The Payments Podcast by Bottomline

Episode Transcript

Owen McDonald (Host): Smart new solutions are changing the real estate invoice to pay process, as in AI smart. Proper ingestion and coding of invoices to the real estate sector can be highly complex involving all manner of invoices to tenants, service providers, contractors, utilities, municipalities, the list goes on. As in many sectors, real estate invoicing and payments are now leapfrogging legacy solutions with AI and machine learning. You could say practitioners are ready to upgrade to a better view. To illuminate this topic, we're very glad to be joined by David Stifter, founder and CEO of Predict AP.

Many of you will know the company as a state-of-the-art cloud-based solution for real estate accounts payable founded in 2020. David's business passion is found at the intersection of real estate, technology, and finance. He previously served as managing director of Digital Bridge, formerly Colony Capital, overseeing data architecture, process improvement, accounting, and finance technology. He holds an MBA from Boston University where he graduated valedictorian. Go Sox. Go Pats. Go Bruins. Go Celtics. David Stifter, CEO and Founder of Predict AP, welcome to The Payments Podcast.

David Stifter: Owen, thank you. That's a great welcome. I appreciate it. You really got my morning off well.

Owen McDonald: You're very welcome, and we couldn't be happier that you're here. To begin, tell us about Predict AP and its unique adaptations of artificial intelligence and machine learning. I'm curious specifically, though, about what business problems you set out to solve and how you feel you innovated.

David Stifter: Yeah. Thanks. So, it's interesting. I think maybe, the path we had was a bit unique in that I didn't go out to solve start a company. I went out to solve a problem.

So, as you mentioned in the start, I was the Head of Technology for Finance and Accounting at Digital Bridge. It's a big global read. And so, I think with a lot of companies, especially real estate, where they're started by deal people, so they're always looking for the next deal and getting bigger and bigger. And I think at some point, they turn around and look back and say, oh, we have all this complexity now, and the operations needs to catch up. And that was the situation we found ourselves in with a number of private equity funds, a number of public companies, a lot of complexity. And, we had struggled just closing the books, and that's where we looked at where is this time, where is the risk, where is the effort going in? And we saw that AP, maybe surprisingly, over half of all accounting entries come through the AP process. Right? There's lots and lots of bills being paid both on that corporate fund structure and on the property structure.

And so, we were looking to get more efficient, and we started on this path. And I always think about AP as kind of three chunks, payment at the end, approval and workflow in the middle. And I think the ingestion including upfront probably doesn't get as much attention. And so, we had solved workflow and payment, and we still had a lot of tools upfront with that, but we had one person, Chrissa. Chrissa had been at the company for thirty years, knew everything about everything. Chrissa retired. We can no longer pay a bill, and that's where this problem was dumped in my lap. Why can't we pay a bill, Dave? We just invested all this money, and now Chrissa is gone, we can't pay a bill.

And that's where I really, I always like to really dig into problems. I started sitting, in the AP group and saying, what are they doing? And there's a lot of complexity, a lot of complexity. I think that the AP folks don't get enough appreciation for how much they have to know. And as you said, I think a lot of times maybe real estate has this impression that maybe they're a little behind in some areas.

But I think the challenge here is actually more acute and more severe because there tends to be a lot of complexity. I know the real estate folks always have these crazy organizational structures with entities and tax blockers and all this crazy stuff the lawyers come up with, very complex. Then on the property, you also have a lot of complexity where your office building, maybe you can recover the expense, and you can pass it through, so you code it one way. Or maybe that industrial, you recover it, you pass it through a different way. Or maybe that corporate bill you have to cover yourself, or maybe you can split amongst these funds.

Layers and layers of complexity at the property level and what you can recover, what you can't, how you do it. It could be different property by property or even lease by lease. So you have a lot of complexity around the structure, a lot of complexity about the operations, and you kind of multiply those two things together. Real estate is a very difficult, complex situation. 

And so that's where we found ourselves. And I said, our team needs good tools. I think so often front of the front of the house gets all the attention with modern tools. And I saw that this was a very difficult thing, and I didn't appreciate how much when the new person came on, Chrissa had trained them or me as a managing director. I'd asked Chrissa, hey, how do we pay this bill? How do I, how do I allocate that? And with her gone, that knowledge walked out the door. 

And so, the problem I went to solve was how do I handle my company's coding well and efficiently with technology, not just relying on yellow Post-it notes or, the travel knowledge being passed down person to person. And so, I spent about a year looking for that solution. I figured, hey, someone must have solved this. Let's go find it. And so, I spent a year kind of being disappointed a little bit and saying some really good solutions out there with payments. As I'm sure you know, there's some good solutions out there with workflow and approval. But, as far as the suggestion and coding piece, there's not a lot of good options.

And that's where, after trying a few things, failing a few times, and realizing that truly this hardcore coding problem isn't being addressed, I'm kind of a geek at heart, and I was learning about these new AI models and ML models. And so, it seems like a really good use case. And that's the problem I wanted to tackle was the ingestion of coding and work of invoices, but I didn't want to change my payment provider I was happy with. I didn't want to change my work provider. I just wanted to give my team good tools to help them with coding.

Owen McDonald: Well, so far so good. We connected recently, and you spoke of unaddressed client pain points with invoicing generally. We're aware of that, of course. But even in electronic invoice processing, unpack that a little bit. Where people are using it, it's not necessarily that well developed. Why not?

David Stifter: Yeah. I think one of the reasons is that payments have gotten a lot of attention, rightly so. Like, if you're looking at where should I maximize the value for my company, and this is where I started. So, I started with payments because the time involved with that, the risk involved, and the opportunity to kind of gain some efficiency and gain some revenue for your company, it's where you want to start. And so, I think most folks, as far as attention, focuses on that end of the process payments. That's what we did. 

And that was that was actually a huge help getting rid of the old checkbook in the Friday check run, and, of course, CFO's signature writing thousands of checks that was it. And I think a lot of companies focus on that part because it is the most impactful as far as operational efficiency and dollars to be saved, from just a pure basis point savings perspective. 

The upfront coding part, I think, is actually the most complicated and painful. But for the AP department, which maybe doesn't get enough of a voice at the table oftentimes. And so, people might think, oh, I'm using some sort of OCR. I'm scraping some simple data because everyone has, every player in the market has some sort of solution that does a basic pass of things. And for some companies, again, this is where real estate, I think, is a little interesting is because, for maybe manufacturing companies or, where there's less complexity, that's probably pretty good. Where you may be a big company, but you're making T-shirts, so you have one entity and buying ink and cotton, maybe that's good enough. You know?

And so, these more broad industry agnostic solutions probably look at that and say, hey we're solving the problem. Not really understanding and appreciating, the nuance of what the real estate folks have to face. And that's what I found was that, hey, that simple industry agnostic solution isn't going to cut it here because it's not going to handle that complicated coding that my company has.

And so that was the kind of the layer of complexity that I wanted to solve around that was that, hey, I understand, that it's unique by company and there's, again, layers and layers and layers of complexity, thousands and thousands of permutations of how this should happen. And that's where ML just shined as a really good solution because I've kind of figured out those three critical things that work well right now with ML, which is good, structured data, and invoices are very rich with that. Right? You have nice structure with it. You have the image. You have the coding. That's great. So, good, structured data. 

You have repetition and patterns. This is another key thing that a lot of times, if you think about your spectrum of invoices coming in, a big chunk of them is going to have a pattern. It may be a very complicated pattern. Like, you may have a service provider that's doing 12 things at 32 properties, and someone has a Post-it note or Excel document that splits that up, but it has nice structure and repetition to pick up on. 

And the third thing, which I think maybe is overlooked a little bit, is you still have a person in the loop. Right? For the payment part, I don't know anyone who's happy to just let it run through a system and actually move money in the real world. So, you have this human in the loop, which lets you, lets the system understand when there's a change. Right? Hey, we just sold a building. You bought a building. Something's different. Accounting's treating something different. So, you still got human.

So, structured data, repetition, and a person in it, is like the perfect use case for ML. And so, after I had tried to find this in the market and couldn't find it, applying these models and techniques, and it's much easier to do for one company than building it as a as a true SaaS product. That's a whole different adventure. But when at the end of that, our AP team was so happy, because the vast majority of it was there, and that let them focus on much more value add things like getting payment terms, like looking for fraud, like looking at different vendor, relationship type things as opposed to just data entry, let us pay bills so much faster, all of that.

But that problem solved really enabled us to do a lot of other things. And with that, some of the CTOs I had asked about, what are you doing to solve this problem? I showed them, and they said, we want that. And that was kind of the transition for me to say, hey, I've got this problem. I know there wasn't a solution. I was actually really worried for six months that I just hadn't found the solution, but I've been pretty confident now that it doesn't exist. But everyone has this problem. Right? And so why not try to solve it?

And that's where that nuance was. Like, this coding part, this true, like, detailed accounting part wasn't being addressed, and that's where the ML model really was the perfect use case even looking back three or four years ago. So, I think we were a little ahead of the curve. Now everyone is looking for ways to use AI, but that was the right tool, at the right time, for the right problem.

Owen McDonald: I would have to agree. You're partnering with a large PSP, which happens to be Bottomline, to fulfill the AP automation and payments part of the process, and we're told it's still faster than a lot of all-in-one platforms. But putting brands aside, talking just trends, what about the workflow and technology makes this combination work better?

David Stifter: Yeah. That's a really good point. And it's interesting because, again, I am not a 19-year-old fresh out of college ready to start a startup. And I think a lot of prop tech start from a technology perspective looking for a solution, looking for a market and may think, oh, real estate's an interesting market. 

But I came from the business. I'm kind of obsessive with solving problems and even having the most perfect tool to solve the problem. If my tool was perfect for every scenario, if it couldn't actually be integrated into the customer solution, then it's kind of worthless. Right? And so that's where, I think, we had a little bit of a distinct approach is saying, hey, from my customer's perspective, how can I make this work as efficiently as possible?

And that's not just AI or ML. Right? That's how does the integration work? Right? How does my customer have confidence that when we sign up that we'll be ready to go in two weeks, not six months?

And I think so often with prop, real estate technology that customer experience integration is an afterthought. Right? They think I have this tech. I solve this problem. It should work like Microsoft Dynamics. It should work like Oracle. It should work like something else, but it doesn't. Right? The reality is these platforms have certain strengths and weaknesses. And so, from my perspective, we made a very early strategic decision to only work with, established platforms that will work with us.

And that maybe is a smart decision or maybe a risky decision, in that typically, prop tech is going to, you're going to find some hacky way to make it work, some nightly FTPs, something. But I think the customer ultimately suffers from that because those things tend to work for six months and then they don't, then it's a disaster. Right? It tends to take very long to get started. It tends to fall apart when there's upgrades and changes and things like that.

And so, with us, we recognize who the market leaders are, and we looked to make a partnership with them. And, again, back to understanding our own strengths and weaknesses like Bottomline, for example, has a very, very strong payment platform. So, from the customer's experience, there's huge opportunity there. And I talked about that's kind of the cake. Right?

That's the big thing you want to do well. I look at us as, like, the cherry on top for best practice to really get your AP flow working really well. And so, for us, it was a pretty easy decision to say, this is one of those predominant platforms and approaching them and saying, hey, let's think about how this will work for your customer. They're already getting a best-in-class payment platform, workflow platform. There are some areas here that we know we can enhance for them. So, let's work together because your customer is going to be ultimately happier with this together. And so that pain around setup and getting it working and getting it maintained long term, that's dealt with our engineers talk together, work together. We have test environments with this.

So, from the customer's experience, it's really invisible. Right? It's really a pain free situation where it's really turnkey. So, when a customer decides to get going with this holistic solution, they just make one choice, and its half an hour work and their time to get going. You know?

So, ultimately, we've made a lot of work to make sure that that integration is as invisible and seamless as possible, and I think, again, for the customer's benefit.

Owen McDonald: Also, I want to follow-up on that point because, when you told me recently that, quote, in prop tech, too often there aren't real relationships and the customer suffers as a result, unquote. So, if I'm a prop tech user and I hear that, what do I think, Dave?

David Stifter: Oh, this is a good question. I think being vocal about it, right? I think setting the expectation. We're the startup in this situation. Right?

So, we're the one trying to establish ourselves, and we're confident in what we can deliver on our core mission of doing really good coding. But I think there should be an expectation across the industry, both from that ERP, PSP provider, and from the add on solution of there that there should be a seamless experience from the customer's perspective. And I think the more customers can be vocal about that and it is more typical, I think, in other industries where there is better cooperation with that. But for whatever reason, it's just different in real estate, I'll say. But at the same time, if there's any prop tech founders out there, I would say don't take that easy excuse.

Right? Just because it's not as easy as other industries doesn't mean you can throw up your hands and tell your customer, go complain to the ERP or the PSP, and it's not our fault. Right? Because the customer doesn't care at the end of the day. Ultimately, you need to have empathy and understanding of what it's going to be like for the customer, and you need to do what you can to make that work well.

So, I think from the customer's perspective, not being satisfied with that hacky solution. Right? Saying, hey, you two need to work together. From the property technology perspective, I think not finding that excuse. Right? You said, go Patriots. That Belichick code always rings my head. If you give someone an excuse, they're going take it. It's hard, but that's not an excuse. Fix it. You know? 

And I think for us, we've made a big investment and, it's interesting. We're backed by RIT Ventures, and there's a summit of companies there. And this is a topic of discussion is these integrations. And I think the message I had there was, like, it should be on us to make this work as well as possible. Right? 

And that's part of our obligation. That's part I'd say probably a quarter of our engineering goes into making those these things work really, really well. Because at the end of the day, the customer doesn't care for AI or not or doing whatever. What they care about is, can my invoice get paid in thirty minutes through instead of three days through. Right? That's what matters.

Owen McDonald: Absolutely. As we wrap, Dave, invoice and AP automation married with quick, secure digital payments is a combination that's changing the face of B2B finance as we know. How does 2025 look for transforming real estate invoicing and payments specifically, David?

David Stifter: Yeah. I think, there's still a lot of people with checks. I'm surprised with which, I would say if you're writing checks, like, your forehand must be hurting. You should look at this. You really should take a look at where is the risk in your company and where is the time being spent.

I think you'd be surprised to find that, again, probably over half of all entries in your accounting system is coming from that AP department. Right? So, that represents a really big opportunity as far as efficiency. It presents an opportunity as far as the money you can make as far as working with a payment provider as far as discounts and other things, and the risk involved. Right?

Payments, I think fraud is becoming a bigger and bigger issue. And the old-fashioned way is fraught with risk. Right? These tools, that we're using for good, things like coding, are also being used, for nefarious reasons, like defrauding you. And I'll tell you the battle of power, the fraudsters are always the first ones to really adopt new technology and adopt it effectively. And so, it's hard for your treasury department that's doing 35 different things to keep up to date on all the new methods and techniques people are using to defraud people. Well, an established player who's a billion-dollar company is focused on payments is allocating tremendous resources and keeping abreast of that. 

So, number one, you need to worry about risk. You need to look at opportunity. You look at efficiency.  And then the second part is, again, these tools are starting to trickle backwards in the flow, like Predict AP and others, to really handle, I'd say, the maybe not so well recognized issue, but the internally painful, problem of ingestion and coding and other things. And so that promise land of a holistic, almost touch free still overview, right? You still want people looking at these things, but touch free solution from hits your inbox, goes all the way through the payments. It's there, right? You're able to do that now with some investment. You just need to think about hitting each piece of that AP chain from payment to approval to ingestion to coding to be a best-in-class solution.

Owen McDonald: Well said. 

And that, folks, is how you fix malingering real estate payment headaches. You start with AI powered invoice ingestion and coding. It's a meaningful innovation for delivering more efficient digital billing and payments. A giant thanks to a great guest, David Stifter, CEO and Founder of Predict AP.

To our incredible audience, thanks for listening. Don't forget to subscribe, and catch us again on your favorite podcast platforms, including Apple, SoundCloud, and Spotify. Bye for now.

The Payments Podcast, from Bottomline.