Skip to playerSkip to main content
Mortgage lenders are facing a growing reality: the complexity of modern lending is outpacing the capabilities of traditional software. In this HousingWire conversation, Zeb Lowe speaks with Michael Kelleher from Sapiens Decision about why AI-driven decisioning is quickly becoming essential infrastructure across mortgage operations. What began as a niche technology is now gaining traction across non-QM lenders, banks, credit unions, and correspondent channels. 

Throughout the discussion, Michael breaks down the forces pushing lenders toward more sophisticated decisioning tools, including regulatory uncertainty, expanding product complexity, and the operational limits of legacy technology.

#DecisioningAI #LendingTech #MortgageInnovation #SapienAI

Category

🗞
News
Transcript
00:06Live from Las Vegas, I'm Zeb Lowe, here with Michael Kelleher. Michael, thank you for joining
00:09me. Thank you, appreciate it. Yeah, so the last time that you spoke with HousingWire,
00:14you were filling us in on Sapien's decision. So what's the latest there?
00:18Yeah, there's been a lot of changes since last time. I think the most important change is
00:22we've gone from what was a concept to a category. So Forrester has now named us a leader in
00:30decisioning AI. And in addition to that, we now have real use cases with banks and credit unions,
00:37non-QM lenders, correspondent lenders, wholesale lenders. So we have a lot of momentum in the
00:44category. I would even say we have an acceleration going into 2026. I think going from concept to
00:50the category is a pretty bold statement. What do you think is driving the acceleration that you
00:55mentioned? Yeah, I think it's two things. One, we have real use cases now. So we have lenders that
01:02understand today's technology has a difficult time managing decisions and more importantly,
01:10keeping up with the changing of decisions. So we have non-QM lenders reaching out to us about
01:15building an AUS. We have banks and credit unions talking to us about portfolio AUS or maybe a HMDA
01:24or compliance problem they're trying to solve. We now have correspondent in wholesale reaching out to
01:29us to try and reduce some of the costs there at that transition. And I think the other thing is
01:35just
01:36when you have a use case in the market and other lenders hear about decisions that used to take three
01:41days or taking three minutes for certain lenders, they come to conferences like this, they start to
01:46talk. When you're able to solve pain and you're able to bring that competitive edge to it, those two
01:52things can be a powerful combination. You mentioned non-QM. We've seen in the past couple of years at
01:56HousingWire, we've done a lot of content over the rise of the non-QM market and how it's growing in
02:03its relevancy and not what it was 10 years ago. So what are you guys seeing in relation to the
02:08non-QM
02:09market? Yeah, non-QM is on fire. I couldn't be more excited about non-QM for the industry as a
02:15whole. And I've been able to channel that through my connections and really showcase to these non-QM
02:22lenders, you need it to increase your volume. But at the same time, there's a lot of manual components
02:28to it. So you have manual underwriters and different underwriters making different decisions,
02:34which produces different exceptions. And the turnaround time starts to get longer. So these
02:40non-QM lenders would come to us and say, how can we have an AUS like Fannie or Freddie? Is
02:49there a way
02:49we as a non-QM lender could possibly have an AUS? And a lot of them said, maybe you can't
02:55have it.
02:55It's art versus math. But we were able to sit down with them and show them that our authoring tool
03:01is
03:01able to take their guidelines, their matrices, and bring it into a world where we're able to
03:08look at eligibility. We're able to look at exception management. We're able to then produce document
03:15checklists and conditions. And overall, the outcome becomes eventually being able to automate it
03:22and govern those decisions and bring an auditability to who's making those decisions, when they're making
03:29them, how they're making it. And so we've brought the AUS to non-QM.
03:34What about banks and credit unions? Is that that much of a different market for you guys? Or is it
03:38really kind of all one and the same? It is for me, I suppose, handling the sales cycle. Those are
03:44much,
03:44much longer sales cycles. But at the end of the day, our lead into those banks and credit unions is
03:50very similar. They have portfolio products, whether it's a jumbo program or a specialty loan,
03:58or maybe it's community lending. But they all typically have the same thing. They have their
04:05guidelines on Excel sheets, or if they're a more advanced bank, they have what's called loan cards.
04:11But these loan score cards are usually very out of date. And so they're coming to us same way,
04:17saying, if we could have an agency solution, how would we build it? And so we look at them and
04:25we
04:25try and figure out how they could take those loan score cards and start to automate or start to build
04:33the decisions where they have a governance to it. They have an auditability trail. They have a
04:39versioning. So they're able to then say, this was the decision made by this person on this version,
04:46and this was the outcome at a click of a button. And the auditors like that. And the banks are
04:51loving
04:52this new AUS.
04:53So traditional software can't handle this sort of workload. Why not? What's the problem?
05:00What's the catch there?
05:03Traditional mortgage lending software is just very linear. We take an application,
05:08and then we try to approve the application. We underwrite it, we approve it,
05:12we close it, and we sell it. And the complexity of decisions has grown so much over time. The ability
05:23to need speed now with it or add AI to it has brought more complexity. I've actually found
05:30building some of those decision rules is the easy part. How do you maintain them over time? How do you
05:37handle the changes? And when the changes are maybe at the state level, or maybe they're at a different
05:42compensation level, and where do you document them? Are you able to version them? So the decision
05:48layer that Sapiens is able to bring allows you to keep your existing software. It listens on the APIs.
05:55It provides a decision layer so that any question you have that you can answer with data, you can build
06:04in Sapiens' decision, and begin to bring those decisions and surface them in the technology you
06:10have today. And today's technology wasn't built to handle that.
06:15What about on the correspondence and the wholesale side?
06:20Yeah, very excited about correspondence. We had a lender come to us to automate
06:25a AUS on a particular loan. They came and they looked at some of the UI that we've made available
06:33across our sectors. And they were able to discover that what they do, which is really the same, they're
06:41looking at investor overlays of multiple investors, investor guidelines, maybe some document checklists,
06:48maybe some dates or certain types of conditions they need. Well, our ability to handle that in batch,
06:55whether it's overnight or in line or both, and then provide exceptions that are needed, allows the seller
07:03and the buyer to have more transparency into what could possibly cause a repurchase and maybe handle
07:12it before they close versus chasing it down the line. And so it's one of the rare times I've
07:17seen where it can be digested on the buyer and seller side. And it allows really at scale,
07:23which is what Sapien's decision is all about, the ability to handle that bulk and provide a better
07:29understanding to all parties involved. Gotcha. Talking about compliance, that's always a headache for
07:35lenders. So what's Sapien's approach to tackling those obstacles?
07:39We wanted customers to realize that we got our feet wet in compliance.
07:44That's where we started. Freddie Mac brought us in to work with them to be really the engine
07:50behind LP. So we understand the importance of having an audit trail, having versioning.
07:57And so we've had a couple of banks reach out to us and say, Humda's a problem. They have a
08:02lot of
08:02full-time employees that work on it throughout the year. Then they rush at certain points of the year.
08:06And it's really checking Humda fields, matching them up with running AMI, area medium income, and then
08:15running it against census tracts. And at any point, if the vernacular on the LAR doesn't match up any of
08:22the fields or the documents, that could be an invitation to come in and audit. So that is what
08:28we're doing. We're running that decision layer where the banks have worked with us or the credit unions,
08:35but they've worked with us to author what decisions should be made in real time and in batch overnight
08:42over their data. And that way there's an audit trail and they're able to show who made what decision,
08:48when they made it, on what version. And if the auditor comes in, they're actually able to press a
08:55button and show that entire audit trail. So it could eliminate, it just lowers your regulatory
09:03risk or exposure. Right. So the picture, what I see is a picture being painted of lenders adopting
09:09this across all channels. Is that a fair assessment? Yeah. That's what's exciting about what I do is
09:16there's no specific niche. If there is a decision to be made that can be solved with data,
09:24it is a great reason to reach out to Sapiens decision. And, you know, that's what we always
09:31say is start with one decision. Right. Well, so Forrester, I saw recognize Sapiens as a leader in
09:37AI decisioning. And so what does that, is that validation? What does that mean? What does that feel
09:43like? It feels great. I know the entire Sapiens decision team worked extremely hard on it. To be called a
09:50leader by Forrester is certainly important. They look at AI decisioning as the combination of the
10:00disciplines that were traditionally siloed. So business decisions, data analytics, and machine
10:08learning. And so to bring it all together into a decisioning layer that you can run orchestrated AI
10:16over is what, or the reason a lot of these banks, credit unions, non-QM lenders, correspondents,
10:24come to Sapiens decision. But I think what makes us most proud is it's not just about the technology
10:31with Forrester. You're judged on vision. You're judged on execution. And with Forrester, you're judged
10:37on the feedback from the customers specifically on what the use cases it's being used for,
10:43the satisfaction with the use cases. And so being awarded that is more than just you have great
10:50technology. It's, we are battle tested. We've been around for 15 years. We're in the public sector.
10:57We're in the government sector. We're in large insurance. We're in large banking. And now we are
11:03making a big effort to bring it down channel, show we're battle tested at scale, but we're ready to work
11:09with the non-QM lenders, the correspondents, the banks, the credit unions, even the IMBs.
11:14I think that actually, that ties into my last question that I have for you. So for any lenders
11:17that are watching is what's the, what's the big message, what's the takeaway that you would like
11:21for them to receive? The big message is this industry believes that technology solves everything in
11:30mortgage. And at the end of the day, mortgage companies do not need to be technology companies.
11:36They win on the decisions they make, the decisions on how accurate or what pricing they should have,
11:45their decisions on how they're going to underwrite files, even their decisions on their willingness
11:49to be riskier or less risky. And so that's where Sapien's decision brings that extra advantage to those
11:57that are ready to embrace it. Because if you're not moving forward in this industry, you're falling
12:02behind. It's moving so fast. So the message is find that one decision in your company, that one piece
12:09that you think is slowing down the process, that one piece that you maybe believe is a real risk
12:18to the company. To put it simply, that one piece that keeps you up at night, find that one decision,
12:25start with us, let us just solve that. And I always say it's like the matrix, you'll start seeing
12:29decisions everywhere. And then you'll wonder how you lived without it. All right, Michael,
12:34thank you so much. Appreciate the time. Thank you.
Comments

Recommended