00:06From Las Vegas, Nevada, I'm Allison LaForgia, and today I am sitting with Optimal Blue's Kevin
00:12Foley, who is the Director of Product Management. Kevin, thank you for joining me today. Yeah,
00:17happy to be here, Allison. So there are no shortage of conversations surrounding AI and
00:23mortgage right now. From your perspective, what are the most important shifts that we should be
00:29paying attention to? So, honestly, the most important shift that I would say is that AI is here
00:36and AI is here to stay. So if lenders are out there and you're still thinking about AI as a
00:44buzzword
00:45or a fad or a distraction from your business, that's the big shift that lenders need to be making
00:51now is to understand that AI is here to stay. And the question in my mind is not whether lenders
00:57will
00:58adopt AI, but how lenders will adapt to the reality of AI becoming ubiquitous.
01:03But I'd make one important point there, which is that we're not talking about replacing humans.
01:09We're talking about AI augmenting human capabilities. So that's something that we call human in the loop,
01:15where AI is serving up insights for humans to ultimately continue to be the decision maker
01:22and ultimately still own that decision. But that's certainly not without its challenges. So
01:28implementing AI is a whole other story and there are lots of real challenges to that.
01:33So I think you've made it very clear that the potential is clear to the industry.
01:38Let's talk a little bit about what challenges going into a little more depth on your last answer that
01:44lenders face when moving from that, just adopting AI to adapting to AI.
01:52So it's a great question because I see, so when I try and get a pulse of where lenders are
01:58at today,
01:59the sense that I get is there are a lot of lenders who are out there in the experimentation phase.
02:04So you might have some more technically minded folks within your organization who have leading
02:09initiatives or helping educate the organization about AI and all that it has to offer. And that's great.
02:16But the real value from AI is going to come not from keeping this a tech focused conversation,
02:24but turning this into a people focused conversation and thinking about process change.
02:30So the real value of AI is going to require a wider set of cross-functional stakeholders within your
02:38organization to get into alignment about how to implement AI, embed it within your workflows,
02:44embed it within your processes, embed it within your governance structures, your accountability
02:48structures. How do you establish trust within your organization? How do you measure the success?
02:54All of that is a people focused conversation. And those are the real challenges for lenders being able to
03:03make the most of their AI investment. It's not simply a tech conversation,
03:07it's a people conversation. Let's dig a little bit more into what makes the mortgage industry
03:13particularly well suited for AI driven insight. So that's a great question. I think we have a lot
03:20of benefits. And one of the main ones is that the mortgage industry, compared to other industries,
03:25we sit on a lot of structured data. And so what do I mean by structured data versus unstructured data?
03:31So unstructured data, that's going to be like the transcript of this, of this conversation,
03:35right? Free flowing, you know, we're talking about a lot of things. Structured data would be
03:40a condensed set of bullet points about what we're talking about, right? And when you translate that
03:45into mortgage, so you have your loan, you have everything that's part of your loan, you know,
03:52one facet of that's going to be your rate lock. And then within your rate lock, you're going to have
03:56your lock date, your lock expiration date. But you have all of this information that's packaged up
04:02in a way where we can understand the objects and relationships within that data. That's very
04:09important, because when we're feeding in structured data into AI, it doesn't need to guess or try and
04:15interpret a lot of those objects or relationships. So we're already a step ahead. And the mortgage industry,
04:21we have a ton of structured data, we have organizations like Mismo, Optimal Blue is a
04:25Mismo partner that are all about data standardization, we have a lot of structured data within Optimal Blue,
04:31and that helps us move the ball forward, it kind of provides us, you know, a few steps beyond the
04:37starting line so that we can kind of make the most of that AI investment. And what it allows us
04:42to do
04:43is, you know, two things, really. One is AI can save people time by generating insights faster.
04:51And the second is that we can generate insights that humans just can't reliably see.
04:56Kevin, let's talk a little bit about what this actually looks like, and what strategic problems
05:02Optimal Blue is looking at tackling with AI, and where you see it actually solving real operational
05:10challenges today. Yeah, well, that's a focus that we take into everything that we're doing,
05:15we have to we have to solve real problems that, you know, our lenders are facing. But they really
05:22fall under two categories. So I'll talk about that. And that ties into my last answer, which is,
05:26number one, we are saving lenders time by generating insights faster. And the second is we are generating
05:33insights that lenders aren't going to reliably see on their own. So I'll give you a couple examples.
05:39When I got started on the tech side of this business, I was helping lenders understand their
05:44hedging positions and how those change day over day. So if there's, you know, big market movement,
05:50or there's a lot of loans coming into your out of your pipeline, you might have your secondary gain
05:55loss that changes more than you would expect, right. And so some some mornings I would come in,
06:01and I would spend like up to two hours figuring out what exactly drove the change in the secondary
06:08gain loss for lenders. So that's, you know, real time. And the thing is, when I was doing that work,
06:14and it was taking me longer than expected, trying to solve these tough problems, I was probably falling
06:19behind in other areas, right? Well, with Optimal Blues position assistance, that that whole process
06:27happens automatically as soon as the pipeline is run. So we have an AI summary showing lenders exactly why
06:34their position changed overnight, and explaining every detail behind that. So that's, that's a whole
06:40set of, you know, work that we don't need to do anymore. And if your lenders were doing this
06:47themselves, now their teams are freed up to work on more productive work. So that's one example of how
06:53we're generating insights faster and saving lenders time. The second example, I'll use originator assistant,
07:00and this is getting back to the point of generating insights that humans can't reliably see themselves.
07:06So originator assistant, it's embedded within our pricing engine. And on every search where we're
07:10looking for eligible products and pricing, we're serving up small changes in the loan scenario or
07:16the borrower profile that can help provide better terms for the borrower over the 30 year life of the
07:22loan. So these are small adjustments that you can make small changes to your loan amount,
07:26your credit score, things like that. And we're serving those up now on every single search that
07:31that is being run. If a loan officer was going to try and do that manually themselves,
07:37it just wouldn't work at scale. They wouldn't be able to reliably do this day over day.
07:42And a loan officer might also have some of their own biases about what product is going to be best
07:48for
07:48this borrower, where they're not getting a second opinion or another perspective on what could be
07:54best for these borrowers terms. So that's another example of how we're helping produce insights that
08:02humans just can't reliably do on their own using AI. Last example, we recently had our summit out in
08:10Scottsdale. It was a great success, great time. And we launched our virtual economists. And our virtual
08:16economists, it's our AI and machine learning based forecasting tool. And that's just another example
08:21of how we're bringing AI insights for lenders. So we have to talk about virtual economists.
08:27You just brought it up. For those who haven't seen it and weren't there, what is virtual economist
08:33and how does it change the way that teams approach market forecasting? So it's a great question. It's
08:40something that I worked on personally, very, very excited about it. And so the virtual economist,
08:46it's our forecasting tool. It helps provide forecasts for interest rates and lock volume.
08:52And it solves three of the major pain points that lenders have if they're trying to generate forecasts
08:56today. So everyone in our industry wants to know what's going to happen with interest rates. That
09:01helps us better understand how do we prepare to grow our business? How do we staff effectively?
09:07There's so much that goes into that. But every time that we go, if we're looking to produce forecasts,
09:14we run into three problems. The first is the maintenance of that. So even just in the last month
09:19between our Optimal Blue Summit and here at ICE experience, the market has changed pretty
09:24substantially, right? Interest rates have gone up. There's geopolitical shocks that are happening.
09:30And if you created a forecast a month ago, that forecast is now out of date. You are in a
09:35very
09:35different place. Exactly. Yes. So the virtual economist solves that problem by having continuously
09:41updated market data. So we're always pulling in the latest of what's happening so that you don't
09:47need to worry about your forecast becoming out of date. The second issue is explainability. So if
09:55you're generating forecasts, you're not always going to be able to generate the why behind the forecast.
09:59Okay, these are the numbers, but why is this what we're seeing? And the context is very important
10:05when you're having these conversations. Exactly. And the virtual economist solves that problem by
10:11just simply being able to provide that explanation. So you can ask it, how did you come up with this
10:17forecast? What goes into it? And then the third issue is interactivity. So unless you have the person in
10:26the room with you who's developed your forecast, you can't ask, well, what if this changes? Or what
10:31if that changes? How did the how did these numbers become adjusted? And the virtual economist, you can
10:37do exactly that. You can ask what happens to interest rates if the price of oil stays high for the
10:43rest of
10:43the year, or what happens to interest rates if the price of oil returns back to a baseline. Those are
10:50things
10:50that the virtual economist is tailor made to help provide those insights. So that's that's kind of
10:56an overview of what it does. And again, we're really excited about it. And looking forward to continuing
11:03to work on it. It sounds very exciting. So I want to end with asking you to get your crystal
11:09ball out and
11:10look at some future items. How do you think that AI will separate the lenders who succeed over those who
11:19struggle over the next few years? Well, I think it honestly goes back to a lot of what we've talked
11:26about here in this conversation today. So the things that are going to separate lenders who, you know,
11:33make the most of their AI investment are going to be the ones who are thinking about adapting, you know,
11:38more so than the the ones who are questioning whether they should adopt. It's going to be the ones who
11:44are
11:45thinking about AI, not just as a tech conversation, but as a people conversation, helping establish trust
11:51within your organization, embedding AI across your processes, your governance structures, your
11:57accountability structures, they're going to be the lenders who are focused on generating insights
12:01faster, or generating insights that humans can't reliably see. Those are all of the things that I
12:08think are going to help separate out lenders, you know, who are going to make the most of their,
12:12their AI investment versus, you know, ones who, who aren't, aren't, you know, going to,
12:18you know, going to be part of that, that same wave. And, you know, finally, all of this ties back
12:24to,
12:25to the bottom line. And I have to make a plug here because, you know, speaking of the bottom line,
12:31Optimal Blue in conjunction with MarketWise, we just did a study on ROI that was published yesterday,
12:38really excited to announce that. And what, what MarketWise looked at was the value that lenders
12:44are getting from using Optimal Blue's products across our whole ecosystem. And what MarketWise
12:51found was that lenders are seeing over a thousand dollars of value per loan by implementing Optimal
12:56Blue's products across, across our ecosystem. That's amazing. We're in an environment where
13:02loan volume isn't necessarily where I think we had hoped it would be because of market conditions
13:07and the cost of origination can still be very high. So that's a very exciting number to see.
13:12Yeah, we're really excited about it. I definitely encourage everyone to go check it out. And you can
13:16see so much more detail around the methodology, all the work that went into producing that. But it's
13:22something that we're very excited about. And again, ties back into the whole conversation of, you know,
13:26impacting the bottom line, how to make the most out of your investment. And, you know,
13:31that number reflects a lot of Optimal Blue's perspective that goes into the AI that we build,
13:36the tools that we roll out, and how we're able to generate that value for lenders. So I appreciate
13:40you letting me get that plug in there. Of course. Kevin, thank you so much for talking to me about
13:45what's going on at Optimal Blue, about a virtual economist. I can't wait to see what's next.
13:49Awesome. Thank you, Austin.
Comments