In this episode of The Connector Podcast, host Koen Vanderhoydonk interviews Johannes, co‑founder of POM, to trace the company’s evolution from a consumer‑focused invoice‑payment app (launched in 2014) to a B2B payment‑enablement platform after a 2017 pivot. POM now equips invoices with QR codes, pay‑links, and SMS/WhatsApp reminders, delivering instant, fully reconciled payments for over 2,500 customers across utilities, government, healthcare, education, and small businesses. The conversation delves into POM’s integration with Mail2Pay, where a five‑year‑old machine‑learning model predicts the optimal timing and channel for invoice reminders based on demographic and behavioural data, personalising outreach for each debtor. Johannes stresses that AI is an augmentation tool, not a replacement for humans, highlighting ethical concerns around keeping humans in the loop for fraud detection and customer service. He explains POM’s cautious technology‑adoption approach: a dedicated AI lab, rigorous testing, and a zero‑tolerance error policy before any feature reaches production, positioning the platform as reliable infrastructure rather than a hype‑driven front‑runner. The episode closes with Johannes inviting listeners to connect via LinkedIn for further discussion.
Chapters
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 another podcast, Fintech Belgium and The Connector together. And today I've got POM, Johannes. Can you please tell me what is POM?
Well, hi Koen. Nice to be here and nice to be invited by you and nice to be in this splendid location here of Fintech Belgium.
Absolutely.
A few years ago, it's like in 2014, we started with the company POM and we is Tom Totté and myself. Tom Totté, my co-founder. And we started with POM and POM stands for Peace of Mind. And our big idea at that time was creating peace of mind for people in the whole invoice payment process. And what we did is actually we had a splendid idea. We wanted to make an app that allowed you to receive your invoices, pay them with one click on the moment they had to be paid, store them automatically, and then share them automatically with whoever needs them, like your accountant or your wife or your husband or whatever. And although it was a splendid idea, and although we worked it out also very nicely, we found out after a few years that convincing a consumer market in downloading your app in the first place and then continue using it, it costs you a lot of effort and especially a lot of money. And what made us make our big pivot, and we made a big pivot in 2017.
Like many of the startups do.
Yeah, I think if you haven't done a pivot, you're not a startup, I would say. So we did our big pivot as well somewhere in 2017 then, and we talked Well, instead of going to the consumer and trying to convince the consumer to use our app, and then by having a lot of app users also trying to convince the businesses, why don't we go directly to the businesses? And that's what we did. And then we told them, we can help you getting your invoices faster paid.
Oh, and that sounds nice.
And that was nice, and it struck an interest, and we were lucky to, well, have as one of our first customers, one of the biggest energy companies in Belgium, and that kicked off all the rest. So yeah, what we did is what we did, what we do now, and is that we add payment tools on invoices. And the payment tool can be a QR code in a paper invoice, can be a Paylink in an email or an SMS or a WhatsApp message. And it allows you just to pay easier.
Which for the companies is really nice because they get the money obviously sooner, I guess.
Yeah.
So they get the money sooner, they get the money always 100% correct. So the message is always correct, the amount is always correct. So there's 100% automatic reconciliation of all the payments that come in. And they have happier customers because, okay, yeah, who doesn't like to just, well, Click and pay instead of having to type over all this boring stuff every month again. And based on that, yeah, we grew to now like we have about 2,500 customers in all kinds of sectors. We work for the government, we work for major energy companies, utility companies, lots of hospitals, lots of schools, lots of sports clubs, lots of small companies as well. So every company working actually towards consumers. And so we have a B2C companies as customers.
Well, Johannes, thank you very much for sharing this story with us. It's a great startup story. Now, in our preparations, we said we also talk about AI, but now I'm wondering, with such a solution, do you really need AI?
Well, it's an intelligent solution on itself already. But anyway, good answer.
Good answer.
Maybe to complete this story and then you'll see the AI coming in. Actually, we We became part of a group. The group was called at that time Mail2Pay back in December 2022. And Mail2Pay was actually specialized in credit management and the creation and follow-up of invoices, outgoing invoices. And it was a— we said at that time, and we continue to say, that it was a match made in heaven. And so we have now— we put the joint product in the market since December '22. And actually, it is in that program that we actually use AI. And well, AI that we use actually, well, it was already invented before the word AI was actually trending. And at the time, we still called it machine learning. But what we actually do is also, first of all, we give peace of mind also in, well, the distribution of invoices and the follow-up of invoices. That's why we have this match made in heaven with MeltoPay. But what I actually do is something we call machine learning. So we developed a— it's like 5 years ago, we developed a model that actually will help you or will decide in your place about the best moment and the best channel to connect to your customer in order to get your invoice paid. And I must say, it was at that time, it was very, I would say, forward-looking and innovative. And so we created a model, we trained it with, we had like a million records, we trained it with 700,000 records, and then we tested it afterwards with the other 300,000 records. And actually, the model does 2 things. It will, first of all, based on your age, the region where you live, will define, will try to calculate how big the chance will be that you will respond to a communication via different channels and on different times. And it will, based on that, will make a prioritization. And based on the group you're in, it will prioritize. For instance, it'll say, okay, we haven't paid your invoice, you will get your next reminder via SMS because your group, your region reacts best on that. But that kind of information, we will also match them with the record of that specific person and say, okay, we know your history, we've seen, we've sent you a lot of SMSes already previously and you never responded on it. So although it is best for your group, it is probably still better to send a letter or an email. And based on that, the system decides when and through which channel a communication is sent.
So actually, you're talking about a very far-fetched personalization.
Yeah.
So clearly an opportunity on AI, but I was wondering, what would you see as challenges specifically then in this whole payment sector, if I may?
Challenges, I would dare to say there's more opportunities. And basic thinking what we have is that AI, well, should not replace humans, should never replace humans, should support humans. And a very nice example, and it just happened to me like yesterday night, Ik deed een transactie, die niet ging door. De dag ik wilde betalen met mijn creditcard en het was geblokkeerd. Maar eigenlijk, ik kreeg een sms dat zei: Oké, wel, we vonden wel een suspicious transaction. Please call to this number and you have a code here to well to skip the line. En dan en eigenlijk, ik was received immediately, dus ik vond. Immediately, ik got somebody on the phone and zei: Ja, wel, we denken dat transactie dat je tried to do of dat was tried to be done yesterday was suspicious. And it blocked my card and the transaction was blocked then. But I had a very good conversation with that person and okay, we could explain everything and the card was re-blocked. So, and I think this is a very big opportunity I see in AI. AI will become better and better in finding suspicious transactions, doing an action on that. But it's in the end, it's finally, it's the person that actually it's a human person that will interpret and will finally take the decision.
Yeah, so I think we're talking more about augmentation of processes then.
Absolutely. I definitely think about augmentation and I don't see us in the near and maybe even the further future not really being replaced by robots or artificial intelligence. I see us being helped a lot by artificial intelligence.
And I would think in this sort of way of thinking that we would be helped, augmented. Do you still believe there's ethical concerns in this whole realm of AI then?
There is ethical concerns if you don't put the human in the front. And if you think that, for instance, your salesperson or your support person is going to be replaced by an AI because the AI does it faster, then you're wrong. I think AI, as good as it is now and can be and will improve, it can still not replace a human person. But it can absolutely improve a lot the service you can give. And it can help you getting rid of the boring tasks, like responding to a mail. It can prepare a mail for you. But you're going to read it and you're going to finally send it through and maybe change here and there a sentence. But it helps you doing the— it helps you with the boring stuff, eliminating the boring stuff. A second thing, it can absolutely help you enormously. Of course, it can process lots and tons and tons of data and much faster than we have. So it can help you getting the insight you need to offer a better service. But it always comes back to the human, I think. It is a supporting tool.
So how does PUM stay ahead in terms of adopting to new technologies and make sure that you're sort of in the forefront of global payment systems?
Yeah, first of all, of course, we keep our eyes open, but I mean, you would be not a good company and not leading value for your company if you don't do it.
That's correct.
We keep our eyes open. We have set up a team as well that is working around AI and evaluating the opportunities around AI, other than, of course, what we're already doing with machine learning.
Mm-hmm.
But what we want to do, we don't necessarily want to be the front-runner. We process millions of transactions even on a monthly basis, and it has to go fluently. And you can't have— you can't accept even 0.1% of errors there because otherwise that adds up to thousands of errors in a month. So what we do, we do test, we see what's in— of course, we follow the market. We will only release things of which we're 100% sure that it's going to work. Mm-hmm.
It almost sounds like an infrastructure.
Yeah, absolutely. Again, we have our labs on the one side, but before something comes from the lab into the actual system, it really needs to be 100% proven.
Makes sense. We're almost at the end of our podcast conversation. I was wondering, who should contact you and how should they contact you?
I'm eager to talk to everybody who's interested in hearing more about what we're doing. So if you want to have a conversation with me, just send me a message through LinkedIn and I promise you I'll answer. And if we find an interesting topic, we'll certainly— are going to sit together.
So also philosophy?
Definitely.
Thank you very much for having you here in the podcast. It was a pleasure. Thank you also for the audience for tuning in and stay tuned. More news from the financial industry. Thank you so much.
Thank you very much, Koen, for having me here.
You're welcome.
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.
