L.D. Salmanson

L.D. (0:04): I've been selling this AI and real estate vision now for you know 9, 10 years and people told me I was crazy for the last nine. So I think the vast majority of real estate operations will be non-human-based in the very near future.

Clara (0:15): We are living in a world where every major business decision, who to sell to, where to expand, what to build next, is only as good as the data behind it. But here's what most people don't realize yet. Not all data is created equal. The companies that figured that out early, they win. I'm Clara. Welcome to Conversations in Data by Tech Accelerator. And today I'm sitting here with someone who decided to go fix an entire industry's problem from scratch. L.D. Salmanson, CEO and co-founder of Cherre. Welcome.

L.D. (0:48): Thanks for having me. It's great to be here.

Clara (0:53): Yeah, absolutely. Thank you so much for accepting our invitation and I'm very excited to chat. I've done a lot of research in the past couple of weeks while we were chatting. So, let's get into it. So, you've built an unique path. At what point did the entrepreneur in you take over?

L.D. (1:05): Yeah, I guess I've always been a bit of a builder. Ben and I started our first company when we were in high school. That went very well. I've been involved building some other companies since then after the army and while school and then later when we sold to Oppenheimer then obviously now with Cherre. I don't know what entrepreneur means. It's one of those words that almost feels hollow. I think of it more like an addiction. It's probably more like a drug in the sense that you can't stop yourself from doing it. It's probably very unhealthy in many ways, but here I am. I keep doing it. So, I don't know. I guess it's a very unhealthy addict. My best way of describing it.

Clara (1:48): That's very that's very interesting. I hadn't heard that before, but I guess it makes sense, right? So you build something, it grows and you want to expand and then you just keep going. And that definitely makes sense to me. So I appreciate your answer. And you know, I've done a lot of research about your company, your employees. You have quite a lot of employees. What is something about you that would surprise them, surprise your team if they heard it here for the first time?

L.D. (2:15): My team, I don't think a lot would surprise me. It would surprise my team. I am abnormally direct and straightforward with everyone in the company. We're an extremely transparent organization. I'm not going to say that we're flat because I don't think any organization is truly flat, but we definitely expect people to have a voice from day one. And across the entire company, I think I sometimes come off as probably a little colder than probably I am. I'm very empathetic towards the people who work for me. I'm very passionate about the people, not just about what we do. And I think that generally just my demeanor is always like cold to a certain degree, but I think those who work very close with me, you know, have seen that over the years and they know me very well.

Clara (2:53): Yeah. No, definitely. I think it's very important for you to, yes, you, you know, hold yourself well. You have a cool demeanor like you said, but at the end of the day, your employees matter. You're giving them I think I saw that you give them the opportunity to come in into the office or not. And then lots of benefits as well. And at the end of the day, those things show, right? So, I feel like they would feel appreciated nonetheless, even if you're a more distinct or colder leader, they still feel seen because of benefits that you may offer, right?

L.D. (3:28): Yeah. I mean, I think I think that this is kind of more philosophical discussion, but I think that employers often have this feeling of if my employees are at home, they're not working. And I think that means one of two things: either you don't know how to measure an employee's output and in which case you probably shouldn't be hiring. Right? If you don't know how to measure output, that's probably a bad role. It's not defined well. Or you know how to define output but you're not able to inspire that employee to deliver more. And that's really on you, right? If you said, "Well, if he was in the office, they would do way more." Well, then you're really admitting that your employee doesn't give a [ __ ] about you and they're just doing whatever the bare minimum is. Again, that's on you, right? So I think either way if you're not able to produce a productive employee at home it's on the employer not on the employee. Now I know that's a general statement. There's some roles where that's not the case and there are some roles you have to be in the office to do. But especially for knowledge workers, I look at our engineers and our product managers and data scientists. Because those are the folks that are often hard to measure because it's very easy to measure folks on a sales and you know commercial side of the house. But on the other side of the house often or people often say that it's hard. I don't think it's hard. There are a lot of ways of measuring that. Those are some of my hardest working employees. And I think that we've found a good way to measure that performance in the same way as we're able to measure the hardest working performers on the commercial side. And I think that's you know it's incumbent on us the managers to be able to do that. It's not incumbent on the employees to figure that out.

Clara (5:00): Yeah. No, I completely agree with you. But let's stop talking about HR and get into a little bit more into the problem and the bet. I feel like in life we face a lot of problems and we take a lot of bets. So I'm excited to chat about this. Cherre was built to solve one of the messiest data problems out there which is real estate. For people who don't live in this world, why is real estate data so uniquely broken?

L.D. (5:19): I don't know if unique is the right word. I think there's a lot of broken data in the world. I think real estate has a few things that are slightly different than other industries. I think one is just the cycle time. So, let's say I knew that something was going to happen and I wanted to show that and I would create some kind of plan to prove that's going to happen. It might take years for me to see that result. Whereas maybe if I'm in a public equity market, you know, a high frequency trading fund, I can get a very quick feedback loop and act on it. There are a lot of logistics is another one where if I change a shipping route, I can see very quickly if I've added value or not. It's a lot harder to do that real estate just because the cycles are longer and slower. So I think part of it's just a structural issue. I think it's also the case that it's a very risk-averse industry. So it leads itself to people who are entrepreneurial but risk-averse entrepreneurial. I mean it's a whole industry. People need to hold asset. They're so risk-averse. It's literally the physical asset holded. And that doesn't lead itself to investing in technology platforms broadly. It's also the case that if you track kind of the evolution of some of these technology solutions over time and the evolution of some of those real estate cycles, they didn't really align. So if you went back to the dot boom and bust that happened just before 2000 and that kind of created this bad taste in the mouth for most of the real estate investors said, "Oh, we're happy we dodged that bullet. Good thing we didn't fall for that nonsense," right? And then in, you know, 2001, 9/11 happens in New York, which is the largest office market in the world, just decimates the industry here, and it takes a long time for things to recover. And then in 2002, three, we have a recession. Again, not a good time to invest. There's a slight recovery period, but real estate's not excited to invest that time. There's maybe one or two companies that form and last within that time period. And that leads up into that 08-09 again global financial crisis where again, the entire real estate industry just gets decimated. And it's not until about 2012-13 the next batch of companies comes to life with you know the CompStak, VTS's of the world that start to come out of that industry and those most of those disappeared. There are very very few companies from that era that even still exist today. So the vast majority of them kind of died out and by that point the real estate industry is already completely siloed. You have these part of the target operating models of real estate. Some are owner operators some just own but don't operate. And even if I operate it might be in one geography. So I have all these different systems that again I never thought of it or planned on it just happened and I had these different geographies and different asset classes with different environments and a lot of siloed tools to do whatever the best-in-class for that market. Then as these platforms started to consolidate also these problems are inherited. So even if I wanted to consolidate, I already have a very messy data environment. And it's not just a system problem, meaning I might have two different accounting systems, more likely 10 and some of our other clients or valuation tools, things like that. I have different definitions. So the meaning of words in different parts of the company mean very different things. I have an ontology problem. I have a master data management problem. So there's a lot of challenges that are inherent. And we took the position unlike a lot of other companies which maybe said we'll create a standard and everybody's going to do what we do. We never thought that was a real thing. I don't believe in that concept. It sounds pretty on paper but you can't tell people what to do. Surely not an industry like real estate where there's not, you know, open to that. So we took a position that we inherit the world as it is and it's our job to work within that world and try and make it a better place. No cliche. There's a great line of Maggie Gyllenhaal and the deuce where like nothing's ever cliche if it's real. And I really like it. So yeah, it's cliche, but it's also real. And I think we can change the world here and I think we've been doing a really good job with that over the years.

Clara (9:06): Yeah. No, absolutely. And I agree that you've been doing a really good job over the years as well. I've taken a look at the stats as well. Totally agree with you on that. So you were talking a little bit, you know, between the years 1999, 2001, 2003, 2008, 9, 12. Where do you think that we stand now? Do you think that this is a good time for real estate investment?

L.D. (9:29): I think it's always a good time generally for everything, right? I think it's a question of when for specific markets or specific opportunities. I think it's never been a better time to invest in technology or there's never been a better time to generate technology firms. So to be a builder right now is a special time. It's really amazing. You can do things that were just impossible before. You can iterate fast. You can create things fast. Most of the barriers that you might have had earlier around cost or skill set, a lot of them have gone away, which is pretty remarkable. So I think that's just incredible. I think it also means that from a real estate standpoint, from the operators, investors, managers, it means that they get to try a lot of things that would have taken a lot more time to do. So now it's scary to them, but they get to play with a lot more things than they could have in the past. I think real estate investing is a very lucrative market right now. It's a core asset and even in office, you know, somewhat challenged because we're trying to figure out what is you we have a lot of extra supply compared to the demand since co this new normal quote unquote looks at around maybe 50% occupancy. That's kind of the new normal and it's going to take some time to reach new equilibrium as you know more supply comes out of the market and the demand picks up. But if you looked at new office towers, they're exploding. So there's a huge premium on new office towers. A lot of new towers being built right now. Boston Properties just inked a new one. Related's building a new one over at Hudson Yards. These are massive, beautiful new towers. Trophy assets in general, like if you look at the MetLife building right over Grand Central. So the big law firms there signed some massive new contracts. The Salesforce tower here next to Bryant Park. So, if you went to the premium corridors, you know, One Vanderbilt or Hudson Yards, those are all the high 200 per square foot record prices. If you look at that Midtown corridor, you know, those record, you know, the top buildings, those are all in the low 200s, really doing well. But then, you know, there's a lot of C-class office buildings, low B's that nobody wants, right? Some of those will be converted to residential if you can. Although it's not as easy as people think. You know, you take this massive floor print floor plan and or floor plate like, "Okay, this will work," and in the middle there's no light. So, I don't know what to do. So, they cut these squares in the middle with a quarter. It's doable. And some do. I think it was like 10 millionish square feet in New York City on in plan for this year to be converted, which is like a record number. And last year was like four or five, which is pretty high number. Not the record, but pretty close. So I think there's a lot of supply coming out of the market which will help but there's a lot of just [ __ ] buildings that nobody wants and I don't know that we have a solution for that. Some will become I think very experiential type places. There's a lot of you know clubs coming up which is an interesting kind of phenomena. It's a new asset class to a certain degree. Retail is going through an evolution as we speak. Retail surprisingly is actually doing way better than people appreciate. Retail's been like the horrible stepchild for a long time in real estate just because way too much built square foot or square footage more than anywhere else in the world by far. Like Canada is number two but like a long way away from New York from US but even that's kind of doing better right now because people are looking for things to do and if you can provide a good experience those are working well. Grocery anchored some of those are doing well. I'm a lot more surprised and encouraged by the way retail is performing right now than what people would expect. I think office has its challenges in this reality.

Clara (12:52): Yeah. No, definitely. Fun fact, I do own an office space in Brazil and it's one of the [ __ ] ones. So, you know, I've had such a a difficult time renting it out. Now, shout out Vicio who just rented it out from me. It was a tough one. And but I did have a question for you about the types of clients and where they're from for Cherre. Are they mainly, you know, US-based New York or is it pretty global?

L.D. (13:25): We're completely global. I would say we started in the US. I don't know if New York specifically, but in the US, but our clients are almost all multiasset class, multi-geography. So, I would say in the early days it was more in the US and taking a with so based out of you know, US, North America with assets all around the world. But, today I would say our clients are everywhere. Europe, EMEA, Asia is big part of our world, but they still say, you know, North America is still a bigger chunk of our client just because that's where we started. But, it's no longer the case that new clients are predominantly US-based.

Clara (13:56): That's good. That's good and exciting. It's good to have global clients. Jumping into another more data focused question. So, you talk a lot about the difference between data accumulation and outcome focused data utilization. What does that mean in practice for the companies that you work with?

L.D. (14:14): So there's this saying that I really hate which is uh data is the new oil. I really hate that statement because data is not oil in any way. Oil is worth money as you take it out of the ground. You just take it out of the ground. It's worth money as is before you even refined it. And it's actually worth money even without taking out the ground. You can literally buy oil in the ground. It's worth money. You don't even have to buy. You can just buy a plot of land that you think might have oil and you can get drilling rights for it. Right. Surely as you take it from drilling rights all the way through different refinements, you know, from heavy mazoots to like light crude oil, it has different values associated with it. But data is the exact opposite. It's more like sand. Meaning it sucks. Like I come back from the beach, it's stuck in my flipflop, you know, it's in my bathing suit and I have to take a shower. It sucks. Like nobody wants sand, right? And it literally costs money to keep the data as it is. It surely costs money to get data into a position where I could use it and I don't even know which of the sand is usable in the first place. Right? So there's a bet on which sand is valuable. So it's a very different analogy in my opinion and I think the more data available the more daunting it is to try and figure out what parts of those data can become glass quote unquote and where do I invest in those processes. And I tend to lean into places where we see pattern recognition from other realms where I know that if I was able to do X, if this was true, I'll be able to do X or the other way around that allows us to be a little more focused on things. So I'll give you some examples. If I knew that I was able to accurately collect the inputs going into a model, whatever those are, and I was able to test those inputs in a reliable manner over time, and that even if I don't know the result is a better result, but I'm just able to do that, I know that that will create some type of foundation for being able to better predict results down the road. Another one might be just operational type things. So, I know that I'm already reporting on certain things to the market manually in a certain way. So, what if I just took that process of collecting that sand that's being collected manually? I already know they can create glass, but I'm collecting it manually. How about I just create a process of collecting that that existing sand a little better? What if I found a way that for the same cost I could collect a little more sand around it, right? Or something like that. Well, now I have a little better environment to work with. So, I think about it that way and I think that's a framework to better test the question of is it worth collecting this data because data is expensive. Is it worth storing this data? Is it worth investing in the process of exploring what can be generated out of this data? And the good news is there are a lot of better tools today to do that last part. So given the amount of given data, can I explore some things around it? Not a lot of great tools to go back to well how do I go and play with the data? Because you'll hear this a lot of well maybe we can skip all these stated LLMs. We don't have to collect and resolve the data. We just throw the LLMs and they work on top of data. It doesn't work that way. Like if you asked the question like you know what's the performance of office buildings in New York City I would have to ask how do you define New York City? How do you define office? Is it based on you know what if you have a mixed-use building? What's the threshold for reporting? Do you count ones that you have a joint venture or not? Like the questions become really iterative and you're not going to be able to have a consistent answer in LLM. It's not going to depend on your definition as an organization. You ask the same question in two years with a different model you'll get a different answer. That's not quality results you would expect in an organization. So I think they're good tools, but they have to be applied in a very different manner and they have to be applied on infrastructure. It's very different than what I think sometimes people perceive.

Clara (17:53): Going back to, you know, your ideas, your entrepreneurial side when you first started pitching this idea to the real estate firms, what was the reaction? Did people get it? Did you get a lot of backlash?

L.D. (18:05): No, it wasn't very good. But I mean I guess I think all founders have to try and figure out how to kind of give their message in a better way. We had a very strong intuition that real estate would look very similar to financial services and it would look very similar to what we see today. And we wrote this long form memo around how data was going to be really influential and be the fuel to this analogy you're talking about to be able to power types of platforms. And we saw high frequency trading funds as a very good example of that pipeline just data-driven decision. And when we went around originally pitching that to asset managers and investors they were very cynical. Either they didn't believe us which I think is a fair general sentiment towards founders. But even those that believe that could potentially you know that we you know they had credibility of us as founders they didn't believe that world mattered. So the investors would tell you that it's all a relationship business, and that, you know, there's no way they'll ever find anything that they don't already know about. The brokers are feeding them the deals. And if you ask an asset manager, they say, "We're already super optimized. There was a way to save money or do things better, we would have already, you know, done it." And they were very, very cynical. So I think more so than other industries for sure. So I think it was very challenging in the early days and it was a very much an uphill battle. It's also the case that our business is dependent on a network of other data vendors and application vendors to come together and work in a way which adds more value to the ecosystem. Meaning that no one vendor alone would lose money by working with Cherre and that everybody who chooses to work with Cherre would make more on their own and everybody makes money together by creating a bigger pie. And today that's obvious but it wasn't obvious back then. The feedback we received was are you going to try and steal our data? Are you going to you know share it with our competitors? It was very very hard to convince people or even if it wasn't nefariously, it's just going to cannibalize our existing business. And by the way, those are all legitimate fears, right? I don't want to belittle them, but all those things happen in the industry here. Like people like to be critical about CoStar's litigious, CoStar's one of the biggest real estate data companies and people like to be very critical of their litigious position. It's also the case that a lot of people tried to steal or actually stole data from them over the years. I mean, court cases have proven. So if I'm sitting in their shoes and somebody like me comes over and pitches idea I could totally see why they say, "Going buddy you're not the first person to come with an idea thank you how about you right?" So I can definitely see why the skepticism exists but I think once we started showing in baby steps with our clients and with them how we can make it work and make money for them and I think once we started showing them that we're creating a rising tides you know rise of all boats rise all ships kind of situation I think everybody started falling in line and today it's the case that most of the vendors they work with come you know the vast majority come to us and look for distribution rather than the other way around so it's been a very interesting journey there.

Clara (21:02): Yeah, definitely and it's always the baby steps right so one client at a time and then it starts networking and it creates a snowball of people that have tried the service liked it next. I appreciate you sharing that with me now. Cherre now powers the management of over $3.3 trillion in assets globally. At that scale, what does data integrity actually look like? And what happens when it breaks?

L.D. (21:27): I think we're like 3.7 or eight right now or something like that. I think even more. I think we'll be above four this quarter already.

Clara (21:32): Congrats.

L.D. (21:32): Thank you. It's massive. Think of everything as a giant knowledge graph. So everything connects together. And that means we understand things like owners and parcels and buildings and addresses and how all these things kind of connect in an intricate manner. The first question is are these things even connected? Right? So it's an entity resolution problem which is not an easy problem to solve. Is this building the same you know is this building in one system the same building in another? Is this building today the same building that was there 5 years ago? Right? Very it's not just a now point in time. There's a temporal aspect to it as well. So that's the first challenge we have is just connecting the data to each other, which is very very complicated and then when your question around data integrity I think of that as a very complex question was a simple question which is did it come out of the place that wherever it originated to where it arrived in a normal manner. I mean did I did something get lost along the way you know think of like a trucking analogy, the truck left did it lose some tomatoes along the way right? Did the tomato that left factory A arrive at the store the same tomato right? Forget the number of tomato, was it the same tomato or did it rot on the way right? So there's a basic integrity of just the pipeline itself you know bringing the stuff across the truck then there's another question of is that what I was expecting meaning I bought tomatoes of some type of species of tomatoes that's probably not the right word I ordered San Marzano and I got something else right? Is it what I expected right? I think of those as business validation rules meaning is it the thing I expected or is it the size I expected. I ordered a certain size but not too big and not too small. Were they in that range? Right? That's a second layer of data integrity. We typically call the first one observability meaning is it what I'm looking at what's going on. The second we typically call business validation rules. But then there's another one which is we talked about it briefly earlier which is you know the ontology. Is the language I'm using here the same language that I mean when I say San Marzano and you say do we even mean the same thing? Because otherwise we're not it doesn't matter if we measure we don't even mean the same thing. And then all the things I just mentioned change over time and they drift. And I have to make sure that I'm that I'm tracking all those things. We handle all of those. And then some others that I'm not going to get too deep into because even if you think about a model, I ask a model, the model changes over time. Do you see model drift? The test drift. There are just so many places where things change. The governance in my mind is a very fluid topic. It's never a point in time and it's never an answer. Whatever question you ask, we always ask oursel what is the answer based on who you know from the person who's asking the question the way they define and we think within that organization within a point of time for a specific purpose that's very different. You and I might have very different answers I'll give you a great one. I can look at the same knowledge graph as you are, you're a debt investor and I do tenant rep so we're looking at the graph and we're trying to unmask the owners, the asset will say, you know, 123 LLC. I don't know what that means. But let's say we have a way to find out what that means. Now, because I'm the retail person, I don't care if it's JP Morgan, JP Morgan Chase. I just want to find out if it's a bank, you know, putting there on the corner. But you may say that really matters to me because I want to know if it's JP Morgan opportunity fund one on that debt or it's JP Morgan, debt fund 2 because the specific entity is the one that might foreclose on that property, get foreclosed on. So even the same graph with the same theory theoretical you and I might have different definitions of how we want to ask the question within the same organization on the same data right that's a more complex question to answer and it's also more exciting one to be able to answer that.

Clara (25:07): Yeah, definitely and I just had a question in my mind so how many data engineers do you guys have at Cherre?

L.D. (25:18): I don't know about data engineers specifically I say our product engineering data science kind of team is in you know the 40s. It's another interesting question of what do we need today in this world? I think a lot of the things that we thought we would need to do manually maybe even six months ago we find that we can do today in a lot more automated fashion. So I think we could probably do a lot more you know we could scale a lot more with these engineers than we did in the past. We still have to hire but I think a lot less compared to the work.

Clara (25:48): It's definitely a great time for technology like you were saying in the beginning. But now let's talk about the AI the real conversation. Cherre just launched Agent.Studio, a platform that lets real estate organizations build and deploy AI agents on top of their own connected data. What's the big vision behind this?

L.D. (26:08): So it's a vision we had for a long time which is the next iteration always of having your data connected and governed and trusted is putting into a business context. And one place of putting into a business context is the classical dashboards, reporting, things like that. I've never liked that part of the world for many reasons, but also because nobody cares about dashboards, right? Something it's a point in time. It's not something that is ongoing. Decisions are made based on. I think generally I'd like to see either some form of automation or some form of augmentation. Meaning either I automate a process and I'm saving money or getting speed or getting I'm able to do more things. I'm able to look at more investments, something along those lines or it's augmentation meaning whatever thing I was doing I'm smarter about the way I can do that those are very different processes and it used to be the case that we would focus more around automation I think today we can focus more on augmentation. Agent.Studio is a natural continuation so if I have all my data connected then it connects to I don't know how many data sources beyond Cherre whatever you want, I want to maybe automate some of those processes there so maybe I want to do variance analysis so budget versus actual I have a monthly budget. It's already connected through Cherre. I have all the actual statements coming in. I just want to ask myself something along the lines of is it, you know, what does the variance? What's the budget versus actual look like? And instead of having a whole army of people go and check that one by one, I might just have an agent that said, show me all your past reports. I already have the budgets and actuals going back from here. Just show me what you've done and I can go back and learn, right? You know the old Arthur Samuel example from you know the mid-50s of the definition of machine learning which is give me the result and give me the data I'll figure out the rules right I don't need to tell me the rules and let me look at that as context in a context graph and say okay I see that you know every single month you do this thing where you take the rent the actual performance of your rent and you benchmark it versus the market and you tend to use this green street metric and then you say you know I'm above or below and you try and explain why you're above or below awesome I see do that every month when I submit it. So, how about what I do is let me try and answer that on my own. Let me take a look and say, "Well, I already know that green state benchmark. That's just data." You don't have to go look for it. Let me just put that in there and say, "Oh, you're 3% above market." And let me try and explain. I see you've explained all these different ways. Sometimes you explain, "Well, it's market driven because whatever." Sometimes you explain, well, the market went down, but our peers went down more. So, I looked at all these different kind of arc types and now I say, "Let me try and test them all and see what happens, right?" Oh, I can see now that what happened was, you know, the market went down by 1%, but you just stayed flat. That's a good thing. So, I'll say, hey, market went down by 1%. We are zero even though we projected plus one. So, we're below our projections by 1%. But we're above the market by one. That's the end of the context. You decide if you want that to go automated or do you want to just have somebody push a button? Those are really good agents that actually function in the market. And we have, I don't know, maybe 300ish agents overall that can be deployed off the shelf. But you can build anything you want. One of the things that we really like is the ability that our clients can build on their own. But also consulting partners can build. You know, consulting partners are just way more creative than we will ever be cuz they're in front of clients all day long with the hardest problems and the weirdest problems that we will never think of on our own. And they can take the system and say, "Oh, I don't care what you thought you want to do with Cherre. Here's something cool that we can do with it." And kind of find some interesting ideas. So for us, it's a great distribution channel. They're really good partners of ours. They can kind of tell us, hey, you know, it'd be great if your system could do and here's another use case we can make or usually we can accommodate. But they're able to really get some good use cases out of it. Our smartest clients tend to want to build with us and not just go and build on their own. Meaning they want to try and push those platforms as far as they can take them and we really love to do that with them.

Clara (29:51): Yeah, definitely. I like the correlation between augmentation and automation. I think they go very hand-in-hand and then when you were talking about the consultants, in talking a little bit more about my side of things and hiring and hiring for a data company as tech accelerator is and data sharp group as well we're all data companies on it. I love hiring consultants. It's just such a fun fun position to hire for because you can really pick their brain and see how they can solve problems. So yeah, definitely agree that they are big assets for the company. Now, what does the data infrastructure a company needs to look like before AI actually works for them? What's the foundation besides maybe augmentation?

L.D. (30:35): I think connected data to core is really important. So having the core data systems that I use in the day-to-day in a single place connected to each other with a single ontology and meaning across the board that's the bare minimum. It really is the bare minimum though, right? On top of that, I have to have some type of control mechanism such that that stays over time and that I have some type of compliance mechanism that I can go back in time and say that that's true. It's the hardest part, no doubt. So that's really the first step. Without that, nothing works. That's just a reality. And I think firms have this vision of, oh, I'll just slap some fun AI, sprinkle dust on top of it, and it will work. They're going to get eaten alive if by those who build really strong systems. I think it's true across all industries candidly. But it's especially true in real estate.

Clara (31:32): Yeah. No, definitely. And we were saying a we were talking a little bit more about AI and how it's been changing every month, every day, every second of every day. It's hard to predict what's going to happen with it, but what's a prediction you have about AI and real estate that most people in the room would push back on right now?

L.D. (31:51): I don't know. I'm not really good at being clairvoyant. I've been selling this AI and real estate vision now for, you know, 9 10 years and people told me I was crazy for the last nine. So, I think the vast majority of real estate operations will be non-human-based in the very near future. I think that if you ask the practitioner, they'll tell you that they have, you know, 10 years before they're automated out and that their relationship business is really important and that their judgment really matters and AI will never fill in the blank. I think these people are in fantasy land. These are the same people who would tell you that driverless cars weren't coming, right? It happens way faster than people appreciate. First you have people sitting in cars with a lot of telemetry and then at some point you let the car do some stuff on its own. So maybe it can only drive on a straight line on its own on a dedicated lane but then everywhere else you have a human with telemetry and then you start letting it do a little more stuff but then a human intervene. So at some point you have Waymo that just drives completely on its own. We're seeing the same thing in code right now. So you can look at platforms like Claude and Codex maybe Cursor if it's not a fake human model underneath. And who you ask. Same thing. First you have humans writing, being tracked, you know, autocomplete. Then you have the model filling some stuff and humans correcting and debugging until it's a completely autonomous platform. Real estate's way simpler than driving or than writing code. And if anybody thinks that this is going to be kept alive by an army of people that couldn't get hired in tech, that's not a real thing. The people at the top are very smart, but the industry is laden with mediocracy and a lot of those people are going to be gone. And I don't think they appreciate that. And I think it would behoove those at the top there who are those really smart people who really do understand where the value lies to stop thinking about this is something that's going to be slow and happen eventually, creates another buzz cycle and realize this is an existential threat to their business. If they don't do something immediate in the immediate future, they're just not going to survive today.

Clara (33:49): On that positive note and um do you have anything you want to share with the audience with me? Anything else today?

L.D. (33:55): To me, this is the most exciting time to be in the industry too, right? It's like one way of looking at is being very negative and saying, "Hey, you know, it's destroying your world." If I'm sitting in our client's shoes, especially in our client's shoes, but also outside of our client base, it's probably the most exciting time to be in the industry. For the last 200 years, nothing's changed in real estate essentially. Maybe some financing innovation, not that some building technology, right? But this is the most exciting time that it's ever been to be part of it. It means that firms can do things they couldn't do before. And that's really exciting. And if they lean in, that means that they have a way to create an angle that their peers can't. It means that whatever it is that they do well today, they can double down on in ways that are just not possible before. They can do way more with hiring less people. It means they can save cost, right? They can operate their assets at higher margins at lower cost. I would argue that those firms that are trying to do something good, this is the best time to be part of it. It's the best time to be a founder in a space. It's the best time to be an operator or investors. I think it's a very exciting period of time. I would look at it as positive, not a negative thing. Unless they choose for it to be negative.

Clara (35:05): Yeah, absolutely. Well, L.D., this has been exactly the type of conversation [clears throat] I wanted to have with you today. It was so honest, so sharp, no fluff. So, I really appreciate your time and thank you so much for being here with me in Conversations in Data.