The Connector.
The Connector.
The Connector Podcast - DFS Digital Finance Summit - AI-Powered Debt Collection, Explained
We explore how AI and behavioural science can make debt collection more humane, more effective, and easier to scale across Europe. PAIR Finance explains reinforcement learning, a single international platform, and why recovery and retention go hand in hand.
• AI-driven collections with behavioural science
• Reinforcement learning optimising tone, timing, channels
• One platform operating in 11+ European markets
• Compliance with EU AI rules and local licenses
• Recovery rates improved while protecting retention
• Inbound automation with a generative AI agent
• Human support for complex and sensitive cases
• Event strategy for partnerships and education
• Unified dashboards and transparency for clients
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
Thank you for tuning into our podcast about global trends in the FinTech industry.
Check out our podcast channel.
Learn more about The Connector.
Follow us on LinkedIn.
Cheers
Koen Vanderhoydonk
koen.vanderhoydonk@jointheconnector.com
#FinTech #RegTech #Scaleup #WealthTech
Welcome to the Connector Podcast, an ongoing conversation connecting fintechs, banks, and regulators worldwide. Join CEO and founder Cohen van der Hoydong as you learn more about the latest available trends and solutions in the market.
SPEAKER_03:Can you please introduce yourself and also your company?
SPEAKER_01:Yes, so I'm Zoe Van Eseldeek, Business Development Manager at Pair Finance in the Binelux team. I've been working here for two years.
SPEAKER_02:Yeah, I'm a Zoe's colleague, and uh I've been also working for two years at uh Pair Finance as a senior business business development manager, and um also happy to be here. And what does the company do?
SPEAKER_01:So Pair Finance is a fintech that specializes in AI-driven debt collection services. I can tell you a little bit more about it. Julius, please. So we noticed that in today's digital world, the need for technology and consumer interaction to go hand in hand is very important, especially in the service industries. So to stay relevant, companies have to adapt to this and be more in line with what the digitally native consumers need. Um Pair Finance was founded to modernize an otherwise very traditional sector. So the debt collection industry still uses a lot of outbound calling and physical letters. This frustrates both the people and the brands that are involved. Um, so our founder, Stefan Sticker, recognized that the data-driven performance marketing was actually a way that could make debt collection much more empathetic and effective because we realized most debtors aren't insolvent. They're actually overwhelmed by poor communication and a lack of digital options. So we use AI in behavioral science to personalize every interaction with over 10 billion data points and 30,000 individual communication paths. So this results in yeah, a more digital customer-centric process with higher recovery rates up to 15% higher than traditional players.
SPEAKER_03:I'd like to zoom in a little bit on the behavioral part. Is that okay?
SPEAKER_01:Yeah, of course.
SPEAKER_03:What does that mean?
SPEAKER_01:So we have our in-house behavioral science team that creates basically uh not only do they study the debtors that we come across, but they also create little template snippets that is then used by the algorithm to approach every debtor in an individualized manner. So this is our behavioral science background, is our team that's doing research, constantly working on better ways to approach debtors.
SPEAKER_03:I also saw that you're active in in more than 11 countries in Europe. And um, what struck me in looking at your website is one international AI platform.
SPEAKER_02:Yeah.
SPEAKER_03:What does that mean?
SPEAKER_02:It means that in compared to uh or in contrary to the tradition traditional debt collection agencies, which often have different legacy systems from uh in different countries, we develop one platform that's scalable across uh all those countries. And um that's it's that's an advantage for international players because they have one platform, one contract, um, and therefore uh it's much easier for them to um work with uh one party because they have one dashboard, uh full transparency of the data, and that's unique for a debt collection agency in Europe.
SPEAKER_03:If you talk about debt collection in Europe, is there also regulation that is applicable for all the countries? And is that the same?
SPEAKER_02:Uh some regulations, like the EU Act on the AI, uh, that's uh for all European countries, but of course, there is also local legislation, and we adapt to that specifically, or for example, payment methods. So in Belgium, you have different preferences in payment methods than in the Netherlands, and we adapt the workflows according to the local methods that are applicable. Uh, also, for example, the regulation on the uh the RAIR letter in Belgium or the VIC letter in uh the Netherlands, we can adapt workflows according to the local practices.
SPEAKER_03:I also read like AI approaches like an adapt collection with 30,000 options and one goal, it's actually to get the money.
SPEAKER_02:Um, how do you use maybe maybe to to to intervene? Yes, of course, to get the money, but also to get in contact with the consumer because in the end it's still the client of our our our clients, so the the consumer, and we want to make the experience as best possible as much or the best possible uh uh for for the for the consumer, but also for the client. So, yes, recovery is important, but also retention is really important.
SPEAKER_03:No, I get that. I think user experience is key in everything you do. Exactly. And and that was also the sort of the question I wanted to go to is uh AI is nice, but how do you bring emotion to the equation? Which is a part that's today still lacking on the AI side.
SPEAKER_01:So, what we see is AI-driven solutions actually offer a lot of potential advantages. It allows us to bring this emotion, what you're talking about, by greater personalization, context aware communication, and also uh yeah, improved adaptability features that are especially relevant in moments of financial friction, which is quite uncomfortable for a lot of people, obviously. Um, so we're using an AI technology called reinforcement learning to do this. We see that uh in using this algorithm, we're able to have it constantly improving itself and it rewards itself whenever certain milestones are reached. So, for example, whenever we are able to uh get somebody to open a message or an email uh or even answer it, that is what we deem a reaction, a positive reaction. So the algorithm will continue to optimize based on different parameters, like the tonality of the message, uh, also the timing, the frequency, uh, the channels used to contact, and also the payment solution itself, which is then fully personalized to each debtor. So an example would be uh based on a certain data profile, we see a business person who we think will react in a certain way at a certain time. We'll send them a message on Monday morning at 8:30 a.m. when we think they'll check their emails. They'll be very happy. Um, rather than a student who will probably open the emails after they roll out of bed at like 11 a.m. or something like this.
SPEAKER_03:I can imagine.
SPEAKER_01:If they were anything like me as a student. So this will be uh a way that our AI algorithms able to um adapt its uh yeah, to each person, adapt the strategy.
SPEAKER_03:Well, I can imagine that traditional financial institutions they also need to do debt collection, but they're probably not as advanced as you guys. So that brings me to the point of outsourcing. They probably use you as a tool. So what would be a good reason to use you guys instead of doing it themselves?
SPEAKER_01:So just to make it clear, other debt collection agencies aren't using our software. We build our software. No, no, it's okay, because this is a question we get quite often. So we build our software and we're also the ones carrying out the debt uh debt collection activities. So let's say uh somebody buys a pair of pants from Zalando, they don't pay, Zalando sends them a few reminders, and they'll end up then they'll send us the claim and we will take over from Zalando to try and recover the money. So we start then with an amicable debt collection phase and try, based on what the algorithm decides, to contact this person in the best way possible.
SPEAKER_03:Okay, I'm learning, so thank you very much. So does that also mean that you are a regulated company?
SPEAKER_01:Yes, of course.
SPEAKER_03:And in what's in what regulation do you follow?
SPEAKER_01:So we have a debt collection license in all the countries in Europe that we're active in.
SPEAKER_03:And that's also a passportable license.
SPEAKER_02:But it's a it's a local uh license. So in every different country we have a regulation that we have to apply to. Uh uh, but uh this can be the debt collection uh legislation, but could it this can be AI legislation, it can be different kinds of legislation depending on uh if it applies to us. But of course, for example, the EU act on AI, we have to comply with that.
SPEAKER_03:Yeah, so that makes it super interesting because um the reason I was asking on passporting is that's a very common practice for banks or PSPs and all of that, but it's not applicable to you. So that I'm even more surprised and amazed that you have a one single platform, to be honest.
SPEAKER_01:Yeah, we see that a lot of traditional players technically they work in different countries, but they have a very fragmented approach because they acquire different systems every time they expand to a new country. So they're not able to have one platform for all of the countries. So their international clients then also get a bit frustrated because they're not using the same software everywhere, right? But since we built our own software, we have made it so that it's much easier for the clients, the the creditors themselves, to access not only the data they have about their consumers, um, but also the data that we've been collecting, the behavioral data that we've been collecting throughout the debt collection process. So they have one login essentially for all countries that they're active in, one contract, one contact person taking care of them.
SPEAKER_03:Well, what brought you actually to Digital Finance Summit?
SPEAKER_02:Well, we uh uh uh when we uh started in the Benelux uh two years ago, uh, we thought it could be uh a good idea to uh connect to other fintech players, and therefore we uh joined uh Belgium FinTech and Holland FinTech. And uh this was one event that they are hosting, so it was obvious that we would participate here.
SPEAKER_03:No, I agree. Uh, and uh I think there's that's definitely a part, maybe that's just me, but that's also an educational level because uh I have been learning a lot in this podcast, so I'm I'm very grateful for that. And I'm sorry that I didn't know.
SPEAKER_02:Yeah, every meeting uh uh uh uh especially this morning, for example, we we met a lot of uh different kinds of people in different areas that we are not aware of, and where we can benefit from uh in terms of uh knowledge, in terms of networks. So it was really, really, really nice already, and it's not even launched yet.
SPEAKER_03:Exactly. We only halfway, so it's only the good start, the good part are only to start. Exactly. Um, well, so you explained me that you need like uh companies come to you. Um, I guess you then do other consumers events as well, uh only fintech events.
SPEAKER_02:Yes, the the point only is that if you go to real consumer-oriented uh events, we went there last year in the Netherlands. Um the companies that want to work with us are often uh coming to special events like the Credit Expo because if they're looking for a solution like us, they will go to that event. And so for us, it makes more sense to go to these kinds of events than to the consumer event. Okay, but for in terms of development, it could be interesting to go there, but in terms of meeting potential clients, it's best to go to the expert uh event.
SPEAKER_03:No, no, I get it, I get it. Well, what I also learned today in the podcast is that you guys are like hyper um focused on on behavior, on AI, on all the latest technologies. So maybe a very sexy question to ask here. How do you see the the future changing in terms of technology applicable on your business?
SPEAKER_01:Well, we started the company in 2016 in Berlin. We have since then grown to a lot of different countries, and we've been using this AI technology since uh almost since the beginning, since 2018, we were able to roll this out for our bond communication. So for quite some years now, we've been working on improving our processes and our software, and our AI algorithm has been learning a lot and becoming smarter thanks to the reinforcement learning. So I think the the moon is the the objective, right? So the we can we can keep on going, keep on building our processes to a very exciting extent. And we see that especially in the debt collection world, which is very traditional. Some other parties are just starting with this now, but we're already almost 10 years ahead of them. So we continue to grow, continue to work on our processes, and from a technological perspective, we're starting to adapt AI a lot more into our internal processes as well. So not just in our core business, outbound communication, but also just to make our internal processes much smoother, faster. Um but then of course, with this uh inbound communication that we do thanks to generative AI, we're also able to continue automating some uh aspects of our um services uh in this sense. So anytime a debtor has a question for us, they will of course be able to contact our human uh customer service team if they wish to do so over the phone.
SPEAKER_03:As long as they are human still.
SPEAKER_01:No, no, they will be. They will be because there's always some aspects by which you need a human to answer the phone, right? Especially since with that collection, it's something that is quite personal and quite emotional for a lot of people. They want to be able to speak to someone if they call. However, if they're sending a message or an email, which is something that we get, as you can imagine, in very large volumes on a day-to-day basis, to better help them. We have created a fully automated AI agent which is able to respond to these recurring messages with very um, let's say, simple uh requests like asking for an installment uh plan or uh saying that they have a dispute on a claim, or um perhaps saying that there's a fraud case. These are questions or comments that we get from debtors constantly. So by having this AI agent that allows us to reduce the backlog, we're able to already scan 80% of the customer inquiries in this way and fully resolve 40%. And our goal for next year is to get this uh to our goal for next year is to get this to 80% by the end of 2026. So 80% fully resolved uh inbound messages by our AI agent. So this is a development step that we're working on.
SPEAKER_02:And the good thing is that our actual uh debt collection agencies that uh uh do the 20% have more time to help the people that have a real need to talk to a real person.
SPEAKER_03:I can imagine. Yeah. Well, we almost at the end of this podcast. Time flies when you're having fun, especially on my end, learning a lot. So I'm very grateful here. Um, how can people contact you?
SPEAKER_01:So you can contact us easily via LinkedIn, Zoe Van Eisoldek or Justin Pesch. Uh, you can also contact us via the website. We have a page for businesses who are in need of debt collection services. You can uh just send us an email and we'll reach out to you uh asking for some more information on what you're looking for.
SPEAKER_03:And we will make sure that we'll put the contact details below in the details of this podcast. So thank you very much for participating. Thank you also to the audience, and uh please stay tuned.
SPEAKER_00:More 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.