Transcription:
Julian Hebron: (
All right. Hello everybody, and welcome to day two of the Digital Mortgage Conference 2021, and welcome to "Transforming the Mortgage Lending Process through Data and AI," where we're gonna get into the trenches on the role of digital for loan officers, fulfillment teams and the consumers they serve. So we've got some of the best in fintech and the lending trenches today to run this down. So I'm gonna do a couple intros and then we're gonna kick it off. We got a lot to do in 30 minutes, so I will call each of you out as I introduce you and we'll say hey, and then we'll get going. So first we have Rajesh Bhat he's the CEO and co-founder of Roostify, which is a pioneer in the POS space that powers digital experiences for consumers and lenders, and a leader in the nuance that it takes to get this right. So good morning, Rajesh.
Rajesh Bhat: (
Good morning. Hi.
Julian Hebron: (
All right. And Rob Clennan, he is the president of Mortgage Solutions Financial, and he's been with MSF for his whole 25-year career, starting as a telemarketer, then a loan officer, then a branch manager, then head of production and now president. So a true, from-the-trenches leader which is gonna be one of the key themes today. We're talking tech, but we're talking about it from the trenches. So, hey, Rob, just sound off to make sure we can hear you.
Rob Clennan: (
Hey, good morning, everybody.
Julian Hebron: (
All right, good morning. And then lastly we have Jesse Lopez, VP of training and development, actually, as you will see, VP of everything, also from Mortgage Solutions Financial. He started 20 years ago as an LO with Wells Fargo, he's led sales and training teams, a few independent mortgage lenders, and for almost the past 11 years has performed these same roles for MSF. So hey Jesse, good morning.
Jesse Lopez: (
Hey, good morning.
Julian Hebron: (
All right. I like the haircut.
Jesse Lopez: (
Yeah. Thanks man.
Julian Hebron: (
Yeah, absolutely.
Jesse Lopez: (
For you, sir, just for you.
Julian Hebron: (
All right. So here's what we're gonna do everybody. We're gonna start with digital perspectives and loan officer priorities. So we're going to get into, everything we're gonna do today is digital, like I said, but it's about execution because that's what's all about in mortgage. So we're gonna talk about daily production and the changing market we're heading into, but we're going to do some setups first from MSF and then from Roostify. So Rob, let's start with you because I just love your resume. It's super rare that people go with one firm their whole career. It's always shorter tenures all the time, and this says a lot about the stability of not just you and your career choices, but MSF. So I wanna start with a two-part question for you. Can you just kind of quickly break down the profile of MSF for everyone? Team size, volume, channels, primary loan types or customer types, and what you believe is your special sauce? Because I'm sure you have an answer for that. And then number two, and I'll remind you of this after we do number one, if you want, but I want to, when it comes to digital experiences for loan officers and your customers, what's your biggest absolute must for 22?
Rob Clennan: (
So we are 598 employees, about 260 originators, we are 57 branches, I believe in 48 states. We do a little over $3 billion in volume. What we focus on primarily are those VA (loans). VA is our, it's what we focus on 100% to be totally honest. We do have the government loans. We do FHA. That's now the big part of our business, and conventional. But everybody does conventional and that's a really big pond to fish in. The kind of loans that we do, other people aren't willing to do. They're capable of doing it, they choose not to. So we choose to do it and we choose to serve the underserved. We choose to serve the vets who other people don't want or necessarily choose to serve. So that's who we are in a nutshell.
Julian Hebron: (
Got it. And what would you say if you have a, I know from talking to you, if you don't offer it, I might offer it, but what would you say is the special sauce of your firm and/or your team?
Rob Clennan: (
It's tenure. Jesse's been here 11 years. Our average tenure, Jesse, is close to five-and-a-half years, right?
Jesse Lopez: (
Yeah, branch managers, yeah.
Rob Clennan: (
I mean, our people stick around because we connect. So our number one core value is connection. The thing that matters to us most is connection. We don't email, we pick up the phone. If there's more than three emails back and forth, the very next step is a phone call. So we try and keep the human element engaged at all times, especially in this era where everybody is spread across the country.
Julian Hebron: (
Yeah. I think that that is, you and I were talking about just full market cycles, right? And how people, you call it binge and purge. Everyone has a slightly different kind of version of it, but it's like, you hire at scale and then you let go at scale. And there are certain players in the space that do very well, and it's always about market share and scale and all these things. I'm like you, I have definitely a full market cycle mindset. And so I just love the fact that not only does that play well for teams and people and being able to plan your career around a firm, but I think it also just goes to how you care for consumers. And we're gonna bring the digital element in here as we move through today. But with that, let's come back to this question. So when you think of digital and technology in your tech stack, as a whole, I know this is, it's hard to narrow it down, but if you could, is there like a certain single absolute must that you have for 22 when it comes to your digital strategy?
Rob Clennan: (
Absolutely. And you know, it's interesting. I wrote an article two years ago called. "The Death of the Mortgage Professional." And it's because technology is working really hard to replace people in the industry. And I'm 100% against technology replacing people. I'm for technology aiding people. So the reason why Roostify works so well for us and Beyond, specifically, is it eliminates touchpoints. It doesn't eliminate people. It eliminates touchpoints.
Julian Hebron: (
Yeah.
Rob Clennan: (
And 68% of the business that we do is manual, so manually underwritten VA and FHA files. So our underwriters are our highest-paid resource in operations. So you eliminate touchpoints, you maximize efficiency and Beyond allows us to do that.
Julian Hebron: (
And I just wanna underline a point for everybody that you just made that underwriters are your highest-paid resource. This is so important. I also kind of came up in production in a different, at a different part of the client spectrum, more like prime jumbo, but ultra-entrepreneurial jumbo in the Bay Area where, you know, dealing with extremely complex tax returns, entrepreneurs, inconsistent income, all that kind of stuff. So I'm with you in terms of, I would add one thing, processing and underwriting are both, you know, underwriter quality processors and just lights-out underwriters were always a big priority. So I completely agree with you on that. So thank you for that setup. And I think that everyone should take away the other phrase that I think you said, "tech eliminates touchpoints." I just love that. So I want to underline that for people as well.
Julian Hebron: (
So with that Rajesh, let's set up your digital perspective for everyone as well. And I'm gonna kind of do, I'm gonna throw out a few questions and we can unpack them. I know our industry-shared language is POS for this category. I said it in the intro. And we're way past this moniker now. So I hope you can kind of talk about that, because I know you have an opinion on it and do so by also then kind of providing your definition of digital mortgages circa 2022. So let's start with those and then I'll throw in one more.
Rajesh Bhat: (
Yeah, sure. Well, I think POS is still a very helpful term for the industry in that it helps lenders understand how technologies like ours fit into their ecosystem. And it's, you know, it aligns tightly with a loan origination system, it's not replacing a loan origination system. And where I think a POS five years ago was an application intake experience, it's now evolved to really being a sophisticated filter for information that's going into the loan origination system and providing the ability to really, you know, cleanse the data and create a loan file that's much more actionable and thereby speeding up the process and closing more loans. So it's helpful in that respect.
Rajesh Bhat: (
But certainly today solutions like ours are doing a lot more. We're supporting document intake, we're supporting decisioning, we're supporting credit pulls and product and pricing pulls, et cetera. And so we're supporting driving more of the experience to the loan officer, to the consumer, and creating more efficiency that way. So I think, you know, in that sense, what's been digitized over the last few years has, it's increased in capability, but I don't think it's necessarily increased in scale across the industry. So lenders today may have the ability to be able to do certain things digitally from product and pricing to decisioning, to eClosing. But we're still not seeing it scale.
Rajesh Bhat: (
And what I love about Mortgage Solutions Financial is they're handling some of the most complex loans that you'll see originated in the country and working, you know, it's great for companies like us to work with them because the way we think the transformation scales across the industry is by solving for those complex cases. And it's very easy to solve for conventional loans that are, the conventional W2 for a conforming loan, but that's not going to enable the industry as a whole to drive the change. And it's not gonna create more accessibility for the homebuying population. And so when we think about digital in 2022, it's really about being able to go end-to-end, digitize the workflow and start to drive some of the real big opportunities around automation that we'll talk about later hopefully.
Julian Hebron: (
Yeah. And so just before we go on to Jesse, one more thing. So if we're saying, you know, the POS vision started with the take-in, the app, and now we're thinking app-to-close and that experience for both borrowers and LOs, again, this is tough to narrow it down. But if there's one thing about that app-to-close process for either LOs or consumers that you're most excited about for 2022, what is it?
Rajesh Bhat: (
Well it's really about driving more of the processing to, really automating more of that processing work so that we can create the ability for the loan officer and the consumer to handle more of that experience. And we can drive out the cost related to that, drive down the cost of the loan to our customers and ultimately to the consumer and just make that process go a lot easier and faster.
Julian Hebron: (
Got it. Thank you. And yes, we are gonna go deeper, but let's go to Jesse now. Again, VP of everything, Jesse and I joked about this, he joked to me about it, but seriously, Jesse, I love it because all the great independent lenders are great because of utility players of which you are one of the key ones. And that's just something that's cool. So let's get to the source of MSF's success, which is not just all employees, but let's zoom in on loan officers for a second. I think that this is a big part of your purview. So I want to kind of run down a few questions with you. Obviously, Rob, jump in here, if you want to. But there's all kinds of LO styles. I think that MSF does have a kind of a particular mold of LO, and if you do, what is it, is it kind of like median producers that are just kind of consistent and steady? Is it super producers, producing managers, BMs who are also running their own production and building little machines around them? If there's a signature for MSF, what is it?
Jesse Lopez: (
Yeah it's that machine right? It's that producing manager or that, you know, two-to-three person origination team surrounded by a support staff. That's just where we seem to find the most success. And you get those, I guess if you will, mega producers and you build a good machine around them and their potential is absolutely unlimited. And we see that on a daily basis.
Julian Hebron: (
And I would assume, just going from my own experience, but I'm guessing it's the same with you is that that then starts to incubate the type of, like an LOA might then be somebody who's like, all right, I want to be the front person now, and then they rise up. But they've been molded in that mold. Is that a fair statement?
Jesse Lopez: (
Absolutely. Yeah, absolutely. It's actually all about the organic growth, right? We have the producers, we have the people that are gonna always be in that top 10%, 20% of production and performance. But everybody's gonna age out at some point in time, even myself and Rob and all of us, right? And so it's constantly looking for those future leaders, identifying people and, you know, I absolutely love finding people that have no experience and we plug them into these machines and they just become, you know, huge producers. I mean, we've got a young lady that's been in this industry, you know, just roughly three years and the volume and what she does is just absolutely incredible.
Julian Hebron: (
Yeah. And so on that note, let's just see if I can connect that to another question I want to ask, which was, whether it's using that person as an example, or just in general, what do your LOs care most about when it comes to the technology that you provide for them? Because I want to build off of Rob's comment too of like, this is designed to enhance them and superpower them, right? So what parts of your tech stack are your LOs, you know, what do they care about most?
Jesse Lopez: (
Ease of use really. It's convenience for their clients as well as for them. The real goal is, you know, we need the sales people out there shaking trees and feeding the funnel. I don't need them chasing W2s, you know, I don't need them chasing the pay stubs. I don't need them having to pilfer through 100 pages of tax returns to find those little anomalies. And that's really where the tech comes into play. And one of the things I love about Beyond and Roostify, what they do for us right, is being able to give that loan officer or production assistant, or anybody in those roles that ability to quickly identify anomalies, to identify any issues, right? And compare that application through their interview process with those documents in a much faster fashion.
Julian Hebron: (
Got it. Yeah. And we're gonna dive super deep into that. I wanna do one more setup thing though, just because I think that again, culturally zooming in on how you do what you do is really interesting. So last one, because it's rare for a shop that, I think you guys are like $3 billion, right? For a $3 billion shop to have, and I get that that's why it's VP of everything, beacuse it's not just training and development. But to even have that role of a firm, of a $3 billion shop is incredibly cool. So what is your playbook for the training and onboarding of LOs, and based on your LO profile, are you really pushing the LOs on tech adoption? Or are you, and I think you kind of touched on this, are you focusing more on the assistants and the processors to kind of drive all that? Or is it kind of all of the above?
Jesse Lopez: (
All of the above. I think I put a little more focus on the assistants, and delivering a cleaner product to the processors and the underwriting of course, to reduce those touchpoints and again, just get a cleaner loan. But again, the LOs, the push is letting them do more with the same amount of effort or less effort, right? Again, I need them out selling, I need them out shaking trees and developing relationships, not chasing down the documents.
Julian Hebron: (
Yeah. All right, well, let's go into it like deeply now. So, I want to kind of build off of Rob's statement before, of this idea of super-powering LOs. And now we're gonna get to the core of this, super-powering LOs with data and AI. And Rajesh, I want to go back to you on this one. So, you've said to me, and publicly, that your goal is to make sure the loan app becomes digital regardless of whether the consumer or the LO prefers or only has paper. So you announced this vision earlier this year, I think around the first quarter. And I just hope that you can kind of break that down, tell everyone what it's called and what it does. And then we're gonna go back to Rob and Jesse and with the three of you, we're really gonna kind of unpack it, but use the MSF context to do that.
Rajesh Bhat: (
Yeah, so we announced last year, a partnership with Google Cloud where we entered into strategic partnership to begin building out models to support the machine reading of the mortgage document set. The first product that we built out, leveraging that capability is called Roostify Document Intelligence. It's a service that we deliver as a service on an API gateway. And the service effectively allows us to take the documents that we ingest on our platform. And we ingest about a million, north of, well over a million a month, and machine read those documents to be able to, on a real-time basis, classify them, identify what type of document it is, validate that it is a machine readable document, and then do some more sophisticated legwork on it. One thing is being able to take the document once classified, match it to the condition that we've established based on the loan file and our interfacing with a loan origination system and determine that the condition matches the document, and then take action on that condition straightaway. And the other thing is extract the data. And with the data extracted, we can do a couple of things. One is we can verify the data and just make sure that the data is complete. And then the second thing is we can actually compare the data to data coming in from other sources that might be consumer self-attested information coming through a loan application. It might be data that we pull in from a third-party system, like an automated asset verification system, or even a servicing platform or a bank depository platform.
Rajesh Bhat: (
And so the whole goal is to do this in a manner where there's confidence in the data coming through and the data's coming through on a real-time basis. So one of the important things here is the concept of removing the human from the loop as we're doing this. So there are a lot of capabilities out there that include a human in the loop to get to 100% confidence level. For us getting north of 90% is super important, but removing the human from the loop is even more important because there's a time value to the data. The ability to be able to do this at the exact instance where the consumer's providing the documentation affords someone like MSF the ability to really one, drive more of the action to the consumer, to the loan officer on a real time basis.
Rajesh Bhat: (
But it also allows for that loan file to be even more pristine before it goes into the loan origination system and become more actionable. And obviously you're taking workload off of the processor's or underwriter's back as well. So we delivered this capability as a service earlier this year and it's available to any third party developers or service providers out there that aren't using our core platform. MSF of course is using our core platform. And so the capability that we rolled out on top of this is called Beyond. And Beyond leverages our own user experience, which is, you know, all good things that, you know, come with being able to deliver the user experience. It's ADA compliant. It's customizable, certain fields, et cetera. And it provides an interface for both the consumer and the loan officer to be able to run these capabilities.
Rajesh Bhat: (
And it provides the administrators within Roostify the ability to really manage the competence levels for not just different document types, but even the different fields within different documents, so that you can really control the quality of the information that's being extracted out. And you can determine ultimately what your workflow should be. So the goal is really to allow our customers to control whether or not they focus on using this technology for handling exceptions or using it as an exception itself as they get more comfortable with really stepping into this new world where we're machine reading the documents real time. And this is really to drive the scale of this sort of capability so that MSF, you know, as an example can get very comfortable with running all their docs through the capability and transform the way they do things.
Julian Hebron: (
Yeah, absolutely. And so Jesse, if I could, or Rob, we can go back and forth, but Jesse, I feel like I want to maybe start with you just because I wanna talk about like rollout and getting this in the hands of your production teams. Forgive me if this is an overly open-ended question, but, where are you on the rollout and what is the methodology? Do you sit down with the assistants and the processors and start getting them into it? Or are they all astute enough to they're like, "Hey, we're already on Roostify. This is just an add-on, and it's killer and it enhances what you are already doing." Like, how are you deploying it?
Jesse Lopez: (
Yeah, it's a bit of a process, right? There's some adoption, things of that nature. And so right now, we're, you know, I would say for lack of better "pilot," we have nine of our producing branches that are utilizing Beyond over the last couple weeks now. We're pretty close to being able to roll it out 100% to the entire production team. But truly the focus has been more on those assistants, because those are the ones I want really handling the documents. That's kind of the gatekeeper before that loan goes into processing. So again, the LOs can be out there shaking those trees. It's just ongoing and their feedback is invaluable. And I think that's a piece and goes back to even our core value of connection and having the relationship where people are actually comfortable, you know, sharing, "Hey, this is working great." You know, "I don't like this, what can we do about it?" And it goes back in a testament to the partnership with Roostify that we can provide that feedback. And as we said earlier, anybody can build a platform for the vanilla conventional loan but dealing with the type of borrowers that we work with, the demographics, you know, the locations of most of our branches, there's a little more handholding that needs to be happening with those borrowers. And Beyond again, allows us to do that a little bit more digitally, right? And allow the focus for the actual employee to be elsewhere.
Julian Hebron: (
I'm glad you underlined that because I couldn't agree more, is that the solution that caters as to, you know, Rajesh said edge cases, I would take it a step further to your point that they're not edge cases. I mean, these are, you know, FHA, VA, there's a few, but let's just focus on that since it's your core business. But I mean, these are, I've always said that they're the most technical files, despite what, you know, if you will, the headlines would say about loan quality or this-or-that, they're the hardest ones to do. And so to have the digital technology to help. So with that said, if I could, what lift so far, if it's not too early, what lift is Roostify Beyond giving to your prior-to-submission files? You said that it's, you know, underwriters can recognize a Roostify versus an LOS-generated file for its superior quality. But also, you know, so what lift are you getting there on prior-to-submission on the packaging of the file, and then, also, are you getting any sort of efficiencies once the loan's approved back-and-forth between processing and underwriting?
Jesse Lopez: (
Yeah, it's a little early in that. We are still working on that bidirectional push from our LOS back to Roostify, which is, once we get that rolling, you know, that'll be a significant lift, I think, for the assistants as well as the borrowers and LOs to see that real-time data coming back into that single-source system. As far as the quality goes, yeah, as you mentioned, you know, our processors, especially, they can visually tell the quality difference of an application that comes through the digital platform versus one that's manually keyed in by that loan originator. That's a huge lift. You know, I think a lift also for the loan officers, right? I mean, we still, you know, really push, you gotta do an interview, strong interviews build lasting relationships. We need to extract that information from the borrower through conversation, but the digital platform gives them a lift and that borrower, they know how to spell their name, they know their Social Security number, let them input that data versus having to ask those questions repeatedly, you know. And then now when the LO gets it, now they're actually validating information and they're not a data entry clerk into the LOS, right? They're truly validating and still going through that interview process. But it gives them a lift and now they can spend, you know, 20, 30 minutes on a file versus maybe two or three hours on a file upfront getting it ready for submission.
Julian Hebron: (
But to be clear, and this is just a really detailed question, but if a borrower is putting in a name that, "Hey, this is what I go by, this is how I fill it out." And then their pay stubs come in and the name is more official and different. This is part of what the system, what Beyond is doing, is showing your processor, the differences in helping them reconcile that digitally, immediately, correct?
Jesse Lopez: (
Absolutely. Yeah. That's a huge lift, and I'm glad you actually mentioned that. Because that is a significant lift for us, right? As soon as that pay stub comes in, we see if their Social Security number doesn't match. We see if they cloud the application with first name, middle, last, versus their pay stub just has first and last, right? We're seeing those anomalies and not only does the system recognize that when we view that document, it actually highlights those anomalies on the physical document. So that, you know, again, we really love to employ people that are new to the industry to train. And so it's also a great training tool. And it gives us a lift there because you know, now they can move quicker and, I'm not gonna say it necessarily shortens the learning curve, but they don't make some of those mistakes that they might, if they didn't have that platform, right? And they're gonna miss things more frequently, if they can't see it, you know. Not only that they learn where they're at on the document, when something's highlighted on the tax return, now they learn where that information is located on a tax form, is a good example.
Julian Hebron: (
That's a great point. I love that point. All right. So we've got a few minutes left and I want to thank you for getting a little bit forensic. It's hard to do in 30 minutes, but I think it's important to point out some of what this technology's actually doing on the ground. And I think we've gotten to some of that. I wanna do, let's kind of do quick hits from Rajesh first and then Rob too on sort of this vision. So Beyond is bridging this digital paper gap, and we just talked a little bit about how, and this is just, you know, we talk about AI, but this is the pragmatic conversation about how. So the industry launched D1C (Day 1 Certainty) as a relief for direct source data, direct to/from source data as a way to drive innovation. That was 2016. It's been a while. So it really took until now to start solving this digital and paper gap. So I hope Rajesh, and then Rob, you can kind of close us out quickly, like maybe, you know, a minute each on how do we speed up now that this data, AI, data extraction technology is finally able to run at scale? And are there other hurdles that you see? So Rajesh and then Rob to close it out.
Rajesh Bhat: (
Yeah. D1C was exciting five years ago. I think the general consensus is that it hasn't seen the adoption that everyone had hoped for, to drive the sort of transformation that everyone wanted to achieve. And so, you know, we ourselves in our platform see roughly, you know, 30% to 40% conversion, depending on the customer of these, of the D1C-type capabilities when they're enabled for asset verification, employment, income verification. Which is great, it's a step forward, but until you get to, you know, north of 80%, 90%, you're not gonna be able to change your overall process, the way you originate loans. Because you're handling those D1C files as exceptions rather than rules and people are running D1C but they're still doing the manual underwrite because they want that fungibility. And they don't really understand just how these loans operate in the secondary market otherwise.
Rajesh Bhat: (
And so, you know, we saw this ourselves. I was, I think, you know, seven years ago, probably one of the people that were knocking at the door of Equifax and TurboTax and Fannie talking about this concept. And I realized that, you know, we were not going to be able to see the adoption at the pace that we needed to realize it. So getting the, being able to extract data from the documents, do so in a confidence level that is very high, where you can trust the data that's coming out of the documents, in order to action it, is very important. And so we are now engaged with the GSEs on being able to accept this as a capability that they can ultimately trust in their own systems and to even look at the processes that it's automating. So being able to automate the matching for example, and what actions are taken from that and understand how that can receive some sort of certification as well. Because all this drives towards larger market acceptance. And obviously, we're still seeing, particularly for the loans that are key-focused to Fannie and Freddie, the LMI loans, the FHA, the VA loans, we're still seeing these be very doc-heavy. And so it's going to be something that we need to solve for.
Rajesh Bhat: (
And, you know, what's important for us is seeing the interest from partners like Google Cloud to invest in this. Because it takes a lot of brains and computing power to be able to perfect this. And perfecting it is what's ultimately going to engender more trust in the secondary market. And as you see that trust developing the secondary market with Fannie and Freddie and other larger investors, you're gonna be able to then start to change the behavior upstream with the originators. And so that's been very important for us. And so that's where we think the opportunity is going to be, to be able to drive the larger transformation that Fannie and Freddie had in mind in 2016 when they launched D1C.
Julian Hebron: (
Thank you, Rajesh. And Rob, if you would bring us home, we're right down to the wire, but if you are always succinct and pithy with your comments. So if you wanna bring us home with some closing remarks about your digital vision, or where you see MSF going in 22, please bring us home.
Rob Clennan: (
Yeah, sure. I see in 2022 that we're going to have Roostify fully implemented and Beyond fully implemented, and it'll keep our originators out of the LOS and out in the field doing what they need to do. If people are out of the LOS, that means LOS runs faster and everything is just more efficient. But what I see this doing for us long term is fixing the high touch nature of our type of business. And we're not gonna deviate from our type of business. Our success is derived from 100% of the time, staying true to who we are and never deviating from our course. So since we don't deviate from our course, since we don't take shortcuts and we don't take the easy way out, this is gonna help us have less touchpoints and create a more efficient process for both our underwriters, processors, borrowers and loan officers.
Julian Hebron: (
So here it is, everybody, AI can fix and enhance the high-touch nature of your business. How's that for one to digest? So thank you, Rob for that. And Jesse and Rajesh, this one's been a lot of fun. So, to be continued in 22.
Jesse Lopez: (
Excellent.
Rob Clennan: (
Thanks.