hey Sean welcome to the channel thanks for the invite glad to be here super excited to have another founder uh that is on the channel as well and love to dive into your product and challenges you see in commercial real estate but before that let's hear a little bit about your background I got my start actually in astrophysics being an idealistic um teenager I wanted to go and do space stuff because space is cool I ended up pivoting for grad school into uh remote sensing so Earth observation uh working with Earth Earth observation satellite data and G is it it sounds like a big jump but it's actually not uh because if you're working with astrophysics data you're working with these uh satellite telescopes like Hubble Space Telescope um now the James web would be a good example even though that wasn't up yet when I was in school if you just have your satellites backwards and instead of taking pictures of space they're taking pictures of the earth well then you have remote sensing so it seemed like actually a natural fit came out of that ended up doing consulting work in geospatial data science did a stint in the Mobility data industry so that's the industry where they have uh that anonymized cell phone data that gives you foot traffic information uh similar to uh companies like Placer placer.ai are kind of in that space then uh more recently I was deciding what I want to do next and I decided it was time to start a start a startup sh decided it was time to start a startup which is seems like an easy thing to say but it's actually a very difficult thing to pursue before we get into the challenges of how starting a company is uh let's first cover what are some of the challenges in commercial real estate that you were looking to tackle I know they're mostly residential Real Estate Investors um and I'm more on the uh focusing on the commercial real estate side uh but I think a lot of the principles carry over one of the so I guess the inspiration for it came out of the work that I was doing at the Mobility data company uh where we were seeing we were selling our data to commercial real estate companies um some of our best clients were commercial real estate firms um relatively big names and the thing that always struck me as odd was the data wasn't being used across the organization it was being used by a small team of maybe three analysts that sat in some centralized office and served products to something like 50,000 Brokers we always had this concern where if one of those analysts left the company we would lose them as a client because we we didn't actually have any expansive coverage in the company we were just selling our data to this tiny team that was serving an incredibly large group and the thing that I was realizing was that if we could get the data products or the actual insights or analytics closer to the business use cases instead of selling them these sort of raw ingredients that are data we could sell across the organization instead of to this tiny analytics team uh so that was largely the inspiration for um jumping into this is trying to bridge that gap between what's happening in Tech and what's being used in commercial real estate taking a step back there was a large amount of data that had to do with uh movement possibly like cell phone data as well and that was not clean data so there only could be access really through data professionals which they then would drive analytics from it and insights That Was Then used by a way larger population so there's this middle person in between um and what you were seeing as a bottleneck was that the end user wasn't able to actually access the data without this intermediary team yeah I think that's a really good summary an issue that we see a lot in Tech when working with various verticals or various Industries is the tech companies they try to be very horizontally focused uh instead of vertical so they're selling like the the raw ingredients that you could use to make dinner and that would be you know this large amounts of data um this would be database Technologies this would be um you know fun um foundational AI models or llms and then what the people in the industry actually need is they need something that's either you know ready to eat or use in their use case or at the very least something that's more like a blue apron where it's giving you the ingredients and it tells you how to kind of put them together and it also like I've seen this throughout other Industries as well where when you only have one person or a couple people who have insight into the data but maybe not as much knowledge about the industry itself the whole process is pretty slow so it's like the data vendor gives a team data then the data team is now asking product managers or other folks what they want to see um but then there's like this Loop this constant Loop that goes on sometimes these projects can take anywhere from 3 6 12 months to complete so our first project um with a client when we started Uzu data we were working with a professional uh sort of brokerage advisory uh firm and that firm had a client who was a broker and that broker had an end client who was a national chain and so every single thing we did uh would go through these stages and so it would you know check with the our client and then he would check with the broker and then maybe if it passed the broker's idea it would check with the end client um and as you can imagine the project uh went much much longer than we had anticipated for actually wrapping it up simply because of that difficulty in the communication Loop yeah that makes sense and taking a step back why is like location data even informative for commercial real estate gez that's an interesting question because the commercial real estate is really all about location but historically speaking uh most of the actual mechanisms around commercial real estate have been done locally or they've been done by people who know the market I I actually borrowed this from someone that I met at uh a recent commercial real estate event called icsc he's sort of one of my my mentors in the space but he was saying that uh you know if a tech company comes and says hey I have this model it's going to make decisions about where you should put a store or you know where you should invest in property and it's going to do it way better than you and you have some guy you know some kind of older guy named Carl who's just been living in that uh market for you know his entire life and he's been doing brokerage for like 20 years uh and the question was who would you bet on would you bet on the model or would you bet on the the guy who has the local expertise and understanding and the sort of consensus is most people would bet on Carl every single time and that's because you know models um well there are a lot of reasons for that models tend to be a little inflexible they tend to need training they tend to make mistakes while they're figuring out how to do things and need to be adjusted and retrained and then also they need to be explainable and auditable by people um one of the sort of founding principles of U data is that we don't build things that make decisions for you we just build things that give you the pieces to make decisions better that's powerful basically providing all of the information analytics modeling so that you can actually drive a decision and a lot of folks they get really nervous about Ai and automation especially in real estate and how it plays a role without removing them entirely and that's one of the greatest Parts about real estate is that it is so hyperlocal that I think there's always still like a human element that's needed so automating everything end to end is probably not always the best case to just make a single decision um where humans and their ability to understand area is still very important yeah I mean I completely agree I think um not just the human understanding of the area but also the human's ability to build relationships manage the mechanisms that need to be done to make things move smoothly um and then also interpret what what they're getting out of the models uh one of the things that I've been saying lately is that commercial real estate and people who get into commercial real estate it's all driven by relationships Brokers are fantastic at maintaining these relationships over their lifetime working with people helping understand what they need helping them get us through a very cumbers process because of the influx of data and technology and the need to make better decisions with data more and more of these people in commercial real estate roles have been forced to become more like data analysts they're spending their days hunched over spreadsheets instead of building outbuilding those relationships that one of the things that we do is you know getting more of those processes automated and leveraging Ai and Technology to get rid of the spreadsheet drudgery or make that part of it automated or easier so that broker and so that people in commercial real estate can spend more time doing the part they're good at which is building the relationships interpreting these things for the clients um and and giving them options refocusing their shift on what is actually going to drive their revenue as well and build relationships so we've talked a little bit about the current problems commercial real estate faces now I expect in the year 2024 that there's probably some solutions out there where are they falling short for what the industry needs and I know we've been jumping around a bit on the topics U these are topics that I think about a lot and so I tend to go off on a little bit of tangents as far as what the industry needs and and what's out there right now there are a lot of scalable Solutions so there are a lot of things where they are a platform um that people can log into and it'll help them walk through some sort of analysis or some sort of work so great examples of this are things like placer.ai they do it really really well for a lot of things uh it's why they're ubiquitous in commercial real estate right now um there companies like sites USA has a really good platform but there are still use cases that these platforms don't meet they meet a lot of use cases if you're doing a very structured analysis that is uh that essentially everybody does but if you have custom inputs if you need to use first-party data if you need to bring in your own stores Revenue data if if you need customization to the way that your company does things um platforms get a little bit more difficult to use and we also all have platform fatigue and no one wants another platform in their life to log into one of the reasons that we've been leaning a little bit more on custom Solutions is the customization required is actually not that heavy to customize it to a company's individual use case and one of the things that's coming with AI now that AI can write some code is you can actually build AI agents to do small levels of customization for you you give an example of that not a very good one right now we we're kind of betting on the fact that um in the next year or two that's going to get better we've done put different versions or or chunks of code into when uh the way that we design things we design them modularly so that we have different chunks of code that can be modified or enabled uh through an AI agent SAS companies have always steered away from customization in their products because it seems like an expense it seems like they'd have to maintain multiple versions of their products that rule is probably going to go away in the near future here uh which is important for commercial real estate because in commercial real estate we have sort of an inverse Network effect if you're familiar with the concept of network effects the more people that you have in your network the more people you have using the platform the more valuable it is that's how Facebook became big right the more people that are using Facebook the more people that are using LinkedIn the more valuable being on that platform is uh in commercial real estate if you have a useful tool it's it's like selling it to an investor the more people have ever using it in some ways the less valuable it is because it's no longer a Competitive Edge something that we've been uh kind of exploring and diving into a bit on our camp since we're more so the the B Toc as to how we restrict access at what levels because if you have 100 plus investors all have access to the same tool for a small Market there may not be enough deal flow but that's a whole other case you have now understood different problems in commercial real estate why customization is really important in the space Also to not just try to solve it with another platform because that is a fatigue for a lot of individuals so where do you actually get the light bulb by DN say I'm going to found my own company to solve this problem a lot of that was discovered after starting USU data um through discussions in commercial real estate we started out more General than that we were just providing geospatial Services um and and solutions uh and Commercial Real Estate was one of the areas that we were targeting and we only recently uh decided to fully pivot into commercial real estate if if we were to go back and say well what is the light bulb idea or what is the the sort of key Insight that led you here um for me it would be understanding how underutilized location intelligence and geospatial analytics are because every industry that's location based so that'd be you know commercial real estate residential real estate um retail you have some GIS function so this is a person with a title that's related to GIS it's um an acronym for geographic information systems uh they often are a very isolated department they use a lot of ezri tools um and they build these one-off reports those don't go very far in in the organization they don't get reused again there there's there's very limited value in that that's actually created despite the gis professionals doing a lot of work so that the main driver was um realizing that there's a lot of potential for geospatial insights to go further than they are so when it comes to geospatial analysis are we talking about analyzing an area based on a county a zip code how detailed does it actually go yeah so it depends a lot on the analysis being done uh so often you will need to do an analysis at the national level where you are looking for markets that a company may want to enter or a uh markets that uh May match a certain demographic criteria area and then usually once you've identified a market we're looking at some Metro region once you're within that market then you need to identify submarkets um depending on some criteria uh so it it varies a lot um you need to be able to scale between them is usually what happens and what are the challenges that these companies face as they're bringing in technology because real estate is one of the later domains to be able to really utilize AI Automation and Tech uh what are challenges that you see when you try to propose some solutions here so that's part of the reason I wanted to jump into commercial real estate even though my background there is relatively limited uh is they've traditionally lagged behind they've traditionally been the last uh sort of group to the table to get access to Innovations in technology and when we were pivoting to commercial real estate a lot of people sort of implied that I shouldn't and they said things like people in commercial real estate don't want technology they said you know Brokers and brokerages they're all Lites they want to do things with a paper and pencil they don't want anything technological and personally I haven't really found that to be true um the more people I talk to the more I discover that there's actually a disconnect between what tech companies are producing and the actual Roi generating use cases that people in commercial real estate need to solve for and so there's a lot of that uh I gave the example earlier of a blackbox model that kind of does something for you which obviously no one wants um it's a little bit like that I've been kind of calling it Tech arrogance which as a tech person I probably should not be saying but um that seems to happen a lot where they're building a solution and they're telling people in commercial real estate this is for you this will make your job better you should use this without really understanding the job that the people in commercial real estate are doing and when you get close to their actual use cases or needs your technology gets adopted super quickly um I think you know I mentioned placer.ai
earlier I think they're actually a good example of this because they got their technology solving and generating reports and getting very close to users needs for specific use cases and now every company in commercial real estate has a Placer subscription so it's not that they're slow to adopt technology it's just that the technology being offered up most of the time just doesn't actually do what they need it to do it doesn't it doesn't solve an actual problem and this actually brings me back a little bit from earlier in my career when we were implementing models of how to identify fraud in the trading markets and there was met with a lot of resistance of oh like I know how to analyze the data if I go into a spreadsheet like if you're telling me that someone's 95% chance of committing fraud like that's still not enough I need more information so we met them Halfway by automating a lot of like the spreadsheet work they were doing as well as having different insights and dashboards based on like what their current workflow was so that when now introducing modeling and and scoring for fraud potential like individuals it was like a progression of all these little steps that made them more comfortable to adopt that technology but also the technology itself now had a lot more context to it rather than just here's the score and just go by whatever the score that very blackbox you were giving them the pieces and the information and letting them make the decision or at least letting them audit the decision before they pulled the trigger that's exactly what's needed from AI like everyone thinks that we want AI to make decisions for us um but we actually don't we want to make the decisions we just want AI to do the research for us and give us you know the options that are presented in a trustworthy and understandable way how do you see it going forward with GE type of company but also just generally in the industry like how much of an impact do you think it will have in the next five years per se well I mean I think we've got a lot of people that are hyping to Heights that may not be justified I think we've also got a lot of people that are uh disparaging a because they've been disillusioned for whatever reason um I think realistically the the real value is in integrating AI with other data and tools and machine learning and building it into these workflows so that it adds tangible value the commercial real estate companies that are using AI uh will have a faster deal flow one thing that I've heard from people in commercial real estate over and over again is that in this industry speed is everything any company in commercial real estate that has a speed advantage generated by AI or um automation or machine learning will be able to pull ahead of their competitors because they will close more deals per year per year per year yes having that information at their fingertips doing research faster how would someone today do the process of analysis for location manually and then how do they use your company to do it automatically we generally build custom workflow solutions for that an example is a client that we're working with right now so they have a manual process where when they're working with their end client which will usually be a retail brand looking for a location uh when they're working with their end client they will take the existing stores of that client they'll do customer segmentation they'll look at Revenue um and they'll build up this model and then based on that they'll look at a number of locations in a market that that brand wants to enter and say of these locations we think these ones match best uh what will generate revenue or what will give you the customer base that you need to be successful this is generally called a market suitability analysis there are a lot of different terms for it though the terms are not very standardized in commercial real estate but that's the one I'm going with this process all in all could take them upwards of you know a few days to a week of work going between the Brokers their GIS team um and you know pulling all the data manually we built them a workflow automation that goes through and once they they just add the locations they can run it in two minutes and get back the answers so it's a pretty quick change but the part that I I I would say I'm almost most proud of is at the end uh and we're still building this out we actually layer in uh some AI automation we layer an AI to create an explanation of what's happening in the workflow we have an explanation characterizing the Brand's customers uh that's all written out as talking points so that their brokers can explain the analysis to the End customer now here's a report you don't really know how we got there um it explains every step of the way in a very understandable way wow that provides so much value right not just saying this is our recommendation but providing the charts the explainability and so that person is empowered to now be the expert off of the data that was reviewed um to their own clients which is very very useful and one one that I hope to see in the AI space in general that as there's more solutions there's a lot more explainability for users as well I'm uh very opposed to the whole hey we're smart trust us approach to an analysis if you understand an analysis then you should be able to explain it and in our case we also trained an AI to explain the analysis which makes things even easier because then they run it they don't have to reference with us they can actually get the explanation even though it's different than the last time it was run well with that Sean I want to do a slight pivot over from you as a founder and it's not easy to start up a company there's a lot of pivots that also occur so has there been a time as You' start of use the data where you have shifted your priorities or change uh the future trajectory of what you plan to do all the time um I I mean we're we're even doing it right now so we have this problem where we're split between two relatively good product offerings that we have one is the sort of automations of these workflows uh the other is mostly coming from conversations with people in commercial real estate we actually built an AI tool that parses leases uh so in commercial real estate unlike residential the leases can be enormous we're talking 100 Pages 200 Pages uh oftentimes it's in this photocopied PDF document that uh you know that the text is completely skewed on one page it's faded on another uh there are handwritten notes in it and so we we you know Dove deep into the world of large language models to build a system and train a system to parse and understand and answer questions about those kinds of lease documents in a process that's generally known as lease abstracting and uh we've had good results we've so far gotten up to where we can parse and handle at least up to 130 pages without any errors uh it does create kind of this fragmentation because we're doing all these geospatial analyses and we've also got this AI lease product and both of them are very compelling to us right now we kind of doing that old western thing where you're riding two horses and you have like one foot on each horse and they're like kind of diverging giant of focus is uh once you dive into one problem you realize that there are many others and it's exciting right to see what can be solved utilizing technology or communication but of course there's always a give and take you have to understand how to prioritize Sean what do you see for the future of proptech and Commercial Real Estate I see obviously a lot more AI adoption is going to happen right now we have a situation where a lot of the platform companies are kind of just jamming AI into their platform um but I think we're going to see more and more AI first products where they they generate more value I think we can go back to the one that I said earlier where you know in commercial real estate speed is everything so commercial real estate companies that are embracing technology and are embracing AI from a levelheaded and and realistic standpoint are going to really pull ahead just because they're going to have a faster deal flow so Sean what is your daily habit yeah so I spend a lot of time kicking things in addition to being a a tech person startup founder I'm a martial artist and so I find that um just getting a good workout working on Kick combos usually puts me in a better mood and and mental state to get things done that's excellent I love being able to work out or even go for a walk sometimes like a math problem or some coding problem that I had I was stuck on part of the day just like it's unlocked it's solved just by doing uh a quick walk so that's great that you get some energy out one of your favorite books I'm skeptical of Love of business books um I I think that a lot of them are either bad advice or it's you know a short idea that they flushed out into a book for no reason when it could have been a blog post uh there are also some that I've read where it seems to just be a long advertisement for someone's consulting company which I'm not really a fan of but as far as actual authors go uh I I do really like Derek cers uh he writes about life and about business I find that he he has kind of a kind of an experimental approach um to every aspect of his life uh that reminds me sort of of Richard Fineman where you just kind of do things without really worrying about whether it's the way it's traditionally done and sometimes you get different results than people get when they do things in the traditional ways yeah he's um I think he's got one called uh I think it's anything you want that is a good one to start with usually it's pretty short we'll definitely check it out and do you have any pieces of advice for those getting into Tech real estate technology yeah I I mentioned earlier that I tend to think advice general advice Ed isn't always useful because even if you're giving someone good advice it might be at the wrong time so they might need to be in a different phase of their life to receive that advice I will go kind of meta and say that a good piece of advice is to only give advice when asked um in terms of recreation I I'm a martial ARS I'm also a swing dancer and so I learn a lot of very complicated dance moves and one of the absolute worst things is when you're in a dance class and you're working with someone and you're trying to figure out this move and they stop and start giving you unsolicited verbal advice while you're still trying to figure out how to get your foot to do this thing so yeah give give people advice when asked but only then I agree with that um also it kind of applies to like language too when you're like learning how to speak a new language and someone's like you know it's good to like throw in some words here and there to help like build a vocabulary but just like stopping the entire FL was not as useful or if someone yeah exactly if you're halfway through a sentence that you're trying to piece together and your brain's just trying to get the puzzle pieces in the right order and they start correcting your pronunciation that completely stops that process and yeah I've started explaining to people in various domains you know I'm I'm not I'm not to the point of refining yet I'm just getting the steps still and that sometimes helps that sometimes helps that's a good point so for those who want advice or want to learn more about use a data how would they reach out to you yeah so I'm pretty active on LinkedIn if you search for sha Knight It's s an k n g HT uh you can also search for Yuzu data on LinkedIn uh or you can uh just email me I'm Sean Sean Yuzu dat.com that's yzd excellent well thank you so much for your time Sean super excited to see how you continue to impact a commercial real estate space and I'm sure we'll have you back appreciate it thanks for so much for having me on the podcast
2024-08-17