AI Implementation Guide for Title Agencies | Title Agents Podcast Ep75

In a world where change is the only constant, Mo Shumil stands at the forefront, guiding title professionals to not just grow their businesses, but to master the art of innovation. With every episode, you're handed the keys to unlock unparalleled growth and stay ahead of the curve. Get ready for a transformative journey. From Bold, founder and president of Antecen Advisors, a firm that helps organizations go beyond AI pilots and into actual performance. Justin's known for cutting through the noise with a refreshingly practical and sometimes contrarian take on what it really takes to implement generative AI in industries like ours.

Whether you're just starting to explore AI or looking to level up your current tools, this is an episode you don't want to miss. Hello, Justin, and welcome to the show. We always start a show by kind of having the guests tell us their background, their story, kind of just get to know you. Yeah, no, and Mo, it's great to be here on the podcast. Thank you so much.

I'll keep it simple and I think relevant to hear your listeners. And so, as Mo said, I'm Justin Trumbull and I'm the president and founder of a company called Antecen Advisors. We do a lot of work with generative AI readiness at the organizational level. This is for larger enterprises, for smaller companies, and it ranges everywhere from you're a large company, you're making a lot of investments and you want to make sure that those investments create value, all the way to, hey, we have a sense of what generative AI is, but we don't really know to start, where to start, or people don't really necessarily know how to use the tools. And so, walking organizations through those basic processes.

Before I started doing that independent work, I spent about a decade working at some larger consulting firms doing, I would say, broadly strategic work and operating model work across a lot of different industries. Before that, I was an academic teacher and researcher in biological sciences for about 15 years. And so, a bit of a winding road to get to where I am today, but certainly excited to talk to your audience in an area that does sit kind of close to home. You and I were talking offline before the call. My wife works in commercial real estate and asset management and investments.

And so, we have a lot of conversations about those topics and the application of generative AI in the real estate ecosystem. And a couple of the clients that I'm working with are in that ecosystem as well. And so, there are a lot of really interesting opportunities and looking forward to discussing those and getting into whatever detail we want to. This podcast is brought to you by Struggling to set appointments and generate leads? What if you had a team working behind the scenes to help you book more meetings and close more deals?

At Alltech National Title, our dedicated inside sales team help you find your pipeline so you can focus on what you do best, building relationships and closing business. Why join us? One, built-in lead generation. Our inside sales team actively sets appointments for you. Two, more time selling.

And less time cold calling and more time closing. Three, proven support system, advanced tech, industry expertise, and marketing tools at your fingertips. And four, unlimited growth potential. No red tape. Just the resources to help you thrive.

Ready to take your career to the next level? Visit alltechnationaltitle.com or DM us today to start the conversation. Awesome. So the Title Agents podcast is really targeted mostly to like small agencies and independent agents. So make sure you understand the audience.

So we're not talking to like Fidelity or First American. They're massive companies that have the whole resource, like more kind of empowering and educating the small title agents. And second, use the word generative AI. So we have, our audience have people that are very savvy and understand AI. We may have beginners, people don't just hear the word and still don't understand what it means and maybe even scared of it.

So to kind of get a little kind of maybe 101, AI 101. We've had a few episodes on AI, but it kind of would love your perspective and use the word generative AI, but it's another level of it. But I kind of just kind of that 101 kind of artificial intelligence kind of education to our audience. Yeah, it's a great question. And I love what you said as well, because a portion of my business, we work a lot with smaller organizations.

I think similar to what you're referencing with smaller title agencies of, hey, we've heard generative AI. Perhaps we have a relative that's using it for fun to book vacations, or we have a relative that works at a larger organization that's talking about these crazy solutions they're trying to develop. And there's often a lot of noise. And so one of the things that we can talk about is, hey, you're on square one, which is what we just described. How do you get to square two, square three, square four in the context of having a little bit of structure around how you're so excited to dig into that?

Generative AI 101, without getting into some of the technical details of it, the way that I typically describe it when I'm talking with a group that might need a little bit of a primer is, think about generative AI as instead of a tool that can perhaps predict outcomes, which is more like traditional AI, it's a tool that can collect large amounts of information, notably text and areas like that, and can then aggregate that information into some useful input. So that could be like what people see with LLMs and chat GPT, you ask it a question, it gives you an answer. Really what that engine underneath it is doing is it's going and creating what we call a probabilistic model that says, based on your question, if you see this word, what's the most likely next word? And so it's a very powerful model, but there are, of course, hurdles with that. And one of the things that we like to talk about a lot is, yes, there are these really advanced solutions that you hear about.

And you even hear about more basic things like co-pilots that people have in Microsoft Office where it's perhaps giving them some guidance of how to do something or doing some formatting. But realistically, for a lot of smaller companies and for individuals, there's an initial step, which is, how can you ask the model or chat GPT or Grok or Gemini? How can you ask it better questions? How can you structure those well? And then how can you focus the model so when it's doing that combination of words and it's putting things together, how can you focus the model on relevant content that you trust?

So if you don't do that, there are a lot of interesting things that have come out of where does chat GPT pull their answers from? I would invite the users to perhaps use chat GPT and ask that or Google it. And it comes up with things like Reddit and go down the list. And if you look at that and think, well, I don't really trust Reddit for professional usage. So then the question becomes, well, how do you direct the model without having to make a lot of investments to then give you a nice, customized answer?

And so it's a tool in the context of work like with title agencies that you can use to take for activities like you're doing title search and examination and different like escrow and settlement and so forth. You know, what are those areas that are very, very repetitive, have very predictable structures and that you have content that you can feed the models and say, hey, here's what good looks like. Here's a good output. Here's what I the details of what I want to accomplish in this specific instance for this specific property. Can you make a first draft for me?

Right. And so you can find those opportunities and for title agencies, they're up and down the value chain of areas where there's a lot of opportunity to use out of the box tools like your chat GPTs to create value now, but then create a longer term view of, well, now that we're getting some useful tools here and there, how can we scale things up? How can we look for the next best solutions? That's awesome. And what I say as kind of more simplistic way for kind of people that are maybe afraid of AI, think of it as a smartest assistant you ever have, like one of the smartest executive assistants you have to help you whatever you need.

And also as a thought partner, I can apply kind of a simplistic foundational way to use AI. Then there's different levels, automation and as we talked, and we're going to talk about it in the title business. I do want to just, just one of these, that, that, that analogy you gave is useful for a simplistic way for kind of people that are maybe afraid of AI, like think of it as a smartest assistant you ever have, like one of the smartest executive assistants you have to help you whatever you need. And also as a thought partner, kind of a simplistic foundational way to use AI. Then there's different levels, automation, and as we talked, we're going to talk about the title business.

I do want to just, just one of these, that acknowledge that you gave is useful for two reasons. One, the reason why you said it, which is here's a nice way to think about it and understand the, from an output perspective, how should you, you know, it's this entity that behaves similar to a very productive assistant or a very productive, you know, researcher or, you know, whatever you, however you want to think about it. But what it also does is it, it puts you in the right mindset to say, consider this a first draft, consider this some advice, but like most things, if you get a draft from a junior person, it might be very good, but you should still take a look and give it, give it that, that once over and, you know, in, for, for smaller title agencies or any title agency for that matter, you know, there's a high cost to making mistakes, right? And so having that mindset to be careful about, about what you take as being true and not, not getting overly reliant on the assistant and treating it like a final version. So that's, I think there's a flip side of that analogy that you use to say it's an assistant.

That's a good way to think about it, but it's also a good way to remember you have to check it. Wow. Absolutely. So what myths or misconceptions do you think title professionals or small businesses have when it comes to adopting AI? Yeah, I, I think the, the, the biggest misconceptions that that come up is that you have to be on the cutting edge.

You have to have the most sophisticated tools and have figured everything out to be able to not just start exploring, but to create value from these solutions. And so one of the things that in our research, that's often the biggest bottleneck to organizations creating value from generative AI is they haven't clearly articulated at least a starting point of how generative AI is going to support their key activities. So this could be a broader vision to say, we want our title agency to be able to you do some activity overall, more effectively, more efficiently, but it could be a very precise part of the activities you do is, as we were alluding to before, you know, if you look up and down some of the main ideas or the main activities, if you think about being able to use AI to assist with title search and document review, or even subcategories within that. So identifying those specific areas where you can start testing. So to bring it back to your question, the misconception is you have to have it all figured out before you start.

And that's just not the case. You have to figure out a starting point and find some of those low hanging fruit that you can empower your people to start testing and say, okay, here's something we really understand how it works. Meaning the traditional process. We look at it and say, okay, we don't know exactly how generative AI would apply to this. There are too many unknowns.

why don't we start with a simple approach where we try to get really good at prompting an LLM, get really good at attaching materials and directing it to materials that are relevant to the analysis you're trying to do and start playing with it and start seeing what it puts out, seeing how you can iterate, improve and expand that analysis. And then you start integrating it into your workflows. And so it looks very similar to the way you would approach any sort of activity that you wanted to improve. So the idea of there's a really high barrier to entry to start using it. There is a barrier to entry and that your people have to be trained on how do you prompt it well?

How do you think about directing it to content that's relevant? And how do you iterate it? So there's a core base knowledge, but it's not that sophisticated of a base knowledge. And then I think empowering people that you just have to jump in and start using it. Once you know those core tenants, you just have to go in and start doing it.

And then you can get more and more structured. You can start seeing what works, what doesn't. You can't really break it. I always like that when people, you know, industries and it's for better or worse, it's an Asian industry. Like it's a, there's definitely a big, like the majority of the tele-Asians kind of the 50s, 60s, 70s.

So we definitely need to recruit a young generation, but to somebody in their 50s or 60s, that's technophobe. And I always ask them, like they use Facebook. They're like, yeah. I'm like, well, if you use Facebook, you can use chat GPT. It's a lot easier.

Like you can just talk to it. Like, you know what I mean? You can't break it. And to tag in on that analogy that you said, I think everyone on this podcast can think about a scenario where, as you said, thinking about it as an assistant, you know, what did you do when you had to start onboarding that junior person? Or perhaps you're a junior person and you had to get your, you hit the people above you on board with what you were doing and why, you know, you had some idea and you are exploring a new way to do something or whatever it is.

It's the same type of thought process. But what it is, is you have to get familiar with some of these core tenants of, like I said, prompting the system. How do you review it? How do you adapt it? But once you do that, you, you think about, okay, I have this problem.

How do I go in and now think about directing somebody in this case, the LLM to start learning how to do it. And at the same time, you start to learn how to do it. And once you develop that muscle to say, okay, I know how to develop a way to accomplish a task or solve a problem using an LLM, you know, like chat GPT, then you can start thinking more freely, like you would if you were onboarding an individual, you know, for some task or partnering with somebody to accomplish a task. Can you walk us through one of the most common failures you see in AI rollouts? What's a smarter way to approach it?

You know, that's a, it's a great question. The biggest challenge, I think we covered one, but I'll give a second one as well. The first one being a fear and uncertainty of where do we start, right? And so that, that idea of starting with the basics, starting with thinking about areas you could improve and testing those. So that fear frees people, you know, they kind of stop or they either think they're past a period where they can learn a new technology.

It's a junior person that maybe thinks I know how to use this, but the people I work with aren't going to embrace it. They won't let me use it. So it's overcoming that fear is a big part of it. The second one is a bit more tactical, which is, well, once you start experimenting and exploring the solutions, the biggest point of failure with. This podcast is made possible by our sponsor.

You have worked so hard to build your business. So why do 70 to 80% of businesses never sell? The truth is most owners wait too long. They don't plan properly or they can't find the right partner. At Alltech National Title, we help title company owners avoid these pitfalls, ensuring that you maximize value, protect your team and transition on your terms.

Whether you're looking to grow, merge or sell, we'll help you craft a strategy that actually works. A good deal is one where everyone wins. Let's talk about how to make that happen for you. Visit alltechnationaltitle.com or DM us to start the conversation. Alltech National Title, smart transitions built to last.

Organizations that have similar dynamics is then changing the way in which they work, where that generative AI process becomes not just part of their workflow, but there's a clear way to take those experiments, take those things that you're testing the solution on within your work and having a path to scale and invest in them when they're showing the type of promise that you expect. Oftentimes, organizations will start playing or people might be playing without other people realizing. They're using it for various things, but it's very informal. It doesn't scale up. That often just has to do with the group didn't take a step back and say, okay, here's how we're going to play and here's how we're going to experiment.

Here's how we're going to track if it's creating value. Now based on tracking that, that you expect. And so oftentimes organizations will start playing or people might be playing without other people realizing, you know, they're using it for various things, but it's very informal. It doesn't scale up. And that often just has to do with the group didn't take a step back and say, okay, here's how we're gonna play.

And here's how we're gonna experiment. And here's how we're gonna track if it's creating value. Now, based on tracking that and showing that it's, you know, has this amount of improved efficiency or it's more effective or we can, you know, we can do more work with less people. We can, let's say, get through some of the requirements involved in what title agencies do faster than thinking, well, we have one person using this. What does it look like to have the whole team using it?

You know, and how do we get everybody up to speed? And what does that solution have to look like? What people do is they start with what the solution could look like. And they don't start with figuring out where those opportunities are first and then thinking about the solution. Do you, do you, you're referring to opportunities is that I usually, when I think of like the, how to use AI, what pain you wanna solve, is that the same thing?

Like do you figure out what's the highest pain point in trying to solve that? Or do you just go for low hanging fruit first? So like, like the one we've kind of, we gradually kind of leveraging AI in my agency. Like we started off with just kind of automatic contracts entry in a system. We use a system called Qualia.

I'm not sure if you've heard of it. They just came up with a big product called Clear AI. We just kind of, it's very powerful and it's still in that beta to beta phase. So when you talk about opportunities, is the same thing as when I think is solving a pain point or instead of thinking of AI as a hammer then everything's a nail. Then you just, you're kind of, you spread out and you're not solving any problem.

Like, hey, it's a phenomenal tool, but it can do anything and everything with it. But how can you be more laser focused and more strategic versus kind of being all scattered and waste time? Yeah, I love what you said. So it's, it is a, being able to adopt a first principles approach to it. Like you do to any problem, which is, what's something that can, what's some gain that can occur?

So more of like the plus side of what you said and where can there be a pain reliever, right? So what are those? And again, not thinking about the tool, but thinking what are the pains? What are the opportunities? And then exploring those.

Okay, now that we've identified those things, which of them are candidates for a generative AI solution? And that's where you can be much more focused to say, it could be a very precise activity that you're exploring. It could be a broader, hey, here's our strategic vision for the group or for whatever we're doing. Where are some areas that, what needs to be true for that vision to come to life? Okay, and now are there some opportunities that we can insert generative AI into that?

What needs to be true, right? So what are the key characteristics? Not the other way around, not tool first and then solution. So it, there's tremendous value in thinking about what you wanna get done. Whether it's an objective or it's a resolution of a pain point.

Think about that first, identify those, then layer another level, which is, okay, which of these problems or which of these opportunities are well suited for generative AI as a solution to that? Awesome. So for a small business title agency starting to experiment with AI, what does a realistic 90-day roadmap look like? To me, the 90-day roadmap, and this gets back to the way that we work through it with organizations. It starts, let's say the first 30 days are, is the training stage.

So this is starting with getting everybody in the room so they understand what's the difference between, hey, I have this document, can you review, it's a title search and document, let's say title search and document review broadly. Can you review it for me? Let me know if there's anything that looks right or update it based on this property, based on this asset we're looking at. How that's different than doing this, asking the same question, but doing it in a very clear structured way, which with lever one, which is the prompts that you put together. So getting up speed on what a good prompt looks like, how you develop it, how do you refine it?

And lever two, which is how do you make the model better by providing it with the right content? And so I think the example that I'll, with some of the real estate companies that I've worked with, if they're trying to, let's say, do contract reviews and they wanna do, let's say like a red flag analysis that goes and says, hey, identify the areas where we should be concerned given our objectives, right? You could ask it that question and you could upload it, but it becomes orders of magnitude more powerful if you have a reference document that identifies and defines what those red flags are and how to identify them. So it's not pulling, it's not creating a red flag checklist from Reddit, let's say, it's creating a red flag checklist that's very specific to what you're trying to do. The same way you would do if you were doing it yourself, right?

You have, whether it's on paper or it's just in your head, you know what you're looking for. So it's giving it that guidance. And so you start with that training. So that's the first 30 days. That training then morphs into an exploration for your people to start identifying those use cases in their work.

Like we just talked about, where are those areas of pain? Where are those points where your processes are well-documented, but they're very manual, there's a lot of text involved, things that just take a lot of time. So finding those points and then giving your people space to start experimenting with those two levers we talked about, the prompting and the external document direction, and have them start experimenting with the tool. And then they start learning more and more. The next 30, let's say 30 to 60 days, that's setting up that structure that we talked about, which is, okay, we wanna be able to track what's being experimented.

And we wanna be able to, in as low of a cumbersome way possible, track how those experiments are going, right? Cause you can get paralyzed by tracking metrics, of course, like in anything. How do we track that? How do we have an open line of communication between the person that's experimenting it and the individuals that could make decisions on if something's worth investing in a solution, like you described earlier, some of the tools you're exploring with. And so you set up that structure for the communication.

And one of the key areas of working at that stage is empowering your people to, what we've seen work really well with large and small organizations are groups that give their team space to say, okay, you have 10 different things you do right now. Can you spend 20% of your time trying to do those more effectively with good use of an LLM? And that works. So developing that working cadence to do it. And then the next 60 to 90 days, I'll use the example you said, but the solution you invested in, it's making those decisions.

Okay, is this something that it's worth doing? We're seeing value creation. Should we invest more money? And the decision becomes, is that investment partnering with an organization that can build you a bespoke, let's say structure, they can do it in a more automated and better way, or is it looking for out of the box solutions like you described? So you look at a solution like that, after you've done that experimentation, all of a sudden you notice there are, you know, 20, 30 different features that align to how people are experimenting.

And then you can start plugging in those individuals into that solution. And they really understand what they're doing, why they're doing it, how to interpret the results because they've been doing that experimentation from the grassroots, right? They really get why it's happening. What we see a lot of is when you start with the tool, you'll have one or two super users perhaps, but most people won't really know how to engage with it because they don't get why it's there and what it's doing and how to use it. So I think those are three good 30 day chunks to think about.

When it comes to automation, what's your take on a virtual assistant, chatbots or document summarization tools? Useful stepping stones or potential distractions? lot of is when you start with the tool, you'll have one or two super users perhaps, but most people won't really know how to engage with it because they just don't get why it's there and what it's doing and how to use it. So I think those are three good 30-day chunks to think about. When it comes to automation, what's your take on virtual assistants, chatbots, or document summarization tools?

Useful stepping stones or potential distractions if not managed well or used properly? Well, I'll broadly say that using generative AI, it's very easy. So I'll go back to when you start training, it's very easy to adopt that hammer versus nail thing. And you start becoming very reliant on the tool and you stop thinking. So the idea of, to answer your questions over something like document review, it can be an enormous time suck, but also introduce a lot of chance to make mistakes if you start using it too indiscriminately.

So let's say you're just throwing everything into it and you're assuming it's right. So that's that challenge that we talked about before. Whereas if you use it a little bit more carefully, you're going to spend a little bit of time on the front end. But then once you start then doing document review in a very careful way, you're confident that you're doing a good analysis and you're confident with what the output does and doesn't provide. Then you're getting into efficiency and you're not wasting your time doing it.

But if you are just throwing things at the wall, even if it's something that it gives you what appears to be a good answer, you can start spending a lot of time analyzing documents and getting outputs that aren't necessarily useful. And that can start to take away time. Chatbots and maybe just to clarify, would these be, are you thinking about more internally faced chatbots or chatbots with your partners, with your customers? I'm thinking about a customer facing. Yeah, chatbots are, my view on chatbots is even though there are a lot of commercially available tools with it, they're, this is going to sound obnoxious, they're helpful insofar as they're helpful, right?

So if the chatbot is actually adding value to whoever you're working with, that's great. But if you think about it, if the chatbot isn't well-tuned and it doesn't, it provides two generic answers and you're always directing people to the chatbot, that has a tremendously, could have a tremendous negative impact on your relationships with those people. So you just, like anything else, if you buy an out of the box chatbot for your title agency, the more precise you can get to say, hey, this is a chatbot for title agencies that we've developed. Okay. Like, hey vendor, give me some specific case studies.

Give me some precise testimonials from companies that have had used it. Give me some quality, give me the metrics to show that it works. And then integrating it in such a way that you aren't by default directing people to it until you have confidence that it's actually helping, you know, and then you can start saying, now let's use it more. But if you, if you don't do those, those basics on the front end, like you would rolling out anything that you do, it does end up being, it does end up causing more harm than good. Whether it's extra time, it's extra pain for your team and people you work with.

And then the impact on relationships. I think chatbot, I'm more thinking more general information on a website, like kind of answers simple questions people may have. Yeah. Like we were using a tool before even chatTPT came about. This kid that I knew, we built a, literally, we scraped the website, our website, all the information, fed it some more information, like literally before LLM even came about, it was like in late 21.

And it's like, now it's because the norm, it's literally just chatbot is more about kind of providing basic information. People can have quick answers. This podcast is proudly sponsored by Need extra hands without extra overhead? Meet Safi Virtual, your on-demand team of virtual assistants trained specifically for the title industry. Our VAs come with foundational knowledge of title insurance, ready to handle admin tasks, data communication, data entry, and more.

So you can focus on closing deals, not paperwork. Scale smarter, work faster, and cut costs without sacrificing quality. Visit safivirtual.com or DM us to find your perfect VA. Safi Virtual, expert support tailored for title professionals. Versus having to look for it that's beyond that and you're correct.

Yeah. Those are usually great. And those fall into the category of a generative AI use case. And so by that, I just mean a way in which you can apply a generative AI tool that's relatively well-established. There are a lot of good out of the box solutions that can query your website and can provide good information and so forth.

So those are lower hanging fruit. I think the question for an agency is, one, what's the impact of people not being able to easily find that information with a self-service tool, right? So is it adding value for them to be able to use a chat bot and go onto your website? And you weigh that with the costs. For an industry like real estate and with title agencies, there are obviously various tiers of what that chat bot can do.

It can be things like, hey, what does our agency do? What are we known for? What are some client testimonials? What are some success stories? What's our value proposition?

Things like that. But then there's a little bit more tactical, which is, hey, we're partners. I need certain documents. I need to know who I should communicate with and how. Then those get into a little bit more where you want to make sure that the models are functioning well in terms of giving people the right information that's actually helpful.

Absolutely. So how can title agency leaders create a culture that's both tech forward and still grounded in client service and compliance? I think we said it well earlier. It's starting with a first principles look at your business. And that falls in those two categories.

It's a strategic category and it's an operational or tactical category. And if you stay grounded on what are those things that you really need to improve, what are those things that you can improve, and then creating a culture from, and small businesses are always organized in unique ways, but I'll just say from the top down, because it's the easiest way to think about it. That might be flatter for some of your listeners or others. But it's creating a culture of collaboration. It's creating a culture of, and that's collaboration between people.

It's creating a culture of openness, which is, hey, here are these new tools, or hey, here's this information I generated from an LLM tool. Are you actually open to seeing that type of feedback and that information and making decisions on it and not defaulting to whoever the oldest person is in the room, making the decision regardless of the information that's put out there? I think broadly, it's what I would call in some sense, a Silicon Valley mindset. So it's a dynamic mindset. It's one that rewards curiosity for people to go in and experiment and look for creative solutions, gives people space for curiosity.

And what that comes down to is back to what we talked about before of, well, how do you set up these processes to not just test and take things through, but to actually have the systems and the operational model where as things are taken through to being developed and being used, that there's a path to be able to show that they work, to be able to scale them, to be able to use the data correctly. I think with the example you said with your tool, how do you create a culture that engages people with that broader, probably very effective commercial generative AI tool? What is it that's keeping your people from using it? And setting up a structure that enables them to not just learn the tool, but learn how to effectively implement it in what you're doing and make it to where decisions can be made or content can be produced from that tool. you said with your tool, how do you create a culture that engages people with that broader, probably very effective commercial generative AI tool?

What is it that's keeping your people from using it and setting up a structure that enables them to not just learn the tool, but learn how to effectively implement it in what you're doing and make it to where decisions can be made or content can be produced from that tool? From a confidentiality and security standpoint, that's where things get, you know, one, first, most commercially available products have security. But if you're using an LLM, make sure it's a private instance, make sure it's an enterprise instance that doesn't train models. And for documents that you're very uncertain about, creating a culture of redacting documents before you integrate them into the model. And there are a lot of commercially available tools to do things like that.

Well, what trends are you keeping an eye on that title agents should start prepping for now? And I know this world is moving very, very fast. We're not talking years, talking about quarters or months. That's how fast AI is moving. Yeah, I think it's, this goes back to that question of, if your listeners think of a time and maybe it happens all the time or not, but when you go into a new area and there are so many unknowns, right?

So many unknowns you can't wrap your head around. Well, should I be keeping up with agentic AI? Should I be thinking about, let's say, artificial generalized intelligence? Should I be thinking about the way in which models are progressing? Which one have the best metrics and so forth?

There are a lot of things you can keep up with. What are big agencies doing? There's a lot of noise is the upshot. And so the trends that people should keep up with, in my view, you can't really think about that until you get a sense of how you can and can't use generative AI. So the point is, is take a step back, start experimenting.

Now, all of a sudden you've locked some of those unknowns. And you, let's say you're one of your main tasks is doing document review. I know we've come to that a lot. It's a simple example. Come up with a way, whether it's checking in a news article from time to time, it's listening to podcasts, it's going and having a subscription to something that gives you newsletters and keep up with what's new in document review and how that's advancing.

How about savvy AI title agents that are ahead of the curve? What's your take? What do you see coming? What should we look for? If you don't, it's okay.

In terms of saying, I just want to make sure, so is it, how does someone that's savvy? Any trends, anything that would you focus on or kind of what's the Asian AI or is it... Yeah. So where are the opportunities as it's emerging? For savvy title agents that are already adopting AI and using it.

Yeah. Well, just as we've talked about before, I would say push the envelope on how you can apply it. In your own work, get more structured, get more precise on how you're using it. Make sure you're using it well. And those are those two levers, the prompting and the external documents.

But a little bit more specifically, there's some opportunities that I've seen and some that I've heard about as well in title agencies, thinking about end-to-end intelligent title file automation. And so just what's emerging out there and they're starting to come out with solutions like this in real estate for title agencies as well as for similar industries, meaning they have a lot of similar processes of what the work looks like if you drill down to the core. So some end-to-end automation tools that help with that. It could be things like instant title commitments for low-risk properties. So where are those ways that you can look for solutions that can, in some ways, arbitrage the easier things and leaving the complex aspects that the models may or may not be able to handle?

Not on commercial transactions, more residential or refinances. Yeah. Or there's too much risk to get it wrong, let's say. And there's always risk. You never want to mess up down to the lowest valued property.

But yeah, I like what you said. Things like digital closing with AI notary and document validation. So streamlining that last step. I think we've all, whether we've been as a buyer or in some similar situation of what that looks like when a deal closes, whether that's an M&A deal or an acquisition or for a piece of property, there's so much time that goes into that. And there's a very real opportunity and solutions that are coming out to streamline that closing process.

Looking from a broader business standpoint, this is a bit more on traditional AI. So things like predictive analytics for pipeline forecasting. I'll let your listeners think pipeline forecasting means a lot of things. But everywhere from how do you plan your own business? How do you do resource allocation?

There's a lot of upside from being able to do that more effectively. And then the last one is one of the challenges that I see in a lot of industries, but in real estate and title agencies is there's so much information and so many different sources and so many different formats. And so one of the challenges that's trying to be addressed with generative AI is how do you take those fragmented data sources or multi-modal data, as you might hear people call it, and how do you unite the clans, so to speak? And so there are tools that are trying to streamline that to get, yeah. And it sounds like, if you think about it this way, it's very hard to do without a generative AI tool.

It's equally hard for the generative AI tool to do it as well. So that same challenge exists for the generative AI. I think the challenge everybody wants control. That's the hardest part. It's like everybody wants that holy grail piece and control the whole thing.

And that's why there's so much fight and so much conflict. It would be very hard to pull off for everybody to adopt that kind of consolidation, if that makes any sense. Well, no, it does. And it reminds me of something that I think is really useful for your listeners to keep in mind, is that I've also done a lot of work with AI solutions in healthcare. And there's always this question of, okay, you're a physician, you're some other healthcare provider, what type of automation are you comfortable with and in what context?

And it's usually an answer like it's automation up to the point where there's a mistake that could be made in a clinical decision, meaning they're okay with automation of administrative tasks, scheduling, certain things like that. You get to the point of making decisions about patient care, or let's say if someone should come in or not, then they want a little bit more at the very least visibility into what's going on. Whether it's like, hey, we're going to prescribe this medication, you've got to check this box first. Don't just submit it. We're going to tell a patient to do X, Y, and Z, sign off on it.

For title agencies, we talked about that triage thing. Maybe there's some things that go through with just some simple validation checks, but for more complex areas, it isn't just about being confident that it's working. It's also about ensuring accuracy. Accuracy, correct. Yeah.

Well, coming to a close of the show, it's been an awesome conversation. I always ask, last two questions is, do you have a favorite quote? Yeah, I do. And it's kind of similar, maybe not, this is perhaps a poor analogy, but which of my kids is the favorite? I don't put quotes on par with my kids, but it's often hard to think of one.

But one that I like a lot, and I think is somewhat relevant to generative AI in every instance is, it's attributed to Mark Twain. And I've seen so many different versions, but I'll give my paraphrase version of it. But it's not what people don't know that's the problem, but what people know with certainty that just ain't so. And so it's, yeah, you can think about a lot of fun ways to apply a quote like that. I don't put quotes on par with my kids, but it's often hard to think of one.

But one that I like a lot, and I think is somewhat relevant to generative AI in every instance, is it's attributed to Mark Twain. And I've seen so many different versions, but I'll give my paraphrase version of it. But it's not what people don't know that's the problem, but what people know with certainty that just ain't so. And so it's, yeah, you can think about a lot of fun ways to apply a quote like that. In generative AI, it's a good mindset of being confidently wrong.

You want to be careful. It's like these models are often confidently wrong in their recommendation. How about a favorite book or one you read recently or listened to recently? Yeah, my favorite book, and I'd say it's a classic. I don't actually know when it was, it's maybe 20 or 30 years old.

It's called Good Strategy, Bad Strategy. Okay. And I have not heard that book. Nice to add to my repertoire. Yeah.

It's more of a business strategy book, but I think it applies to any business and not just large organizations. It has this idea of, well, what's the kernel or the foundation of a strategy? What does that look like? Now you blow up from that. What is the roadmap and the milestones to achieve that strategy?

And then how do you know when to adjust it? And so I read it along, I may be misrepresenting the rings exactly, but it's a really simple first principles way of thinking about your business. And I think it applies to deploying AI and generative AI for title agencies. Let's start with the core. Like you said, what's the pain?

Okay. Now, what does it look like to test to see if generative AI is a tool to solve that pain? And then what does that look like to implement it in your business? It's awesome. Well, Justin, thank you so much.

And I truly appreciate your input. And it's been a great conversation. Thank you. Yeah. Thanks so much, Mo.

I appreciate it. Huge thanks to Justin Trumbull for joining us today and sharing the roadmap for real-world AI success in the title industry. If we took away one thing, let it be this. It's not about chasing the latest AI tools. It's about aligning them with your business vision and client experience.

If you enjoyed this episode, please give us a five-star review and make sure to subscribe to the show. Until next time. And that's a wrap on today's journey with Mo Shamil. From the Title Agents Podcast, reminding you that mastering the art of innovation is key in the title industry's fast-paced world. If you're finding it tough to keep up with the changes and challenges, remember, you're not alone.

Our calendar is open for you. Find the link in the show notes and let's connect. Make sure to hit subscribe to not miss out on strategies that elevate and insights that empower. Together, we'll navigate the future of the industry. I look forward to our next meeting in the upcoming episode.