Skip to playerSkip to main content
  • 2 days ago
What happens when AI stops being a feature and starts reshaping the foundation of mortgage technology? In this conversation, Craig Rebmann of Dark Matter Technologies spoke with Allison LaForgia to explore how AI is transforming loan origination systems into the brain coordinating people, processes and technology.

Category

šŸ—ž
News
Transcript
00:06I'm Allison LaForgia with HousingWire and today I'm sitting with Craig Rebman,
00:12the Managing Director and Product Evangelist at Dark Matter. Craig, thank you for joining me.
00:17Thank you for having me. It's a pleasure.
00:19Let's start with your role. How do you think loan origination systems are evolving as lenders
00:25integrate AI strategies? Yeah, so loan origination systems play a really critical role. They really
00:31have to help facilitate that strategy. Whether a lender is using artificial intelligence capabilities
00:37brought on by the LLS provider, or maybe bringing their own innovations into play, or maybe they've
00:42got a favorite third party that has some great innovations that they want to bring into that mix.
00:46It's really important that the loan origination system is able to facilitate all of that.
00:50And part of my role is to help lenders kind of navigate that ecosystem and understand how
00:54best to make that work. Now, in a recent demo, you showed how Empower has evolved to coordinate
01:01increasingly automated processes, but mortgage manufacturing is still deeply human. How are
01:08you approaching the balance between humans and technology? That is a great question,
01:12and it's something that is often overlooked when trying to tie a process together. It's a combination of
01:20the machine and people, automation, and increasingly AI agents all playing a role. So the loan origination
01:29system really needs to be able to orchestrate that process and coordinate the work between all of those
01:34parties. Breaking work down into tasks facilitates that. And then really, you're just managing task
01:39assignment. Is it going to a resource that's human? Is it going to a resource that's agentic? It's really about
01:45getting the right skills to bear at the right time on a particular loan problem.
01:52You just mentioned something that I want to spend a little more time on, and that is tasks and process
01:57orchestration and a very important capability. But do you see process orchestration as a fundamental
02:05responsibility of the LOS moving forward?
02:09That's, you know, there are a lot of kind of external capabilities that might play the role
02:16of control tower and help manage work across systems. But really, the LOS is the foundation
02:22and the fundamental operating system for the lender within their origination workshop. If the lender can
02:30manage the process appropriately through the LOS, and if the LOS has a sufficiently robust orchestration engine,
02:38which we've been very fortunate to have, then that can be done from within the loan origination system.
02:44And the loan origination system itself can be that control tower, manage that process across
02:51systems and across people in a really efficient and effective way.
02:55Going back to AI for a moment, AI has been around for a second now, which means that everybody has
03:03their
03:03own strategies and there's increasing customization even as people are rolling out AI technology.
03:12What does that mean for an LOS to truly support a lender's differentiation rather than standardizing it?
03:21Yeah, it means really the LOS has to be able to facilitate lender choice.
03:24And sometimes that's going to be the agents that the LOS brings to bear.
03:28Sometimes it's going to be innovation that the lender is bringing themselves.
03:34So model context protocol is a big part of that.
03:37We introduced model context protocol to Empower to facilitate making it easier for lenders to bring
03:43their agents to work or bring their favorite third parties agents to work with Empower.
03:48So it's important for the LOS to facilitate that and make it easy for the lender to use the agents
03:55of their choice.
03:56A couple of times in this interview, you've mentioned the word foundation, control tower.
04:04What responsibilities come with that in terms of data governance, transparency and control
04:12in an AI driven environment where we're concerned about all of those things?
04:18Exactly. So the loan origination system is still the system of record.
04:21So data governance, access controls and ensuring that all of that information is being properly managed is
04:30critical. The loan origination system has to be able to do that. And loan origination systems historically
04:34are very good at doing that. Facilitating that kind of governance while also allowing the access that's
04:40necessary that allows for agents to come in and do the work that they need to do,
04:45particularly autonomous agents. It's really around managing the access controls in a really responsible
04:52way. And that's one of the great things about model context protocol. It helps to facilitate that.
04:57Many lenders are exploring their own agentic AI capabilities. In your view,
05:03what's the right balance between platform led innovation and lender driven experimentation?
05:10So what we've always found works best is open dialogue and collaboration. We work very closely
05:16with our lender clients on what their vision is. We share our vision with them and collaborate a lot
05:23on kind of who's doing what, what are people thinking? And is there something that maybe we can take off
05:30their plate of individual innovation that they're trying to build and build it for the ecosystem and all of
05:37our clients benefit from that? And in other cases, there are some secret sauce that they want to bring
05:41that's really their own and would be unique to them. But having that open dialogue, having a collaborative
05:48relationship between lender and LOS provider is really critical. And being able to have those trusted
05:56parties that you can kind of share your visions with makes all the difference. Looking ahead, if you look
06:01at your crystal ball and AI progresses, should lenders rethink the role of core systems like the LOS
06:09shaping not just efficiency, but competitive advantage? I think that what we're going to see is that
06:16the role of the LOS is going to change quite a bit, and it is going to become much more
06:20about command
06:22and control and coordination, data governance, and system of record. Those fundamental core capabilities are
06:30critical for the loan origination system. There's going to be so much innovation occurring outside of
06:35LOS's that an open ecosystem is really important to that. And so I think, again, it comes back to
06:43making it easy for lenders to bring whatever capabilities they want in to the LOS, but it's also
06:50important for that LOS to be providing their own capabilities back out to lenders, agentic capabilities in
06:57particular, command centers, and persona based assistance, those sorts of things that can
07:03really help cement the user's experience in a way that helps them understand what's going on on that
07:11loan, even when there's a whole variety of automation occurring. It seems like there's so much happening.
07:18The pace of innovation in the industry is truly unprecedented. I can't wait to see what's next.
07:26Likewise. Thank you so much for joining me, Craig. Thank you very much, Alison. I really appreciate it.
Comments

Recommended