Is New Servicing Tech Reinventing The Originations Game?

Is the originations game being reinvented by servicing tech? If not, should it be? We know originations acquire customers and servicing retains them – first through engagement, and ultimately through new refi or purchase loans. So why isn't originations and servicing tech more unified? What does it take to get there? And which lenders/servicers are leading the way on this unified customer service model? Come learn from them. 


Transcription:

Chris Bixby (00:09):

Hi everyone. My name's Chris Bixby. I'm a Managing Director of Venture Capital Strategies at Rice Park Capital. We're a private investment firm out of Minneapolis that invest in a variety of asset classes including agency MSR, private lending, and then direct investing in venture capital and companies like Capacity and Candor, which is an automated underwrite writing technology. I'm here today to talk about what may be a provocative subject, which I'm hoping to draw out of both of these very nice guys, but the tension between origination and servicing portfolios, especially as we go through a potential refi boom that's impending, especially as rates are starting to drop and the Fed may continue to reduce rates over the next couple of quarters. So I'm going to let these two introduce themselves and then we'll jump into it. Jeff, maybe you would start.

Jeff Bode (01:04):

Hi, I'm Jeff Bode with Click n' Close Mortgage. Also have a technology group Mortgage Machine Services. Click n' Close. Started in 1959 since I acquired the company in 2010. We're kind of a niche player in the market. We do the Section 180 4 program. We're the largest lender in that space. We have about 54% market share in that market. We also do a down payment assistance program.

Chris Bixby (01:34):

Sorry, I'm just going to pause a second. The chatter in the back's actually fairly distracting in the front, so it's a fairly loud room, so appreciate it. Sorry, Jeff.

Jeff Bode (01:44):

We do a down payment assistance program where we originate the first mortgage loan and the Rosebud Economic Development Corporation does a second mortgage, which allows for the consumer to get into a home with no down payment, and that's been fairly popular, been a big part of our product stack for the past two years. We service about 9,000 loans and the strategy is not necessarily to build up servicing, it's just where's the highest rate of return for us. We've got the Ying and the Yang and the push and the pull of the originators that think every loan that we have should be refinanced and the servicing that thinks we need to retain every loan we've got.

Paul Akinmade (02:30):

Great. Sure. Paul, Paul Akinmade Chief Strategy Officer for CMG Financial, within CMG marketing, BI technology, and bringing servicing and house roll up. To me, we have currently two sub servicers now with about 230,000 loans under management, and we're looking at ways to change the consumer experience for servicing while at the same time considering how do we service that customer for life. So I think that's probably where our tech focus is,

Chris Bixby (02:57):

And both of these gentlemen are very, very humble. I've heard good things. I did not know them before the conference, but I've heard they're both very kind people, knowledgeable and Jeff knows how to make money is what I've heard from a variety of different people. So Jeff, let's start with you on that one. How have you gone through since you've acquired the business, how have you gone through the decision around servicing your book in your portfolio versus allowing someone else to do it or selling your MSR? How do you make those decisions

Jeff Bode (03:26):

As opposed to sub-servicing? That was a no-brainer for us. Once we eclipsed I think the 5 billion mark on our servicing portfolio, we decided to retain it. We have a high degree, a high percentage of government loans and our originations, we originate 80% government loans, so we are kind of a unicorn that we don't do that much conventional business and those loans default at a higher rate, and we found that when we used a subservicer that the subservicer just didn't give it the same attention that we did and the loans were just didn't get through the Lost Met solution as fast and just issues with dealing with Jenny May. We had to fix that. So bringing it in-house was the way to do it. As far as which loans do we retain versus which ones we sell service and release, it's purely a function of internal rate of return on what my investment is and typically, unless there is a need, like the Section 180 4 business that we have a commitment to the tribes that we're going to retain those loans or the down payment assistance where we're servicing both the first and the second, pretty much any loan we've got, if we need a little more cash, we will sell the servicing off and we've kind of been in the market to sell servicing the past year and probably going to go back in the spring.

Chris Bixby (04:39):

Great. Paul, maybe the same question for you as you sit in your seat, which spans a lot within CMG, how have you seen the decisions being made around servicing, and maybe you can talk a little bit about what your strategy is today and where it's going to go.

Paul Akinmade (04:54):

Sure. So I think it's opportunistic. Again, depending on where the loan has come from, we acquire service and acquire MSR, but I think from a strategy perspective, when I look at what we're going to do going forward, the areas I've asked to focus on outside of just the internal rate of return is making sure that we account for tangible relationships. Because again, every loan officer wants to maintain a relationship with their customer, and you have to account for the fiduciary duty to the MSR owner that you just sold your servicing to. So I think from our perspective, we like having servicing. We're going to continue to grow that, but we're going to be opportunistic about what we keep versus what we sell.

Chris Bixby (05:28):

Yeah. How about we've seen in the market, Wells Fargo portfolio sold off last year, beginning of this year we saw Flagstar kind of go through off a sale to Mr. Cooper. If you're sitting both in your shoes but also in the shoes of kind of an IMB owner, how are you thinking about what that's going to do to the market and how that's going to impact either your retention strategies, the ability to either sell into the market, but what you've seen with some of these large consolidations of MSR pools?

Jeff Bode (05:58):

I noticed that the price really went down when Wells Fargo came out with the announcement that they were going to sell a lot of their servicing. They didn't dump it all at once, so it didn't just beat the market up pretty badly, but we did find that it was a little soft for a period of time when Wells Fargo made that announcement. We've since then had a sale of some servicing and thought we did fairly well with it. But I think another thing that's probably going to happen is now that Freddie Mac has got their second mortgage program, that's going to lead to a great deal of opportunities to refinance some of those Freddie Mac loans. So that's probably going to be a big value if you've got a Freddie portfolio, which we have of our small conventional piece, Freddie Mac is a larger part, so I think that's going to be of value too.

Chris Bixby (06:42):

Great.

Paul Akinmade (06:44):

Yeah, I think on my end, I really rely on our capital markets group where I don't have oversight of when we're deciding and what to sell. I think, as I said, the big thing that I focus on is trying to steer a couple things around what types of loans are we selling, where I think there's opportunistic places for us to engage as well as how we structure our contracts, because again, what is our outreach strategy allowed to be?

Chris Bixby (07:08):

Yeah. So Paul, I'll go to you and maybe you can share a little bit more about how you're thinking about servicing Sub-servicing and kind of this future strategy. Maybe just pause there and let you share just a little bit more about how you're thinking about that, and then I'd like to go into a little bit more, we're at Digital Mortgage, people are kind of excited to talk about the tech side of it as well. And so maybe as a fall on we'll talk a little bit more of how you think about your technology in this transition.

Paul Akinmade (07:33):

Yeah, sure. So I think when COVID happened and one of our sub-services is lar, so obviously there's a big strain on the subserving industry. At that time, we really were at a position where we were getting to critical mass where there's a conversation about is it more economically advantageous for us to take care of the customer ourself just from a financial perspective. And then we started looking at how our customers were being treated and it was out of sight of our control. We really wanted to step in into trying to figure out how do we do that. So from my perspective, I try to align with and find the best and brightest, and we were very fortunate with getting to work with Courtney Thompson over at Sage, and at that time she came on board and she's building our in-house servicing. So that's going to be leveraging not the same tools that everybody else has.

(08:15):

I think what we're trying to do is do something a little bit different where we can self-manage and allow the consumer to self-regulate some of the things that they normally have to have a human intervention for, which drives up the labor cost. So we're using components of Sagent, partnering with Aspen Grower BPN, using Salesforce for the customer service platform, which is also on the same stack as Consumer Direct. We're building 'EM unified so we can pass data back and forth. And then we're working with Industry Vault and having them develop some of the normalization, normalization practices and strategies to make sure all of our data is accessible while building out our own inside of Snowflake. That's kind of the stack that we're looking at as we're approaching this. We should be ready to start taking flow business in April and should hopefully have all of our portfolios in-house by end of summer next year.

Chris Bixby (08:59):

Great. Jeff, you've been on Sagent for your portfolio in terms of the MSR and servicing that you retained, any advice you can give to Paul as you go through this process or anything that you've learned kind of from that relationship and servicing it with Sagent as a platform and a partner?

Jeff Bode (09:16):

We have been pleased with our relationship with Sage, and I think they've really made a commitment to move forward on the technology that I think that was the right decision, and certainly pulling away from the sub-servicer was just, it was painful when we did it. I mean, I don't think I had a worse day in my professional career than the day we brought our servicing. We thought we were prepared for everything, and then a flu virus broke out and half of our customer service relationships were out the day with the loans came in. It couldn't have gone any worse, but the sub-servicer had actually made a mistake on the goodbye letter and it truncated at 20,000 loans. So out of all of our loans, we had only half of them came in at the right time and we were going to have to backtrack and go backwards on that, but we move forward, be sure you're ready day one because we weren't, and I think you're smart for looking at your data to make sure it lines up with your LOS. We have our own LOS, so we control that data and it seems like the servicing systems, I think Sagent moving forward on that. But historically it was tough to get data out of there, and I think it is now make sure you can see a snapshot of what it's supposed to look like at any given time.

Chris Bixby (10:38):

So if it doesn't go well, which none of us are hoping for, don't blame Courtney is what I'm hearing. It's a process that everyone goes through in terms of taking it in-house.

Paul Akinmade (10:48):

I think Mike Tyson said it best. Everyone has a plan to get punched in the face.

Chris Bixby (10:51):

Exactly.

Paul Akinmade (10:52):

I think we're taking steps to try to be ready for that. So again, you can simulate everything. Some of the things that I'm trying to do ahead of time is create a bifurcated environment where if I own identity management now I'm already taking part of the problem away now got them registered into my environment so I can pass them over sequentially so we can have a better control of the flow. So there's things that we've discussed and how to mitigate some of that risk and spread it out so that way it's not happening all at once and stage it and incrementally grow. So we can learn from those early mistakes, but no, it's, it's not going to be fun either way.

Chris Bixby (11:24):

Great. So I want to go back to a little bit of the tech stack kind of within Sage in and where they're delivering, where they're maybe under, and this is by the way, not a sage in advertisement and or not, but it's I think helpful to get into some of the details as we talked originators that go through this game, but the tension in the businesses and part of this new servicing tech, reinventing the originations game is going back to how these two systems talk together. And so from my relatively early career in mortgage, one of the points that we continue to see is systems of record on the origination side being a CRM and really LOS then on the servicing business, really being the servicing platform, whether it's MS Ps agent or another option. Paul, as you think about this process where you're both over kind of the front end and the back end in terms of the servicing and the originations game, how do you think about Sagent in that process and how do you think about either the communication between your platforms or setting it up so you can have a more seamless experience between origination and servicing?

Paul Akinmade (12:32):

Sure. So I think on the last presentation somebody mentioned the source of truth, and I think there's something to add to that, which is a source of fact because you can have systems override each other, so you're going to have to have a good data structure for normalization and a good process to manage how data flows in and out, and then a way to recall and look at the record of that. So you have to identify what is your fact and then of that facts, what is your truth? And that's going to change based on what your intent is and what system's looking at it. So when I look at Sagent, I'm not going to make Sagent my system of record completely. It is with respect to servicing, but we need data to live elsewhere because we needed to do other things. So that's how we're approaching it is making sure that we can create, I don't want to use the word data lake or data warehouse, but we're going to build an area that allows us to normalize the data, aggregate the data, and then utilize other applications that can access the data to deploy experiences that we need to happen, whether that's in lost mid for some levels of automation or whether it's for a consumer journey and so forth.

Chris Bixby (13:29):

Great. Jeff, you've been on Sagent, you've probably had the same challenges and opportunities. You mentioned analytics and data as one of the things that it sounded like you'd like more of. How do you navigate that in terms of your organization or how does your team navigate data across your platforms in the front end and the back?

Jeff Bode (13:50):

As mentioned, the servicing system is the system of record for servicing the loans. We do a pushback of certain elements of data back into mortgage machine, so we can market to those people from mortgage machine through CRMs, but we do use the servicing system and we don't mess with it. I mean, we just can't afford to corrupt that data. So I think it's pretty much the same solution that Paul has.

Chris Bixby (14:17):

Okay. We've talked a little bit about the front end journey. We had a chat this morning about the ECOs process and where that's shifted in your business over the last couple of years. Maybe that's a helpful piece to start. And then Paul, I'll go to you a little bit more, and again, connecting the dots between the origination side and the servicing side, but maybe Jeff, you want to talk a little bit more about that and Click n' Close?

Jeff Bode (14:39):

Yeah, the ECOs decision, we made it real early. We moved in 2015 to start e-closing and best decision I've ever made, it's a no brainer. We were able to stop a lot of errors that occurred, and I'm not going to blame the title company or our closers, but the problem was you didn't have one set of data or one set of information that the consumer had to sign. We would have maybe a tax return document that wasn't part of the information that went to the closing table. So the title company would get an email, say, these are the closing docs they'd printed out and they'd get those signed. But unbeknownst to them, we sent them an email saying, we need to get these tax returns signed as well. And they were busy with 30 other closings that they had to do, so they missed that.

(15:33):

So we found that we had to spend, so many people had to chase down those documents from the title company or the borrowers or whatever. Sometimes we just bypass the title company, we got information from the consumer, but instead we've decided with the ELOs, we'll just load that up in our platform where they sign the document so it can't be missed. So we were not putting the title company in peril by doing that and not putting ourselves in peril by missing those things at funding. So we know that whenever somebody goes into our signing room and every item is turned green, that means we can fund that loan because everything's executed to our desire and not only the title company pages stick together and sometimes they didn't get signatures. So weird things like that were problematic, and we found that the e-closing really saved us a lot of labor.

(16:24):

I made an acquisition in 2019 to buy a company that was about the same size, maybe a little bit smaller than ours. We had three people in post-closing, and our average loan sale date was in eight days. The company that we acquired had 26 people in post-closing, and their average turn time was 25 days and stuff stayed on the line so long and it was just a bit of a mess with that group. And we got to the ELO and things shaped up pretty fast. So that was a big deal. I know that Notarized did a study and said it was like a $440 savings to do e-closing versus paper closing, and we found that right off the bat and picked up some nickels and dimes on the way When Fannie and Freddie made changes on the no under occupied loan level price adjusters, we were able to do Enos and sell those the day of closing. So if it was the 30th of April, we delivered those loans in, and I think those adjustments were pretty significant about a point. So it was a pretty big lift for us to do that. And handling the e-note, we've never lost an e-note. We buy loans on a correspondent basis, and when that bridge went out in Philadelphia, we had 20 million worth of loans kind of stuck on our line because they couldn't find the notes and they were at the FedEx in Atlanta. So you never lose an e-note and it's always where it's supposed to be.

Chris Bixby (18:01):

Yeah, that's great. Paul, you've also gone through acquisitions or I think there's an S on that, but at least an acquisition that we talked about, and one of the things that allowed you to do is look at the tech stack in terms of what they had and then evaluate it compared to what you have on the origination side. Maybe just talk a little bit about the process of either the integration and or identifying POS and LOS and then we're going to finally connect the dots between origination and servicing and go to some of the tech companies that are even here at this conference.

Paul Akinmade (18:32):

Sure. So I think the acquisition you're talking about is when we acquired who's on Blue Sage, and I think the challenges, certain systems are good for certain companies that have certain proclivities around what type of services they offer. I think when I look at it, some systems are great if you're, I always describe us as the Cheesecake Factory versus Chipotle. So at Chipotle it's pick your one of three proteins, what your toppings are, call it a day and move out. For us, it's like what country do you want to be served first before you decide what you're going to eat? And I think with that, you need to have high configurability. So when we were doing that initial assessment, we looked at Blue Sage, we looked at what it would take to integrate just our products around things like our All-in-one, which is a first-lien HELOC.

(19:18):

And the amount of time it would take to do that made it completely untenable. So with the LOS that we have now Bite, it's our system of record. It's really highly configurable, and I built my own wrapper around it. We took some work that was done by our prior company that needed to be completely revamped, which it's been a fun long process, but our POS was going to be a single pane of glass between our CRM, our servicing, data origination data, and my goal is for it to have LOS parity so people don't need to go into the LOS. So that's really what our journey's been on from a tech stack perspective, and I think it's been one that we're starting to see the light at the end of the tunnel, and it was funny hearing the conversation earlier about what it's like at the beginning, and the beginning is not fun, but we are in a stage where we're having a lot of fun right now. We get to do a lot of fun things over the last year where we're able to create differentiators because we own the environment so we can own what LO experience looks like.

Chris Bixby (20:11):

Great. So Paul, we have some likely vendors in the room, technology companies. We have some that you may work with, some that you may work with in the future. Can you talk a little bit more around the technology that you've used or utilized specifically around the retrench strategy? So I want to start going into a little bit more. We've talked about servicing, we've talked about the platforms you've been on, how you evaluate servicing. Talked a little bit about your MSR strategy. I'm going to circle back on that a little bit later. And we've talked about the origination side. Now let's talk about the recapture side, the actual ability to recapture these loans. I guess what tech are you using? What do you want to use? And maybe you could start with what does a perfect journey look like for you guys to think about as you're starting to do this shift into owning your servicing portfolio? What do you want it to look like and then how do you use tech part of that process?

Paul Akinmade (21:07):

Well, I think when I'm looking at what the journey is, it's going to be what do I imagine what I would want it to be? For me, I don't want to have to go talk to somebody and say, Hey, what are today's rates? Where do I sit? I want a system to tell me that. I want it to tell me that, Hey, today if you're interested, today's rate based on this profile is X. And that for me would be something I would tie into that journey. I would like to be surprised by the fact that it says, oh, you have 30,000 revolving debt. Your score is a seven 40 if you try to HELOC. I think doing that programmatically and in a way where people can check back in is advantageous from an experience perspective. I think when I'm looking at the technology that sits out there, a lot of people are kind of attempting that.

(21:48):

Some are going partially into one of the segments, like a VM is the easiest one. What can you do with your equity, right? You've got home bought HomeIQ, which we have partnerships with both of them. We're starting up on with House Canary for a VM and home search. So that way I can show you properties that are like yours, but I'm doing that all custom built, but leveraging their technology. When I look at what I think is going to happen in the future, one of the things I mentioned earlier is I wanted to own identity of the consumer. So I want to own the gateway in which they go to servicing. Even if we weren't bringing servicing in house, I would want to own the place that they come into log to go into make their payment. If I own that, then I can direct them and I can create that environment to do what I want it to do.

(22:28):

Inside digital services and inside of care, you have limitations. Once you pass them in there, you can stick a banner there and sell one product, but you can't sell a dynamic price. You can't sell something that's unique to them. So you can't pass those variables into those systems easily. I think when I'm looking at the space that I carry and I control, I can make that whatever I want it to be. So what we did last year was we took the data sets of our servicing, we ran credit against them, which is not a cheap endeavor, but we figured it's going to get our money back. And I know as a loan officer, a pass loan officer, you may have a four and a half percent interest rate, but if you qualify for a six and you've got $60,000 in revolving bet and I can save you money, I've got an opportunity for a conversation.

(23:07):

I think the challenge is if you don't put all that data together, you have no way of instructing your LO to identify that. The thing that we've been testing over the last year is how do we automate that process? We've been testing it for our inactive los to make sure we get the content, the delivery system right, the frequency, so that way we don't oversaturate or oversaturate our portfolio pool. That's something that we're going to be turning on here in the next couple of months as we get ready, as Mr. George says, are you ready? Get ready for when the market does turn. So our LOs can say, okay, we have identified this customer. There's five opportunities. We're in the money. This is the sequence we're going to do it. Here's where the highest opportunity is. There's waiting on each one and they can opt in for that automation.

(23:44):

So it makes my LO seem like they're driving a great experience and if I can send them somewhere where they can self-manage and self update, now I'm giving that consumer more control. So I think that's what the experience should be. I don't think anyone has mastered that yet. I think a lot of people are getting close, but I think there's elements that are involved around pricing around credit that are tough to do from a vendor perspective to bring it all together. And there's a few out there that are getting close to it. So I think it's something that people should be striving towards, but how you integrate that is going to be interesting and it's going to be an effect of where you see your ability to retain customers.

Chris Bixby (24:17):

Yeah. Jeff, you've talked a little bit about your portfolio being a little bit different than agency, given it's FHA and there's portions of second liens in there. How have you thought about a future retention strategy? And I think there's been some tech that you've used in the past that have given you early indicators that maybe you could share with the audience?

Jeff Bode (24:35):

Listening to Paul, I got an idea.

Chris Bixby (24:38):

Great.

Jeff Bode (24:39):

In our LOS, we have a pricing engine. So I think what would be smart of us is to just run that pricing engine, just run a query every night. So in the morning, the LO can see this loan's in the money could be refinanced. We like to use everybody, the trigger leads. That's a winner that is being hit on by somebody else. There was a product that was out there that was predictive, and when a consumer was going to possibly purchase or refinance, what they did is they kind of took their cell phone number and their email address and kind of monitored where they were on the internet, what they were doing. So if they were going to realtor.com or something, we would get a trigger telling us that they're looking to do something, whether they're trying to refinance or they're trying to purchase something. And we found that that was fairly effective in what we were doing. We ceased buying that product. When interest rates drop down, a big part of our portfolio is somewhat impaired with a prepayment impairment on it. So the consumer's got a second mortgage behind it, so they less likely to refinance than just a straight 96.5% L-T-V-F-H-A. So we are not focused on that as much, but we need to get there and great products that were discussed today and yesterday in the Shark Tank session that just really do an amazing job.

Chris Bixby (26:11):

So we think about it from our end as given we're MSR investors and we're kind of developing relationships with originators to work through what recapture looks like and actually allowing some of the or originators that we purchase pools from to retain the ability to recapture those products. And it's actually a very symbiotic relationship. We think about the data as a level one data, which is on the borrower themselves. So their credit, their home, their equity, actually the profile of who they are and what their loan looks like, and then what their home looks like. The level two being kind of this activity based information that I think we both have talked about. Number three is predictive. And so the level three is that predictive that's actually not based on actions. It's not based on actually what the home looks like or what sort of pricing that you can pull from a pricing engine. How have you both, or have you thought about ingesting predictive data into this or have you seen other groups think about it or start to utilize it?

Jeff Bode (27:12):

I think it's very easy to predict when somebody's going to refinance. The purchase issue is a different deal, but having the data of what the home value is worth, the A VM solutions, you can buy that data to know exactly what your portfolio does. And obviously if they've been delinquent, they're less likely to go away because they've got a problem they can't refinance. But also credit scoring, you can take your portfolio and match it up with what their credit score is fairly inexpensively. So you know what that borrower can do, not only what they will do, and typically if it's in the consumer's best interest, you need to be reaching out to them and giving that opportunity to refinance.

Chris Bixby (27:59):

Great. Paul, have you thought about it?

Paul Akinmade (28:01):

Yeah. Are

Chris Bixby (28:02):

You thinking about it?

Paul Akinmade (28:03):

Yeah.

Chris Bixby (28:03):

You think about everything. I've come to know that you're always thinking about something.

Paul Akinmade (28:07):

So I would say it's the combination of all that data and how they react to it. So you can go buy predictive data from various vendors, credit bureaus sell it to you, but again, it's a prediction. So I would much rather control my own prediction, meaning that I can see, okay, what are the inputs that we have that are data elements we can look at? You can look at pre shopper data by hash tagging your email address and seeing what they bounce on sites. You can look at MLS for purchase. That's the leading indicator there. I think it's proactive versus reactive. I think you become reactive when you have the triggers or the payoff demands and you have to go to becoming a low cost provider. So your margins are tax, you have to have a special lo follow up with that. And when you start adding in your second level data that you have around credit and stuff like that, it's the spread between how much in the money they are and what's that relevant to them.

(28:57):

Meaning that is $200 savings relevant to somebody makes a hundred grand, maybe is it relevant to somebody makes 50 grand more likely, but who are you taking all the elements of the loan and doing a statistical analysis. And that's where I kind of nerd out with Minitab where I will go ahead and look at all the variables, find correlations and say, okay, alright, we found something. Let's identify in our portfolio, everybody who has this combination of attributes because they have a 65% chance of saying yes or 90% chance of saying yes. I think that's where you have a big opportunity for data to kind of step in and look at everything and draw the lines between what the connectors are that will say this person will have a high likelihood of converting. So when I look at predictive or I guess the third level, I don't trust anyone else technically to do it other than myself unless somebody came out with something that took all those inputs from me and then spit it back to me. Got it. Which is a good economies of scale. I would rather not be the data scientist on that one because I think somebody can do it way better. Get there. If you have enough data, you can do it and

Chris Bixby (29:56):

Do the right team. So we promised provocative, I don't know if this is really going to be that for mortgage, but maybe it's a little bit more tied into our theme between origination and recapture. What's the role of the LO in a recapture process or a refinance? So you have the data on your consumer, you have the predictive information, or you have the insights and you generally can put them through the funnel assuming you have the right tech stack to pull them through. What's the role of the LO in that process? Silence.

Jeff Bode (30:27):

I expected Jeff, sorry, you go first. Well, they need to work. I think it's to make sure that they're getting the right data and that they do reach out, make sure that your CRM system has got that call queued up for them to dial and talk to the customer at that time, be reaching out through whatever sources they've got to communicate with that customer. One thing we try to structure it where if the LO is no longer there, we try to push those loans to an LO with lower comps, so that way it's a more competitive pricing opportunity for the borrower. And since we sold off our retail in 2022, that's not a hard thing for us to do. And if we're using brokers, we try to push those back out to the brokers because they've got the relationship with the client and work with them on that. But the LO has just got to be on it and the right compensation is significant. But with the LO comp with Dodd-Frank, the person that originated that loan may not be the best one to provide the best pricing for that consumer going forward. And if they're in a refinances are typically more of a price battle than a purchase transaction is, it's just completely elastic on price there.

Chris Bixby (31:47):

Got it.

Paul Akinmade (31:48):

So I would take somewhat of a contrary statement on one thing. I think when I look at our retention rate for how many, when we look at our payoffs, how much do we get back? My active LOs have doubled the retention rate than I do with the consumer direct channel that's hitting my inactive LOs and my correspondent wholesale stuff. But it's from my perspective, where are you attacking them in the funnel? Are you being reactive? Are you being proactive when you get into the margin play of, Hey, let's drive that price down because you're trying to recoup that. Now we have a salvage LO program for payoff demand statement. So I agree 100% there. I do think that there's opportunities to move up in that funnel because from my perspective, if you look at somebody who bought their first house, they had that person as the advisor bring them through that.

(32:30):

So I think they have a significant role in it, and I think it needs to be one where they need to opt in. Do you want to do the refinance or do you want to focus on growing your purchase business because that's an annuity versus a one-time pop when the market changes because they staff accordingly and differently for that. So I think what I want to do for distributor retail is I want to give my LOs the opportunity to kind of make a choice of what's right for them. One of the things that I have to do first is I have to make sure I have the automation in place. So I do that proactive outreach with all the signals that I have, which is I said what we're going to be rolling out this year, but at some point in time if they don't do that at outreach, a phone call has to be made.

(33:04):

And if it goes into being a reactive mode, they may lose an opportunity because we will lose the business if I leave it with them based on their current comp level. I do think that there's a blend and we're launching consumer direct for this is where you have LOs who opt in, and a few companies do that today where you have a lower cost loan officer follow up. But when they follow up, from my perspective, they should be following up on behalf of that loan officer. I'm on Chris's team. My name's Paul. I support him on when we're doing refinance transactions, are you thinking about doing another purchase? I give the purchase deal back to the distributed retail lo, and then I go ahead and do the refinance there. Then I've got the best of both worlds and that hybrid is a delicate balance to walk.

Chris Bixby (33:41):

That's great. Alright, we have about five minutes left. I'm going to open it up for questions. If there's anyone in the audience that has any, I still have a few more, so don't feel pressure, but would love if there's any questions in the audience of anything that we talked about or other questions you have for Paul or Jeff, someone. Alright. Alright. So Paul, I understand that you just went to Disney and you admitted to me that you planned your entire trip using ChatGPT. So first of all, I think you should share with everyone because I think it's a pretty interesting way to plan a trip, but then I'm going to dovetail, how do you think about AI within CMG?

Paul Akinmade (34:21):

Sure. So the wife was bugging me about what I wanted to do in Disney on each specific day, and I didn't want to apply the cognitive load to figure that out, stare at a map, do research. So I did ask ChatGPT to write me an ideal efficient schedule with an extra 10 minutes buffer per ride for a five and 2-year-old, and which lightning passes I should buy. When I did the lightning passes, some did hit at the time I wanted, some didn't. So I readjusted the prompting to adjust for that and I spent probably less than two minutes planning an entire Disney trip for a five and 2-year-old at Magic Kingdom and Animal Kingdom.

Chris Bixby (34:55):

So mic drop. That was enough information that we could ever get out of this. But Paul, you're in a seat where you really do look at a lot of ai. I know you work with one of our companies which we invested in is capacity and then there's a whole different variety of companies that you continue to look at in terms of whether it's chat, whether it's guideline reviews, whether it's workflows, whether it's tapping into various systems internally. Maybe if you could share any just high level kind of work that you're doing at this point from an AI standpoint.

Paul Akinmade (35:24):

So I do agree. I never want to communicate directly with my consumer directly with ai. I think there has to be a copilot in place. Too much can go wrong. I do think from a marketing perspective, there's a whole lot of opportunity to create different variations of content that can inspire different sentiments. So inside marketing we use it on a regular basis, especially with ad-based testing and especially with content generation inside CRMs. What has higher performing values to get the anticipated click rate that you want? When I'm looking at on the tech side, one of the things that we saw is one of our analysts, our bas was writing a BRD for a specific difficult process that we were trying to build out and they were working on it for six weeks. We tried chat to tv, got 90% of the way there in under a minute. So I think there's huge opportunities there. I'm actually kind of pushing back on my BSAs to say, you have to get better. We've spent money bringing in people to run workshops to about how to leverage ChatGPT. We have enterprise ChatGPT, we have a tech partner, CINT, which has a guided flow for various natural languages which bring in Bar Gemini and ChatGPT.

(36:55):

So with that,

Chris Bixby (36:57):

Sure. Yeah,

Paul Akinmade (36:58):

I dunno if we have to exit.

Chris Bixby (36:59):

Did anyone understand?

Paul Akinmade (37:00):

I didn't hear any of that. So we're just going to assume we're safe. If we see smoke we'll run. But I think from my perspective, I'm going to be pushing to be an early adopter of it. I think it can speed up workflows. I think you can get higher underwriting efficiency. It has an interesting way of finding patterns in complex areas. So you could probably ask specific questions around large data sets, like who's likely to have a payoff? Here's all my past payoff data and all the attributes identify for me. And I don't think it'll pick it for you, but it'll help give you a compass as to what you should be thinking about so that way you can build it out. I do think there's room for it in qc, but it's going to be an area that I think the world is going to evolve around it. So are you going to be helping cut your way through that jungle and benefit first, or are you going to just wait on the path when it's developed for you?

Chris Bixby (37:47):

That's great. Alright, we'll wrap up here, especially given that announcement we heard, but Jeff, you've been in business for a while. It seems like you've been very successful. Is it luck or is it smarts? Oh, it's all luck. Well, thank you so much for being here, everyone. There's obviously more conversation. I think what I was excited to see was the amount of really vendors and technology companies that were here kind of focused on that servicing and retention strategy. There's a lot out there kind of in the hallways. I recommend you talking with them, kind of talking with the members up here and they're more always happy to talk about our venture strategy or MSR strategy. And it's great to be here and thanks to Heidi and the team and Julian, et cetera for putting this on and having us.

Paul Akinmade (38:33):

Thanks.