. As I said earlier, currently this is a very cumbersome process. It takes It takes a lot of time. You need to give them an amazing amount of data. The joke goes that you actually almost have to give them the color of your underwear before you can become a customer of them. I mean, so, but it's— we want to move this, which is commonly known as onboarding, to no-boarding, in which we mean that it's both for the customer and for the company themselves, it actually becomes a very seamless and very fluid process to become a customer of them. Because, I mean, the problem is not only because it's a hassle for the customers, they don't do this intentionally. I mean, the fintech companies, it's not by intent that they make your life miserable. Definitely not. But it's just a very, very big flow and a lot of data that needs to be acquired. So, our mission is to go from onboarding to no-boarding by applying technology, UX, and a lot of other components to make this really better.
Yeah. Now, 2 questions with this. First of all, maybe a little bit more deep dive, because you said a few times that you're specialized in the payment industry. So what, what makes you so special? And what are the jack of trades that you provide on the payment side? So I guess we talk merchants, we talk acquirers, we talk payment facilitators. So what is so special about your offering?
Well, the payment industry is one of the regulated industries. Both my and my co-founders, our history and our experience is within the payments industry. So that's why we start with the payments industry. But I want to be clear, we don't only focus merely on payment industry.
Okay.
The private equity market is another one, and the B2B banking market also is another one. However, The problem for the payment industry is enormous. I mean, they have a very high amount, a very large amount of merchants, just like banks, just like B2B banks. And they need to even gather more information than other companies need to do. Why is this? Because they actually pay out the money. So, if you're in the transactional flow, an acquirer receives the money, and overnight, they pay out the money to the merchants. This means you need to know where the money goes to. That's where the regulator sits. So your national bank, which could be NBB for Belgium or DNB for Netherlands, they need to be assured that the process you installed to check and validate who is your customer when you onboard him, but also during the lifecycle of the merchant. So continuously, you need to be sure that where the money goes to. To make like an example, if a customer of yours is on a Russian sanction list, You need to know this because you cannot give money to somebody who is on the Russian sanction list, and that person could be the owner of the company who owns your customer. So the UBO, as we call it, the ultimate beneficiary. So you need to know this, and especially in the payments because there's so much money going through that industry from the, as I said, anti-money laundry and counter-terrorism financing. You really need to know where the money goes. So this industry is particularly vetted. by its central authorities.
Yeah, so you have high regulations, but earlier you said that you use UX/UI. So that was my other question, because I'm always wondering, how do you strike the right balance then between UX/UI and regulatory obligations?
Yeah, I mean, on the UX side, it's pretty easy. That's just like, say, you make the things more fluid to work with, you make them more conversational. So you redesign the flows as they exist in a more user-friendly experience that is more intuitive to work with. The biggest balance that we need to strike is how you apply technology in this. The good thing that we see coming is that we're actually on the crossroads now of a couple of things colliding. One of them is European legislation is, although slowly, but it is unifying. For example, there are some new AMLD 6, so Directive 6 is coming in, which says that all the European countries need to have an open, accessible UBO register. Unfortunately, it's not going to be one API.
Finally.
Yeah, finally. But unfortunately, it's not going to be one API. It's going to be multiple APIs. So, the complexity will still be there. But where currently Belgium is already doing a really good job, Absolutely. Other countries will need to do the same. Second term there is, for example, the e-identity, where we have itsme in Belgium, very well known, works very well. Other countries, same again, will need to comply and have an electronic identity accessible via technology. So that's what we need. Voilà, exactly. Again, unfortunately, it's not going to be one standard, but it's going to be at least a step, a massive step forward. Second thing that we see is that a lot of this data is already accessible. It could be through company registers, through public registers of the— so the questions that are asked to a lot of merchants, actually, you can already find this data. This data is accessible.
Mm-hmm.
I mean, if you have a VAT number, we at NABU, if you give us a VAT number, we give you 90% of your profile. You don't need to fill it in anymore. We know who you are. We know where you live. We know your company. We know your website. We know everything about you. of your business. I mean, not of your person, of course. But, uh, I mean, but, and then some aspects we let you confirm, and that's where the compliance comes in. Like, sometimes you need to self-declare some stuff. We challenge that to a big extent. We say, why do you need to self-declare if we can find it in the official registers? UBOs is an example where you need to declare the UBOs, but we actually can extract UBOs in certain countries. So, What we do now, or what we propose, if the compliance officer agrees of the company, we make a suggestion. This is the UBOs we find, please confirm. If they say no, okay, we change the UX, we still do a self-declaration, and then AI compares it. Or that's not deep AI, of course, but the engine compares what you declared to what we actually find.
Mhm.
And then we make a report out of this, of the consistency of that data. So the biggest balance we have to strike is where you put in, where you where we put in the technology and to replace the manual work and how compliance goes about this.
I think I better understand now also API orchestration, what's also heavily on your website. I think that's like a combination of bringing things together. You also mentioned their deep KYC. What do you mean with that?
Yeah, so we have a KYC on multiple levels. One is the KYC, which is on structured data. That's where you just extract data. That's the API orchestration you refer to is a very good point. We extract data from multiple registers. That can be from an UBO register, from your identity register, that can be from your company registers, and so on. That's pretty much plain data, public data structure. We refer to it as structured data. The thing with an API orchestrator is I think for every country you need at least 3 or 4, and the usual payments company is active in minimum 5 countries. Usually, our recent projects are in 17 countries. Yeah, no surprise. So, the multiple of APIs for them, us providing, as I joke about it, like one API to rule them all. That's what we try to provide. They do one pipe into us. Now, in that pipe, you have the structured data, but also the deep KYC. And that's where unstructured data comes in. We, for example, we know by experience that the address on their socials usually is more accurate than the address in the registers. If they move the shop 2 streets forward or 2 numbers in the street to the left, they're going to first change it on their Facebook page. They're going to first change it on Insta, and then probably somewhere in the end of the year, they're going to give it to their accountant to change the registers. So we double-check. That's where we go deep. We double-check the data there. Like, is it consistent with what we found across the board? Same goes for activity. If you're in your register, you say, I'm still a cafe or a bar, but actually we find that you evolved. I mean, over time you evolved and you became a hotel, or your bed and breakfast became successful and you became a hotel. That's fine and that's excellent for you, but it's a different activity. So, in all the registers, you will be mentioned as a cafe, but we will say, no, this actually, we checked it out, it's a hotel.
Behavior tells differently.
Yeah. So, and again, we don't decide. We just give the analyst a report with a red flag or an orange flag saying, look, check this out because this doesn't seem to match. Please check.
Interesting. So you do a lot of checks, and you do that on non-structured data. Is that where AI is coming in?
Yeah. So the AI is especially on the unstructured side of the story. That's especially where it comes in. For example, another example that I'd like to give is a lot of checks happen if you have to onboard a shoe store, which is in a certain street. A lot of our partners now go to Google Street View and check it out if that's actually in that street. Like, does it exist? The reason for that is anti-money laundering, because a lot of VAT numbers don't actually exist. They just create revenue, and then that's the way you launder the money. They go and check. Our system can go fully automated. Our agents, it's called a verification agent, it just gives you the link, And we say, this is the picture we found. In the picture, we can read that it's a shoe store. This is the name. We think it's okay. And we give it a scoring of the accuracy.
Nice.
It's not 100%, of course, because sometimes it's very hard to see.
But it's better than not checked.
Yes, absolutely. So instead of going manual, they just— we give them the pictures, we describe the pictures, and we say what we see. But we also give them the link. If the operator says, look, I don't agree with this, this doesn't look to be like what I want it to be, He can still just click the link and he immediately goes to the right place. So that's where AI comes in, checking pictures, reading pictures, and so on. That's an example.
I have one more question before we actually go to the final one, which is basically asking you where to find you if you need— if people need you. What about fintech for good? Because that was a big theme on the conference. How do you position yourself against this Heavy-loaded term.
Yeah, it is a very heavy-loaded term. And I mean, I'm very wary of using it too lightly. So I won't say that we're an extreme fintech for good. However, what I really think is a difference we can make in the industry is this onboarding mess with a lot of companies. I mean, again, it's not intentional, but it's there. This makes Makes it very hard for a lot of people to enter into services. Getting a bank account, getting a payment terminal, getting a certain service is just a hassle if you're a little bit digital illiterate or you struggle digitally or you want to switch services. The lockup and the emotional hurdle of switching from one provider to another one and thus getting a better rate, getting a better service, getting something that fits you better is just exceptional because of this onboarding mess.
Yeah.
In a lot of cases. We think that we can democratize, if I can say, or open up a lot of financial services for a lot more people if you would just make it much more easier to enter into services. So I think that's especially where we fit in. You can make the service cheaper, you can get the service faster, but especially it's much more accessible for much more people.
So to make the impossible possible, how do they contact you? Where do they find you?
Oh, I mean, you can go to nabutech.ai or you can find us on LinkedIn. I know, or nabutech.eu, sorry, and on LinkedIn, of course. But I mean, just reach out to us. We can definitely— what we first always do, and that's what we like to do, is we do a shadow session with the companies we come in. Shadow sessions mean we just follow with you for a couple of hours your different processes. Again, we give under NDA, of course, but we give to the people the assurance like we don't judge. We never— we're not cynical. It's a very hard project and it's a very hard process at many of our customers, actually all of them. So we just follow along and that's where we just make a first advice on where we could do better.
Well, thank you very much, Karel Lodewijk, for your participation today. Thank you also to the audience and please stay tuned as more news of fintech will come to you. Thank you very much.
Thank you very much. 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.
