AI in Commercial Real Estate: A $100B Opportunity with Locate.ai

Andy Sack (Computer) (00:01.969) This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI innovation and business. I'm Andy Sack and alongside my co-host, Adam Brotman. Each episode, we bring you candid conversations with the business leaders transforming their businesses with AI. No fluff, just real talk and actual use cases and insights for you. Welcome, Joe Lee from Locate AI. Glad to have you on. Joe Lee (00:32.28) Glad to be here. Thanks, Andy and Adam. Andy Sack (Computer) (00:35.723) Let's get, what I want to do is cover three things. I want to cover your business, you know, as it has existed for the last seven years, I think, and then talk about your own transition to what you're doing now, really becoming an agentic verticalized commercial real estate company and what that means for you, that transformation. And then lastly, talk about how what advice you have for other CEOs undertaking AI transformation in their businesses. So with that general outline, let's get right into it. If you would start, Joe, with just a brief background on yourself and a brief background on Locate AI. Joe Lee (01:13.432) Yeah, yeah, first. Joe Lee (01:23.886) Yeah, for sure. I'm Joe Lee, founder, CEO of Locate AI. My background in technology really comes from my Stanford days where I studied computer science and AI. And what Locate AI does is we're an AI native leasing solution for national franchises. So if a franchisor is looking to grow, they'll typically get franchisees and try to sell licenses and open real estate around the country. So where we help... is we take a deep dive analysis of the franchise brand. We understand key drivers, like what is driving a successful store location, one versus the other. And we can actually create an AI model that can forecast where that brand will be successful anywhere in the country. We can do revenue forecasting, we can do market analysis, we can do site selection, site evaluation. And over time, you know, what we've done with our businesses, we've vertically integrated our business. So when a franchisee is looking to get a signed lease from start to finish, it's actually a real estate service that they need. So what we've done is we've integrated our AI model and our AI intelligence platform so that the franchisee can work with our end-to-end integrated service and we'll take them from start to finish, start to signed lease. using AI in every step of the way. Andy Sack (Computer) (02:50.553) And can you touch on how, give an example to the extent that you can, if you're willing to talk about one of your customers, give us a real example of a customer and what the business value is that you offer to that company. Joe Lee (03:05.742) Yeah. Yeah. I mean, at the core of our value prop, it's opening sites in high performing locations. That's the core of it. Right? So if a franchise business is, you know, if they're looking to open 50, 100, 200 locations over the next few years, you know, how much revenue each site is doing, that matters. That matters to the franchisee. That matters to the brand. That matters for royalties. That matters for momentum. matters for fundraising, just matters across the board. So what we've done with our clients is we've been able to demonstrate that by using AI to do site selection in a much more intelligent way, they can open stores that generate higher revenue levels than stores that they had on average prior. So many of our clients actually generate 10 to 15 % higher average unit volumes, average revenue per store, after working with Locate. compared to stories that they had before working with Locke-Kidd on normalize on a per market basis. Andy Sack (Computer) (04:09.965) That's great. want to come back to that in just a moment. of course, Adam, feel free to jump in at any moment. I'm going to. Joe Lee (04:12.013) Yeah. Adam Brotman (04:19.938) I'm excited to jump in. want you to ask your follow-up question because I want to talk about it's fascinating how Joe's describing how they have been using AI in their core proposition for differentiating as other brokers and helping to find higher performing real estate. than if you don't use Locate AI. But I can't wait to hear about how you're using AI internally as well. So there's both the application, almost the client facing element of what you're doing, as well as the internal use of it. I want to hear about both, but Andy, go ahead. Andy Sack (Computer) (04:53.517) Yeah, and that's actually a great call out, the sort of external client facing versus the internal use. And I think to give that some color before we dive into that distinction, Joe, if you could just talk a moment about as CEO of Locate, your own personal sort of aha moment with generative AI, because I think you are operating as an AI company for external. And then you add sort of an aha. Can you talk about that? What the aha was? Joe Lee (05:27.15) Yeah, for sure. it was actually Andy when, uh, it was shortly after we had dinner. Um, I forget maybe a year and a half ago was it. We had some nice sushi dinner in New York city. Um, and you know, I, I've tried chat GPT and, you know, I tried the first few versions. Um, but when the O one came out, uh, with the ability to, you know, quote unquote reason, its way through providing answers and thoughtful input. I think that changed the game. So it was sparked by our intellectually stimulating, you know, conversation over sushi and drinks. But really afterwards, I sat down and actually wrestled with, okay, like, let's give this a whirl, another whirl, and let's see, let's start asking, you know, very creative, thoughtful, difficult questions. that I'm struggling with on a day to day basis, whatever I'm working with or working on on any given week that changes that varies week to week. So, you know, that was a moment and you know, without a doubt, like the first few projects I used 014, it performed tasks that would normally take me a couple of weeks. It would do it in like 15 seconds, right? And like, I needed that personal like, like, I used to do this myself. And now this thing just did it in 15 seconds. Just like I needed to like juxtapose those two things side by side for me to like actually, you know, emotionally feel the impact that generative AI is having on the world and also what I could leverage it for. Andy Sack (Computer) (07:10.425) I don't know if it's true, but I think that was six months ago, not 18 months ago. I think we had sushi dinner in October of 2024. Joe Lee (07:16.246) Okay. Yeah. Joe Lee (07:21.196) Yeah, yeah, you're right. Yeah. Yeah, I'm getting the years next up. Yeah, six months ago. Andy Sack (Computer) (07:23.897) You had already used, it's interesting, you had already used ChatGPT and we're using it some, but it was really October, that dinner on October. All right, with that, Adam, you want to go back to your distinction and... Adam Brotman (07:36.482) Well, yeah, I I'd love to dive in for a second, Joe, because I'm, first of all, on your point about 01. couldn't agree more. And it's fascinating, that moment. I'm relating to what you're saying that like, I'm having that moment myself right now with O3 through the deep research project or product that OpenAI has. But Andy and I had it with O1 and O1 Pro before that, to your point. And it's sort of interesting because you sort of need to, as you said, it juxtapose what you thought the AI was capable of. Joe Lee (07:56.696) Mm-hmm. Adam Brotman (08:13.622) with what it is capable of. And it's hard to do until you actually just try a lot of different experiments. And there's no benchmark or test or whatever that's going to do that for you. But I have a question for you along the same lines, because you're in a really interesting situation. Unlike most of our podcasts, you are a computer science Stanford entrepreneur that started and founded an AI company. But you were, you were using what I'd almost call like old fashioned AI. You were using, you know, predictive models and other models and, machine learning before neural network based gen AI was something that you were playing around with. think, keep me honest. If that's not true, let me know. Cause my question for you is, have you actually gone back and I'm curious and, like taken some of the stuff that you guys have built models for, for your core? Joe Lee (08:46.712) Yeah. Adam Brotman (09:11.746) customer or client facing stuff and tried to see if Gen.ai can do it as well or even try to do it or is that am I touching on a third rail and if I am I apologize. Joe Lee (09:23.628) No, no, no, it's okay. Like I'll share with you like how like the framework that we're using to think about it. So I think you call it old fashioned AI. You know, so like the previously like machine learning models, what they were really good at was picking up on patterns, right? And being able to understand patterns in a way that we like humanly cannot by looking at Excel spreadsheets, or just we don't have the capacity in our heads, right? And what that led to was a surge of intelligence and predictability that we didn't have prior to machine learning. And machine learning is, know, it's decades old, but like really the modern machine learning, you know, that people started really picking up on since 2010. I think with this generative AI, we're looking at it slightly differently. We're looking at it more from an automation lens, right? So, so there's, there's really two pillars within Locate that we're focused on. So one is intelligence. It's still the core. machine learning models that we're serving petabytes of data and it's picking up on patterns. And it's still really good at that. And it's really good at regression thinking, like predicting how much revenue a store can generate. It's still really good at that. But I think we're generative AI, the upside with generative AI, LLMs and also specifically AI agents is the interface between technology and the human. Right? So like, Like, let's say I want to go figure out how much a site is going to do in revenue for a particular brand. That's a very common question we get. How do you actually do that? How do you actually do that for a client? Well, you read the report, you get a number, but you can't just tell them the number. You can't just be like, dude, this is going to do $2 million a year at the site. You want to sign a lease? You can't do that. You've got to have supporting evidence. You've got to have Adam Brotman (11:01.516) Right. Adam Brotman (11:16.002) Great. Joe Lee (11:20.526) data, you got to tell a story, right? You got to build reports, you got to modify the reports, you got to have pictures, you got to have maps, you got to tell a visual story that's analytical, but also like it strikes emotional chord. So like you can imagine the entire workflow that is generated there just to get to a revenue forecast number, right? How good is the site? So for us, way that we think about it, the entire workflow process and the ability to sift through that. is where we're applying generative AI, LLMs, AI agents, to actually go automate the entire workflow. But the intelligence, the old fashioned AI that you mentioned, is still just as powerful. And it's getting even smarter by the week. Yeah. Adam Brotman (11:54.177) Yeah. Adam Brotman (12:05.416) I get it. in other words, in other words, one of the ways you've been thinking about this as someone who's been in the AI world, even before the generative AI is that before generative AI, was, it was, and still is very good at a particular vertical thing that you've trained it and taught it how to do machine learning on and understand the patterns and, that's working well. But then you have to surround that with a bunch of like knowledge work. That's not super scalable. That's kind of manual. That's not even your secret sauce, but is important to provide the context for it. And that's all work that you're like, wait a minute. These, these neural network LLMs, they could automate that for us. And so there's like, you're blending sort of like a, like specialty, native machine learning. Old fashioned AI, as I called it. And then you're like surrounding that potentially with like a Gentic or like automated workflows that a, that a gen AI can do. Cause it's more general. can like do what a human does in a lot of ways. Is that versus like a human's not going to be able to do the, the vertical hardcore AI part, but a human might be able to like present it and place it in a certain way. But you're like, wait, I could actually automate that part of it with LLMs. Joe Lee (13:10.083) Yeah. Joe Lee (13:23.862) Yeah, yeah, yeah, totally, totally. And just to maybe like double click and just give like more specific example, like when we're building complex models for our clients that can forecast revenue, like, like the revenue forecasting, the AI models, that's, that's the cool part. But like, there's like, like 95 % of the human labor actually goes into setting that up. Like communicating with the client, getting the data, formatting the data, making sure that Adam Brotman (13:52.002) Yeah Joe Lee (13:53.058) Hey, the site ID is missing here. Like we need to know that it's missing and we need to cross reference it with another spreadsheet that has a slightly differently formatted ID and it needs to know and cross reference and match and put it together. And then it needs to send it off to our data scientists and they need to look at it. They need to put it in the pipeline. And, and, know, that's just on the internal side of things. Right. So like, like you, break, you break down like. the ultimate user experience and what happens behind the scenes to make everything go. It's all the labor that I think is involved. And I think that's the biggest impact that we're seeing personally with LLMs. Andy Sack (Computer) (14:35.213) Joe, would you talk now for a moment, as opposed to the old fashioned AI, talk for a moment about your vision for what you're doing at Locate going forward and the move to an agentic commercial real estate brokerage. Joe Lee (14:51.084) Yeah, yeah, super interesting. like, so ultimately, the value prop that we provide our clients is getting assigned lease at the best location. Right. So, so we, are a licensed brokerage. have real estate agents that are on staff at Locate. They're representing clients. And there is a very extensive workflow if you're a commercial real estate agent. and you know, from prospecting clients to scheduling meetings, to going back and forth, following up on, you know, rescheduling meetings to doing the initial site search, to pulling reports, to, to inputting information, uploading sites that we're going to go tour, creating the tour books, modifying the tour books. Like if you, if you look at the the workflow of what a real estate agent does today. You know, we're in the process of mapping this out and this is actually an ongoing exercise, this dynamic, but we expect 80 to 90 % of what a real estate agent does today will eventually be automated by AI agents. If you just look at, if you just break down and piecemeal the specific tasks. Now, of course, you're not going to displace the agent altogether, the real estate agent altogether. you still need a human touch to get a deal done. Sometimes there's difficult negotiations where you need to pick up the phone and talk to the landlord. And maybe you need to meet him in person and say, okay, like, let's get this deal. Let's get these deal points hammered home. Right. And that's difficult to do. BIA AI agents may be in the future, but I don't think we're there today. or if the client gets upset and they want to just. That. You know, they are like. You know, Andy, if I want to vent to you, I want to see you on a Zoom call and I want to give it to you. know, like, are just kind of simple examples, but the real estate agent is still going to be critical to the actual success of getting a deal done. The question is, what is the role going to be over time? And how is the role going to change over time? How's the persona of the real estate agent going to change over time? Joe Lee (17:15.762) What's the interaction between the real estate agent and the AI agents and how is that going to change over time? Like these are the questions that we're struggling and really wrestling with today. Yeah. Adam Brotman (17:25.878) I gotta say, just to jump in, Joe, what's so cool about what you're saying, and tell me if you see it the same way. And by the way, this could apply to a lot of other businesses as well, so tell me if this is true. And in your case, it's fascinating, because you're saying, look, at the end of the day, locate, we have a job to be done. Our job is to help a franchisor, or through their franchisee, whichever one it may be, find the optimal location. for their concept amongst choices in a certain geographic area. And optimal for a number of reasons, but primarily optimizing for revenue. So, okay, great. So you, that's your, you're clear. Like, you know what your job to be done is. And you bring two major things to the table. A model that like runs and finds patterns and is the old fashioned as I'm calling it, tongue in cheek. And you've also got an agent that has to provide the human touch around Andy Sack (Computer) (18:10.425) you Adam Brotman (18:22.026) understanding what the client needs, being there to answer their questions and be empathetic, to finalize negotiations. So there's a unique things that this technology, old fashioned AI and the human can do. But surrounding both of those things you've just described on this episode are a bunch of things that are not unique to either the technology or the human necessarily. They're like things that need to be done. but they're not like the magic of either the technology you guys originally developed plus this person that's the agent. And so you can surround those two pillars with a bunch of automation and smarter decisions and content production and a bunch of administrative stuff. And many of the times that's not the highest and best use of either that technology or the person. So it's fascinating to hear like when you describe it, you're like, it's this real augmentation that like really allows you, your technology and the humans that work with you, the agents, do, like, frankly, be way more productive because it's not, you're not, they're not being bottlenecked or slowed down by a bunch of administrative stuff that either takes time or you don't have the resources for or whatever. And in the end, you end up getting a theoretically, a much more efficient operation that either allows you to pass those cost savings onto your client and or make more EBITDA for your firm. Joe Lee (19:46.104) Yeah. Yeah. And I think, you know, getting a little more specific to our industry, which maybe the audience, maybe akin to or not, but, know, if you look at commercial real estate, you look at most people know commercial real estate brokers or real estate brokers in their life. The top producers, eventually what they'll do is they'll have junior agents under them, right? To do the menial tasks, the repetitive tasks like scheduling or pulling reports or do an initial data entry, whatever. In the beginning, all brokers will do A to Z, but eventually they'll compartmentalize their work. And then they'll say, I'm gonna be external facing, I'm gonna do the human things, the way that we've been talking about, and the agentifiable things, I'm gonna pass them by. And they don't know, that wasn't part of their language, right? But that's what they were essentially doing with the junior agents. So over time, an interesting question that... it comes up when I, when I sit at some of these conversations and talks with other leaders in commercial real estate and they're thinking about implementing AI and, and LLMs it's like, what's going to happen to the rising junior real estate agents, right? Like that's an interesting question, right? Adam Brotman (21:00.93) Yeah, Joe, say interesting question. I've heard that same question from law firms and accounting firms. Joe Lee (21:07.694) engineers. I'm an engineer by trade, right? Like, like, you know, like, it's not like when I graduated from Stanford, with a CS degree, you know, like, it's a great time to be a 22 year old software engineer, right? And you are anywhere in the US. But, but now it's like, when I want to talk to my other founder friends, we're hiring engineers, like, they're not looking for a junior talent, they're looking for, you know, minimum, you know, minimum five to seven years of experience that could harness, you know, Andy Sack (Computer) (21:10.041) Yeah. Joe Lee (21:36.578) just faster coding, like more scalable coding using LLMs. And it's like almost putting like downward pressure, right on these junior guys. And I think it's kind of funny, you're, we're going to see something similar with in the commercial real estate, right? So that's part of our also our child's cause we also do want to keep encouraging the younger folks to join and we want to empower them, right? So maybe it's a tool for them to scale up their careers faster, right? Adam Brotman (21:38.882) Right. Joe Lee (22:04.909) Maybe they can also apply the tools and level up in ways that the senior producers took years to do. Maybe they can do it in months. So there are silver linings that you can find there, but it is an interesting dynamic that's at play in commercial real estate, but probably across all industries. Andy Sack (Computer) (22:22.617) Joe, I'm going to take the conversation in this slightly different direction, which is you're an old-fashioned AI. You're now a virtualized, agentic AI commercial real estate company. What advice do you have for either the CEOs of some of franchises that you're... What advice do you have for them about embarking on an AI transformation journey? Joe Lee (22:29.944) Yeah. Whole-fascinating. I love that. Joe Lee (22:49.614) you gotta be excited about it. Like the executives, the leaders gotta be excited about it. Like they need to preach it. Andy Sack (Computer) (22:56.353) And what makes you excited about it? Because you you got excited about it. Joe Lee (23:01.678) was, it's the aha moments, you know, it's like, it's, like, it's bigger than myself. It's like, it's like, this is like, this is a next wave in technology. Like we're at a frontier. Right? Like, like this is like the internet. It's like the cloud. It's like, it's like, it's like AI. Andy Sack (Computer) (23:16.537) In practical business terms, what had you go, reflect on October and that had you, it was the reasoning moment. What had you go, this is an unlock for locator? Joe Lee (23:35.31) You know, I think here's the way I would add to that. think a lot of the times when you work on business problems, you start with a business problem, right? You're like, hey, we need to increase profit margin. And then you work backwards from there. And then you're like, okay, like, then how do we apply the tech? Or even like when you're trying to figure out when you're starting a venture, you're like, you start with a customer pain point, and you work backwards from there, right? And you say, okay, how do we, how do we create a solution to address the pain point? I think In rare occasions, the innovation is so strong that you start with the innovation and you get inspired by it. And then you look, look at the world around you and you say, okay, my data activities, how can this be impacted by this innovation? And I think those moments are rare, but I think this is one of those paradigms where, you know, yeah. Andy Sack (Computer) (24:24.409) That's well said, I'm conscious of time. I think we covered the main points, which was you gave a brief background on Locate and old fashioned AI. You talked about how you're transitioning to becoming an agentic commercial real estate company. And you gave some advice to CEOs. I'm sort of curious, Adam, what's your takeaway from this conversation with Joe? Adam Brotman (24:54.966) You know, I'm going to say something a little bit, maybe unexpected or tangential, but my takeaway is what's so interesting and unique about Joe and all kidding aside about old fashioned AI versus gen AI. Like in all reality, Joe's our first guest that I think we've ever had on this pod. And maybe one of the only people that we've talked to of late Andy who was already working with AI prior to neural network, gen AI coming along. And what's interesting is Joe's job to be done. Joe, you said start with a business problem. had another guest on and said, start with a business problem, figure out the data. Like this is a, there was a, there was a kind of a one, two, three, like what's your business problem? What's the data related to it? What's the output use AI to help you solve that. Ultimately it's making better decisions faster. And what's fascinating about this conversation with Joe, Andy, is that Joe was already using AI to help his clients make better decisions faster about locations. Right. But the difference is it wasn't democratized. It was like they had to work really hard at building sophisticated, specific pattern recognizing models. And now there's this like general purpose nature and easy access as Joe said to AI that can do all these other administrative agentifiable things. Most people are at that point from day one, Joe was already using AI to make better decisions faster for his core business, but now realizes he can amplify that and be more efficient at doing that by using a surrounding with AI is very meta, but it's actually really cool to see Joe kind of layer the gen AI on top of the humans and the core AI system to create a much more efficient ecosystem for himself. And I actually think a lot of businesses and I hope leaders that are listening to this have that same takeaway that I just had, which was like, yeah, I'm. Joe Lee (26:30.7) Yeah. Yeah. Adam Brotman (26:50.282) In the same boat, I might not be natively an AI company, but I can borrow from Joe's principles to sort of help me figure out what are the problems I'm trying to solve? How could I, what's the data around that? And how can I use AI to make better decisions faster? Joe, you just were like starting from a native AI perspective and sort of surrounded your agent humans and your core technology with that. it's really inspiring and it was fun to talk you about it today. Joe Lee (27:16.782) Yeah, yeah, it's new AI powering the old AI. Adam Brotman (27:20.29) That's great. Andy Sack (Computer) (27:21.657) My big takeaway from this conversation is one that I've stated in prior recordings and interviews, which I just am continually reminded about how important it is for the CEO to get an aha moment with Gen.ai themselves. And when that aha happens, things start to click. then they look at their business and their business problem, which in Joe's case was, how can I how can I use this new AI to unlock gross margin in my business and do this more efficiently? This could be really transformative to the underlying operations and financial performance of the company, particularly with agents. Like it's just a it's a total game changer to be able to employ both humans and AI agents as a as a commercial real estate brokerage. And the and fundamentally the financial picture that that produces is different than if you were just operating in either old AI or just operating in a commercial real estate brokerage without AI, without gen AI. So the role of the CEO and the growth mindset, everyone that's listening, get on chat GBT, start using these models. And number two, when you focus on your business from a financial perspective, Joe Lee (28:28.738) Yeah, yeah. Andy Sack (Computer) (28:46.125) position. This is a major where the innovation of the technology so great that you actually need to follow it. This is this is that moment. So, Joe, thank you so much for being on. Always great to see you. And for all those all those folks listening, thanks for listening to AI First with Adam and Andy. For more resources on how to become AI First, you can go to our website, forum3.com. Joe Lee (29:02.252) Yeah, you too. Andy Sack (Computer) (29:15.223) You can download case studies, research, broke out briefings and executive summaries, lots of great materials on our website for leaders looking to expand their knowledge. And of course, as always, we don't believe that you can over invest in AI learning. Onward. Adam Brotman (29:33.41) you