From Vision to Reality: Joe Tyrrell on Embracing AI and Automation | Market Insights | July 14, 2025

Jim (: 00:18

Welcome to Optimal Insights, your weekly source for timely market analysis and expert commentary from Optimal Blue. I'm your host, Jim Glennon, Vice President of Hedging and Trading Client Services at Optimal Blue. Our clients and industry partners have long relied on Optimal Blue for trusted insights and commentary, and these podcasts are an evolution of our commitment to keeping the industry informed. Let's dive into today's episode. Hi, welcome everybody. Welcome to this week's episode. It is...

July 14th, hard to believe we're already a good part of the way through the summer. as always, there's a ton going on. So in the interest of keeping you informed, we'll talk about some really market updates today and some stuff that's going on that affects our industry. So a few things that don't. later on, we're going to have a big interview as well. Vimi and I are going to interview Optimal Blue CEO Joe Tyrell.

We will talk everything from, know, is AI going to kill us someday to industry outlook to some real applications of things like AI and automation across the business world, not just in our industry. Just real quick, before we get into the market update with Alex, a little bit of data, OBMMI. So conventional 30 year fixed right about 6.7. So not a ton of relief there. We had a little bit of a dip during the brief.

war with Iran, but still back up near that 6.7, 6.8 number. So something to keep an eye on. Volume is still resilient though, seeing double digit increases year over year. In fact, there's been some really good articles written recently about how the purchase market this summer is quite a bit better off than a lot of folks would have you think. We'll probably do a whole episode on that at some point here later in the summer and kind of look back and see how we performed. Otherwise.

a lot going on in the industry, a lot going on with the Fed and have continued resilience in our economy, those of us in this industry, kind of low for negative numbers, at least at some point, so we get a little bit of reprieve with interest rates. Anyhow, let's go check in with Alex, see what's going on in the market.

Jim (: 02:30

All right, welcome Alex Hebner. Thanks for being here, sir.

Alex Hebner (: 02:34

Hey, good morning, Jim.

Jim (: 02:35

Good morning. Good morning. All right. Lots to talk about today. We do want to cover a few things, including some of the drama going on with the Fed and some things in the media. And we have some very particular viewpoints on that, but we'll just start with some did we see last week in terms of data and what should we be looking forward to this week?

Alex Hebner (: 02:56

Yeah. Yeah. Last week, a pretty quiet one as the middle of the month almost always is, from a data perspective, initial jobless claims as I kind of shout out each week was kind of the highlight of the week for me, um, looked fairly resilient, 227,000 expectation was two 35 and that's, that's well off the, two 50, um, which is kind of the, breaking point for, for the labor market in our books, is, kind of a downward trajectory.

So, it seems okay for the moment. but again, definitely something we want to be keeping an eye on as we just see that the effects of tariffs, roll through the economy, potentially the effects of, of AI automation, the economy, you know, that one's on a, on a years long trajectory. just all things that are, that are, you know, headwinds for the, for the labor market. looking into this week, it's inflation week. So we're getting both CPI and PPI. CPI is expected at 3%.

see where it lands. anything south of there, think pretty decent shape for, from an inflation perspective, especially considering we're starting to see the of tariffs being accounted for, ⁓ headlines last week of they already raised a hundred billion dollars, tariff revenues, this year, a sharp peak on the graph at the moment. but we'll just have to see how that continues to and as these prices are.

potentially pushed forward. ⁓ We've had a lot of conversations on this podcast about when do companies pass the cost along to the What's the breaking point? It probably comes down to the individual company. I remember James Cahill a couple of months ago made a really good point about how companies like Walmart, these big sellers that can bear the brunt of the cost for a period of time, they don't want to be

In the eyes of the consumer, the first one to pass the cost onto them and become an enemy in their eyes and maybe lose customers something that ends up being a very temporary measure. So it doesn't seem that costs have been fully pushed through yet. You you see little dramatic headlines here and there, but by and large, it doesn't seem to have hit inflation too much yet. being said, you know, a lot of the economists were saying, you know, midsummer is when we expect to see the cost start to flow through. didn't really see that in June.

but July, August is definitely time to keep an eye on these numbers. as they continue to know, jumping into, to kind of the headlines from the past week outside the data. I think tariffs are really driving the market narrative right now. the administration seems to have put on kind of a maximum pressure approach. slapped tariffs on about a dozen or more countries at this with an August 1st effective date.

August 1st being the date that they kicked the can on from, ⁓ July, July 7th, believe it was, they only had a handful of trade deals done, but they're saying August 1st is going to be the final day that, ⁓ they can get a trade deal done before the tariffs go into effect. I think the, most notable, and Saturday, we saw announcements, ⁓ of between 30 and 35 % tariffs on goods, from Canada, Mexico, and the EU. Those are, those are massive trading partners.

so, so, so I think the pressure is biggest there to get a, a deal done. and then just, just to kind of at least the administration's, expectation for this particular tariff, but a 50 % tariff on copper to boost, ⁓ domestic production of copper. ⁓ and that was done under the section two 32 tariffs, which I've highlighted before as, much harder for the courts to.

reversed just because the section two 32 tariffs are under national security grounds. So they can do pinpointed tariffs on individual which I could see being the, uh, the strategy moving forward.

Jim (: 06:19

Mm-hmm.

All right, pretty strategic and maybe, hopefully it doesn't lead to theft of copper like we saw, I don't know, was five, six years ago when the price of copper expanded and people were losing the catalytic converters off the bottom of their automobiles. I think that's probably great.

Alex Hebner (: 06:43

Right, right. I don't know if that problem ever fully subsided, ⁓

no, it could, it could definitely lead to a little spike in petty crime such as that.

Jim (: 06:55

Yeah. All right. So I mean, to summarize, we had once again, resilient jobs numbers, which, you know, despite our wishes, jobless claims are still coming in pretty well. We saw, you know, we'll have inflation this week where we've not yet seen the full extent of tariffs. And there's still just or this unknown of where along the line will these costs be passed on?

So that $100 billion that the treasury saw in tariff income in the last few months, had to come from somewhere, right? And it's coming from somewhere along that chain. And so far, the consumer has not bore the brunt of that, which is interesting. Kind of goes against what some of the economists would have thought, I think, back in January when a lot of this started really becoming reality. So I suppose we'll see as we go that bleeds into the inflation numbers. May even see this week if we see something over 3%.

And that, you know, begs to bring up the drama conversation of some of the rhetoric going on in the media between, really, I mean, most of it's coming from the administration, let's be honest, right? The administration has kind of put full pressure on the Fed at this point and has basically lashed out and said, you need to lower rates. And some of the figures have been probably a bit outlandish, a bit exaggerated, but you know, 3%.

cut rates 3%, cut us back down to something around one. And I don't know, at first blush, you would say, the administration knows something we don't. They're looking to, they see cracks in the labor market that we don't see, or they see that tariffs are going to are going to be bore by the businesses that are subject to them rather than the consumer. There's some kind of disagreement there.

But anyway, the more you think through it, the more you see some of these tweets, you wonder, maybe the administration knows more than we think they do about this and they probably do. You've got smart people across the board. You've got folks who kind of understand the economy. They understand the ramifications of jeopardizing the Fed's sovereignty could do. I mean, some of this, is all of it just showmanship at this point? It feels like the administration knows that the Fed

won't drop rates 3 % overnight, that would never happen. In fact, that would cause a kind of a horrible repercussions if the entire world suddenly realizes that the Fed is controlled by the executive branch and it's no longer autonomous, right? What are we missing here or is it as simple as that? The White House is kind of screaming that they're coming to the rescue, but they fully understand that

There's no way the fed's going to cave on this. Even if they replaced every governor on that board, the board still has to look at actual economic data to make their decisions. And they're not going to be become a puppet regime.

Alex Hebner (: 09:57

Right. Right. Yeah. What you're referring to here, Jim is, the kind of the headlines are hidden. The, the tape this morning of Trump calling for, for a 1 % target rate for the federal funds rate. you know, we're currently sitting at four and a quarter. so in an immediate cut, which, which again, he doesn't get into details when he posts these sorts of things, but, know, an immediate cut of 3 % would be double the emergency cuts that we saw headed into COVID in March of 2020. So you'd have to imagine.

Jim (: 10:22

Mm-hmm.

Alex Hebner (: 10:25

It would require a situation, you know, to, to, get this done in the democratic fashion that's done at the, at the Eccles building with the federal reserve board. it would require them to be in a state of fear that is potentially let's call it double that existential double that of what we were facing down in March of 2020. and I, don't even want to imagine what that might be. Um, so I'm with you and invade. Yeah. Yeah. They're yeah.

Jim (: 10:38

existential.

Alien invasion. Alien invasion.

the flying saucers have landed.

Alex Hebner (: 10:52

I'm with you. think it's a bit of showmanship, which I think is in line with the personalities that we're talking about here. And I'm with you that I think that there cooler heads in the I can see Trump getting wound up and then maybe Besant talking him off the edge. If you ever watch an interview of Besant, he's monotone, very straightforward, but you can tell he's been around the block once or regards to, especially currency trading, which

You know, an emergency 3 % cut would be, I think, pretty catastrophic to the value of the U S dollar. so I'm with you. I don't think 1 % is the true target rate. I think that they're throwing numbers out there, get people riled up, in the hopes that they can get maybe a slightly accelerated pace of cuts. and as you said, all this rhetoric is coming from one side. Pal staying pretty quiet. He's getting bombarded with questions every time he has public appearance, but he's really just sticking to his talking points. He's got.

you know, handful of bullet points where he says, I'm not going to comment on that. and the federal reserve is going to continue to be a data dependent I mean, that, that, brings us to, to where we are right now, you know, September expecting a rate cut, but you know, we were here a month and a half ago in regards to, June and July rate cuts. And I think, you know, if we see the jobless claims remain muted, we see non-farm payrolls.

Continue to be decently strong with no, you know, outlying data factors, just, just tame inflation numbers. I don't, I continue not to see from a general economy standpoint, why there would be a reason for a rate cut.

Jim (: 12:18

Same. No, it does at this point. It's more and more likely that it's performative. And even the administration has to know if you were to cut rates 3 % today, inflation would almost certainly come back, even if we didn't have the tariff issue. We do have the tariff issue as well that has yet to rear its head. I feel like the president has appeared at least open to when he...

does make outlandish comments or demands sometimes. I do believe someone within the administration either corrects him or just, or they've had these conversations and he kind of knows what the effect is going to be. And some of it is show. I even wondered whether he's maybe on the side, you know, texting or calling pal and saying, man, just play along. Like I'm going to call you some names and it's going to get a little bit nasty, but just do what you feel is right.

but we'll all kind of have our position in this game.

Alex Hebner (: 13:15

Right. I could see that happening. also, with the headlines last week about PALS potential or at least rather they're blaming it on PALS, but potential overspending on a remodeling of the Federal Reserve building in DC. That to me felt like a, I don't want to call it grasping at straws, but you know, putting everything on the table of how can we get this guy out of Again, they'd have to clear the entire Federal Reserve board to get unanimous decisions that align with those of the executive branch.

but it starts with how I think, those headlines were a little you know, I don't want to make excuses if there was any potential graft and they, they added some fancy marble pillars in a potential lobby or something, but that to me doesn't feel like a fireball offense. and I also feel as if I don't think Powell's the one that's an architect drawing up the plans for the remodeling. know, he's got an economics degree, not an architecture degree.

Jim (: 14:04

Right.

Alex Hebner (: 14:08

So that just felt like a weird one to hit the again, just kind of felt like throwing something at the wall, seeing what could stick, seeing if they could get public opinion onto their side to get them out of there.

Jim (: 14:18

Yeah. And to your earlier point, Jerome Powell is not the only one who makes rate decisions. He cannot do this independently. There are a ton of others that have to vote that way for rates to get cut. And if by some miracle, all of the Fed governors were ousted and replaced and they all suddenly started voting to deeply cut interest rates, I think that would send an awful message to the rest of the world, right? That could send rates higher, I would think, because suddenly

The Fed is no longer autonomous. It's controlled by the executive branch of the US government and folks lose all faith in dollar denominated assets. I think again, Bess and Trump, they all know this and they know it's not going to go down that way, but yet the rhetoric continues.

Alex Hebner (: 15:07

Yeah. I mean, just look at how when the drama first began between Trump and Powell and how the, 10 year reacted to that, you know, you know, the bond market could easily revolt if they, you know, if something actually came to fruition again, the federal funds rate, that's just the overnight lending rate between banks. all the other rates, you know, that's an input that federal funds rate, but there's a lot of other inputs, including, and a major one that they're thinking about right now is inflation expectations way down the line. that could.

The inflation expectations alone, put aside value of the dollar global reserve currency status, that all feeds into it as well. But inflation expectations alone could keep rates as high as they currently are or higher.

Jim (: 15:47

Yeah, that's a good point. mean, I would say the average consumer may not realize what actually affects things like their mortgage rate or the rate on their car. I think the rhetoric has been very general out there where the tweets will say, know, Powell needs to lower the rates immediately. The rates. It's not called the overnight rate, it's just the rates. I think there's probably a perception that Powell holds the strings to mortgage rates, for instance. If he lowers

The rate 3%, suddenly we have 3.5 % mortgages again. That's definitely not the truth. Definitely not how it would go down. It could go the opposite. Like you said, the bond market revolts, right? The short-term rate, the overnight rate goes down to 1%. Mortgages could go to eight or nine. There's no law that says that can't happen because it's all based on supply and demand. And demand could be shaken if suddenly the Fed is compromised.

Alex Hebner (: 16:41

Yep, no, there's a historical precedent for them to move in tandem, but there's no hard and fast rule in the world of economics that says that they will move in tandem.

Jim (: 16:50

Right. Okay. Anything else going on that we should be aware of Alex?

Alex Hebner (: 16:56

too much. Just keep, like I said, like we've been talking about here, I think, keep an eye on the headline, see how tariffs continue to evolve. We are two weeks out from the August 1st deadline, which they have said is the final deadline. we'll see if they kick it again. Again, we've, we've been in wait and see mode now for three, four months. but they, they've said this is the final, deadline and it seems like, you know, people just expect Trump to chicken out. It's a whole thing. People are saying, you know, Trump always chickens out taco.

that feels like something that now that he's aware of it, he won't be chickening out. There's a lot of rules in the financial world that once you're aware of something, it's no longer a hard and fast rule. then again, they have to deliver on these promises at a certain point. ⁓ Otherwise they do lose their teeth and the rest of the world will know it. So ⁓ yeah, keep an on the headlines. We'll get PCE here in two weeks and keep an eye on those inflation metrics.

Jim (: 17:27

Right here.

Mm-hmm.

Right. that's the, fed meeting comes right before the tariff deadline as well. Right. There's no cut expected in the July meeting, but I think it's July 30th.

Alex Hebner (: 17:54

I believe so.

No cut expected, but yes,

30th of July there is a Federal Reserve meeting.

Jim (: 18:02

All right. Looking forward to all that. Great. Thanks so much, Alex. Good catch up. Take care.

Alex Hebner (: 18:06

Thanks Jim. See ya.

Jim (: 18:09

All right, everybody, as promised, got Joe Tyrell with us today, CEO of Optimal Blue. has been with us for just over a year now. And in that time, just made some really strong impressions across OB, across the industry, and just made some very positive changes to make OB better. And he's got some great ideas and vision of the future, not just for OB, but of the mortgage industry and just business in general. So anyhow, we wanted to have Joe.

on this podcast to just have a general discussion about where the industry is going and really where the business universe is going. Welcome, Joe. Good to have you.

Joe Tyrrell (: 18:44

Thanks Jim, it's

great to be with you and Vimi

Jim (: 18:47

Right, so I mean, just to kick it off, can't go anywhere right now in our industry or in any industry or even read internet news without hearing about AI and what that future looks like. you can go anywhere from the general assistant uses that the industry and others are using it for now. We all have used it in instant messages or to help us reformat emails or put together a presentation.

And then you can read books about how someday AI will kill us all. But then there's a lot in between. I mean, just generally speaking, where do you see AI in the next five to 10 years? And how is it being used in our industry now? How is it going to help automate things moving into the future? And what should we be thinking about as we try to figure out use cases for, like, practical uses for AI?

Joe Tyrrell (: 19:42

Yeah, well, first of all, I don't believe there's any chance that AI is going to kill us because I'm convinced one of my four daughters is going to do that through the course of my life. So I'm not worried about that. know, it's funny when people talk about AI, I've been deploying technology for a little over 25 years, and the majority of that time has been in deploying mortgage technology.

Vimi Vasudeva (: 19:53

you

Joe Tyrrell (: 20:09

And when you talk about AI, mean, there's forms of AI that have been around for decades. It was just called something different. It was called automation. And then in the last five years before generative AI really became kind of the phrase to encompass all things, technology based tasks and actions being taken. The focus was on a word called orchestration.

Jim (: 20:19

Mm-hmm.

Joe Tyrrell (: 20:36

which was basically taking all the automation capabilities that existed and just putting the rules in place to fire that automation off at certain points in a process or when certain events take place. And so now you have generative AI is really kind of the catchall phrase that people use to talk about automation these days. But from my perspective, it's been around for quite a while. And even artificial intelligence in its most current definition,

has still been used for over a decade. It's just our industry has typically been kind of the last bastion to adopt innovation. And it's understandable why that is because the stakes are so high if you get it wrong, right? The majority of your revenue as a lender right now is coming on your ability to strategically execute loans in the secondary market, right? Regulation has taken away a lot of the front end.

Jim (: 21:15

Mm-hmm.

Joe Tyrrell (: 21:35

revenue that you were able to obtain either through charging points or competition has made it so even administrative fees have really been compressed down. So all the revenue now is coming from, you know, secondary market execution. That coupled with the fact that if you make mistakes in this industry, one, you've got, you know, seven to 15 different regulatory bodies that are going to be quick to jump on you.

and come at you with fines or other penalties. But the biggest issue is if you have to purchase that loan back for any reason because you didn't perfectly cross that T or symmetrically dot that I, that's a risk that a lot of lenders can't take. And so the thought of allowing someone else to make decisions or take actions that could lead to the possibility that me as a lender, I get stuck with now having to

Jim (: 22:22

Right.

Joe Tyrrell (: 22:32

figure out how do I service this loan or manage a scratch and dent portfolio, ⁓ that's a big risk. So I get it. I get why it's been slow to get adopted here. think spending a year and a half outside of this industry recently and deploying generative AI and machine learning to some of the biggest brands in the world like Apple and Meta and Walmart and Marriott and Mercedes-Benz.

gave me a real appreciation for just how quickly things are actually moving outside of our industry and how people are absolutely looking to adopt AI and innovation as a competitive advantage, specifically as to being able to personalize an experience for a consumer. And so I think there's a lot of opportunities where there is very, very low to almost no risk for a lender.

with tremendous levels of value that they could deliver and true competitive advantages. But you got to know where to start. You have to have a strategy for how you're going to leverage automation and AI, where you want to deploy it and be very use case specific because it'll allow you to do two things. It'll allow you to get a ton of value without adding a lot of risk, but it'll also allow you to build your confidence in how you can trust that the technology can do.

the things that you needed to do without introducing unintended consequences.

Jim (: 24:01

Yeah, I had not heard it articulated that way before. That's interesting that it's just very difficult, especially in our industry for people to accept letting go of the wheel, you know, the steering wheel that they want. They want to be driving. They want to double check everything. And if you do kind of stay in that mode, you lose some of the advantages of having something like AI that's supposed to make your life easier, make the process less expensive to undertake. Right. So, I mean, what are the, you know, what are the early or the

next pieces of the origination process that would make sense to automate using AI.

Joe Tyrrell (: 24:39

Well, think you've got to start at the very, very beginning. One of the greatest attributes of leveraging generative AI is if you're doing it the right way, it not only can make you efficient, no question about that, but it can make you more effective, which is even more important. So think about, know, one of the biggest concerns a lender has is if I deploy AI, is it going to introduce bias into my process?

Am I going to be leveraging algorithms that are based upon historical data, which have a natural bias built into them? So the only thing that's worse than a bad process is an automated bad process. So you don't want that. So if you look at how we've been thinking about deploying AI, it's going and attacking those specific use cases. So where the whole origination process gets started,

is typically with an originator having a conversation with the consumer, right? But think about that process and think about the natural bias, the human bias that's already present in that conversation. Years and years ago, and I won't tell you how long, I was an originator. And so what happens is as an originator, you learn

a couple of products, loan products really, really well. Those become kind of your bread and butter products that you can always go to. You know them inside and out. And you start with those. And so when you sit down, whether at a Starbucks or in someone's living room or over a Zoom call with a potential borrower and you start having a conversation with them, those natural biases, those human biases automatically creep in. I think, okay, well,

This borrower is telling me that they have about this much down and they're telling me they make about this much a year. And so I start to default to, well, I've got this conventional loan program I know really well, and I got this FHA loan program I know really well. I immediately think, okay, well, they're not going to qualify for the conventional. So I automatically switched FHA. So now I've just allowed my experiences, my bias to determine what sort of a loan program that this consumer's going to get.

then I stop because once I've got them qualified for one of those few loan programs that I know really well, I don't keep going. I don't keep looking for the seventh loan program they could qualify for or the 10th, because I got them qualified. And that was the goal for me as an originator and that was their goal as a consumer. However, if you look across our industry, there are so many different unique and niche products that exist out there, whether through the agencies,

if you happen to be a firefighter or a school teacher the various programs that each of the states make available. So there's all of these different criteria that if met could unlock additional loan products. But what I just described to you of a practice that I did years and years ago, unfortunately and ironically, sounds pretty familiar to what a lot of originators do still today.

So we deploy generative AI through what's called an originator assistant to help the originator be aware of not only all of the loan programs that that consumer might be eligible for. So you don't just stop with your human bias of the one or two or four or six you know really well, but you look across the entire portfolio of loan products that are available to you so you can make sure you're giving them the very best loan program that they're eligible for.

But we took it a step further. We also looked at the near misses. So what are all the loan programs that that bar were just missed out on? That normally as an originator, I would never bring up because they don't qualify for. But now with generative AI, I can tell the originator in a very natural language, know, way of communicating that if this consumer were to do this, close this trade, consolidate this credit card, improve their...

their credit score by seven points, they would then qualify for these additional six or 10 or 20 loan products. So now you're really putting the long-term best interests of the consumer first and foremost beyond just how do I get to the quickest yes possible? So that's a way that you're using generative AI as an anti-bias tool to help that originator be able to get to yes,

Jim (: 29:09

Mm. ⁓

Joe Tyrrell (: 29:13

not only more often, so not just quicker and more efficiently, but finding a more effective long-term loan program for that consumer. So if you can do that at the very beginning of the process to make that conversation more meaningful and more impactful for the consumer, imagine all of the other touch points across that origination process where generative AI could really make a meaningful difference.

Jim (: 29:39

Wow. yeah, I mean, basically eliminates the bias and makes the process potentially more approachable to maybe by a newer originator, originator that maybe has only dabbled in certain corners of the loan product matrix, right?

Joe Tyrrell (: 29:55

That's right.

Vimi Vasudeva (: 29:56

And this is a really interesting point to bring up during a time right now where there's been a lot of focus on first time home buyers in our industry, just to encourage home ownership, get volume where we can. And so it's interesting because you mentioned the human touch and the human bias, but do you feel like the use of technology has made the mortgage process easier for first time home buyers or slightly more confusing?

Joe Tyrrell (: 30:22

You know, it's a great question, Vimi, and I think the answer is yes, which means it's done both, right? I mean, if you think back decades ago when the GSEs introduced automated underwriting, you know, that made it so much faster and more efficient to be able to determine whether or not you could underwrite someone. But it also took a little bit of the art away. I mean, I remember

You know, again, I'm dating myself, but years and years ago as a as an underwriter myself when and I won't tell you anything other than I had a three digit chums number. So that tells you how old I am. But, you know, doing FHA loans like you had all of this opportunity to be creative and looking at how you could kind of create a credit profile for first time home buyers where one didn't exist on a credit report, but you could demonstrate where they paid.

Jim (: 31:01

You

Joe Tyrrell (: 31:18

their utility bills on time and all these other things. So in some ways, it's made it easier for those that fit nice and neatly into the box. But what we're trying to do with Originator Assistant is go after the art, is to be able to make sure that we're not just trying to fit people into this small box and if you don't fit, sorry, we can't help you, but really make the box truly representative of all of the opportunities that do exist. There is a

lot of programs for first time home buyers. The challenge is that if you don't have an originator who's aware of all of these programs, and sometimes it's a combination of multiple programs, right? Sometimes it's a bond program at a state level plus a down payment assistant plus something that a GSE might offer. If you're not aware of how to bring those together, then you're not able to say yes to as many people that are actually eligible to hear that yes. So I think

In some respect, it's helped immensely. If you look at appraisals, there's another area, right? Where if you look at years and years ago when AVMs were introduced, so that was great if you had a lower LTV, if you didn't need to get every dollar of value out of that property to qualify. Again, it made the process more efficient. But when it comes to making it more effective to be able to say yes to more people, to be able to make more accurate decisions,

That's where I think we have a huge opportunity in this next kind of round of adoption and that round maybe two years, maybe five years, maybe seven years, where people are at least interested in understanding how technology can help them be more effective in this origination process.

Vimi Vasudeva (: 33:06

And you mentioned that there will be several rounds of adoption of technology. And you've also mentioned some of the hesitation from the past. So how would you advise lenders to kind get past the common mistakes that have been made that have made technology harder to adopt?

Joe Tyrrell (: 33:23

Yeah, in this industry, it's interesting because it's not only cyclical, right? Driven by what happens from a rate environment, but as we all know, it's seasonal. So the end of the year is a slower time. Historically, spring to summer is a busier time. And what happens with lenders invariably is, you come to a period where when rates are low or you're in that high peak season,

You're too busy to think about adopting something new. You just need to close as many loans as you possibly can and take advantage of the volume while it's there. And so when you're in slower periods, whether it's because of seasonality or interest rates are higher, so there's lower demand in the market, you think, okay, well, that's the time. That's when you're like, all right, let's look at every technology out there so that we're ready to scale next time we're in one of those peak seasons or peak rate environments.

so that we can scale without having to add a bunch of people. But at the end of the day, if you look at the cost of origination in our industry, it has basically remained fairly constant after we got through the housing kind of bubble that burst and all the new regulation came into our industry. Because historically, the cost to originate alone was around $3,000. And then post all of the regulations being introduced after 07-08, ⁓

That number jumped to like, you know, pretty steadily to five, seven, eight, nine. And now it, you know, it'll change, goes to 10, drops to nine, goes to 11, but it's all determined by volume. So what that tells me is, is that there hasn't really been any meaningful embracing of technology to eliminate the dependency on all of the human resources and human capital that were still required in this industry.

And so for me, I think that the opportunity for lenders is to say, okay, I need to do something different if I want to get a different outcome. Otherwise, my cost of origination is only going to go down when volume goes up. And when volume goes down, my cost of origination is going to go right back up. Now, there's certainly, it's not everybody. There are a lot of lenders that are out there that are really forward thinking and have really embraced this philosophy.

But on average, you're still seeing a lot of dependency on people to get this process done. And so for Optimal Blue, we're not looking deploy a bunch of technology that eliminates jobs or eliminates people. We're trying to deploy technology that assists individuals that are required to make critical decisions in this mortgage process so that when the volume comes back,

A lender doesn't have to go out and figure out how are going to find 20 or 30 or 50 more people. They can do it with the exact staff that they have right now because they're doing it in a smarter, more effective way, leveraging things like generative AI to do menial tasks or high value, low that enable me to scale as a person. So I can do more for the lender and the lender doesn't have to worry about keeping the same margins or even shrinking the margins when the volume's there.

because they've had to go out and recruit a bunch of people and pay a bunch of bonuses only to then worry how they're going to keep those people employed seven months later when the volume goes back down.

Jim (: 36:58

Right. And it shouldn't take much, right? As you said, we're a cyclical business. You know, we're seeing maybe a 10 to 12 % increase in volume this year. If we do though, if we see a significant drop in rates, you it's not unusual for this industry to experience 25, 30, 40 % increases year over year in volume, but just automating a few of the lower risk areas of your business can allow you to be a ton more scalable without, as you said, without adding additional people. When you see that and,

It makes mathematical sense that obviously your cost to originate goes down by every new loan that comes in where you don't have to hire a person. With tools like Originator Assistant, putting smart tools upfront, you're also ensuring that you're going to see the maximum amount of volume that is available to you as well. You're not going to be turning away loans could be contributing to improving your cost to originate.

Joe Tyrrell (: 37:55

Yeah, and I think, Jim, in our industry, the thing that's even more problematic is when volume starts to increase, it happens slowly. It happens, you know, a quarter of a point at a time, right? And after, you know, a month after a month after a month with each fair open market committee of the Fed taking place. But for interest rates to rise and for volume to drop,

That can happen quickly. And so if you're a lender, you've got to be thinking about, my opportunity to get to much higher volume is going to be a slower, more gradual climb. But there's a cliff that is at the end of every single high volume period we've had in this industry. So it's not like I get to ramp up slowly and I get to ramp down slowly. I ramp up slowly and then it just drops.

Jim (: 38:52

Mm-hmm.

Joe Tyrrell (: 38:52

And so

leveraging technology, especially at a period right now, is one of the advantages lenders can create for themselves six months from now or a year from now. So not only do they be able to manage that gradual increase in volume and capitalize on every dollar of margin that they could get, but they're actually much better positioned for when the cliff comes.

Jim (: 39:14

Mm-hmm.

Vimi Vasudeva (: 39:16

And right now, I imagine a lot of lenders are being approached by vendors with new and exciting technology and innovations. So with so much opportunity out there in so many different parts of the industry and potential technology, how would you suggest that lenders evaluate if a new tool is worth investing in, especially in a tight margin environment?

Joe Tyrrell (: 39:37

Yeah, so I'd say there's two things there, Vimi. One is there's two types of companies that are introducing solutions into the market. There's those that actually have developed the technology, that have it working in production, that can make an absolute incredible impact for a lender. If that company can show them specifically where in their business they would recommend them deploying it to create that high level of value.

note a low level of risk. So that's one type of company. The second is the company that comes in and says that they have all those capabilities and promises all these incredible return on investments and periods and huge efficiency gains, but they actually don't have that technology. They have components of it. They have pieces of it, but they're really going to build it out or connect it all once you sign their contract.

And they will go to any lengths to get you to sign that contract. They'll be so desperate, you a year for free just to get you to sign the contract. And then they'll go and deliver. And unfortunately, you've seen some lenders that have found themselves in that position where they've signed on the dotted line only to realize through implementation, well, these things you said you had, you actually don't have, or they're in what they call a beta period, which means

that they don't really have it. So the first thing a lender has to do is really determine what is real and what's not real. Because there is a lot of promises being made out in our industry and it's not backed by actual technology. So that's the first step I would go to. The second is I'd want to understand what's the true cost. There should be a very obvious value in deploying new technology.

You should be able to point to, if I leverage this technology and I'm spending a dollar for it, I should be able to very clearly see where this saves me $1.25 or $1.50. And I shouldn't see that it saves me 25 cents or 50 cents because now what I've done is I've just added to my cost. So you've got to be able to very clearly identify where the profitability comes from making that sort of investment.

One of the things that we've really pivoted to at Optimal Blue last year was focusing on being able to articulate the ROI of everything that we deploy, which is one of the reasons why we've deployed so much generative AI and automation at no cost to the customer. Because then it's a very easy calculation. If I deploy this and I no longer have to do that, then that's 100 % margin to me. And so...

making sure first and foremost that what people are promising you on PowerPoints or vaporware is actually real is the first step. And then the second is make sure that you can clearly identify what and quantify what the actual dollar savings are going to be to you as a lender. Because if I'm a CFO, know, people have come to me for years saying, well, we want to invest in this because it's going to improve the experience, approve the satisfaction of the customer.

And it's so squishy to try to figure out, you know, what is the return on satisfaction? When you're deploying automation, it's very easy. If I spend a dollar here, I should be able to show clearly where I'm going to save a dollar 25, a dollar 50, or even better, five or $10 over here. So if you can do those two things, then you should absolutely be trying to embrace that new technology from any provider who can actually deliver it.

Jim (: 43:30

Right. it's something, yeah, something just jumped in my head as you were saying that, and just in general, technology is a risky business, right? A risky investment. Even in our own lives, you know, in your 401k, if you choose one of the tech-heavy mutual funds in your 401k, it's going to say this is high risk, but potentially high return, right? So there is even a lot more at stake when you take your business and invest in another business as a partner. We've certainly seen

you know, in our industry and every industry over the years, you'll see these logos pop up. You know, we have a ton of them on our website that we integrate with, for instance, and they're all great partners. But every once in a while, you know, they don't all make it right. There's, there's just a risk in this industry that you do kind of start out having to promise things to people. And then at some point the costs outweigh what you're bringing in and it doesn't work out or the technology doesn't pan out the way that, that it was advertised.

So that's a really good way of, think, looking at this. And that is the way that companies should look at implementing any sort of AI in their business that they plan on having around for five, 10 plus years.

Joe Tyrrell (: 44:36

Well, and what's funny, Jim, is that I've heard lenders tell me directly that, yeah, this tech company came in and made all these promises. And then at the end of the pitch, they said, and if you want to, we can also give you the ability to invest in us. I'm thinking, ⁓ my gosh, like, if I'm a lender and I hear that I am running for the hills, I mean.

Jim (: 44:53

yeah?

Joe Tyrrell (: 45:01

incredible about the position that we're in at Optimal Blue, obviously, is we're backed by an $85 billion public company, but we get to run, you know, in a very autonomous, you know, fast-paced tech company startup way. So you get the best of both worlds with Optimal Blue in the fact that we don't need to ask you for money so that we can keep running our business. very well established, very well funded.

Vimi Vasudeva (: 45:22

Thank

Joe Tyrrell (: 45:28

And our focus is on just investing in value that we can deliver for you and technology that's not only to make you more efficient, but make you more effective as a link.

Jim (: 45:38

Right. Well said. got a bunch more questions. think one just to tie out the AI discussion, if AI is not going to kill us, which hopefully it will not, ⁓ but where is it just across the globe? How quickly is this exponential growth that people talk about, do you think? know that nobody can know, but this exponential growth of the power of AI, like the programming.

Vimi Vasudeva (: 45:49

Mm-hmm.

Jim (: 46:04

strength of it, but also how it proliferates across businesses and how it's adopted. Where do you see that in the coming years and decades?

Joe Tyrrell (: 46:14

Yeah, the way I look at it is I think you'll see periods, right? And this next period that we've just entered AI could last three years, could last seven years. But what's going to happen within that period of time is, and let's just talk about our industry specifically, lenders are going to get more comfortable deploying it in very use case specific scenarios.

and it's going to enable them to build trust in the technology. And so I don't see in this next period, this next, whether it's three or seven years, I don't see a lot of lenders running to deploy agentic AI, for example. I don't see them making a rush to say, I want to get rid of my underwriters and I want agentic AI to make all of the underwriting decisions. Because again, there's a period of trust that has to be built.

Same thing that happened years ago when the GSCs introduced automated underwriting or when AVMs first came out. think we're entering that period now where they're going to start to think about deploying it in far more parts of their processes, more use cases, so that they can gain trust. And I think you'll also see it's going to be people that have historically not leveraged AI. So perfect example is if you look at Optimal Blue's profitability assistant.

So this is generative AI targeted to somebody within a, that every lender has, but nobody's really focused on delivering value for them, which is the CFO. So the CFO, you know, is, on a daily basis, whether they're incredibly, you know, astute at managing spreadsheets or not, which typically CFOs are, it still can take a CFO anywhere from 45 minutes to three hours just to reconcile

Jim (: 47:49

Mm-hmm.

Joe Tyrrell (: 48:08

What was their loss or gain on the previous day's loan sales? Because they're pulling data out of multiple systems or downloading reports. They're staring and comparing across three different screens to make sure it all adds up. Profitability Assistant is a tool that will go in and in real time instantly tell you, here's what your loss or gain was in a very conversational way. So again, using natural language processing, LLMs to be able to come back and say, your loss or gain yesterday was this.

And here's the reasons why, and then present additional data either graphically or in a table, whatever they prefer, that allows them to click in and then getting to deeper levels of granularity. Like that 45 minutes to three hour savings allows that CFO to now think strategically about what do I do to change the outcome of today? And so I think there'll be periods of time where you're going to find these great use cases, you'll deploy the technology, you'll build up trust.

And then that'll allow you to want to adopt more and more and more. And so I think this next period is going to be all about specific use cases, proving it to the minds of the lenders, building trust in the technology. And then I think that second period, whether that's another three or seven years, you will see a lot of deployment. You'll see a lot of the manufacturing of loans look a lot more like the manufacturing of cars.

Right? The manufacturing of cars is almost entirely done through robotics. So that's, you know, there's not a lot of folks that are on the front line still pounding rivets into cars. It's robots that are doing that. And so I do think there'll be a period of time when the technology will be able to evaluate structured and unstructured data very similarly. So being able to look at W-2s and pay stubs, which is structured.

Jim (: 49:40

Mm-hmm.

Joe Tyrrell (: 50:03

and a divorce decree which is unstructured and get to the same conclusion on both of whether or not it meets eligibility. That will unlock a level of scalability in this industry that we have struggled with forever. It will take your underwriter from doing one or two loans to being able to do 20 to 30 to 40 loans. And so that's a level of both efficiency and effectiveness that I frankly can't wait for.

Because I think all that does is it just makes the American dream of home ownership a reality to far more people, far faster than we can get there just trying to do it on our own.

Jim (: 50:41

Right. No, it's another good point. that's, you you mentioned we've just, we've scratched the surface of how optimal blue is implemented some AI, right? But the, you know, the profitability assistant for, for instance, is near and dear to Vimi and me for sure. It's, in our, our Compass Edge platform. And it's another one of those tools that it's not, you know, it's not meant to eliminate people, but eliminate work, right? To say, we want our folks on the desk spending a lot more time advising clients working.

Joe Tyrrell (: 50:48

Yeah.

Jim (: 51:10

directly with clients versus kind of troubleshooting in the morning. ⁓ it allows us to do something similar that that CFO is doing, is automate that so I have a lot more time to think strategically.

Joe Tyrrell (: 51:14

Yeah.

Yeah, and maybe one last thing that you just reminded me of something going back to Vimi's question on how should lenders be evaluating. There is a third thing that you just reminded me of is, if you're coming in and you've got all this cool new technology and you're like, solves all these problems, but you've actually never sat in the chair and had to solve those problems yourself because you've never been in capital markets for a lender. You've never managed secondary marking. You've never been an underwriter. You've never been an originator.

You know, I would be very skeptical of how do you actually know you're solving the right problems and how do you truly know all of the intricacies and granularity of what that problem is unless you've actually done it. So what I love about being here at Optimal Blue is you've got this incredible domain expertise. Like we have people that have run capital markets. We have people that have run lock desks. We have people that have run production, origination.

servicing, secondary marketing, who've been originators, and they're the ones building the technology. So it's this incredible marriage between this expertise, this domain expertise, and cutting edge innovative technology. And when you bring those two together, that is something that if you're a lender, you can trust. Because the people that are building it have had to live with the consequences themselves.

Jim (: 52:52

Indeed. Indeed. yeah, when we certainly on our desk, we're very, kind of hand in hand with, the product development folks. And there is a feedback loop there that, yeah, I don't think you can build this kind of technology without the people who actually solve those problems working through it. And in iterations too, it doesn't work on the first day. You know, we've, we spent a lot of time these assistants before we actually rolled them out to, to clients. Cause we knew at that point that they passed our test. So they're going to pass the, you know, pass the client's test.

Joe Tyrrell (: 53:20

That's right.

Jim (: 53:22

All right. So before we wrap it up, I'd be crazy not to not to ask while we have you just general industry stuff at this point, right? Where do you see this kind of long period of industry consolidation going? Where does that end? Does it end? Does it continue as you know, do we continue to see large players in the space consolidate? And then just generally, of course I got to ask, even though nobody has a crystal ball, you know, where do you see rates and volume going over this next couple of years?

Joe Tyrrell (: 53:52

Yeah, well, let's start with the rates first. I think there was a lot of hope at the beginning of the year that in the April, May timeframe, we'd start to see some semblance of kind of a normal, whatever that means, normal spring that we're used to seeing in the industry. And it just, didn't, it didn't happen. It didn't materialize. You know, we've, as you've mentioned, you know, rates have come down, but they tend to go back up again. So.

We're really not in that period where we've started the nice gradual climb towards higher volume. I think, you know, obviously what's happening from an overall inflationary perspective is still having an impact. I think what you're seeing with consumers, general concern over the cost of everyday goods is causing a lot of folks to sit on the sidelines. Even things like, you know, just the availability of inventory with so many

investors acquiring homes instead of just traditional, you know, consumers that are using it as their primary residence. All of those things are causing kind of the continued constriction in our industry, in our industry specifically. So that normally would mean in any other year, as we're now in the midst of, you know, smack in the middle of summer, that you kind of missed your window and now you got to wait until next summer or next spring.

to start to hopefully see some of those traditional signs that the industries and the market's gonna break loose a little bit. I hope that's not the case. I hope that we get maybe more of an Indian summer and we can see into September and October loosening a little bit from an interest rate perspective. As you know, we're not that far away from a good percentage of current homeowners being back in the money.

for those that kind of got in towards the tail end the low rates. So I'm hopeful that there'll be those opportunities, the realist in me expects that it's gonna be still a little bit of a slog for a while. And why I'm such a proponent of take advantage of this time to really evaluate where you can deploy more innovation so you are prepared.

Jim (: 56:03

Get ready.

Joe Tyrrell (: 56:08

so that when that gradual increase comes, you're able to capitalize on every dollar of margin that you possibly can. then in the broader, answer to the, or my perspective rather on the other question regarding broader market consolidation, I think consolidation is gonna continue to happen. I think what's really interesting is what you saw with Redfin. I mean, I think there's gonna be some interesting,

models that are going to start to take place, where people are going to look at how do we really strengthen strategic relationships at the point of thought, to the point of sale, to the point of origination, to the point of delivery. And I think you'll start to see a lot of companies want to think creatively about what does our ecosystem look like? One that we control, one that we have more ability to, you

have an impact in delivering more in the front end to us and then having a cleaner delivery on the back end. think what we've seen take place in this industry is you have these big events that take place, these big consolidations, and then it has this ripple effect down to the next couple of levels of lenders where they're all figuring, okay, well, what does our version of that look like? And that typically involves further consolidation.

Jim (: 57:26

Mm-hmm.

Joe Tyrrell (: 57:30

I think you'll see more of that through probably the rest of the summer into the fall. it'll be interesting to see what happens and what comes out the other side of this and if the model truly changes or if it's really just kind of marketing connectivity that doesn't really deliver a true change for a consumer or for a lender. So for me, a lot to wait and see.

Jim (: 57:56

Yeah, interesting. Everybody's taking a bit of a risk to try to build new models right now and what's going to take and what, or will we just revert back to the mean, which is what's worked for many years, at least in general.

Joe Tyrrell (: 58:09

That's right.

Jim (: 58:11

All right. Vimi, any other questions on your mind today?

Vimi Vasudeva (: 58:16

many others, but I'm out of respect for Joe's time. I will wait till our next podcast where we have Joe as the guest.

Jim (: 58:22

Part two.

All right. Well, yes. Thank you so much, Joe, for joining us today and talking with us. Great interview. Thank you, Vimi. Good discussion. And we'll talk again soon.

Vimi Vasudeva (: 58:31

Yeah, thank you, Joe.

Joe Tyrrell (: 58:31

Great, thanks again.

Jim (: 58:34

All right, let's wrap this thing up. Thanks everybody for listening today. Thank you, Alex, Joe, Vimi, really good show today. Join us next week for another episode of Optimal Insights, where we'll continue to provide you with the latest market analysis and insights to help you stay ahead. Check out our full videos on YouTube. You can also find each episode on all major podcast platforms. Thank you for tuning in to Optimal Insights.