Episode Transcript
[00:00:00] Speaker A: Foreign.
[00:00:09] Speaker B: Well, hello, listeners, and welcome to the Bottom Line Me podcast, where we like to spend some time exploring ideas, challenges, and innovations in the title and real estate industry. I'm your host, Ann Allard. Our topic today is artificial intelligence.
I think we all know that AI is reshaping industries across the board.
So today we're going to take a look at it and how it works and how it's profoundly affecting title and escrow to help us understand this new technology. I'm very excited to have a conversation today with Nate Baker, co founder and CEO of Qualia. Nate, welcome to the Bottom Line Me podcast. Thank you so much for taking the time to join me today.
[00:00:55] Speaker A: Thank you for having me on, Anne. Appreciate it.
[00:00:57] Speaker B: Well, I'm looking forward to this conversation, but before we dive in, tell the audience a little bit about yourself, about Qualia, and how Qualia has been working with Old Republic Title to help our customers navigate this new technology.
[00:01:11] Speaker A: So Qualia started about 10 years ago.
We have actually been working with Old Republic since maybe around that time, maybe nine years ago.
So we've had a long partnership. Initially, it was just an integration partnership and working with your agents to help issue policies and calculate rates and all that.
[00:01:34] Speaker C: Thank you.
[00:01:34] Speaker A: But increasingly recently, we've expanded the partnership a lot.
So earlier this year, jointly, we announced the acquisition of RamQuest and eclosing by Qualia, which were Old Republic technology platforms, and that Old Republic would be transitioning to Qualia, our software. And a big part of the stated reason for that is the excitement around AI and the changes that are coming over the next few years. And, yeah, we're working on it a lot together right now.
[00:02:07] Speaker B: Excellent. Great. I was having a conversation earlier today with someone about just telling them that I was going to be talking with you today. And I said, you know, I hate to age myself, but I remember back in the days when we were rolling these things through a typewriter and you had somebody really actually proofreading what you. What your work was. And we've certainly come a long way since then.
[00:02:32] Speaker A: So things have definitely changed a lot over the past few years.
I think one of the things which is interesting to me about the title industry, which is surprising to a lot of outsiders, is that Tidal is actually an early adopter of technology in a lot of different ways. I, over the past few years, have interviewed the founders of a lot of the software companies, the prior generation of software companies in our space.
[00:02:57] Speaker C: Thank you.
[00:02:58] Speaker A: They were the people bringing computers to Tidal for the first time or bringing shared systems, shared back office or Web capabilities.
And so it's interesting. I think people have this sense that title might be slow to adopt, but I think the reality is actually quite different. And we're seeing that with AI as well.
[00:03:20] Speaker B: I love it. We're on the cutting edge. I love it.
Yeah, yeah. You know, one of the things I thought about that I mentioned at the beginning, is that it is on the mind of a lot of people today, but maybe you could just kind of talk a little bit about, you know, layman's at a higher level, what the layman's terms are and what you. Where you think it stands right now with respect to our industry. And you know what could professionals outside the industry or outside the tech part of this industry really benefit from learning from this part of this conversation.
[00:04:01] Speaker C: Thank you.
[00:04:01] Speaker A: So two years ago, as I think most of your listeners would know, ChatGPT came out and that was really the first time that the general public was using large language models. LLMs, which we'll talk about more.
And over the last two years, they've just gotten way, way, way better.
And even as recently as two days ago, Claude, which is one of the three main LLM providers, released a much better product and it's so much more capable. So the first thing I'd recommend to everybody, if you haven't done it yet, is thank you for sign up for Chat GPT, sign up for Claude, sign up for Gemini, something like that, and start asking just the most complicated title questions to it that you can think of.
I was sitting with a client a couple days ago, and he hadn't really used the tools much, and he asked it a really tricky railroad easement question and just was stunning how good the LLM was. It actually went through and read a bunch of case law, it read all of the state statutes, and it synthesized this information together and answered it really elegantly. So the number one thing which has been going on over the past two years is the LLMs are getting way, way smarter, out of the box. And the question that we all have to be asking ourselves as people who work in business is, how do you actually take these tools and use them productively in your day?
[00:05:46] Speaker B: Exactly.
We'll talk a little bit about that because I think that is the question for everybody. This is all interesting. It's sometimes it can be fun, it can be certainly educational, but how do we. How does it apply to me in this industry?
[00:06:06] Speaker A: The first thing which is most difficult about these tools is that they are fundamentally unreliable sometimes.
[00:06:16] Speaker C: Thank you for.
[00:06:17] Speaker A: And so the entire way you have to approach them is with an assumption that sometimes they're going to be wrong.
That's similar to an employee that you might work with, some junior employee that you're training in something, but they have the ability, the LLMs, to be catastrophically wrong sometimes.
[00:06:36] Speaker C: Thank you.
[00:06:38] Speaker A: But that is not to say that they can't be incredibly valuable.
And so the entire question is, how do you build tools, software and business processes around the fact that it might not always be perfect?
[00:06:53] Speaker C: Thank you for.
[00:06:54] Speaker A: And if you can do that, that's where a tremendous amount of value can be built.
Very concretely.
Today we launched a product called Qualioclear that actually can.
You can say, hey, please do a QC check before I go to closing on my document package. And all of the data in the file. And it actually goes through and reads everything. It looks at the documents, it compares the names, it compares the dates, it reads the email threads, the messaging threads, synthesizes it all together and highlights to you any potential errors.
There was a customer that was using this a couple days ago and they noticed they had actually overpaid the real estate agent by a certain amount and they had shorted the HoA by a certain amount. The same amount and stuff like that, where quality checks is actually a pretty compelling way to use these tools.
Another really interesting one is doing a title examination.
I expect that in 12 months you'll be able to upload a title search to chatgpt and it will produce a commitment.
[00:08:14] Speaker C: Thank you for having.
[00:08:16] Speaker A: With no additional training and already it's getting close and we have a product that helps substantially with this.
And so I think the question that we all need to be asking ourselves is how do you harness this and how do you learn how to use it really well? Because it's important.
[00:08:37] Speaker C: Thank you.
[00:08:37] Speaker A: Individually to people to know how to use these tools.
[00:08:42] Speaker C: Thank you for.
[00:08:43] Speaker A: It's something that we think about every day in our own business because I think your expectation as a customer is going to go up 100x in terms of what we can need to deliver to continue earning the business.
So I think that's how everyone needs to shift their mentality.
[00:09:02] Speaker B: You know, it's so interesting because when I was sharing some of this information with a colleague earlier, you're talking about how this now dots the I's and crosses the t's.
And I can remember days when you had to physically sit in front of the loan officer. It was like being in front of the principal while they did all of what you're saying now that you know a Software can do it can make sure that the names are correct, that the amounts are correct, that everything's properly executed and was recorded.
That's just fascinating. Fascinating.
[00:09:36] Speaker A: Yeah, it's going to be, it's going to be very interesting. And I think there are big changes coming in how we all do our work.
[00:09:44] Speaker C: Thank you.
[00:09:45] Speaker A: You know, we're a software company and it's a lot easier today to build software than ever before because of these AI tools. And so we have to be thinking about that and incorporating that into how we build stuff as well.
[00:09:59] Speaker B: Well, you know, it leads me to.
It's all fascinating and I think certainly it makes things far more efficient. But how does it actually affect, you know, the escrow officer, the business owner, even the sales team in an agency.
[00:10:16] Speaker C: Thank you for having.
[00:10:17] Speaker A: One of the most compelling things we're seeing on the sales side is that.
[00:10:21] Speaker C: Thank you.
[00:10:22] Speaker A: So part of this product, we actually are able to say the AI system can watch all the transactions you're doing, but also all the transactions that are happening out in the market.
And so it can say to you, hey, there's a real estate agent who just listed a property yesterday who has been sending you a lot of business, but they fell off the last three months. Your wallet share with that real estate agent has declined. Here's a recommended email outreach to them highlighting some of their feedback that clients of theirs have previously given you on your closings. Would you like to send this?
And you just hit send.
So the level of capability and kind of sales outreach is going to be completely different here.
[00:11:12] Speaker B: Interesting.
[00:11:13] Speaker A: Yeah. So I think as a sales rep, as someone who's kind of going out and winning business, it needs to be a rethinking of how on top of it you are with your clients and other ways you can add value to them.
[00:11:29] Speaker C: Thank you.
[00:11:30] Speaker A: Because the bar is going to go up well.
[00:11:33] Speaker B: And I'm assuming that if it suggests sending out an email like that, then it probably also will prompt you to do a follow up or take the, you know, in the sales process, take the next step.
Am I correct in that?
How the heck does it know all that?
[00:11:52] Speaker A: It's crazy how good these tools are getting. And the thing I would say is that.
[00:11:57] Speaker C: Thank you.
[00:11:58] Speaker A: This is new for us too.
It's only very recently that this is possible.
What was possible 12 months ago is trivial today. And I expect that a lot of what we are offering in our AI products today is considered trivial in six months.
That's the rate at which these things are improving.
[00:12:24] Speaker C: Thank you for. Thank you for having.
[00:12:25] Speaker B: It's amazing. Well, tell us a little bit more. I'm sure our audience would be interested in hearing a little bit about what Qualia Clear actually is. I went onto your website and took a look, but you can probably explain it far better than I ever could.
[00:12:40] Speaker A: So Qualia Clear is a.
It's a agentic system that basically wants to help you close real estate transactions.
[00:12:50] Speaker C: Thank you for.
[00:12:50] Speaker A: The really unique thing about Qualia is because we are the system that a title company uses to manage the transaction, and because we're deeply integrated with email, we have good access to the state space of basically everything that's happening in the real estate transaction.
[00:13:09] Speaker C: Thank you.
[00:13:09] Speaker A: All the documents, all the communications, all the numbers live in Qualia.
[00:13:14] Speaker C: Thank you for having me.
[00:13:15] Speaker A: Additionally, we have deep knowledge about how to do closings.
We're a large provider of title software all over the country. And so we have a lot of experience with that, a lot of training. All the regional nuance, Southern California is like this, North Carolina's like this.
[00:13:33] Speaker B: Exactly.
[00:13:35] Speaker A: And then we're able to convert that using these AI systems into a recommended action.
[00:13:42] Speaker C: Thank you.
[00:13:43] Speaker A: So that's what's so cool and that's what Qualit Clear is. So a very specific example is.
[00:13:50] Speaker C: Thank you for having me.
[00:13:51] Speaker A: Let's say you open up your computer in the morning, you're about to start work and you're an escrow officer and you get a document in, maybe it's something back from the hoa.
Qualia Clear is actually able to say to you, hey, based on this document that you got back, you should save it to this file.
[00:14:14] Speaker C: Thank you for having me.
[00:14:15] Speaker A: You should extract this data from the document, you should send this reply to the person who sent you that document and it presents it all and says, would you like to do this?
And you just hit yes, thank you for. And so that's where this is all going. And the next step of that is saying, okay, instead of it just being a button, you should be able to get comfortable enough that for this type of document, always just save it to the file.
[00:14:47] Speaker C: Thank you for that.
[00:14:48] Speaker A: Or for this type of document, always just extract the information output in the file and move on.
[00:14:54] Speaker C: Thank you.
[00:14:55] Speaker A: And so I think it's going to become more about how do you work with these tools, because like I said at the beginning, they're not perfect, they're fundamentally flawed in some ways. And so it needs to be about how do you build the correct controls and QC process around these tools that are so powerful.
[00:15:15] Speaker C: Thank you for having Me.
[00:15:16] Speaker B: Wow.
It's fascinating, but it also sounds very exciting, you know, so let's, let's get a little tactical here for the moment. What are. Give us, give me some of the do's and don'ts when it comes to adopting AI tools.
[00:15:31] Speaker C: Thank you for.
[00:15:32] Speaker A: So the number one thing I would recommend to everyone listening to the podcast is if you do not personally have a ChatGPT account, Claude account, or Google Gemini account, thank you for. It is worth spending a little bit of money and getting a subscription because this is probably the single most important invention that has ever happened. You know, it's more important than electricity.
It's going to be a bigger deal than the industrial revolution. This is like alien intelligence coming down from the sky that you can get a subscription to for, you know, $10 or $20 a month. It's crazy. And so do that and then try to just work on how do you get that into your day to day. Whether it's helping you write notes, helping you write marketing content, helping you write emails, helping you review contracts.
[00:16:28] Speaker C: Thank you.
[00:16:28] Speaker A: For once you're starting to understand it better, you'll get better at prompting, get better at using the tools, and once you're there, that's when you can start going to other systems that are maybe more tied into your business. But I think that'll give you a base level of understanding of what's going on.
[00:16:48] Speaker B: Yeah, you got to dip your toe into the pool is what you're saying. Exactly, exactly. Yeah.
[00:16:52] Speaker A: Because the rate of change is so high right now, it's.
[00:16:57] Speaker C: Thank you for asking. Thank you.
[00:16:59] Speaker A: Just the announcements that have come out over the last month, ChatGPT5 and Anthropics Sonnet 4.5 are groundbreaking versus where things were month ago.
And so that's the, the benefit of starting to use these tools more is you'll start to understand that actually, if you ask ChatGPT something today, it's on par with some of the best underwriter counsel in the country right now in terms of ability to grasp nuance.
[00:17:33] Speaker C: Thank you.
[00:17:33] Speaker A: And so it's just an incredible thought, partner.
[00:17:37] Speaker C: Wow.
[00:17:38] Speaker B: Well, and I think as you said earlier too, one of the things to always keep top of mind is that think of it just as you would think of a colleague or an employee, that we're all human and we make mistakes. And even though this isn't human, it can make mistakes. So, you know, you know, I think that's really, really important for people to, you know, myself included, to remember that, that you gotta, you gotta double check some of this stuff or all of this?
[00:18:01] Speaker A: Totally all of it. And one of the things which is most interesting is there's. You've probably heard about hallucination in AI, which is where the AI will just make something up.
And it often does that if it just doesn't have the information it needs to make a decision.
If you look, one of the big changes that happened last month was the predominant model from ChatGPT was called O3 about a month and a half ago and it had a 20% hallucination rate, meaning it made something up in its answer 20% of the time, which is a lot.
And GPT5, which is the new version, makes something up 5% of the time.
[00:18:45] Speaker C: Thank you for. Thank you for.
[00:18:46] Speaker A: That is a huge improvement in reliability and it's easy to paint the picture of where those charts are going and the implications as it makes fewer errors and it's more reliable are profound.
[00:19:03] Speaker C: Thank you for.
Thank you for.
[00:19:05] Speaker B: Incredible.
Wow. So you gave us a couple of tips on how to get started. What are some of the common pitfalls or things that you need to really pay attention to so that you don't get caught in a hallucination?
[00:19:24] Speaker A: I think that the number one thing is to be skeptical of the output and treat it like you would treat someone that you work with who just started, who doesn't really know title.
They're one of the smartest people you've ever met, but they don't know anything about the end market.
Just double check what's going on. But increasingly these tools will give you links to the statute or it'll give you a strong reasoning on its analysis and they're getting more and more reliable.
[00:19:54] Speaker C: Thank you.
[00:19:55] Speaker A: So I think it's at the level where you really can include this as a core part of your workflow.
[00:20:02] Speaker C: Thank you for that.
[00:20:04] Speaker B: Incredible. Anything else you'd like to share before we get to the bottom line?
[00:20:08] Speaker C: Thank you.
[00:20:09] Speaker A: Yes. I actually have.
One other thing is this Old Republic made a big bet in partnering with us to power the core technology of Old Republic earlier this year.
[00:20:24] Speaker C: And thank you for having me.
[00:20:26] Speaker A: I'm very excited about it and I think Old Republic is in a position to really take advantage of the absolute cutting edge technology here.
[00:20:35] Speaker C: Thank you.
[00:20:35] Speaker A: You are the only title insurer at this type of scale who's going to have tools like this.
[00:20:42] Speaker C: Thank you.
[00:20:42] Speaker A: And it will be an incredible competitive advantage for you.
[00:20:47] Speaker C: Thank you for having.
[00:20:47] Speaker B: That's great. I mean that almost could be the bottom line, but I've got to have you. Bottom line. Us on this.
If our listeners remember one thing from our conversation today, what do you think it should be?
[00:21:01] Speaker C: Thank you for. Thank you for having me. Thank you.
[00:21:03] Speaker A: It should be that this technology, meaning LLMs, is more important than electricity, and it is changing at an incredible rate, and it will have profound implications not in 10 years, but in 10 months.
[00:21:21] Speaker C: Thank you for. Thank you for having me. Thank you for having.
[00:21:24] Speaker B: Well, that's quite a bottom line. I know that it's more important than electricity because.
Wow.
[00:21:29] Speaker A: Electricity is pretty important.
[00:21:30] Speaker B: Well, it is. Yeah.
It's running neck and neck with electricity. Huh?
[00:21:34] Speaker C: Okay.
[00:21:35] Speaker A: Yes. Yes.
[00:21:36] Speaker B: All right. Well, Nate, thank you so much for this really enlightening conversation. I learned a lot, and I'm sure those that are listening also appreciate all of your suggestions, advice, and just giving us some idea on where this all started and where it's going. So, again, thank you so much. We'll have to have you back to keep us up to speed on what's happening.
[00:21:59] Speaker A: Absolutely. Thank you for your time, Anne. I really appreciate it.
[00:22:02] Speaker B: Great. And to our audience, as always, thank you so much for taking the time to join us. And until we meet again, let's all keep learning, growing, and prospering. Thank you.
