- 1 hour ago
AI adoption in mortgage has moved past FOMO, but most lenders are still poking at the basics. Tela Mathias believes that the real opportunity isn’t automating the steps you already have, it’s rethinking the process: controlled autonomy that runs work in parallel, strips out repetitive tasks, and makes the borrower experience faster and more personal. Her advice for lenders is practical. Start with use cases you can actually ship, govern them carefully, and measure them against business outcomes rather than activity. She also gets into what this does to the people doing the work, and why protecting room for the human moments matters as more of the file runs itself.
Related to the episode:
Zeb Lowe’s LinkedIn
https://www.linkedin.com/in/zebulon-lowe-a02353a4/
Tela Mathias LinkedIn
https://www.linkedin.com/in/telamathias/
Want more from Zeb? Don’t forget to subscribe to LendingLife.
https://www.housingwire.com/newsletter/
The Power House podcast brings the biggest names in housing to answer hard-hitting questions about industry trends, operational and growth strategy, and leadership. Join HousingWire’s Zeb Lowe every Thursday morning for candid conversations with industry leaders to learn how they’re differentiating themselves from the competition. Hosted and produced by the HousingWire Content Studio.
Related to the episode:
Zeb Lowe’s LinkedIn
https://www.linkedin.com/in/zebulon-lowe-a02353a4/
Tela Mathias LinkedIn
https://www.linkedin.com/in/telamathias/
Want more from Zeb? Don’t forget to subscribe to LendingLife.
https://www.housingwire.com/newsletter/
The Power House podcast brings the biggest names in housing to answer hard-hitting questions about industry trends, operational and growth strategy, and leadership. Join HousingWire’s Zeb Lowe every Thursday morning for candid conversations with industry leaders to learn how they’re differentiating themselves from the competition. Hosted and produced by the HousingWire Content Studio.
Category
🗞
NewsTranscript
00:00Hello, and welcome to Powerhouse. I am not your host, Zeb Lowe. I am Rebecca, the producer,
00:05tapping in once again for the intro. So back in October at the Mortgage Banking Summit,
00:10Zeb sat down with Tila Mathias to talk about AI and the future of mortgage. Everything from where
00:16lenders should start to how automation could reshape the industry and the moments that still
00:21call for a human touch. Let's get into it. Please enjoy this episode of Powerhouse.
00:36Hi, Tila. Thank you for joining me. Thanks for having me. Yeah, of course. So we're at the
00:39George Bush Library, and you joined us on stage for a session about the economics of AI, which was,
00:45I would say, a little more ahead of your conversation than we're going to have right now.
00:50I want to zoom back a bit and talk about your perspective on the state of AI in industry and
00:57the adoption curve, where people are getting it right and getting it wrong. So I guess to jump
01:03off into that conversation, where do you see that we are actually at on the AI adoption curve? Are we
01:10further along than people say that we are? Are we lagging behind? Big picture question. I know there's
01:16a lot of generalities here. I think big picture what's been settled is what I would call the FOMO or
01:23the fear of missing out strategy, the table stakes strategy. So I think what's happened in the last
01:2718 months or so is that it's become clear what table stakes is. So things like call summarization,
01:36sentiment analysis, even voice agents are becoming really much more commonplace.
01:40And so I'd say what we've settled on is kind of what is everybody doing or are they going to
01:48do in
01:48the next say year or so? I think it is still fairly rare to see really transformative implementations
01:58of generative artificial intelligence where a part of the process has been fundamentally reimagined
02:05to do in a new way. And I think that continues to be really confounding because it requires us all
02:11to
02:11approach things so differently. And the metaphor that I use on stage is if you think about air travel
02:18before we had airplanes, right? It would have been crazy to go somewhere in an afternoon that you
02:23can now do for effectively zero. And so these previously unimaginable things are really what
02:30we want to tap into. And I don't think we're really quite seeing that yet in mortgage.
02:34Where do you think that that, again, I kind of struggle with the language in asking these things,
02:39but the origination process in and of itself is a linear thing, right? And a chronological,
02:45it's followed this happens, then this, then this, then this. And do you see this transformation being
02:52tied to whenever you have the front end kind of talking to the back end at the same time,
02:59right? You have systems speaking to each other, talking to each other in real time, which was
03:04never, you know, previously, it just hasn't been possible before. To me, there's a recipe or a
03:11formula for, again, this transformation that you can't quite imagine because it's never even been
03:16teased at or do where do you see like another area or another opportunity beyond what I'm talking
03:25about that would yield this, this kind of transformation?
03:28I think that when I look out three to five years from now, I work backwards from what do I
03:36think
03:36it looks like? And call it three to five years from now, I see a much leaner, more autonomous
03:46mortgage organization that has leaned into what I would call controlled autonomy. And so I think that
03:55one of the questions that I would ask is we talk about mortgage as a linear process. Well,
03:59and I think you alluded to this, does it have to be linear, right? How many of these things could
04:02actually be done in parallel? How many of these things don't need to be done at all? And so I
04:06think when we look ahead, you know, we see what I see is I see, you know, a call center
04:12that is
04:13basically 15% the size that it is today. I see hyper-personalized engagements with customers on
04:19their channel of choice. I see a situation where, you know, what even is the point of a mortgage
04:25organization in a world where the home buyer of the future can have instant access to the most
04:31competitive, most appropriate to their unique needs, you know, loan without ever having to
04:39engage with anyone if they don't want to. And then transitioning that relationship in much the
04:45same way, you know, in a world where we have personal financial agents, uh, that are generative,
04:51that have instant access to everything you need to know, what that is the value of a mortgage company.
04:57And so it's those types of things that I, that I start to think about, uh, and that, that, uh,
05:04I think, you know, there, there is an opportunity to differentiate thinking along those lines in a way
05:13that, um, I don't know that a lot of companies are thinking today.
05:16Yeah. Okay. So, so far, and you, I think you kind of touched on this in your first answer,
05:21but what do you see, uh, lenders, like in generally speaking, what do you see them getting right? Is it,
05:28um, what you were talking about before, like the, the, uh, recording, recording the calls and the,
05:33like that's, that's what they're doing right. But where do you see themselves or where do you see us
05:37getting ahead of ourselves? Yeah. And I think it's, I think there's, it's not really getting
05:42things right or wrong because even when we get things wrong, when we learn from that experience,
05:48you know, then we got it right. And I think that this AI business is really difficult. It's really
05:54hard to implement generative technologies at scale in a way that's safe, in a way that's responsible,
06:00in a way that moves the needle, uh, from a return on investment perspective. Uh, I think frankly,
06:06the industry is still figuring it out. I, I, I really, you know, and I see, you know,
06:11thousands of people in my classroom, you know, dozens of, um, of clients that we engage with.
06:16Uh, and I do think that we are really still figuring it out. And, you know, in theory, my,
06:21my company is one of the most, uh, AI forward AI native that there are, and we're still figuring
06:26things out. You know, this morning we were on stage with Brian and they're still figuring it out.
06:30Uh, and so, you know, I think that requires, you know, a degree of intestinal
06:36fortitude that you just have to kind of lean in and, and keep going. Um, so I, I really don't
06:42know that there are any clear winners or losers right now. Uh, I think that it's really just
06:47beginning. Yeah. So I talked about FOMO earlier for, for, for people that, uh, are, I don't know,
06:54lenders that feel like they're slow in adopting AI. Do you, I mean, do you think that the, the,
07:01perhaps the biggest caution, you know, for them is adapting or trying to adopt too quickly or
07:06adopting anything at all without, uh, without entering into the situation, knowing that they're
07:13going to have to fundamentally change, you know, aspects of their, uh, of their, of their organization?
07:18I do think, uh, it's a good question. And I do think that for companies that are doing nothing,
07:24you know, they, they really are behind, uh, because the reality is to stay competitive.
07:30People are going, are already seeing a competitive edge from again, what I call the table stakes
07:35strategy. Um, I think getting started for a company that's just getting started, there's
07:41so much to learn that the, the table stakes is a great place to start, go with the basics,
07:46right? You got to implement, let's say a Microsoft co-pilot for the company. You got to get,
07:50if you have call centers, you got to get call summarization and, and things of those
07:54nature and you're going to learn along the way, what it takes to do this stuff and what it takes
07:58to govern it and what it takes to be compliant. Um, you know, the, the, the concepts that have
08:05been introduced from a state law perspective around what we're calling consequential decisions.
08:09It's very important to understand if you have use cases that are touching consequential decisions
08:13and what to, how to appropriately govern and be able to, uh, to, to do that safely in a way
08:19that is,
08:19you know, transparent for consumers and good for business as well. So
08:23I think that, uh, getting started is important. Starting somewhere, you do have to pay attention
08:29to, to governance and safety. Uh, but, uh, it isn't, I wouldn't start trying to, I personally
08:35wouldn't start for existing companies. Great. I'm just going to reimagine the whole shooting match
08:41and, and, and start from there. I just think it's too steep of a learning curve.
08:44Yeah. I think, I mean, the elephant in the room really is job loss. Right. And, um, you know,
08:52I'm curious to see, to know what you see changing the most, you know, what jobs do you see? Because
08:59obviously jobs are going to, there are going to be these hybridized roles and, um, and perhaps even
09:06new roles that, uh, you know, that they come about, but what jobs do you see being the most at,
09:13uh, at
09:14risk, you know, at, uh, what level or for the organization would you most encourage people to
09:18begin to invest in understanding AI and making the, the most out of it besides, I mean, obviously the
09:24LO, but where else throughout an organization? Yeah. And when we look at jobs, when I look at
09:30jobs and when we look at jobs that are, that are, or can be automated, uh, those are all at
09:38risk for
09:39sure. And we do have a lot of jobs that, um, continue to be focused on, um, repetitive, uh,
09:48manual tasks, uh, in the process that simply haven't been incorporated into the core technology.
09:53The closer to data entry you are. That's right. That's right. And, and that's the reality. The,
09:58the reality is that everybody should be looking at, uh, at the use of generative artificial
10:04intelligence to aid in their, their daily lives, in their work lives, in their personal lives,
10:10because the reality is the toothpaste is out of the tube. It's not going back in the tube.
10:14Yes. There are a lot of existential questions around this technology. It can do enormous good.
10:19It also has the potential to, to do harm and that that's real. Uh, and the people who will
10:27be able to, I don't want to say future-proof, but I'll just use the word future-proof their careers
10:33are going to be those people that lean into this technology, that lean into this way of thinking,
10:39this new way of working, where you have this effectively a collaborative thought partner,
10:43uh, that can help you, uh, think through, you know, really your own, uh, career path, uh, the,
10:51the way that you're working today and, and start to change the way that you think and work.
10:55Uh, it has been incredibly, uh, expensive in our own company and in the industry to help our existing
11:04workforces, uh, unlearn, uh, because it is much easier and cheaper to learn than it is to unlearn.
11:10And what I'm finding is that when we look at, um, professionals that have been in the workforce
11:17for some time there, it's very hard for them to pick some of these things up. And even the most
11:21AI savvy, AI native can still really get stuck in a thinking of, okay, I'm going to use this
11:27to make a single thing go faster as opposed to, I'm going to use this to make a thousand things
11:33go
11:33faster all at once so that I am now free to do a completely different type of work. Uh, and
11:38that
11:39I'm actually seeing being a lot easier for folks that are just getting their careers started.
11:44Uh, and so I think that, um, the, the investment in the existing workforce as well
11:51as the investment in, uh, up and coming emerging professionals is the right, uh, talent strategy.
11:58They'd really have to look at both ends.
12:00So where, what do you think, or maybe, maybe there's, there's no, uh, sacred space, but what
12:07areas do you think should not be automated where we shouldn't be leaning into AI? I mean, the, at the,
12:12at the end of the day, I mean, the relationship is built on, well, business is built around a
12:18relationship and relationship is a human thing, right? So do you think that there are any sacred
12:22spaces or sacred areas that just, we don't go there? We keep this, we keep this human.
12:27We don't automate this. We don't lean into X, just your perspective. Cause that every person's
12:33going to have to decide that themselves. Every organization is going to have to decide that
12:36obviously for themselves. I'm just curious on your perspective there.
12:39My perspective is that there are absolutely aspects of relationships that are innately
12:46human. Uh, there are things, uh, that, that will remain uniquely human. The question of
12:54should is an interesting question. Are there things that we should just designate as, as,
12:59you know, out of bounds? Right. Uh, I think those things are more in kind of militarized
13:06functions and things and things like that. I think when it comes to a business like mortgage
13:11that has traditionally been based on, um, so relationship intense, uh, you know, people are
13:17having relationships, uh, as they experienced them with generative artificial intelligence.
13:23Sure. Uh, and those can be, you know, really helpful. Uh, you know, think about,
13:29you know, having a mental health crisis, for example, and not having access to high quality,
13:33you know, therapy, right? Like these are things that you can get from a machine as, as, you know,
13:39potentially radical and as, and uncomfortable as that may make us. When I look at what remains
13:45uniquely human in our industry, um, it truly is the moments that matter for those people
13:51who wish to work through those moments with other people. That's not everyone. Uh, and it's never been
13:57everyone. There will always be, I believe, um, some humans that simply want to work through things
14:03with humans. Uh, more and more, I think we're going to see that fall away. And then it's really
14:08is around creating more of the moments that matter for the humans that want to engage with humans,
14:12because the reality is we are going to see, um, the virtual complete automation of, uh, you know,
14:19the total mortgage ecosystem. Uh, it's going to happen. Uh, we're going to have to face the question of,
14:27you know, can an, an AI agent, uh, make a better decision than a human at scale, uh, reliably.
14:35And when that happens, we will face the legal question of whether an AI agent can be licensed
14:41as an MLO. These things are all going to happen. Uh, and so we have to lean into that, I
14:48believe. Uh,
14:49I think we have to be open to what's new and different and what I did with my own company,
14:53you know, three years ago when we made, or four years ago when we made this pivot
14:57was I kind of decided and I had to decide for us all. All right, great. Are we going to
15:02decide
15:02what the future is going to be? Or are we going to sit around and wait for someone else to
15:06decide for me?
15:07Okay. So let me ask you this. I've been thinking about this, uh, for, uh, I guess it was last
15:12week
15:12or so. I mean, you see this happen every, it's in the news every, every couple of weeks where you
15:17have
15:17a, uh, an executive, a, uh, coder, uh, uh, resigned from Anthropic or from Google over,
15:27over their concerns about the implications on AI and, and the rest of society, culture, you know,
15:33replacing jobs and all that. And I don't see that in like, to me that those people, they're at the,
15:41they're at the source, right? They're creating the thing and they're, you know, it's, we don't have it in,
15:47in our industry, people that are tied to AI in our industry resigning because they're overly
15:53concerned about the implications. Right. Right. But you do see it quite at the source. Yeah. But
15:57it's not just our industry. It's, it's any, it's any other music production industry that's using AI.
16:02You don't have the AI music engineers resigning over fear of whatever as it's tied to the rest of
16:10their industry, society. Why do you think, why do you think that is? So I think let's look at
16:17uh, once it was possible to use, uh, nuclear technology for, uh, weaponization. Once that
16:26was possible, it was then done. Right. So once a thing can be done, it will be done. That's kind
16:30of just human nature. Yeah. Uh, and, uh, I think that that's probably the closest proxy to the time
16:36that we're in right now in terms of looking at what happened during that age, uh, and what the reactions,
16:43what the reactions were, there are absolutely existential questions here. And once a thing
16:48can be done, it will be done. Right. So the weaponization of generative artificial intelligence
16:52is a very real thing. I'm sure it's already happening. Uh, and we do have to be, uh, we have
16:58to be incredibly cautious about the types of technologies that can now be enabled. My personal
17:04belief there is that, uh, we have to lean into the federal government, uh, to regulate this technology
17:11like we regulate nuclear technology or in a way that is akin to that. In terms of people leaving,
17:17it's not new that we've had, um, very ferocious debates about, uh, this concept of super alignment
17:25or alignment, which is this idea that we can and should ensure that generative technologies are
17:31aligned to human values. Right. And whose human values, whose human value, whose definition of
17:36human value and, and how does that alignment get achieved? Um, one of the earliest was Leopold
17:41Austin Brenner, who, who, you know, fairly famously wrote a 187 page paper on this exact topic. And that
17:48was years ago at this point, good old days of like 22 or 23. Um, and so, you know, I,
17:54um, I am glad
17:56that we have people that are very concerned about the existential threat associated with these
18:03technologies. And I'm glad that there are people that are so principled that they'd be willing to
18:07stake their livelihood on it. Um, you know, we don't, um, the, about the most existential of
18:14decisions that we have, and they're very important in mortgage, you know, are, who are we going to give
18:19credit to? That's a very important decision and that needs to be done, uh, fairly. Uh, the, the
18:26stakes are a little bit different in our industry than they are in, uh, say, you know, film. Yeah,
18:32sure. Um, you know, but still be good, like, you know, if, uh, the, what, whatever the decision
18:36making process is, you know, are the decisions that are being made, is it more beneficial to the
18:40lender? Is it more beneficial to the home buyer? Right. Right. You mean, are the, what do you mean?
18:45Well, if you're relying on, uh, if you're, if you're using AI, this, this future state, right,
18:52to, uh, to intercept. Right. So who gets the benefit? Yeah. Right. Is it, is the, is the loan,
18:57is the loan product, is the, is the, the, the structure of the loan, is it more beneficial to
19:02the, to the lender? Is it more beneficial to the, the, the home buyer? And does that come from the
19:08person that's, that's, that's, that's creating the program with AI? That's where, where does that
19:14direction come from? How, what's the regular, regulatory oversight? I get kind of curious
19:19whenever you can't throw somebody in jail, right? Like that's what, that was always kind of like
19:22my redneck, you know, logic on like, well, it'll be a long, it'll be a long time before there's a
19:28fully licensed AILO because you can't throw it in jail. Right. Right. At the end of the day,
19:33something goes wrong. You're going to have to, somebody's got to go to jail. Right. Fair enough.
19:38And I think that's going to be one of the questions that's going to have to be answered. You know,
19:42this idea that we can throw a person in jail, you know, the, the person is accountable. Ultimately,
19:47the organization is accountable, right? And the organization is effectively a fictitious creation
19:51anyway, right? Right. This idea, if you read, if you read Sapiens, you know, it's got a very good,
19:57um, discussion of, of how we organize and why we organize the way that we do around these effectively
20:03fictitious, you know, organizations. In terms of who benefits, when you look at the core underlying
20:09technology of a foundation, large language model, uh, the values that are, uh, implemented in that
20:17model are the values of the company that is the purveyor of that model. And we have to care. I
20:22care
20:23very much about who I do business with, uh, and especially with this foundational technology. Now,
20:29these technologies can be steered to align to how we want things to be. They can be steered by an
20:37individual consumer who doesn't want to have emojis and exclamation points. They can be steered by a
20:44mortgage company who, um, wants to ensure there's a really high quality, um, consumer experience.
20:50Uh, and so all of those things come into play. And I think it's a fair question that I haven't
20:56considered too, too much. I think that ultimately we're going to see a mortgage organization where
21:01there's a mortgage agent and there's a consumer agent, and those agents are going to work together
21:06to do business, um, to make a deal. And at the end of the day, I, you know, I'm not
21:10in the camp of,
21:11you know, money is going to become irrelevant. Uh, and I think the, the, you know, the, the goal here
21:17for us all should be to reduce the cost of debt, right? We should create, we should be creating more
21:24wealth, um, for humans so that humans can live a more meaningful life. Uh, and when we can take cost
21:31out of the process of originating and servicing alone, I think that is both good for, uh, lenders
21:39and servicers and it's good for consumers because now the cost of debt has been reduced. And I think
21:44if we look big picture and we look at all the types of debt and the cost of servicing that
21:49debt
21:49in the country, um, taking that cost down is a good thing. And so I think that there will be
21:55a balance
21:56ultimately, you know, getting a, um, a human or an organization to part with their time,
22:03to part with their dollars, there has to be value. And so it's a, it's a, it's a handshake,
22:08right? The person who has the dollars to part with and the organization that, that seeks to earn those
22:14dollars that has to be mutually beneficial. Right. I got, so, uh, we were, I guess had a little,
22:21been having a little bit deeper conversation that I originally had, uh, no, no, no, no,
22:26this is, uh, this is, uh, this is my jam. Um, but, uh, you were talking about, um, uh, you
22:32know,
22:32not in this future state world of, uh, one, reducing debt, not that money will be, you know,
22:37meaningless, but wanting to reduce that debt and make things easier and then homeownership easier.
22:41And one of the things that I was thinking about when you were talking about is that, you know,
22:44for the, for the majority of the history of man, um, your, your meaning, your purpose is tied to
22:53money. It's like, it's how you earn, how do you, you know, what do you do for a living? How
22:57do you,
22:58you know, provide for yourself and provide for your family, your purpose and your meaning is tied,
23:01you know, to that. And so, but that's, that is one of the big existential, uh, questions about,
23:06about AI and AI taking over so many, uh, jobs. Like what do people do whenever the thing that
23:15historically, especially in our country, United States, like you ask, like, you know,
23:19tell me about yourself. It's your job. Like that's the, I'm a, my job title. Right. And so like the,
23:24kind of the, the, the, the search for meaning and purpose, whenever much of that has changed,
23:29I'm curious to see how, what you think about that and its implications into the mortgage space,
23:34like as a loan originator, right? Like your job's going to fundamentally change.
23:39And so much of like a loan originator, their job is like problem solving. Right. Right. And they
23:43derive pleasure from that and like that. So you'd ask a loan of it, you go to at happy hour
23:48here,
23:49or you go to any like sales summit, you're going to find originators having a few drinks and they're
23:53going to walk through the intricacies and the complexities of a non-QM file or a manual underwriter.
24:00And they like the joy that they get beyond the transactions to solving the problem.
24:04And so, so much of the things that originators derive their purpose from, and they find meaning in,
24:13is going to be removed for them or from them, right? It's going to be done by AI. So what
24:18would,
24:18like, do you have any thoughts or guidance on how originators should start to
24:28discover an alternate purpose or find meaning in a very different world than they've ever
24:34before they've ever had before?
24:36Well, and I think if we look at, as opposed to the history of mankind, humankind, we'll say,
24:42you know, the, we're in a, especially in kind of the mortgage industry, this place of
24:50abundance rather than survival.
24:52Yeah.
24:52Um, now that is obviously coming from a position of privilege. Uh, and so I don't take that,
24:57that lightly. Uh, but when we talk about, you know, how do I derive a sense of meaning that
25:02comes from a place of being very high on kind of the, the hierarchy of needs, right?
25:07So when we're talking about professionals that already have, you know, housing and food and
25:14safety and all those types of things solved, then, then we can look at this quest for meaning.
25:19Um, and I think we all derive meaning in different ways. Um, I, you know, certainly there are many of
25:25us who derive a great deal of purpose and a sense of meaning from our jobs. Uh, and that's, you
25:31know,
25:31that is one dimension, um, of, of creating purpose in life. Um, I think that we, it really is
25:42to us to settle on, um, um, what, you know, you know, what is that future going to look like
25:54and where will that meaning come from? People who derive meaning from their work will continue
25:58to derive meaning from their work. It will be different work. What that work is going to be
26:03can in large part be shaped by that individual loan officer, right? That individual loan officer
26:09can already lean in to whatever the technology is, uh, and automate so much of, of the work should,
26:16you know, he, she, they choose to do so. Uh, I think one of the ways that we can prepare
26:24ourselves
26:25for this future, frankly, is leaning into things that can't be automated at all.
26:29Yeah. Um, so we were talking earlier, I'm learning to play the banjo. Um, and I've taken five lessons
26:34and I've been at this about a month and I can play one song, not very well, right? There's
26:39nothing about this process that can be automated. Right. It can't be accelerated. There's no,
26:43you know, artificial intelligence for banjo. There just isn't. Right. And I think that unlocks
26:48a part of our thinking and our brain and it brings us back to earth. Um, and it takes us
26:53outside of
26:54this artificial world that we're in and into this world of, you know, of, of humanity and,
27:00you know, sure you can have a machine, uh,
27:02a weird juxtaposition to where the more, the more artificial the world that also allows you
27:07on the other side of that allows you to become, allows you to become more human.
27:09I, I think so that that's the only advice that I have. And I think, I think that there is
27:14no
27:15shortage of creativity in human spirit, right? We will find things to do. Yeah. We will find ways to
27:22create. We will find problems to solve. There will always be more problems to solve. We just may not
27:27know what those are right now. Okay. I got, I'm going to take this way back and ground this into
27:32something that specifically for, for lenders and, uh, and organizations. If how can, uh, how would you
27:39recommend that a lender, uh, or even a branch, whatever, uh, a lending organization assess whether
27:49or not they are, uh, they are truly integrating AI into their, into their organization versus their,
27:57you know, I don't know, they're, they got a chat GPT pro account. Right. That they made the,
28:03their admin start using and LOs. Right. What's the, what's the framework to kind of like, no,
28:09we're, we, we're actually being serious about this and we're taking the correct steps to be
28:13an AI integrated company versus just bolting on stuff. Yeah. And I think it's not so bad to bolt
28:19things on because I think it's a place to start and starting as part of the process. We definitely
28:23can't finish if we don't start. Uh, I think it's important to have a set of outcomes in mind.
28:30Uh, even if you don't know how to get to those outcomes and maybe it's accelerating customer
28:34satisfaction or increasing, you know, creating happier customers. Yeah. Great. Well, we think
28:39we can make happier customers if we are faster to resolve their problem. Awesome. Well, we've got
28:46these types of problems and a subset of these, we think we can use generative artificial intelligence
28:51to help us solve. Great. So then go and do that and then measure against that metric. Right. So I
28:57think it, it's, can be really foundational. These are things we're already measuring today in, in many
29:02cases. Um, they may even be things that, you know, we might contemplate, um, uh, automating through
29:08traditional technologies and maybe that's still the path. So I think AI strategy is really just
29:13technology strategy and technology strategy is just strategy. And so artificial intelligence is but
29:18a means to an end to move the things that are meaningful in our organization today. And as
29:23long as we have that lens and we're not just looking for, you know, we're not just having AI as,
29:28you know, we're looking for a reason to use it. Um, you know, then I think that's a, a more
29:34beneficial approach. Right. Excellent. Excellent answer. Thank you so much, Sheila. I really appreciate it.
29:38Yeah. Thanks for having me. Yeah. We love it.