00:00From the George Bush Center in Dallas, Texas, I'm Allison Laforja, managing editor of HousingWire's
00:12content studio, and I'm with Joe Wellu, the CEO of Total Expert. Joe, thank you for joining me
00:18today. Thanks, Allison, for having me. I'm really excited to talk about how far Total
00:23Expert's AI has come. I mean, the last time that you spoke with HousingWire was back in February,
00:28so the roadmap has changed a lot since then. Take me through it. Yeah, so when we last spoke,
00:34we were just kicking off phase one of our first AI agent pilot with a customer, a top five lender,
00:42and the initial use cases we were going after was really about qualifying a lead, taking outbound
00:50and inbound calls, and really teeing those up to the point where they are ready for advice on loan
00:57options. And so I am blown away by the progress both our teams have made and the collaboration
01:07we've been able to do with our customer. So we anticipated when we first launched the voice
01:12AI agent, we anticipated that it would probably take maybe nine months to a year to be able to get to
01:19parity with a human that's really serving that inside call center type of role. And we're blown
01:27away. It happened within two weeks. And it's been incredible to watch it evolve. We're doing over
01:332 million calls a month now. We have multiple lenders that are in various phases of scaling,
01:39scaling, and we are rapidly executing on a bunch of new use cases for agentic AI.
01:46Well, congratulations. It sounds like the progress has been rapid and fast.
01:53Yeah. What I would say largely about the pace of innovation, it's like nothing anyone's ever seen
02:00before, right? I mean, the models I spoke a bit earlier about, there's hundreds of billions,
02:07if not soon trillions, being invested in infrastructure and large language models.
02:13And as an industry, we get the ability and opportunity to build on top of those capabilities
02:19and reimagine what we can do in our industry. And when you have that kind of pace of innovation
02:26happening, it just really provides this environment to where you can move incredibly quickly.
02:32So talk to me a little bit about this pilot program and some of the lessons that you've
02:38learned from the implementation of it now that it's live.
02:41Yeah. So there are all, like everything, it's iterative. And we've just been incredibly fortunate
02:49to be blessed with some amazing customers that are great partners. And they think about it in terms of
02:57a very tight feedback loop, meaning we're going to iterate through things. So as we learn, or maybe
03:02in some cases, do things incorrectly, it gives us information and we make adjustments and then
03:08things improve, right? So I think ultimately, some of the decisions were enabling the teams inside our
03:16customers correctly, making sure that they had clarity on what goals we were shooting for,
03:22right? So a key thing that we've learned with AI projects in general and AI transformation in your
03:29business, get really super clear about what it is in the business that you are improving. Give them
03:35that target. And then that allows you to see if you're successful or not successful. So it's not
03:42this ambiguous sort of black box. You have clear defined targets.
03:47I think that's so interesting because I feel like that's some of the confusion that happens in this
03:54conversation about AI, that undefined target, right? Yeah. It's sort of like, I did a blog post
04:00recently and the people that are not close to technology, they'll ask me about AI and what it's
04:06doing, what it's like. And it's sort of like humanity just discovered fire and you have teams just
04:15wanting to light everything on fire all the time without necessarily thinking about where can we
04:20unlock the most value quickly? Where's the lowest friction, lowest risk of execution? How do we
04:27control the blast radius from a compliance standpoint? So you can take risks but do it in a
04:34very controlled way to protect the organization, right? So you have to really re-enable your organization
04:42from our experience. So let's talk about some of the lift that you're seeing in terms of both lead
04:48conversion or operational efficiency. So we are seeing in some cases 10% to 20% better conversions.
04:56Okay. And if you think about agentic AI and an AI agent and this concept of digital labor,
05:03well, the cool part is it's very human-like. It actually has empathy, if you can believe that.
05:10But it is precise. It always remembers to call people back. It never goes off script. It never
05:17calls in sick. It doesn't take PTO. It works weekends, seven days a week, evenings. So you just
05:23have this sheer horsepower as an organization that you can tap into. And it provides incredible leverage,
05:32leverage that you would have otherwise had to get from humans and training and hiring and firing.
05:40And it just allows you to take your great people and amplify what's possible for them and have them
05:48doing the most highly productive work possible. So on that highly productive work note, how are you
05:56seeing this impact the borrower experience? That's a great question. One of the pleasant surprises that
06:05I've noticed about agentic AI and voice AI in particular is the quality of the experience to
06:11the end consumer. And many times it's up to the lender if they want to disclose upfront that they're
06:19talking to a virtual assistant. Or if they don't, there's some ambiguity on that. But every time that
06:26I've listened to a call where the consumer understands they're talking to an AI agent, the quality of the
06:33interaction is so high, they continue down the path. And if you think about that in comparison to
06:40experiences in service-related industries where somebody's having a bad day, the mood is maybe off,
06:47this is always in a perfect mood, the right tone, it has the ability to match sort of the tempo of the
06:54conversation. And so the end consumer experience, in my opinion, is becoming elevated because of the
07:00consistency, the ability to tap into data and immediately have context about all of the things
07:10we know about the customer. So if you think about that in comparison to an average person that may be
07:16on a call and you receive an inbound lead or I'm calling out to somebody, I'm not going to
07:21immediately in the moment retain vast amounts of data. And I can't infuse that into the conversation
07:28and personalize that conversation. AI gives you the ability to do that. Now, I think something that
07:34is really interesting that you've alluded to in this conversation and that I know that you've spoken
07:38about previously is that total expert uniquely positions itself as having AI that supercharges
07:46the loan officer. And that there's a lot of clarity saying that this is not a replacement.
07:52How are you ensuring that that message lands with your user?
07:56Yeah. So, well, it doesn't always, right? I think when people see the power of the technology,
08:01sometimes people are threatened by that. What I firmly believe is that in our industry,
08:09home ownership in America, there are still moments where consumers want high quality advice.
08:18And the goal is to take a loan officer and put them in a position where they are spending the most
08:26of their time, the majority of their time, having the highest quality conversations in the
08:31moment where consumers are needing advice, which is where they provide the most value.
08:35And so if I think about abstracting away things that don't add value, well, it doesn't add value
08:41to the consumer or to the loan officer if I'm calling them asking if they're actually ready to move forward
08:47or I'm touching base with them on if they have any questions or if I'm checking in post-sale.
08:53Those things don't necessarily add value. But what does add value is the moment where they're
08:59deciding on loan options and how those different options impact their financial future.
09:05And we believe that really highly trained advisors, loan officers have the unique ability to walk
09:11people through that and provide that advice. And so we think about it as really this human plus
09:17technology force multiplier, hence superpowers.
09:21I love that. It's allowing them to do what they're best at, to make the most impact.
09:29And to your point earlier, have that superpower to retain all of that information and pull it through
09:36their whole conversation and really set them up to maximize that impact.
09:41Yeah, absolutely. It is really about impact.
09:43So you've mentioned that there's a lot of emphasis on building intentional, scalable AI.
09:51There's no randomness over a total expert. How does that philosophy influence your roadmap
09:59and decision-making process?
10:02So first of all, we believe that when you're an enterprise in lending or banking,
10:09you're going to be making decisions on partners. You're going to build some things. You're going
10:14to partner on some things. And when you do choose partners, we believe having partners that are
10:20focused uniquely and specifically in the vertical of lending, meaning they are obsessively focused
10:26on those use cases, on your compliance, on your guardrails, really on understanding the data set.
10:33You're going to get better velocity of innovation when you use those partners.
10:38And so we also then think through and collaborate. In fact, I was collaborating with multiple executives
10:45already today. And we were talking about how to prioritize some of the next use cases that we
10:51believe this is going to solve. And it's really about where can we get value quickly, but also
10:58where can we most likely execute? And where can we, where do we have the least amount of risk of
11:05execution? Meaning what are the barriers to going live? We heard the CTO from Rocket talk a minute
11:12ago and he talked about speed. And I bet he said speed three or four times. And I believe it's
11:19absolutely critical for lenders that, you know, every day that you are not in action making progress on
11:26these initiatives, you have competitors that are. And so we think about getting super clear on those,
11:32on those targets, on those use cases and partnering with people that are going to be as obsessive as
11:39you are about making it great, making a great solution. I want to stay there for a moment. What is
11:45next for Total Expert? What excites you the most about the next 12 months? So, so many things and
11:52probably don't have time to share all of it with you, but in short on really unlocking the power
11:58of our platform and all of the partnerships and data that we have, our customers have in giving them
12:05additional capability. So we have a lot of capabilities that are going to be delivered to
12:10our customer base in the next few quarters from customer intelligence to lead management. And then
12:16of course, Sygentic AI, we have multiple different use cases that we are making progress on there. And
12:22I think seeing it all come together and allowing our customers to just really run and take advantage
12:31of this powerful technology is what gets me up excited every day. Well, I'm excited to see what's
12:37next. Joe, thank you so much for joining me today. It seems like there's so much that's happened since
12:42February until our next conversation. Thank you, Alison. All right. Appreciate it.
Comments