5.22.26 Basel III Capital Requirements; Insellerate's Josh Friend on Increasing Efficiency; Higher For Longer

In today’s episode, we go through MBA's recommendations about Basel III capital requirements. Plus, Robbie sits down with Insellerate’s Josh Friend for a discussion on increasing loan officer and sales manager efficiency. And we close by looking at bond market movements from across the globe.

Welcome to The Chrisman Commentary, your go-to daily mortgage news podcast, where industry insights meet expert analysis. Hosted by Robbie Chrisman, this podcast delivers the latest updates on mortgage rates, capital markets, and the forces shaping the housing finance landscape. Whether you're a seasoned professional or just looking to stay informed, you'll get clear, concise breakdowns of market trends and economic shifts that impact the mortgage world.

This week’s podcasts are sponsored by TransUnion. Discover how data-driven mortgage intelligence is helping lenders identify in-market borrowers, strengthen portfolio performance, personalize outreach, retain customers, and drive smarter growth in an increasingly competitive housing market.

Transcript

SPEAKER_01 0:02

Welcome to the Crispin Commentary, Daily Mortgage News Podcast. I'm your host, Robbie Crispin. Topics on today's episode include continued news from New York and Basel 3. Why are rates higher for longer, or at least the latest narrative around that? And my interview with Incelerates Josh Friend on increasing loan officer and sales manager efficiency. Here, take a listen to a quick preview. How do you find the balance of this person wants to be contacted? But you know, maybe maybe we don't want to contact them too much. How do we find the right amount of contact point they want? And then also for those people that say that like my dad or maybe myself that call an airline and they get an automated person, you put a zero until you get an actual person. Maybe the the answer is that it's in the loan officer's voice and they've they've been recorded. But how do you manage kind of these client expectations for the amount of touch points and who they want to be interacting with?

SPEAKER_00 0:55

Yeah, you know, it's interesting. Um I mean, there's best practices of how often you should call a customer. You know, just plain and simple. And we have some of those we share this with our customers. So here's kind of the uh the ideal call strategy or email or text strategy to reach out to your customer. Um, what we're finding though with the AI, surprisingly, people like to actually interact with the AI if they can and get as much information from the AI before they talk to a human, generally speaking. So, yes, typically, you know, um the old IVR systems, you know, say what you want, press one to do this, press two to talk to service in, press three to talk to customer support, press four to book your reservation, press five if you have a question about your reserve. I mean, yeah, you hit zero because you really can't get much information out of those systems, right? So it's just like just give me to someone so I can actually tell them what I want. AI is a little bit more advanced than that, right? And that's where, you know, honestly, using AI voice, you don't typically want to use AI voice to take a complete application of a customer unless they know, hey, this is it's Saturday, no loan officers are available, but if you like, I could go out and take your application. This is an AI assistant. And we actually find people interact with it. And and the reason is anytime that someone's applying for a loan, for a lot of people, not all, but like for a lot of people, there's always a hesitation of fear of rejection. We're humans like that. Fear of rejection is just a real thing, always across the board. So when you talk to AI and we're finding this, and we actually in 2018 had an AI bot that could get someone pre-approved for a home loan and let them know uh how much it should go, what value of home they should go buy. And what we found was people aren't afraid to get rejected by AI. Oh, wow. Because it's not a real person, right? So there's more of an aptitude to go ahead and like interact with this AI and kind of disclose more because you don't feel like you're gonna get, you know, this human rejection on the other side or you're not gonna get judged, right? Like if you tell the AI, you know, I just filed bankruptcy a year ago, can you still help me? I may not feel judged. But if I tell a cuss, a human, yeah, you know, we filed bankruptcy a year ago, you know, some people are gonna feel judged. So that kind of fear of rejection or judgment goes away. And so AI does allow for humans to kind of feel a little bit more open to talk about and discuss what's going on in their lives.

SPEAKER_01 3:19

Thanks to this week's podcast sponsor, TransUnion. Discover how data-driven mortgage intelligence is helping lenders identify in-market borrowers, strengthen portfolio performance, personalize outreach, retain customers, and drive smarter growth in an increasingly competitive housing market. To learn more, visit transunion.com. I recently got on an airplane and there were no other passengers. A flight attendant walked up and chittered, you got the whole plane to yourself. A large group that was going to a psychics convention, all canceled at the last minute. Want to know the new buzz acronym? It's residential transition loans, or RTL. That's financing for properties in transition, including fix and flip, bridge, and rehab deals. Catch the wave. Now, if only more income guidelines by the agencies matched how incomes have changed. Yes, change is constant. Someone wrote to me saying if you think inflation is bad, home insurance price increases and many parts of the nation are worse. When is the government going to address that? I don't know the answer to that. The government certainly is involved in lending, but not much so with insurance companies. Interestingly, insurance companies' use of data analytics is way above that of the mortgage industries, although servicers and others are certainly catching up on risk-based pricing. Mortgage prices are a result of supply and demand, more specifically the demand by MBS, institutional and portfolio investors, and banks and credit unions. Critics say a town in Switzerland shouldn't impact the U.S. mortgage market, but it does. Especially when it comes to servicing rights and weighing the risk. Regarding Basel, our MBA writes quote, the current rule raised the risk weight on MSRs from 100 to 250% with no empirical justification. As a result, banks looked at the capital cost and started walking away from mortgage servicing. The new proposal recognizes that 250% might not be the appropriate risk weight and requests input from stakeholders on what the appropriate number should be. MBA will be recommending that the risk weight for MSRs return to 100%. That should increase bank MSA activity and MSA values. To put this in perspective, a 25 basis point increase in servicing value is a 25 basis point reduction in closing costs or $1,000 in savings on a $400,000 loan. The MBA says it's also striving to improve the treatment of warehouse lending. Under the current system, warehouse lines carry a 100% risk weight, but that defies logic. If an IMB fails to repay, a bank gets the whole loan, but only at a 50% risk weight. The risk weight shouldn't improve when a counterparty fails. The risk weighting should be fully aligned. The current policy makes it more difficult for banks to participate in the lending and servicing business, and MBA says we need a policy that fully recognizes and respects the role that banks play in providing vital liquidity to IMBs who make the bulk of affordable loans to American families. MBA is strongly calling on agencies to change the capital requirements for the banks that have loans on portfolio that have private mortgage insurance. They're obviously not as risky, and the rules should reflect that. Bottom line, there are several places in Basel III that still need fine-tuning. For today's interview, I wanted to welcome back to the show Encelarate's Josh Friend to talk about increasing loan officer and sales manager efficiency. He's CEO and founder of Encelarate, which began with a mission to transform the way lenders engage and interact with customers. His goal was to help lender clients have a lasting impact on the lives of borrowers and referral partners they work with. Let's talk efficiency today. Josh, I want to talk how AI voice callers factor into this. I want to talk about how we can get more applications completed. I want to talk about how we can get more app completed applications to closing. And so let's just talk high level before before any of those specifics. Where is technology in the second quarter here of 2026 when it comes to identifying eligible borrowers, understanding their propensity or desire to actually get to a funded level? Like where are we with technology?

SPEAKER_00 7:43

I mean, I can tell you from uh all the studies we've done with our AI and our platform, uh, it's pretty wild how accurate um AI can actually be. So your your ability to actually understand who's going to transact, um, who's gonna close alone, and not necessarily just alone with you, the lender who's talking to the customer, but alone with somebody within the next 60 days, um, it's pretty substantial. Um, just to kind of give you an idea on the math of it. With our what we call a deal, we do a lot of scoring, but one of the scores we have is called a deal possibility score. And what we're finding is if we if we score a lead, um, a call, an eight to or a nine, our average lender is gonna fund 13.65% of those. However, the eights and nines that that lender does not fund are on average, 33% of those get another loan with another customer within 60 days. So you're talking about and scoring someone eight or nine, it's almost a 50-50 shot they're gonna get a mortgage versus I score someone a one to four, it's like a two and a half percent shot they're gonna get a mortgage with somebody, you or another lender within 60 days. So, I mean, we're talking, you know, 20 times more accurate than just you know a random guess. So the math is really um uh it really stands behind. And it's if you really think about it, it's I mean, what one of the things AI does is allows us to do um good quality, complicated work at scale. It's really what it is, right? So, you know, what do I mean by that? I mean, if you're a sales manager or or someone, if you were to listen to every call that every loan officer took, you could probably pick up details that this person's needs a loan. They say they have a 7% rate and rates are 6%, they need a loan. They they're trying to buy a house and they're trying to close within the next 30 days, they made the offers accept it, they need a loan. They have a bunch of credit card debt and they qualify to pay it off and say $500 a month, they need a loan. So it's really not like wild to think um, you know, how it works, but since AI can actually listen at scale and do that quality work that a sales manager doesn't have time to do, um, that's where it really comes into play. And that's where, you know, I think just not just in mortgage, but in everything we're seeing, AI just brings huge efficiencies.

SPEAKER_01 9:55

So my my mind initially went to well, let me ask what AI is picking up on, and and or they can tell tone of voice. But that's not the question to be asked. I don't care how the back-end engineering and code of DoorDash works, as long as my food gets to my door. Like, I don't I don't even know.

SPEAKER_00 10:11

You don't care how the sausage is made, you just want to see the sausage.

SPEAKER_01 10:13

I just want that, I I just want that Adele's chicken apple, yeah, or whatever. That's right. Johnsonville Brawurst. And hopefully the mortgage we're delivering is tastier than a Johnsonville Brawst. Anyways, I digress. The the question that I would like to ask is once companies identify the propensity of a client to to close, or or they're told, hey, this is a this is a one you should be pursuing, what can they do to better usher that along the process or massage that file through the process?

SPEAKER_00 10:46

I mean, it's not rocket science, it's communication. It's hey, listen, you know, you only have so many hours in a day, you only have so many people you can follow up with or talk or second voice or manager can step in and help sell a deal. So it's really about, you know, if you have hundreds of of customers you've spoken to that week, how do you just make sure you're talking to the ones that actually matter? And that's where the efficiency comes in. So once you know who should close, it's really about giving them good service, calling them back when you say you're gonna call them back, reaching back out to them, um, offering some you know, real advice. Again, this is a service business. We're in, we're in the service business, right? That's where financial services. It's explaining to that customer how you can save them money, how you can help them buy their first house. Should they take a 15 year? Should they take a 30 year? So then it's just about actually how do you provide you know good service to the customer and communicate and make sure you focus on those customers. It's it's pretty wild that we'll see. We'll see a lender, large lenders, and we'll look at, we'll go back and we do, we're doing studies all the time. We're looking at deals they close, don't close, are they closing elsewhere, how many times they're calling them, what's the ideal time to call them? And a lot of these deals don't get called back. Someone talks to them, they there's a deal there, then it can go a week and there's no follow-up from the loan office or to this customer. So if that happens, of course, you're never going to close that transaction, right? You know, you don't, if you're not calling your customer, someone else is gonna call them, right? Right, this is a commodity. There's plenty of opportunity out there for plenty of lenders that want to speak to your customer and follow up with them. So if you're not following up, then someone else will.

SPEAKER_01 12:21

What was good or great in 2018, 2019, 2020, and what is great now when it comes to what these what the technology can do?

SPEAKER_00 12:32

So here's here's the punchline the 2018 bot we had. Um we actually had it at which is kind of interesting. So we we we shared and debuted it at the LE May show when there's still LE May in San Francisco in 2018. And what we found, and so what's great today is people are ready for it and expecting it. And customers are ready for it and expecting it. In 2018, when we showed it to people, I got two responses. Number one, my bank credit union lender doesn't allow we're not able to use AI. We're not gonna, we're never gonna use AI. Number two, that's not real. It's never gonna work. That was like, so, and that was in 2018. So at that point, we said, all right, listen, we're gonna go and shelf this because I'm not gonna try to convince the market that this is real and you should use it and you should be ready. We'll wait till the market's ready. And and now every customer we talk to, they ask, Do you have an AI voice bot that can talk to my customers? I mean, so quickly that changed. Now it's eight years ago, but you know, starting probably two years ago, everyone's asking for it. So, really, what's changed now is the market's ready for it. Right? You know, in 2018, the market really wasn't quite ready for it. They were unsure of it, they're afraid of it. Now you it's you you no one can deny that AI is real. No one now can deny that AI is absolutely going to change everything that we do. Everyone sees it now. So people are ready and more willing to interact with it and are open to it. Now, the voice clarity, voice quality has all gotten better too. Voice recognition has gotten better. Um, NVIDIA has come out with a new dual process and chip that allows you to have a conversation so AI can talk and listen at the same time. So our engineers have been playing with that. That's brand new technology. So it makes the conversation even more realistic. So as you're talking and you laugh or talk to me, I can hear it, the AI can hear it and respond versus right now. It's kind of this start-stop thing. You talk, it listens, it talks, you listen, and if you interrupt it, it loses its train of thought. That's all changing this year as well. So the technology has definitely gotten better.

SPEAKER_01 14:34

If the acceptance of the general public is there, we're talking about dual processing technology. I'm going to be so greedy here asking you this, but where does it go from here? How does it continue to get better? What do you see? And I ask that because as a technologist, you are constantly forced to keep reinventing yourself and advancing and pushing forward. That's right.

SPEAKER_00 14:55

So we're we're playing with right now, um, we built a new loan presentation tool within our platform. And it's not just a loan presentation tool, it's now AI-powered loan presentation tool. So um, what we actually have, the the first beta of it, which we didn't release because uh Fannie Freddie Mac now really is requiring a human in the loop. If someone sells a Fannie Freddy and they're given loan advice and given loan options, there needs to be a human in the loop. But what we had built was you send someone a presentation, a customer can actually ask the AI, well, hey, I want more cash out. What could I qualify for? And it could actually go, you know what, I can give you this much out, but I can't give you this much because your DTI is too high. If you want to take more out, you're gonna have to pay this credit card off. So it's actually moving into doing even more complicated work. So now we're giving that tool to the loan officers. So when they take an application, they don't have to do much thinking. They can just take the application and the AI is gonna say, hey, for Josh Friend, the best option is for him to take a 30-year fix, pay a quarter point because he's gonna save this much money, pay off these two credit cards because it gets his DTI down, and now he saves this much money, which saves him $6,000 a year, $18,000 in the next three years, and over the life of the loan or early payoff, or apply it to his principal balance. He can save nine years on his mortgage or $180,000 in payments. So the AI is now getting more intelligent, right, to actually do more of the job. You know, I was a loan officer years ago, and you know, I was when I stopped doing loans in uh 2002, I was averaging 44 loans a month, which in today's market's still a heck of a lot of loans. And one of the things I was able to do in um is I'm good with math. So if I talked to a customer, I could figure out what loan you qualified for, what your best options were. I could put a presentation together really quickly. I can explain to you why you don't want to pay the point or do want to pay the point to buy the rate down because the payoff is three years or seven years, and why paying this credit card off makes sense and how this affects your overall long-term value, and how if you can, you know, I you know, I was good with math, but the average loan officer doesn't have that skill set. It's a lot of work. So, now how do we make the average loan officer just a genius? So the average customer now gets like a really good financial picture, gets the best advice possible. And that's where AI is going to right now, as far as that's the side we're working on is how do we help consumers get the best options possible and how we put loan officers in the driver's seat to allow them to present those options to those custom consumers.

SPEAKER_01 17:21

I want to close by by asking kind of where your mind has been lately. Any things that have been kind of your your north star here are just little tidbits you've been thinking about or ways that technology is interacting with people. And I mean this in a very open-ended sense.

SPEAKER_00 17:38

What we're finding, and we're doing this internally, is we're at a place right now in the market that if you want something built or something done, or you have a use case that you really understand how to solve it, it's easily to be you can easily do it now. What I mean by that is we built something called our uh um app enablement hub. So I'm sure if you heard of like Replit or Lovable, these Vive coding platforms where you can pretty much just go build whatever kind of application you possibly want. It's it's really wild. I mean, I have, I don't know, I probably have built 15 or 20 applications, everything, and just all kinds of random applications. And you know, I was a programmer 30 years ago, I don't program anymore, but I'm able to write and build some pretty insane applications like within a few hours. And it's really about subject matter expertise. So if you have that subject matter expertise and you know the use case, you can define it, you could actually give it to AI, it can build you really great tools. And so what we do with the Apple Enablement Hub is that gets you about 80% of the way there. You still need about 20%, you need data management, data security, API stuff. There's some other work you need to do. We just built a platform now that lets lenders go in there and actually build off of our APIs, which connect to LEMA and all the LOSs out there, price and engines, phone systems. So, really, if you have a real need or desire for some type of piece of technology, whether it's a processor reader board that tells you how efficient your processors are or whatever it is, if you just describe it, write it out in detail what you want, you can easily get it built today, which is before it could take months and hundreds of thousands of dollars and tons of mental brain work and a lot of meetings. That stuff has now gotten shortened to like within hours or days. So technology is much, much more accessible today than it's ever been.

SPEAKER_01 19:31

The industry should be formally put on notice because Incelerate already has obviously AI powered technology, agent connect, CRM and lead management, sales enable enablement, engagement, data intelligence. So I'm I'm excited to see what you'll roll out next. For people looking for more info, go to Incelerate.com. Josh, always a pleasure, man. Thank you very much. Thanks, Robbie. Markets experienced a sharp reversal Thursday when reports surfaced that Pakistan might have helped broker a potential ceasefire framework between the US and Iran, easing fears of further escalation. Oil prices quickly fell back below $100 a barrel, which helped calm inflation expectations, stabilize bond markets, and lift equities. Longer dated treasuries recovered to finish stronger on the day yesterday, while mortgage-backed securities also rebounded after early weakness, allowing some lenders to modestly improve mortgage pricing. Despite the relief rally, investors remained cautious over and about elevated energy prices, persistent inflation pressures, sovereign debt issuance and demand, and structurally higher interest rates. Stepping back slightly, expectations have been scaled back for a near-term rate cut as investors increasingly accept a higher-for-longer policy environment that could extend well into 2027. Longer-term treasury yields have risen to levels not seen since 2007 without sparking a major equity sell-off, suggesting a gradual repricing to a world of permanently higher capital costs. I've also heard growing concerns that yields approaching key psychological levels could eventually trigger broader risk asset repricing. And rising gas prices are being or beginning to pressure consumers, with recent spending data showing household budgets increasingly strained outside of fuel purchases. It's the sole release of note today, though markets will also receive remarks from Fed Governor Waller. We began the Friday before a three-day weekend with agency MBS prices better than Thursday's close by an eighth to a quarter, the two-year yielding four point zero eight, and the ten-year yielding four point five five after closing yesterday at four point five nine percent. Let's wrap up with a joke and some housekeeping. Well, well, well, if it isn't the consequences of my own actions. Thanks again to today's podcast sponsor, TransUnion. Discover how data-driven mortgage intelligence is helping lenders identify and market borrowers, strengthen portfolio performance, personalize outreach, retain customers, and drive smarter growth in an increasingly competitive housing market at transunion.