00:12From Chicago with House & Wire, I'm Allison LaHorja and I am sitting down with the Consolidated
00:17Analytics COO, Rudy Zabrin. Rudy, thank you so much for joining me again.
00:23Great to see you again this time in Chicago.
00:24It's great to see you in Chicago. So as the COO of Consolidated Analytics, what experiences
00:32have shaped your approach to leadership and growth in the mortgage industry?
00:37So you know what's interesting, I've developed leadership not just in the mortgage industry.
00:42I've also been a football coach for 23, 24 years. So I think that shaped a lot of my leadership
00:48background. Set the standard. Practice the script. Execute on game day. And that's how I approach
00:56my teams. They have to know what the standard is that we set forth as a company. Our clients
01:01depend on us for accuracy and intelligence and new ways to do things. And we deploy that methodology
01:09or that mindset to our workforce. And the practices, the day-to-day things, the fundamentals, the
01:15blocking and tackling, if you will, knowing the data, knowing the industry, how things work.
01:20And you know, we take that to our teams through training and various other types of experiences.
01:24And I think, you know, when the teams know what the expectations are, they can execute upon it.
01:28And you know, I bring a lot of that from my football background. And I think, you know,
01:32they see that in me. There's a standard. There's an intensity. And I think that comes out,
01:37you know, in the day-to-day kind of operations at Consolidated Analytics.
01:40Now, as mortgage companies navigate a changing market, which is top of mind right now,
01:46where do you see the biggest opportunities to operate more efficiently and proactively?
01:55More efficiently, it comes down to how we operate with data and documents, right?
01:59The mortgage industry is unique in that, or maybe not unique. Healthcare might be similar. But,
02:04you know, we take all this digital data and we wrap it in PDFs all the time. And we've become
02:09addicted to these PDF documents, right? So where we see a lot of opportunity is the product suite
02:14that we offer through LoanDNA. Our technology, it's, you know, it's a purpose-built technology
02:19for the mortgage landscape. We identify documents, scrape the data, and then can process that data in
02:25a variety of different ways. And I think that's instrumental. At some point, maybe we're more API
02:29native and more AI native and all these types of things. Right now, we're still wrapping data in PDF
02:34documents. And, you know, our LoanDNA technology is like what we're deploying to try to help
02:38originators. On the intake, maybe it's building a loan application faster. On the capital market
02:44side, maybe it's identifying and classifying documents and running rules against it faster
02:48and more efficiently. You know, volumes in the mortgage space are variable. So your cost structure
02:54has to be variable. If you don't fix that, you know, the market, I think, will fix it for you.
02:58And,
02:58you know, we're at the cutting edge of trying to provide, you know, technologies and services and
03:03things like that through LoanDNA and consolidated analytics that will help people just be more
03:07efficient and, you know, reduce some of the human capital required and be, you know, just more cost
03:12efficient in the process. You mentioned a hot word, AI. AI adoption is moving from experimentation to
03:20implementation. What do you think separates organizations that successfully integrate AI from
03:27those that struggle? So you have to start with your governance, right? You know, how you're going
03:32to manage and adapt AI. And I think it has to be purpose-built. I think, you know, a lot
03:36of people out
03:36there are toying with AI solutions. And, you know, I don't know if they've taken really a regimented
03:43approach to trying to identify the right solutions and implement them. It seems to me that there's a
03:48lot of like fragmented, you know, systems and things like that. And I think when you have a lot of
03:54fragmented systems, it can be, you know, death by a thousand cuts, if you will, on the cost side,
03:58right? You're building, you have one tool and you add another tool and another tool. And now it's,
04:02you know, an API toll gate into your LOS. And now it's a vendor fee from your, from your, you
04:08know,
04:08technology adoption. And, you know, once you get six or seven or eight of those bolted together,
04:12it becomes, you know, not cost efficient. So we feel like there's an ecosystem that needs to be
04:18evolved. And we're doing that again in the LoanDNA technology and offering, you know,
04:22kind of a whole solution of products that will help, you know, mortgage originators
04:26and capital markets groups and servicing teams like operate a little bit more effectively.
04:31AI native is, you know, it's the ability to adapt or adopt, connect your LOS and your,
04:36you know, your rate lock platform and all these, you know, capabilities. But they're the,
04:42really, we use AI to help us develop faster. And then the technology solutions behind like LoanDNA
04:49document intelligence or LLMs and a variety of other things. So finding the right tooling
04:54in the right ecosystem, I think is really important in, you know, not going out and
04:59experimenting with maybe 20 different vendors that are going to lead to, you know, a faster throughput
05:04of a more garbage product. I think you have to be thoughtful and really intentional around how
05:08you're building these things.
05:10So not, so people should be thoughtful and intentional and really strategically look at
05:14their AI adoption. Now, there's another side of this that we know is very much a reality and that
05:19is the human side. How should lenders approach the human side of AI adoption, especially when
05:25employees worry about changing workflows or even job displacement?
05:30Yeah, I mean, I think you have to be open to employees using it. And that governance policy,
05:35I think is, you know, at the forefront of it. You have to have that in place before you just
05:38start
05:38rolling AI out to the workforce. The reality is, if you don't have that, people are going to use it
05:44anyways. And then you have the shadow use, which is dangerous. Like somebody is using their ChatGPT
05:48account and loading, you know, some borrower's loan application to, you know, the internet,
05:53you don't want that. So you have to enable them with the tools to be able to use it and
05:57show them
05:57different ways. Like we, we build out skills within our organization and allow, you know,
06:01people to do it. Not all of it's sexy, right? Like a lot of the ways we find a lot
06:05of opportunity
06:06with AI, let's take it separate and apart from like the loan DNA thing, just how people use it daily.
06:11You know, it's the formatting reports. It's, you know, how we're doing our legal document reviews
06:16and a lot of these other things. It's allowing us to be faster than I think operate better. But
06:21we've got the right guardrails ultimately around it. Like I said, you have to give them the technology
06:25to use it. If you don't give them the technology, they're going to figure out a way to use it
06:28anyways.
06:29So, you know, building the skills, educating people on the skills that exist and how to use them.
06:34We'll do 30 minute sessions with our teams where I'll show them something that I've built with AI and,
06:39you know, give them an opportunity to kind of think about it, ponder it and, uh, you know,
06:43get the wheels turning on things that they may think might be useful. And then, you know, though,
06:47there's a task force in each of our businesses and they submit ideas and things like that back to it.
06:51And we find opportunities then to take those ideas and build new skills around them that hopefully make
06:56people, you know, more efficient and some of the products a little bit better. We need to be
07:00responsive to customers faster. That's ultimately what we're trying to do.
07:03I want to take a second. You said that it might not be the sexiest use case.
07:07Yes.
07:08But making people more productive and efficient and giving them tasks that better improve your
07:15company's operations is sexy. It is for me as a COO.
07:20Now, Rudy, what specific areas are you seeing technology make a measurable difference in
07:25in mortgage operations today? Uh, you know, again, go back to the PDF. Um, you know, we've at any,
07:33on any given month, about 60,000 loans flow through our shop, whether it's due diligence,
07:38valuations, fulfillment, and it's all super document heavy. Uh, so for us getting the document,
07:44kind of assembling the file, understanding what the documents are and extrapolating, you know, the,
07:47the correct data and being able to deploy rules against that data is where we find, you know,
07:52the highest benefit. If I'm a mortgage originator and you know, I'm building a loan application,
07:56I can take a borrower's ID and a pay stub and these types of things and allow me to
07:59fill out a loan application faster. You know, there's an opportunity there, uh, on the front
08:04end for those guys. If I'm a capital markets desk and I need more throughput, you get like
08:082000 loans that are being sold or securitized. I need that to go through cost efficiently and I
08:13need it to happen quickly. The, the trade cycles tend to be pretty tight. Uh, so for us, the,
08:18again, the ability to classify documents or extract data and then run rules against it
08:22is where we see the, the biggest lift in our operations. Uh, and we use the loan DNA technology
08:28to take that, those tools that we've built purpose purposefully in house and offer them to clients as
08:33a service. Now with fraud and security remaining major concerns, how can mortgage leaders build
08:40processes that are both more secure and efficient? More secure and efficient. I think there's,
08:46I think there's a lot to come in AI in that regard. I think we're, where we see challenges from
08:51time
08:51to time is our clients are, while they'll ask us, they approach us and say, Hey, are you going to
08:56use
08:56AI to support us? And these fraud reviews at this, at the next turn, it might be their governance teams
09:02or their security teams saying, hold on, you can't, you know, use AI to process our documents.
09:06I think we'll ultimately get there. Uh, there's a lot of information underlying a mortgage loan file,
09:11you know, again, from the borrower's ID to the pay stubs, the disclosures that they make,
09:14um, you know, public records and all these types of things. And I think there's going to be a
09:19tremendous amount of opportunity for cross-reference or cross-mortgage documents that should uncover
09:23things that otherwise would be very difficult for a human reviewer to see, right? So we're doing that
09:29today. We'll, we'll take pay stub data. We'll match it to bank statement records and make sure,
09:34you know, cash flows are coming in and out. You can see the movement of cash. I think that's one
09:38of the,
09:38one of the key areas where we're deploying it right now. I think there's a lot of opportunity for clients
09:42to do that also.
09:43Now looking ahead, what opportunities do you see for mortgage companies that are willing to rethink
09:48how they operate and invest in the next generation of technology?
09:53With, you know, volumes being what they are, I think you need to scale and you need to,
09:57well, you need to be able to pivot and scale and move and with the ebb and flows of like
10:01volume cycles
10:02and interest rates and these types of things. And I talked a little bit earlier about, you know,
10:06kind of the variable nature of volume in the space. And I think you have to have an operating model
10:10that can
10:10sustain and through your peaks without adding a ton of human capital each time, like one of these
10:15volume strikes happen. So picking the technology partners and resources that can help you do that
10:21without, you know, having to hire a hundred people overnight is I think where, you know,
10:25that's going to set the guys apart that really can respond the next time the next volume cycle is
10:30there. It's really some, we've seen some clients build some nifty things, I guess, internally,
10:36but there's a vendor network out there, loan DNA and a number of other like high quality vendors
10:40in the market that can really help these guys keep that cost of human capital at a reasonable
10:46level and be able to respond when volume ticks back up. Rudy, thank you for sitting down with me
10:53again. It's always a pleasure getting to talk to you and thank you for walking us through the
10:57opportunities to lean in, AI adoption and what Consolidated Analytics is up to. Always a pleasure to be with HousingWare.
11:06so
11:10you
