How Proprise is creating an Unfair Advantage in Self-Storage

THE SELF STORAGE REPORT — EPISODE TRANSCRIPT Episode: How Proprise is creating an Unfair Advantage in Self-Storage Guest: Jean Michael Da — Founder and CEO, Proprise Host: Chris Berg — Abernathey Development Recorded: April 23, 2025 Video: https://www.youtube.com/watch?v=fmph5lvnL24 Key topics: AI hallucinations and context-window limits in CRE underwriting; OpenAI o3 and o4-mini; Carnegie Mellon AI masters and Purdue computer science; Y Combinator; Proprise hotspots micro-market screening; 55,000 self-storage facilities tracked with rental rates refreshed every four to seven days; 500,000-plus commercial listings scraped daily; a 3-mile buy box (40,000 population, $50,000 household income, 10x10 climate-controlled rates above $150, supply under 9); housing-starts filter for forecast population growth; the AI listings agent scoring roughly 40 raw-land listings in the Sunbelt; composite heat maps and square feet per capita; zoning-code accuracy and the omitted-listings tab; Olympus Ventures going from 6 LOIs in 2024 to 4 in Q1 2025 and covering two times more markets; Ghost Storage closing an off-market East Coast parcel and saving 75% of deal-sourcing time; owner entity data without contact info; roadmap for development costs and deeper zoning Note: Speaker attribution reconstructed from raw captions. Light cleanup of transcription errors only; wording preserved. Timestamps and YouTube chapter markers removed. Turns marked [attribution inferred] could not be attributed with certainty. ————————————————————————————— Chris Berg: Welcome to the CRE Report. I'm your host, Chris Berg. We continue to hear about how AI is dominating so many different areas of industry. So, we're going to talk about how AI is continuing to penetrate commercial real estate on today's podcast. More specifically, self-storage. Very special guest joining us today. He's the founder and CEO of Proprise. Love their tagline, the unfair advantage in self-storage. Jean Michael Da. Jean Michael, it's great to have you with us, man. Really looking forward to this conversation. Jean Michael: Chris, thank you so much for having me on, uh, the podcast today. I'm really excited to have this conversation and, uh, yeah, looking forward to diving in. Chris Berg: I mean, it's incredible to, one of the things that jumped out to me, because you and I obviously spoke a little bit before this, and I just want to start here, is that you talked about, hey Chris, I've been really surprised. Because you've got such a great background, which you talk about in a minute when it comes to computer science, but you told me how surprised you've been on how many CRE people are starting to utilize AI. Why did that surprise you, and what's some of the use cases that you're seeing so far? Jean Michael: Yeah, you know, um, yeah, there's a couple things here, but you know, I think the overarching theme is that historically commercial real estate is an archaic industry. Things move slow here. The adoption of technology is, uh, you know, really really lagging behind. Um, and so just the underlying consensus is, you know, there's not a lot of technology innovations in commercial real estate. Um, and you know, in some cases that might be true, but I think now in 2025, um, we're seeing a lot of these institutional firms and even just everyday people, folks in real estate, are starting to see how AI can have a meaningful impact into their businesses and transform the way that they're doing and just thinking about real estate investing. Um, and so, like, yeah, a lot, most of our clients, some of them even have internal mandates where they're, uh, you know, making sure that everyone has a ChatGPT or another type of large model on their computer in some fashion so they can start using those every single day, and just to figure out how we could use AI to, um, you know, improve all aspects of their business. So that's just fascinating to see, um, this adoption of technology, and that, yeah, from the, uh, the top down being implemented at these, uh, very large companies. Chris Berg: You know, one of the things, just because of your background, which I'll talk about in a minute again, but that I want to share with you that just happened to me yesterday. I want to get your feedback on. And so I uploaded an OM who was in, you know, Houston, Texas, was asking him to do some deep dive questions for me around that marketplace within a 3, 3.5 mile radius. And this was with ChatGPT, and it started spitting that information to me in Calabasas, California. So I mean, one of the things I'm concerned about that I want to get your take on, is I think a lot of people just sort of fall prey, like, oh it's AI, it must be giving me great info. But I got to be frank, like I'm not seeing that even when it comes to underwriting. I mean, something as simple as an OM that clearly says Houston on it, and it's giving me data on, like, what's going on there? Like, how much should we be trusting the actual information that we're getting from AI? Jean Michael: Yeah. So, one of the biggest flaws about these AIs, AI, what I'm referring to is large language models, the way that they're built is that they are, um, it's not perfect. The outputs are not always perfect. And so they fall prey to what we call hallucinations, just outputs that just don't really match what you would expect, um, you know, a rational human being to output. Um, and so, um, it definitely is a common problem. And as we're seeing just advancements and improvements to these, uh, to these models, we're seeing those, uh, hallucinations, those error rates go down over time. And definitely, um, if you notice, you know, like even just this week, uh, OpenAI released two amazing models, o3 and, um, o4-mini, that just have, uh, you know, dramatically lower hallucination rates. They're more, you know, uh, reasonable. They're the smartest models that are out on the market today. And so using some of these higher-end models definitely reduce the risk of, um, uh, missing information. But it's always, you know, these are not perfect, uh, models. So there's always going to be flaws. Um, so right now you kind of have to double check the work that it's outputting, and kind of wrangle a little bit to make sure it, uh, you know, outputs correctly for you. Chris Berg: I mean, I feel like I'm getting a pretty specific prompt. So how does it hallucinate, read in a Houston OM and give me Calabasas information? Jean Michael: That, I'm not sure what their internal processes are. Um, yeah, I'll try to, you know, try running it again. Maybe it didn't even, you know, read the OM in any way. Or if there's significant amounts of information that was sent to it in that OM, maybe it can have, it probably couldn't have processed every single data point that was relevant to, um, what you were trying to ask. And because of that, you know, this is called the context window. The context window is maximized. It's not able to fully ingest all the data that's relevant, and, um, you know, just kind of sees the snapshot of that information, and it kind of starts hallucinating everything else that it doesn't know about because it's not in that window of, of, of, uh, their frame for that model. So those are common reasons why we see some of these, uh, issues with large language models today. And so it's still really important for people to be out there and critically think, know what's spitting back at you. Don't just trust it blindly, because, um, it's definitely not giving you always the best information. Chris Berg: So let's talk about your background. The reason I wanted to have you on here today, and then obviously your product as well and what problem you're solving, but just to give people an idea of your expertise in this arena, like, share with us, you went to Carnegie Mellon, computer science. Give us the lowdown. Jean Michael: Yeah. So, a little background, like how I even ended up here. You know, my family, both of them worked at Intel for 20-plus years. Both computer engineers, uh, worked in the Bay Area, and, um, you know, a lot of their, um, you know, their interest in money, you know, they started investing in, mostly in real estate. And I just grew up along that path. You know, they wanted me to do a PhD. I kind of refused that. I, you know, there's not enough time in my life to do a PhD. So I kind of paused at the masters. But you know, I started learning, uh, computer science at the age of 16, 17, um, really loved it. And I went to school, at first I was at Purdue for undergrad studying computer science, and that's where I learned more about, um, artificial intelligence and AI. And initially I wanted to create an AI that could trade stocks for me, you know, make me money automatically. Um, this was 2015. Realized that, you know, there's PhDs, you know, math wizards, and, uh, that are being paid millions of dollars at Citadel and these high frequency trading companies that do that. And, um, you know, as an 18-year-old kid, probably could not compete. Uh, but that just introduced me to how AI is used in many aspects of the world, whether that's self-driving or, uh, you know, doing, uh, auto-translating text from English to French or Spanish, and just used all around the world around us. Um, and so I wanted to pursue more. I got very interested in AI and how it could help us in the world, and I decided to go to Carnegie Mellon to get my masters, where I really was focused on how do we learn more about these, you know, teaching more models and to do more, um, you know, insights that are, um, above the human level. And, um, it's really amazing to see the advancements of AI now and how it's just permeating throughout essentially all of the culture across the world, um, and how it's being used in almost everyone's daily lives today. It's really fascinating to watch. Chris Berg: Not to get too far off track here, but recently I saw a piece from Eric Schmidt, you know, the former CEO of Google, talking about, hey look, in six years the AI is going to be smarter than all the humans combined. And we don't have even the language to describe what that's going to look like for us in a world. One, do you agree with that timeline? And do you have any language to help describe it? Jean Michael: So I think he's probably on the right track. So, you know, this whole wave of AI started back, I mean, when ChatGPT launched, which was what, November 2022, I believe. Um, and that's, you know, not even three years ago. Um, and so we're only three years into this. So I can't even imagine what, you know, six years down the road would look like. Um, and you know, these models are blowing me away every single day. I'm having most fun just touching these models, interacting with them, using them for my business. Um, and so I'm really excited to see how, uh, you know, all these investments, on every single day there's constant constant improvements. And, um, I definitely see that this rate of growth, uh, increasing over time as well. Um, and then in terms of like a language, uh, unfortunately there's not much I have to answer for that. Um, I think, um, you know, right now it's all around prompt engineering. So essentially I think as we're going forward, we're trying to realize like, how are these models trained? And because of the way that they're trained, there's a certain way to interact with them to get the right information out of them. Um, it can't just be, hey, you know, get me this information. It's more of, you know, think like this type of analyst, and this analyst is a perfect, is, you know, 10 years of experience, and that type of framing makes it, uh, you know, essentially put on the hat of an analyst and start analyzing things for you in that realm. Otherwise it's going to be very broad and narrow. So I think at this time there's still a lot of tons of experimentation on how to properly, you know, talk to these models to get the output that you want. Um, but at the moment, it's not perfect science just yet. Chris Berg: No, far from it, based just some of the results I got yesterday. But it is definitely a fascinating field. So, um, your family's with, you know, in tech, you're studying computer science, a little bit of real estate. How do you land on self-storage? Jean Michael: Yeah, it's, uh, it's an interesting journey. So, you know, it all started from my interest in, um, in real estate as well. So, you know, when I got out of school, I knew that I wanted to, um, start acquiring assets for myself. Um, this was, uh, 2023, and sorry, not 2023, this was 2020. And I, uh, you know, it was a very competitive market at the time. Interest rates were rock bottom, competitive, uh, real estate was very competitive at the time. And so what I was doing is I was just building up my own machine learning models that would identify, um, opportunities. Right now these are single family homes that were going on the market that fit my buy box. And I would start, uh, my goal was to be the first offer in all these, on these, uh, houses so that I could, um, you know, be the first to get that property. And, uh, as I learned more about real estate investing, uh, and the different asset classes out there outside of single family homes, I learned about, you know, folks who are investing in parking lots, uh, mobile home parks, self-storage. And I didn't know at the time that those are investments you could make, but definitely opened my eyes around, you know, okay, real estate is very expansive. There's a lot of opportunities in, you know, any structure out there in the world. Um, and that led me down to just investigating self-storage a bit more. Um, there isn't, uh, you know, you're not cleaning bathrooms, uh, you know, as significantly you might would have to. And these are resilient, uh, this is a very resilient asset class, and we're just been seeing tons of institutional capital, um, being funneled into self-storage as it's a very strong asset that, um, very rarely, um, goes under. So, um, that led me to just figuring out, you know, how can we make sure that, uh, the folks who are in this asset class are well equipped for investing in this, uh, in this market. Um, and, you know, let us, uh, further iterating along with customers and clients, um, who are doing the job day in and day out to find and acquire, uh, assets for self-storage. Chris Berg: So you're part of Y Combinator. As you are fairly new to, like, self-storage and real estate investing, you've got this great tech background, like, what would you say is the problem that you're solving with Proprise? What was the initial one, at least? Jean Michael: Yep. So initially, you know, when we got into this market, um, we didn't have much exposure to self-storage at the time. I mean, yes, I had a single family, uh, investing background, so I understood a little bit about real estate investing. Um, but we took it on us for, you know, we took essentially a month of talking to every single, um, person that we could get a hands on, to just to learn about their process, um, and, you know, what they do for, uh, for their daily lives in self-storage. So these are brokers, lenders, acquisition teams, developers, and just understanding, you know, what do they do and where is there a significant amounts of pain. And we identified just, like, one very, um, critical piece for developers in particular, how they were looking for pieces of land or potentially assets to then redevelop or, um, convert into self-storage. And they're just going through so many listings and opportunities that are being sent to them, and they have to do significant amounts of, uh, checking to make sure that this piece of property makes sense for their investment thesis. You know, whether the population is increasing, or the rental rates where they should be to allow for development. Um, you know, what are the zoning for this piece of land? Um, so there's a significant amount of checks that they're doing, and we saw that there's a lot of time wasted, um, or is just spent on analyzing opportunities that just resulted in a no from just a quick check. And what we're trying to do is, how do we automate that process so they can find more yeses way faster than ever before. Chris Berg: So let's dive into that. Let's do a demo here, if you want to use me as a guinea pig, or if you want to just use some other use cases that you're familiar with. But I'll bring up, uh, your screen, um, and just kind of walk us through, I guess, what we're seeing, and then walk us through a use case as a developer. Hey, I've got to go find some land that we want to build on. How does Proprise help me solve that problem and do it with efficiency? Jean Michael: Yeah. Yeah, happy to show you. So just for some context, Proprise is, uh, you know, we're a data analytics platform that is AI powered. And so right now you see a bunch of dots on the map all across the United States. And, uh, these all dots represent some form of property. So if I zoom in, let's say we go to California. Um, we could hover over any one of these purple dots and you'll see that these are, you know, self-storage opportunities. Um, just existing self-storage out there in the world. We're tracking the how many floors it is, when it was built, the address, the net rentable square feet for this property. Um, as well as some, we have these orange dots which are, uh, these are self-storage projects. So in the pipeline, whether at their planning stages, under construction, or nearing completion, we're tracking this all across the United States. So today we have over 55,000 self-storage facilities that we're tracking, along with their rental rates, um, as well as tens of thousands of, uh, um, uh, pipeline, uh, self-storage source in the pipeline. Chris Berg: So, do you mind if I interrupt you at times, because just, I'll kind of be the ombudsman here. Okay. So you say that you're tracking 55,000 self-storage units, the rates, things. Where are you pulling all this data from? And how often are you pulling it? Jean Michael: So for the rates, we're collecting that data every week. As somewhere between four to seven days we're collecting the rental rate information to stay, um, uh, up to date. And, um, we're getting that data from their website. So, you know, we built AI models that can go and crawl any website and collect that data, fetch that data for us automatically. Of course, like we've mentioned, AI is not always perfect. And so we have a team of reviewers, uh, you know, data quality analysts, what we call them, to, um, you know, triple check, triple verify all the rent information, the size of the parcels of the property, um, and to just make sure that we have the highest quality of data on the platform today. Chris Berg: Nice. Jean Michael: Um, and so yeah, so we have all this information, but you know, this is great to get a sense of what's, you know, self-storage as out there outlook. But what we're really focused on is finding, it's just the deal sourcing aspect. How do we find opportunities that make sense to, uh, you know, further invest time and money into? And so one of the unique things we started off with Proprise was we call them our hotspots feature. So what we're trying to find is find micro markets. So little pockets within cities that are extremely, um, opportunistic for self-storage. And, um, if it's okay, I can kind of walk through a demo of how, uh, we can do this. Chris Berg: Yeah, please. Jean Michael: Okay, so you know, I'm putting on my self-storage developer hat. Um, let's say I really love the Sunbelt. I want to find opportunities in the Sunbelt region. What I'm going to do is just open up, oops, let me zoom back out. Actually clicked on the map. Um, I'm going to find hotspots up here, and we're going to start building up our buy box, our investment thesis. So typically in self-storage, um, clients typically work along a 3-mile, uh, radius or circle. And so we're saying, hey, in a 3-mile, um, in the Sunbelt, you know, we want population minimum, we want some activity. So maybe 40,000 people in that 3-mile. Um, we want non-negative population growth, so we'll just start out with zero at the minimum. Um, and then let's say household incomes, we want that to be $50,000. So this information comes from the census. Um, and so that's how we're able to get this demographic information. And then for the rents, you know, we can add multiple different types of rents and layers onto this. But you know, we're tracking the 10x10 climate rates, the online, the in-store rates. We can even look at 5x5s or even different amenity types as well. Um, in this example, I'm just going to do the 10x10 climate. And let's just say I'm looking for, uh, rates above $150 for the monthly 10x10 climate control rates. And then lastly, supply is an importance. Um, so we can start filtering out, uh, for areas that have, you know, maybe low supply, and we can include, you know, we can just have the existing self-storage, or we could even go down to include all projects that are, you know, from the planning stages all the way down to existing as well. In this case, let's say, um, we want to be realistic. Let's just do the, um, so group capital with, you know, actually upcoming construction, and we'll say that's under a 9, which is below the, uh, roughly around the average of the US average. Um, the last thing I wanted to call out is this, our housing starts. So you know, we're collecting information on every single housing development project across the United States, whether this is single family or multifamily projects. And, um, we could start filtering for, hey, I want to find areas that, you know, if all the projects got completed tomorrow, how much percentage would the population increase by? This is another way to just figure out, uh, you know, forecasting growth in a market. So let's say, you know, we want 4% population growth based on these development projects, and we can start a search like that. If I zoom back out, um, you know, now we have these pink areas across the US, um, all across the Sunbelt region that, um, are telling us, you know, they fit all these criteria. And if I zoom in, I'm going to turn off facilities here. If I zoom in, let's say, to Phoenix, um, we can start to see that, you know, there's, you know, these aren't, you know, uh, normal shapes. These are very unique custom shapes that Proprise is giving you. And essentially what it's trying to say is, hey, in these areas, you know, if I put my, you know, my 3-mile pin in this market, um, it will always fit my parameters that I've set right here. So anywhere in these, uh, pink shapes will find exactly, uh, markets that fit this. Um, does all this make sense? Chris Berg: 100%. So then the other thing that's nice too, if I remember from our earlier conversation, is that I can, I think I can somehow maybe circle that area, and then if a listing pops up, it'll automatically email me. Or walk me through that. Jean Michael: Yeah. So, um, another data point we're collecting on a daily basis is we're collecting listings from all over the internet. Um, and so, you know, today we have over 500,000 commercial real estate listings, from data centers all the way down to multifamily properties and even land. And so, you know, we could layer those listings on the map and start to see, you know, where that is, um, you know, across the whole United States. Um, so this is great, but there's just a ton of dots, and these are, uh, you know, you can kind of be overwhelming to go through every single one of these. And so we have a functionality where we can filter on certain listing types and the size and multiple other factors. And then we can even have automations to email you when new listings that come online that fit your criteria, um, come to you. But what I wanted to show off today was how AI can actually analyze all those listings for us automatically and then create a report and help us rank all these listings for us, um, in just a few minutes. So I'm going to quickly show that here. So you know, we have all these listings, and I really only care about listings in my, you know, my pink areas. And because I'm a selfish developer, I'm looking to 2 to 5 acres of land. I don't care about data centers. I don't care about housing. I'm looking for land in this case. So let's tell AI a little bit about this. And so we'll just have a way to analyze listing. Um, so I click this button and it's able to just, you know, it just tells me, hey, these are your criteria. Um, and it's asking what kind of, uh, listing am I looking for? In this case, I'll just say raw land, um, 2 to 5 acres for self-storage development. It'll start thinking, um, understands I'm looking for land in the Sunbelt area. And then it's going to start kicking off what we call our listings agent, and it'll go through all these listings and start analyzing this. So this will take maybe a minute to process, I think. Um, but, uh, yeah, happy to answer any questions as we're going through. Chris Berg: Yeah. So as it's doing that, is it also, I mean, do you have the opportunity to go in and, and because the big thing we always bump up against, of course, is zoning, right? So does your software and or this agent analyze the zoning as well to confirm that, hey, this thing is zoned for self-storage? Jean Michael: Uh, yes. Exa, exactly. So, um, actually took faster, uh, than I expected. So we can just view this report and we'll see what popped up. So in this case, it went through, try to do the math here, uh, roughly 40 listings, and it scores them. And so what it's telling us here, you know, it gave us an opinion for all these listings. It tells us the address. If I zoom in a bit, um, you know, the first thing it calls out is the zoning. Um, this is 2-acre, uh, I think I-1 zoned lot in Florida. And it's great because of its strategic location, high visibility and zoning. Um, because it understands, you know, we took, we gave it a satellite image. It knows that it's on a very high, uh, throughput, uh, you know, road. And so all the highlights are, you know, it's a 2-acre lot, has light industrial zoning. Um, and you know, it's in a prime location. And you know, it gives us some cons. Of course, not everything is perfect. Um, so talked about a high purchase price and, you know, potential development costs in this market. You know, and there's, you know, every single one of these listings that I found in our, on our hotspots, um, it has ranked and categorized them for numerous reasons. Um, yeah, go ahead. Chris Berg: Does it have the ability as well? Is your software reading city and/or county documents so it can tell me the FAR, how high I can build, setbacks? Are you guys that far advanced yet, or? Jean Michael: Not yet. It doesn't go that detailed, um, at the moment, but it's something we're always, like, working on, is just, you know, getting more clarity on what can actually be built here in this market. Um, so yeah, so that's kind of how it is today. Um, but yeah, that's a great question. Chris Berg: And then again, I don't know if your tech is here, but I'm assuming you're going here. Do you also have an idea? I think this one is in, what it says, Sunrise, Florida or something like that. Um, do you have an idea, like, what construction costs are in that particular area? So that way I can start to take that data, and will it even help me underwrite what it looks like to build a class A self-storage facility on 2 acres in this area? Jean Michael: Yeah, excellent question. So this is all on the road map right now. Or this is, you know, we just launched this yesterday, and so it's, you know, how do we actually get this list of listings? Maybe there's 10, maybe there's 100 of those. Let's rank them based on potential, uh, you know, viability for self-storage development. Um, and then of course, the next things we have to add on are, uh, development costs, deeper zoning information, and that would definitely rerank all this on this list and just get a more higher clarity on what actually makes sense for a real self-storage development. Um, but yes, everything's, like you're mentioning, is on the road map and does not exist just yet. Chris Berg: Wow, it's incredible. Um, how accurate are you finding that your zoning is data, the zoning data? Jean Michael: It's quite accurate. Um, you know, if I go on Proprise and, like, investigate this listing, um, you know, we're going to be able to pull the zoning code, uh, as well as, you know, go to the actual document themselves, and you just cross check, uh, you know, what we're saying here with the raw material that, uh, you know, the local jurisdiction has put out. Um, but we've seen significant accuracy in what we're, uh, proposing today. Um, just while I'm here, the last thing I wanted to show off is like this omitted tab. So you know, we matched a lot of them, but also, um, some of these listings are, you know, have a very low score and they, and the AI tells us it's not recommended for a few reasons. And so like in this case, you know, it calls that there's an existing warehouse and that it's already leased as a daycare center, may not be optimal to, you know, purchase this asset and clear the tenants to do self-storage. And you know, does all this for us automatically. So, you know, we can always, we could prioritize the opportunities actually make sense for me. And we're always iterating on this. I think there's, um, many things we could add as inputs to help the AI understand our investment strategies. Um, but I, this will keep, we're going to keep refining and iterating this over time. Chris Berg: It's really incredible, Jean Michael. I mean, very, very well done. Anything else you want to share as far as, um, what you're doing currently, or what else is available within the software? Jean Michael: I think that's essentially it. I guess one more thing I wanted to call out was our heat maps feature. Um, that's a great way just to understand markets, um, deeper. So let me pull that up quickly. Go back to this page. Um, and so, um, let me remove all the listings. Um, so we've, you know, gone ahead and did our hotspots. Um, but we also can overlay, you know, heat maps onto these pages. So if we overlay the population heat map, you can start to see, you know, high populations in, you know, Glendale and Phoenix area, and we're seeing less population outside in red. Um, but I think what's actually interesting is, is if I go down a bit, we call them our composite heat maps. And so we can start adding on, uh, interesting, I would say, data points to then create our own custom heat map that our users use. So maybe in this case, some of our users think population is important, but maybe not as important as others. So we'll add in, um, square foot per capita. So supply is a very critical importance for them. And then, um, also, like I say, rents for 10x10 is important maybe, but maybe not as important as a square feet per capita. And so we can apply that, and now visually on a map we can start to see where there's holes, and, and like even how that relates to our hotspots. And we start to see, you know, actually these heat maps make sense for me based on, you know, what I value as an investor for the criteria that I'm looking for. Um, so that's a very powerful feature that a lot of our clients really really enjoy using. Chris Berg: Yeah. So walk me through the colors here, because what the yellow equates to what, and then like the pinkish equates to what? So what am I looking at here? Jean Michael: Yeah. So, um, red means, you know, it's a low score. So we created our custom heat maps, is we're getting low scores, and then green areas and dark green areas are high scores for our heat map. So we want to be looking for areas that are in the high green areas. Um, it doesn't look like there's any, uh, in Phoenix, but we're seeing some lighter green areas in our hotspots above here in the Paradise Valley. Um, and, uh, you know, we can start to see some areas where may not be as great for self-storage developments, like Peoria area and things like that. Chris Berg: Are there ways as well, since you're scraping every four to seven days, where I can look at, okay, these are areas that I like, maybe the rates aren't quite there yet, but hey, can I get alerts that, hey, rates have gone up 10%, 15%, so I can circle back to that particular area in the future and be like, okay, now maybe this underwrites. Jean Michael: Yeah, it's a great idea. So right now we have a section in the app that goes deeper into like the rate trends over time. Um, one thing we're looking to add soon, as we're getting into, um, significant AI features, is a way to track markets and say, hey, I really like Glendale, for example, but, um, I just want to start tracking it and have AI send me reports, um, whenever there's an interesting thing that pops up. So maybe, okay, we've noticed 5x5, uh, are significantly, um, uh, you know, close to a 10x10 rate. Um, that might be helpful to know. Or, uh, 5x5s have been increasing, you know, 5% month over month, or 10% month over month. Um, and just get me those insights so I'm not constantly digging through these markets every day to analyze that. Um, so we're working on ways to just start tracking and have essentially AI do that work for us, um, and pull those insights that we may have missed otherwise. Chris Berg: You know, one of the things I enjoy about what you're doing, Jean Michael, is you seem to be very responsive. Like, I can come to you and go, hey, I've got this problem, can you guys code this? And I'm presuming you're going to do everything you can to help solve that problem for your clients. So, couple things. One, if you can share with us some powerful anecdotes of current clients that are using this. I know you've got, um, how do I want to say it without disclosing? Some very well-known clients maybe is a good way to say it. So, any anecdotes that you can share? And if you can disclose who these clients are, that would be great. And then also, what are you hearing? Like, what's a common theme, common pattern you're hearing from developers? You're like, man, I need you to solve this problem for us. Jean Michael: Yeah. Yeah. Great. And is it okay if I stop sharing here? Chris Berg: 100%, man. I can help you with that. Yeah, that'll be good. Jean Michael: Is that good? [attribution inferred] Chris Berg: Yeah. [attribution inferred] Jean Michael: Excellent. Yeah. Thanks. Uh, thanks, Chris. So in terms of like the clients, yeah. So we're working with numerous, I would say, top 100 client, uh, top 100 operators in the self-storage, uh, industry. Um, and so most of our clients are, you know, in those top 100s. Some of them are, you know, recognizable publicly traded REITs. Um, and so we're helping them essentially identify these micro markets, these hotspots that they really want to dive deeper into. Um, as well as, uh, being able to, like, look at listings and look at and analyze properties deeper in rich detail, um, to come to conclusions and make better strategic decisions, um, across the board. Um, there's just so many, since we're operating nationally, there's just so many, uh, little holes they have to cover, and it's very impossible for even a large team to cover every single thing, and AI definitely makes that process, um, much more palatable and easier to do. Um, and so a common thing that we're seeing, uh, you know, some quotes that we're able to see is, um, working with clients like Olympus, Olympus Ventures out in Minnesota. Um, and, you know, they've seen that their, uh, the letters of intent have increased. Um, I believe they said it was in 2024 they did 6 letters of intent, and, uh, in this year, 2025, they're already at, you know, just the first quarter, they're at 4. And they attribute a lot of that because of, uh, you know, of what we're doing at Proprise. They're able to set alerts for listings, identify markets that they would have never found before. So they actually are now covering, you know, two times more markets than they ever did before Proprise. Um, and that's getting them more deal flow, being able to access more opportunities that, uh, they didn't have the bandwidth to before. Um, additionally, we had, uh, you know, Ghost Storage use it for, within a couple weeks of using Proprise, identify a plot of land that was actually off market that, uh, you know, was on the East Coast, that they, um, overlooked at the time. And then they got a broker to just to call through a list of, uh, you know, zoned parcels and got in contact with the owner, and, um, got that deal done within a couple weeks of Proprise. It saved them 75% of deal sourcing time. And so it's just a significant, uh, improvements on the current processes, and we're going to keep pushing and see if we can get that 75% down to zero, get that down to 100%. But, um, uh, so yeah, so always always iterating and talking with a lot of our clients to just fine-tune these AI models and help start doing actual work, and, um, get them closer to a yes for properties that it would be very hard to find otherwise. Chris Berg: Thank you for bringing that up. I did want to talk about the off-market piece. Um, this has been great, by the way. Thank you, Jean Michael. So for example, some of those pink areas that were the hotspots, um, if there's no listings there, I think the beauty as well is at least now it gives me some very specific areas I can start diving into and going to call, you know, land owners and go, hey, are you open to selling, and get some off-market deals done? Jean Michael: Yes. Yeah. Yeah. Exactly. So, um, I can't imagine a world without this. Um, you would have to be digging through a bunch of areas, finding if, you know, the rents are where there should be, and we'll say same with the supply, and then, um, start just, you know, pick and, like, look around for every parcel in the area, and then match those to the proper zoning. Um, and so that just must take a significant amount of time, and almost, uh, almost impossible actually to do this regularly. And, um, we've seen a lot of our clients not really interested in doing a task like that, you know, before Proprise. But now that we do, we can give them, you know, an easy curated list of, hey, here are the pieces of property that are off the market today, here's what the owner's information, and here's, you know, it's between 2 to 5 acres and is zoned correctly. Definitely go and pursue those leads. And that's actually getting them, you know, better action, uh, based on, rather than going on the market, which there might be a little bit more competition there. Chris Berg: So last time we spoke, um, and maybe this has changed, but you said, yeah Chris, I can definitely show you, you know, the hotspots, and then obviously that'll give you the areas that you want to focus maybe on some off-market deals, and then you will show me if it's zoned for self-storage or not. The one thing you didn't have is you didn't have the actual owner's contact info, correct? Do you have that now? Jean Michael: So we know the entity that owns it. Um, we don't have the actual contact information yet. Um, brokers are remarkable for this, and they can get that done. They're, uh, you know, private investigators, um, almost. Um, but, uh, at this moment contact information is not yet available on Proprise. Chris Berg: But I still can say, okay, this is off market, it's zoned correctly, and I can start digging and scraping and trying to find that contact, and then do what I need to do to get it done. Jean Michael: Yeah. Right. Right. Exactly. Chris Berg: Wow. Fantastic. All right, my friend. I thought this was going to go like 20 minutes. We're well over that. So I'd love to have you back just to continue to talk about what you're doing and obviously the advancements in prop tech. Um, I want to wrap up with two things. One, where do you see, like, let's just go out 12 to 24 months. If you can go 36, great. I mean, where do you see Proprise in 12 to 36 months, knowing how fast this tech is changing and moving? Jean Michael: Yeah, look, I mean, we're really interested in self-storage. I think it's a great market, and, um, we definitely learned a lot in this industry, and we want to continue serving, um, developers. One thing that we're definitely, um, we want to start serving as well is just the brokers and acquisitions teams, which, um, we may have neglected over the past year of being existence. And so helping them, understanding their pain points and how can they start using AI and their workflows to, um, automate a lot of the tasks or just generate insights that would be very hard to do otherwise. So developers got a lot of love, um, but we want to see, you know, can we get the whole ecosystem involved as well. And then it naturally, um, you know, what we're doing doesn't just apply to self-storage. Um, we see this going across, you know, industrial, retail, multifamily as well. And so we have big visions for the future to expand all across commercial real estate and help everyone in the investment world, um, just find opportunities that, um, they would have otherwise missed. Chris Berg: All right, my friend. I want to give you the last word. Anything else you want to add or share that I haven't asked you? Then most importantly, if people want to reach out, how can they get a hold of you? Jean Michael: Yeah. So, um, you can find me on LinkedIn, Jean Michael Da. Very unique name. Um, but, uh, also if you're interested in, you know, learning more about Proprise, please visit us at proprise.ai. Um, happy to have conversations, um, with anyone that's interested, even if you're not a self-storage operator. I like talking to folks. Um, this is how I learn, and, um, anything that we learn, we try to implement into the product to make it better. Chris Berg: And I've got your email here. Are you okay if I put that up on the screen for people? Jean Michael: And yeah, please do. Chris Berg: So if you want to email him, too. It's just jmd@proprise.io. Again, jmd@proprise.io. Check out their site. Obviously, great stuff. It's fun to see where prop tech is heading, and I really think you're doing a great job kind of leading that charge, um, Jean Michael. So I appreciate the time, and we look forward to having you back. Jean Michael: Thanks so much, Chris. It was a blast talking to you, and always will be, bro. Chris Berg: That was fantastic. All right, this is the CRE Report. I'm your host, Chris Berg. Thank you for joining us. Please share this podcast. — END OF TRANSCRIPT —