In this Triple Win LIVE recording we bring you a panel discussion on the current state of AI property management. Our panel of experts, Wolfgang Croskey (The Perfect Tenant), Tom McGarry (Second Nature), and Ray Hespen (Property Meld) share their insights on how you can leverage AI to improve your property management business.
Throughout this episode, you’ll learn what's working right now, what's next, and what might be tougher to implement. We also dive into real-world use cases including listings, content marketing, screening, recommendation engines, leasing, maintenance, and more.
If you're a property manager looking to stay ahead of the game and maximize your business's potential, this is the podcast for you.
Join us on March 15, 2023, for a live event: The Modern Property Management Tech Stack
Register for TWLX on May 16, 2023, specifically for Broker Owners.
Got questions? Email us.
Join our private Facebook group just for professional property managers.
Be sure to follow the Triple Win Property Management podcast by Second Nature and never miss an episode!
Hosted by Andrew Smallwood and Laura Mac
Featuring Wolfgang Croskey, Tom McGarry, Ray Hespen
Produced by Andrew Smallwood, Laura Mac, and Carol Housel
Edited by Isaac Balachandran
Related: Check out the other property management podcasts we recommend for single-family property managers.
Laura Mac
Hello professional Property Managers. Welcome to the Triple Win Property Management Podcast. My name is Laura Mac and I am your co-host today. We have a great episode for you. Our Triple Win events are typically recorded live on Zoom in front of an audience, and for this episode we had our biggest turn out to date where a couple hundred PMs plus some joined us live to talk about AI and Property Management. So, really excited to listen to this episode. We dive in deep, so in the beginning you might feel a little out of your depth but just hang in there. We very quickly get to some practical applications that you can apply today in your Property Management company, and we actually got some feedback that we needed more. So you guys wanted more, We delivered. We got on a follow up call with Wolfgang Croskey, who's one of our panelists and spoke to him for over an hour on practical applications of AI and Property Management. So that is being recorded and produced as a follow up podcast episode that will be released on March the 9th. So be sure to look out for that one. So without further ado, please enjoy this conversation with Andrew Smallwood. Ray Hespen, Tom McGarry and Wolfgang Croskey. Enjoy!
Andrew Smallwood
My name is Andrew Smallwood with Second Nature and we've got Laura Mac here, who is co-host of Triple Win Live, where we say there are no rules, the points don't matter or something like that. We don't actually have a tagline, but we just like to get properly mangers together, to talk about topics that are interesting here, create some value for each other, put in a little effort to organize, bring some expert panelists in and just create as much valuable content and connection and collaboration as we can for you guys. We’ll have a couple exciting announcements at the end of the call about upcoming events and opportunities for you to continue to participate and keep the conversation going. But we are here today to talk about AI in Property Management, AI in action, and I’m very excited to introduce our expert panel that we have here. So with that, here's what I'm going to do. I'm going to bring up Tom McGarry to the Zoom stage first, Laura, Second Nature CTO and Tom, if you don’t mind just sharing with folks who haven’t had a chance to meet you, a little bit about yourself.
Tom McGarry
Great to meet everybody. My name is Tom McGarry. As Andrew said, I'm chief technology officer at Second Nature. I've been doing big software for 30 years. Company building for 20 years, big software engineering, started in a desktop, moved over to big web applications, enterprise at one point as well. So built, a lot, done a lot, and recently in the past 10 years really data has been sort of my mantra, and what I realized through the years is that we built big systems, but we never took care of the data. Data came as an afterthought and data strategy came in much later and was always a disaster and was also always very difficult to implement after the fact. So I've been on the soapbox to make sure you have a data strategy. I've been involved in data and much more in the engineering side and the plumbing side, bridge building side. But just recently I've obviously taken an interest in artificial intelligence, machine learning, and really in an academic sense for the time being. But I'm very excited about bringing it into bringing data science into our own ecosystem and leveraging it where we can, mostly at this point, predictive modeling, really see where we are being reactionary, historical data mining, really to see where we can look forward, try to predict things before and try to hit the market before things really move. So really, really excited. I hope to share some technical aspects of not too technical, hopefully, but, you know, shed some light on some things that, you know, that aren't completely apparent, so I’m excited to be a part of this. And thank you very much.
Andrew Smallwood
Awesome. All right, Tom, thanks for that. And coming in from Charleston, South Carolina, area as well, I know we've got a couple of folks on here with Paul Meadows probably, somewhere nearby. Next up is Ray Hespen CEO, co-founder of Property Meld, and Winston Churchill with him as well I think so there's Ray. Ray, for those who don't know you, do you mind giving a little introduction about who you are, who Property Meld is?
Ray Hespen
So I'm going to go and baseline, set expectations right here, I’m the least qualified panelist in the room. So just everything that comes out of my mouth is probably going to sound a little less intelligent than the peers that are on this panel, but super thrilled to be part of this. As a lot of you guys know, I co-founded a software company called Property Meld. We specialize in maintenance coordination. My background, oddly, I’m a degreed mining engineer, not super relevant to real estate, but the thing that I've always done is I've worked in operations prior to coming in this business, and so I've learned the superpower of what automation and data can do in decision-making and lean margin businesses. So I'd like to think that we've taken some of that and moved that over into maintenance. And the thing that I love much, like Tom, I'm a huge data fiend in this business and I recognize the power that it brings. So a lot of you guys probably saw on LinkedIn I tend to surface a lot of Property Meld data, so I'm super excited to talk about AI, the mechanism that ultimately gets to use some of this information today. So thanks for having me on Andrew.
Andrew Smallwood
Yeah, and Ray, I think for probably the benefit of mental health amongst other things, I don't think you're on Facebook the way you used to be anymore, but Ray's a great follow on LinkedIn. I'd highly recommend everybody on this call. Go find Ray. I love following him on LinkedIn and a lot of the insights he regularly shares from the unique data set that Property Meld has, so, I just encourage folks to go do that. Okay. Hey, that means we've got Wolfgang himself. Bring him up on stage, please. Wolfgang Croskey there he is.
Wolfgang Croskey
Good afternoon, everybody. Wolfgang Croskey here, our family has a real estate company in the San Francisco Bay Area. And honestly, I'm probably the least qualified in here when it comes to this tech in everything. So I'm really going to be your translator. I'm going to turn AI to toilets. So that’s kind of the goal today and how we can use AI for dealing with those things that we have in Property Management, in real estate, but really excited to be part of this, not only to share, but to learn and to look at how we can use AI much greater than just asking ChatGPT some stupid questions and seeing what it spits out. But for me, why I have a passion for the automation and AI in tech is we're a small independent company and you see these large franchises and these VC-backed behemoths and nothing against that model. It's just for me, technology is the great equalizer, and it allows us to compete with these nationwide companies and to provide not only the same level of service but to be able to pivot and adapt much quicker than those larger companies can. So for me, you're a smaller company, AI, Automation Tech is that equalizer that's going to allow you to shine just as well as these larger companies.
Andrew Smallwood
Awesome, all right. Well, hey, that gives you a little bit of an introduction. With that said, we're going to get this kicked off. And, you know, Tom, I'll come to you. But I was telling these guys, my goal here is to throw some questions out and just let these guys chat for a little bit, you know? But I think a good place to start with all three of you is, you know, context around what frameworks or definitions do you feel like are important so that we can all kind of have a similar understanding? You know, here when I hear AI, I feel like for me, you know, it's meant a lot of different things or a lot of things have been called AI. And so something I've always wanted is a little more clarity on how we might be defining that in this conversation, you know, where that can go. I know Wolf was saying not just ChatGPT, right? It's other things here. So things like AI, machine learning, you know, or what’s training, inference or, you know, what are these different key terms we may need to help us through the conversation? So, Tom, I'll let you get us kicked off here and then we'll go to Ray and Wolf.
Tom McGarry
Yes. So real quick, Ray, I'm a classically trained aerospace engineer and material scientist so we got to figure out, get together, see what went wrong there. But let's start with some basics, right? So just to get some definitions out so try to separate the learning world into two aspects. So there’s machine learning and then there's what's called artificial intelligence. Let's put that to the side. So machine learning, let me just I'm going to try to keep it simple. Statistics on steroids. A lot of this is stuff that a lot of taking stats courses, even the simplest courses, you would be familiar with. So I'll spare you the details of what's supervised versus unsupervised, for example, and all the buzzwords. But if you think about it, machine learning is using processes, statistical methods to deal with larges amounts of data. And there's basically two let's say two things you can really do with this. One is, you probably heard of regression, linear regression. You have a bunch of data on a graph, plot a line through it, and then you figure out what you know. If you have another input, you have that particular pathway of that line. You can figure out where that target variable is going to be. And then you have other other machine learning models where you can cluster data, you can take large amounts of data, start to find patterns, right? That's more of the unsupervised versus regression. The point is machine learning it’s really advanced statistical methods with a high computing power and having to deal with high volumes of data. So put that to the side and then you have a new concept which is deep learning, they call it a lot of buzzwords, artificial intelligence and it's a little bit different. So there's no predefined algorithms or statistical processes, actually when you break it down, artificial intelligence to its bare minimum is pretty simple. Right? You have these in individuals they're called neurons, right? And formally, they're called perceptrons, I don't want to get too into the weeds. But they’re very simple, logistical little points in this structure and what you do is like I said don't go into the weeds, but each one of these nodes has a very simple job. It's either two pieces of data that are coming in and then figuring out how close it's going to get to an output at the next node. Now let's just keep it at that, right? So the whole point is that you need, in an artificial neural network is just to put billions of these together and pump gigantic training data sets through it. So what's important to notice, you have this data, and then you know what it's going to be. That's why it's training. So, you know, the output of what it should be and then, I'll spare the gory details of how this works through the network, but by assembling billions of these nodes, you could pass data through and then using statistical methods inside each one of those nodes you can approximate at each one of those points, with some sort of accuracy. So in the end, what's happening is the model in itself is very simplistic. There's no actual algorithm. There's just a set of points with a weight factor across the entire neural network that you would establish this massive set of parameters that becomes your model and you store your model and then obviously you adjust these parameters and tune these parameters to get as close to accuracy any training data possibly can. And once you have that, you have your model ready to go and then you can run data through it and see if you can approximate and find out what targets you want. Now that's very technical. I apologize if that got too much in the weeds. But the point I want to get to at a higher level is that it in itself, an artificial network is composed of very actually not very intelligent things, but assembling massive amounts of these, you can tune a very gigantic set of data to get results that you expect from a training data set. So I'll leave it at that. I just wanted to throw that out as the basics and start there. We can go eventually into what different models there are in each one of those, but that's in its basic form, those two components are the building blocks of learning, of machine-assisted and then artificial learning.
Andrew Smallwood
So I think that was great. I see people appreciating that in the chat. I also know there are some folks like me that just heard Perceptron for the first time. Yeah. So appreciate that. That's great. Yeah. Ray and Wolf, what would you add here of like how you guys think about a framework or key definitions that we should have here before we get started?
Wolfgang Croskey
I think I should've brought Coke for this meeting. I think I'll break it down for a fifth-grader. Basically just take a whole bunch of crap and language and stuff, put it in the computer, and a computer tells you how accurate it is and how fit our assumptions are and yeah, I mean, to me, that's what it is. It just takes all kinds of information and it tries to act humanlike and it tries to act upon that. And what's interesting, though, is those that have been kind of following AI, there have been things surfacing, oh, AI’s got bias. And people have. No, it doesn't have bias. It just depends on all the information that was thrown in there. So if you only throw in certain bits of information, did you create the bias or does the AI have bias? And to me, that's interesting because I've seen questions come up about, well, how's AI going to help Property Managers? I think the number one place we're going to see this real quick is tenant’s training and coming up with why or why not, you should rent to a tenant, which is probably going to make those fair housing ambulance-chasing attorneys… their minds are going to be blown because, well, the computer did it. It wasn't me, there’s no bias. So it's going to be interesting to see how it does that. But I think if I had to explain AI, actually I asked ChatGPT right now what was it a perceptron? Sounds like a Marvel movie. But basically the interesting was recognizing patterns in data, making predictions and processing large amounts of data and adjusting the algorithms just constantly. You know, that's the learning part. So I think it's just that those large data sets and having a compnay like Second Nature and Property Meld on here, who have I'm going to say they're large data sets and it's not like Google and it's new what they come up with some crazy names, not Brad it's like barred or something is their competitor. Those are but you know to look at it's basically just taking this data and making predictions off of it.
Ray Hespen
I'm going to go and live out what I originally anchored here at the beginning. I think if Tom and I got into a conversation about AI, in about 30 seconds he would school me and I'd shut down. But the best way that I heard somebody describe it and this is like the most articulate version and I think it was so well done in doing it. Where you try and help these large data sets, you train models on to make individual decision points that ultimately create an output that you're trying to go for. The best way that I ever heard ChatGPT as a thing, it was a calculator for words. So where you basically try and put in some information much like you do a normal numerical calculator, but it's a calculator for words, how do I come to this output? And so the machine is sitting there in the middle, right? Your calculator has been defined on metrics and logic and you take something like human language that is much more complex and you're basically creating a model between you and your input and what kind of output you're getting. I think that was the best way that I thought was described to me, that a Neanderthal like myself could ultimately understand. But I think the big underlying thing is, the data that's required to produce these things of any accuracy is like the foundation of all of it. I know we're going to jump into that.
Andrew Smallwood
Yeah, I think I heard all of you say the importance is the data set, the size of it, the quality, the diversity of it, right? That's going to determine what training all these, I don't know if neurons is the right word, get and then it really seems to me, maybe I'll mis-explain AI and you guys can either tell me this is right or help correct it in a way that would give people the right nuance. But, you know, I think about automation, which has been a big word in technology, right? And that's like people deciding, what should happen and then, you know, the technology is kind of like doing that. As we move into AI, it's actually starting to make proposals, suggestions, recommendations, right? And ultimately even decisions, right? And that's kind of like the distinction of what's happening here as opposed to, and I think, Tom, you made a mention of like there's supervised and unsupervised, like maybe there's more nuance than that to be a lot of like, what people tend to hear as the, you know, part of the big potential of radically changing how work is done. Is that fair, guys, or does that need correcting?
Ray Hespen
So I think the big thing that as I was kind of like writing down notes, you know, if I'm anybody sitting in this room and sitting there going like I'm seeing this brilliance of ChatGPT and I'm sitting there trying to understanding what kind of implications in my world it's going to be, I think, understanding why ChatGPT exists, and what information it has and like, what are some of the carryovers that are part of that. Now, if you guys all remember, I remember probably even five years ago and use the term automation. But I remember if you really wanted to like make a splash in the industry, it was integration, API. I'm trying to remember what the other buzzwords were, and I can tell you at that same time, if you want to go pitch somebody on why they should invest in your company, it was AIML and you had to say it in that order like buzzwords that go, All right, well, what does that mean, one then just start putting it into practice. Like APIs require work, like it's connecting points and most people just don't understand that. Right. And a lot of people do. But it's it's end points and you got to connect it over, and how does it work? Artificial intelligence. Machine learning. Awesome. That's a big buzzword. But once people get into it, it's how does it work? I think the biggest thing that's been super interesting to kind of like explore what are the implications of this in the industry is to understand like, first of all, foundationally you know, like troves and troves of clean data. Now ChatGPT basically, and I don't know what data it gets to train on, but it gets the entire Internet has tons of communication and gets to see everything that's out there. But now let's take that sort of thought process to like your business and making a leasing decision. Where is that data? How well is it store? What's the cost of not making the right decision? And I think one of the big things that I’m seeing is that we're kind of seeing the gaps of ChatGPT and I'm curious on Tom's and Wolfgang's thoughts is like we're just starting to surface massive amounts of data. There's an exciting product coming out in June and I'm sitting here going like, until you have this really nailed down, the application becomes so limited, or at least the accuracy in those Tom was walking through of that output being the right decision. So that way as Wolfgang was saying, even in a leasing conversation, if you don’t have enough information built into the models, like some of these things become very risky, that you could turn out a perfectly great prospective resident or let a terrible one in and it all comes down to what data is available to make some of the decisions in our industry, and who has it, because that's a. Mm hmm. I don't know if I'm oversimplifying it.
Tom McGarry
No, but just jumping in here, I want to make a couple of distinctions about the applications and things. We have ChatGPT. Obviously, it's really based on a natural language processing model, right? That said, you know, terabytes upon terabytes upon terabytes of text data, volumes of books, etc., like basically textual information. And it's really it's just a sequentially modeled way to take an input from a previous word and figure out what the next word should be. And just so you know, this application is really for, at this point in time, language, obviously, you can figure out all the different applications you could use for that like we use now, like obviously chat and content filtering and content generation, whatnot. But then there’s other aspects of AI and machine learning where like I said before, predictive capabilities and I think obviously like Wolfgang said, screening is a huge one. But then there's other applications of you could take data that you might that be very valuable, leasing data, housing data, and then you can use that to make predictions across all different types, what are called features. And you know, most of what's called numbers like parameters, variables, and in the AI world they're called features across, for example, geographic markets. You might want to spot patterns in there within different models of neural networks for image recognition that we know very well. Obviously we know that it's used for facial recognition and image recognition and whatnot to do different applications. There's the operational side of things where you could do something like chat and natural language processing. But there's also this analytical capability, which is where I want to take a lot of this inside Second Nature. And to use that to be ahead of the market, to analyze patterns in this data and use that to be more intelligent. An intelligent approach to hit the market. And I think that separating these contexts of applications of AI in the right context is probably where the biggest challenges. There's a plentitude of models and free models and free frameworks. I didn't really want to get into the specifics of frameworks yet because that's very Propellerhead ish but there’s plenty of different tools that you can use, but applying them to the right context is one of the biggest challenges, besides data curation and data hygiene and data quality. So I just wanted to sort of throw that out there. The topic point.
Andrew Smallwood
I’d love to throw this on its way to Wolf, which is maybe in the chat, even if there's folks on this call who have started to use tools, you know, whether it's ChatGPT or Jasper. I know there's a number of tools for content generation, whether it be for listings or for SEO content to attract owners or you know, I'd love to hear like where are people using some of this now and our panel even asked, they were curious about this We'd love to hear that or see that in the chat, you know, where you might be using it in using it now or considering using it next. And Wolf, what do you see there as far as potential practical applications and/or where it's being used now? I know you predicted screening earlier. What else do you see?
Wolfgang Croskey
So I don’t know where this speech was taken, and quite honestly, now seeing what AI can do, I don't know if he really said it. I don't trust anything on the Internet anymore, but supposedly the gentleman in this video is the current CEO of OpenAI or ChatGPT, I guess. He's a current CEO. And he says, you know, when AI was coming out, we expected AI to replace all of the like, blue collar jobs and then worked its way up and that the last ones to be replaced would be the creative type jobs and professions. And what they're finding is it's the complete opposite that AI has pretty much put out of business the creatives. Literally using Synthesia, Dub Masters, you literally can create a phenomenal video all from, you know, AI it looks like a real person. They're talking, they have movements, they have tone of voice, etc. of all from your computer. As far as realistic applications that you can do right now and you should be doing, if you are still wasting your time trying to draft up creative Property Marketing descriptions, you need to stop. Nobody reads them anyway. It’s like, I'll buy you lunch if you actually leased a Property because somebody says, You know what, That marketing description was phenomenal. That's why I want to lease this house. Never said anybody that so why spend a lot of time on it? Put it into ChatGPT. I have a four bedroom, three bath house, 745 Railroad Avenue, give it three descriptors and tell it to keep it under 300 words and you're done. So how can you automate that? So we're a lead simple user, we love it. And so I’m building a zap right now that when we click a task it's going to take some fields in lead simple, the bedroom count, bathroom count and a new field that we call three characteristics. It then goes to ChatGPT, it creates this phenomenal description, brings it back as a note and then now can go out into the marketing. And why waste your time with the social media stuff? You can have it create a Instagram caption with proper hashtags and emoji, and a Facebook. So if you are if you're spending time on that, just stop. Nobody reads them anyways, but your owners require it. It's got to be done. So the marketing descriptions, fliers, in my spare time I run the Pittsburgh Chamber of Commerce and we're doing something with manufacturers and it's called advanced manufacturing. What is that? So put it into AI and it came out with next generation manufacturing. Okay, so we asked it to define that. Then we told it to write a blog post, then, because we're saying this to kids, we asked it to do a ten question, multiple choice tests and then we asked it to translate all of that in Spanish. And then we asked it to create a radio ad and it even told us when to put in sound effects and all these different things. So the creative side, I would really use AI for that right now as a Property Manager, the next thing that I would use AI for is your other content for blogs and things. I saw a couple comments about, well, you know, is it copyrighted? Who owns it? I will tell you this out of my own experiment, I put into ChatGPT the same prompt ten different times and got ten different answers, somewhat similar, but crafted differently. From my previous life as an educator, I still have access to some plagiarism tools. Threw it in there and it passed every time. So it is genuine content. There are definitely similarities. Like if you ask it to draft up an email for you, it's always going to say, I hope this email finds you well. I don't know. That's it. That's the line that it likes. But if you're struggling, creating stuff, you're using a tool like lead simple, whatever that has the email templates and you know what it wants to say, but you're just not mighty at word. Put it in there. Have it create the template for you. Even if you don't like it 100%, it's better than starting from scratch. So now you can build your email templates. If you have an email that comes in and you're not quite sure how you would respond, give it a try. You can tell it ‘how would you respond to this email, but with like an uplifting tone’ and it'll go from there. I hope at one point you know, using inbox that the AI will be able to… the email will come in and it will say, you know what, normally this email should go to Carlos since they're the maintenance coordinator. But this person sounds really upset and it just automatically escalates the mail to me because it detected that this person was upset rather than getting lost in the back and forth. So I think that the communication and marketing piece is a place you can start with right now. There are free tools and then there are some paid tools. But that's where I would start today.
Ray Hespen
Wolfgang I think that's super interesting. And I think the biggest thing that I heard from over on your end was it's taking some of the things where it's not necessarily decision making and kind of an aspect. It's more like content curation that maybe somebody might outsource to a, you know, a fiber or somewhere there and further reducing some of the cost, maybe enhancing the speed of kind of the return. I think as we've been looking at maintenance like in general in the application of AI, by the way, we've got to an engineering manager here that was so mad that he's not on here. I said, first of all you’ve ruined the podcast because he would… if you think propeller head was not a word, it would be a word by the time we're done. He's amazing. But, you know, the thing that's interesting about AI is like, what is the cost of making the wrong decision? And you know, he was a former AWS Amazon employee. And, you know, that's one of the things, even if you get a 96% right, what's the cost of getting it wrong 4% of the time? And you have to like really weigh those risks. And sometimes AI can be very complex, if it just makes a wrong decision, what's the impact of that? So we've been thinking about maintenance, and is maintenance coordination going to go away? All that sort of stuff. You know, we're just starting to tap into the world that is understanding the data of the real world and creating a digital twin to some extent, which is basically we have data that properly reflects what's happening in the real world. And then once you do that, you can start to understand what are those interesting correlations, what levers people can move to ultimately make an outcome that they want, And then you can get to this point where I think the idea is the distance between even where we're at a Property Meld, which I think we're the ones who got the biggest headstart on this. The difference between that and ultimately making a decision, if a resident submits a maintenance request and like a leaky toilet, the ability to understand if it’s water leaking on the ground or leaking in the tank and those implications and automatically run everything, schedules everything, we're so far out. But where it feasibly can get in the more short term in terms of maintenance is what it's called like. And Tom, correct me if I'm wrong is AI assisted decision making where it's not making your decision, much like Wolfgang is talking about you input something, you get an output, you're accepting the decision, your control+C, control+V somewhere. But like the cost of getting wrong on that is you know, you say in regards or whatever, at the top of the end, the risk is light. But when it starts to become when it's maintenance and it's somebody home and it's potential $20,000 and there's potential litigation involved and all that sort of stuff, this becomes really important. So the concept of where data can go and once you get clean data and you can essentially train outcomes and all this, you can do what's called AI assisted decision making, which is basically when that request comes in, maybe it's suggesting a next step, but a human being is backing it up. It's not like… you still need a maintenance coordinator that knows what a toilet is and knows all that's happening. But when it says, Hey, do you want to send somebody out as an emergency repair or somebody not, we're going to suggest this because of the information and the AI is suggesting it, but it's not actually making that decision. And I will tell you, even from that element, I think it's going to take years to get it very robust to where it is injected in every aspect of the maintenance process. Just as a context from what we're seeing on data and what we're seeing in the industry.
Wolfgang Croskey
So, Ray, I mean, do you think you'll get to the point where AI will say, okay, your properties are located in Northern California and based on our data set, we recommend that these properties in this area get this maintenance done based on our data, that we will make those recommendations or predictions.
Ray Hespen
Yeah, and I think Tom really kind of set it out earlier. The difference between AI and machine learning like the ability to take large data sets like, you know, Property Meld captures what's happening in terms of repair when it's scheduled, how happy the resident is, the cost when it happened, what location. And so you can start to do some of those things. And if you're using some of our Property care plus preventative programs, it can start mapping out who's doing what kinds of programs and what's the impact, ultimately having all the maintenance costs of the unit or the frequency of repairs. So I absolutely think what you're talking about, Wolfgang, is being able to make suggestive recommendations for preventative programs and some of the other things. And sort of that machine learning is very real. But I think having somebody replace them and make the decision for is the I think that's the big difference in such a leap away from where we're currently are, from a data aspect.
Tom McGarry
Ray to jump in. You said recommendations, I’m just going to add a little bit more technical layers. Like what you're referring to, right, is a recommendation system, right? So we were all familiar with Netflix and Amazon and clobber filtering, but this is where you take this recommendation systems to the next level. It's sort of an advanced topic in deep learning where, you know, you just don't use it for like shopping, for example, and getting, you know, presented the next best thing. This is where you can use practical application reccomendation systems in real world operational context like Ray was mentioning. And if you want to do more research and look it up reccomendation systems and obviously you might get in the weeds as far as technical aspects of it, but in the non-technical, this is where we're going to go and assisted decision making is going to be leveraging heavily, the evolutions recommendation system. So that's going to be on the horizon and the forefront and along with natural language processing and GPT by OpenAI.
Andrew Smallwood
So before we move to something different here, I just want to open it up. Like I feel like we've heard a couple of things of, hey, people have been hearing about AI and specifically AI in real estate and specifically in rental real estate and Property Management for a long time. I thik there’sa number of people who feel like there have been some like broken promises, I think about like looking back on it with what you've shared now, I think about vendors who are getting started and saying like, we're bringing AI to this, but they have no customers, and they have no data set, right? And so it's like, how much value can you drive without that sizable data set is, I feel like an important insight. You know, that was shared here and will continue to be a challenge and a barrier for folks moving forward. You know, I like what you said about, you know, hey, considering the risks, you know, the small tail risk and impact of that and where's the ticky tack maintenance of resetting a GFI outlet or, ressetting a disposal, you know, or some of these things that are quality of life. It's not that they're unimportant, you know, but, you know, they don't have the gravity of making the wrong predictive decision of like changing out in HVC unit on, you know, how that's going to impact residents or investors or other folks about helping Property Managers on this call think about what's going to be fast to move in this direction, what's going to be slow to move in this direction? What am I looking for and paying attention so that I understand, you know, my strategy and the decisions I'm making, I feel confident as a business owner? I feel like we've given some good context in addition to practical applications here, where I want to go next and if there's anything else you guys want to cover, just ignore anything I'm saying and talk about whatever you want, right?
Wolfgang Croskey
Bacon, you said bacon.
Andrew Smallwood
We got to make our way to bacon at some point here. You know, I was talking to a couple folks before, you know, this event and what they'd want to, you know, learn about and think about. I think a core question for this audience is what's your opinion? And we'll call it that. It's an opinion. What's your opinion on it? Is this good for professional Property Managers like the ones who are on this call. Right. Is what's happening going to be good for a self-managing landlord? We talk about this gap between the professional and the accidental landlord. Are there forces at play here that shrink this gap? Are there opportunities that actually expand this gap? Is the answer both, right, and dependent on something else? I'd love to just, you know, hear your guys’ opinion. Wolf, if I'd like to start with you, if you don't mind, and we'd love to get a take from each of you.
Wolfgang Croskey
Being in Property Management, you always think the worst of people because everybody's a liar that walks through the front door till he proves innosence. So, for me the self-managing and the smaller I think the biggest fear, it goes back to the data set, right. He who controls the data rules the world. So when you have these large national trade organisations for the real estate industry that sell out to private companies, they have the control of the data set, right. You know, the MLS, all these different things. That is the data set. So if they then sell their data to companies that then can create great AI and now make it available to everybody, it really reduces what I can offer to differentiate myself. And it's going to come down to which that Second Nature is known for that customer, that client experience. So sure AI is going to get better. It's going to allow the DIY landlord to do more on their own. I remember, you know, like 15 years ago the differentiator was, well, we can run your tenants credit. Now, anybody can check a tenant's credit, that's no big deal. So those tech things, those tools are those things that you thought were so great. It's going to come down to one thing. How do you make people feel? And those that feel good, are going to write you checks. If you don't make them feel good, you're not going to get a check. The tech tools you have or you know, everybody's going to have access to them. But my fear is some of these national associations are going to sell their data set because that's their most valuable asset. And especially with some of the competition laws out there, the seeling of the data is going to really make it harder for smaller companies to just be average. You're going to have to be great if you want to stay in business.
Ray Hespen
I think one of the things that I have to imagine now, I wasn't there, but when the calculator first came out, I have to imagine that was considered cheating by a lot of people. And it was like, but why haven't used the… whatever the thing with the wooden dowels and you move the… abacus and everybody was sitting there going, This is witchcraft. And you know, just abacus. As you learn the foundational elements of what we're doing and that's what makes you exceptional. There's value in still understanding fundamental math. But I think that the thing that AI is doing to the value chain is moving just like the calculator, it's like and, you know, I saw some comments and some notes about people being able to write papers, maybe the value chain of human beings is not about writing papers anymore, and that's not the value that you're doing. And we consider in that concept, is it good or is it good for the industry? Is probably a better question asked. Will it go there regardless? And the idea and thinking through real estate, I can really move down to two different things; predictable NOI and do you hang onto your residents like those are probably it. And so then you can go back and go, Can AI help me improve those two things? And if it can, boy, is it going to happen. If it can't, it probably won't unless the reward function of real estate changes somehow. And so I think AI in Property Management is probably going find it’s ways in what Wolfgang was talking about, which is what sort of things that you normally hire a fiver you got a virtual assistant somewhere or anything like that. I think more complexity making a decision about how to do a repair or whether to repair or replace the things or make other decisions, there's value in there. But are any of these tools going to be really beneficial by AI and machine learning that are going to just adjust those two metrics? And if they do, I would say it's not, Is it good for the industry? It's just, How do I need get involved in this, because ultimately my investors care about that.
Tom McGarry
I’ll just throw in something here, it's more like philosophizing about, you know, what's happening with AI is what you guys have been mentioning. It's the commoditization of technology, right? The democratization of technology, putting it in the hands of many. This has been an age old problem with technology and it forces competitions, it actually liberates a lot of companies to do what they do best, just taking care of their customers. Right. So, it's going to force competition, force you to do what you do best and puts you at a better advantage which is your core mission. And the democratization of technology in other people's hands. It's going to be inevitable fact and like ChatGPT would it if you think about it, well, it's going to put all the content writers out of business. But is it, you know, like so this is where it's going to force in that particular industry some real reevaluation of what your differentiation is. So that's just an example what ChatGPT could do. But all these different examples you're mentioning, it's just going to force all the tech companies to become more competitive and also, like I said, alleviate and liberate them to do what they do best.
Ray Hespen
So can I say one other thing on there and I'm sorry for just jumping in here, you know. I was doing this talk at our user summit and the thing is this talking about the value chain moving. And if you look at the cars like in the seventies, there was a built-in light to basically work on your vehicle. That was the value, right? They were adding value there. And then eventually it moved on to where you had a reliable vehicle and that was like the Honda. And then the value chain moves to where it's like, you know, it's safe and then the value chain moves and it's driving itself like you tend to commoditize. Nobody tends to win like all alone, to commoditize it. Like Tom said, it forces you to move. So I think for us, making sure we're not putting all of our eggs in the basket of saying we're just really good at reading a prospective resident's application, that's probably not a very defensible place where you're ultimately going to have a lot of value. And so continuing to move down the value chain of putting more emphasis, I'm a big one on maintenance, I'm a little biased, but it's got to be one of those other things that are harder to do that these are going to take off and potentially become pretty even in the industry and keep moving because I think that's ultimately the long term defense ability and it does free it up. But I think hanging on to that personal touch is a great one that I've talked to a lot of people about. You can hang on to it, but if it's not where the value is for people and you're hanging onto it, it's going to be a problem. And I'm not saying personal touch is not it, that’s not my point some people really want to have a phone call to talk to somebody, that exists for a certain place, not 100% of the calls.
Andrew Smallwood
Yeah I think, great points, and I feel like there's definitely been a theme of people have listened to a lot of different podcast episodes or part of events or just a thread through a lot of the conversation in this industry is, occasionally checking in and thinking about what's being commoditized or what's being more easily duplicated, right, in Property Management, right? And what are the unique strengths, skills, capabilities that I can create in my company that are going to be durable and valuable, not things that people like, right, but that people are also willing to pay for. You know, I would take to Ray’s last point, you know, okay, I like it. But am I also willing to pay, you know, so I can build a business on. And so creating those experiences, people may want different experiences and there's opportunities to segment on different customers. But, it's interesting to think about what's going to happen to the rate of change in certain areas, as it relates to commoditization and the barrier. And just thinking about in the workforce at large, you know, attendance based compensation. Like when I think about roles like, you know, the bellhop or whatever it might beor maybe that's not even the best example, somebody who's there to stand and greet or whatever it might be or you know. Well, if the goal is just to be there to pull a lever, you know, or just sitthere, a parking lot attendant. Those are the kind of jobs like a parking lot attendant that gets replaced right by technology, you know, more quickly than other jobs where making a difference is more scarce. Right. And how do we create that kind of value? It's going to be interesting to see how that happens in real estate here.
Wolfgang Croskey
I think it's also like, how do we use this tech? I'm seeing some comments, you know, from people that are concerned, like I'm barely holding on to, you know, what I'm learning now. One of the things that we tried to when I was in middle school teacher type teach the kids that the skills of the 22nd century is the ability to learn, unlearn and relearn as quickly as possible, because that's you know, something that you learned five years ago is probably not as useful. You know, now it's being able to learn and unlearn and relearn the new is the skill that you need to have. And so it's how can I use these tools, whether it's AI, machine learning, zapier, lead simple, Property Meld, whatever it may be, how do I use those so that I can free up time, so that I can focus on those items that have moved with the value chain, as Ray was saying. So like one example for us, you know, what are things that we can have the computer do and then what are we going to do with that free time? So one thing that we've built into our what we call our weekly process is because we've been able to have the computer do more, is we now have a Property Manager and our client success coordinator they each; so the Property Manager does owners, client success coordinator on tenants, they call one a day just to say, Hey, what's up? They now have that time to do that becase we've been able to have the computer do other things and because if the only relationship you have with your owners is only around, something's broken and I need money, it's not going to be a very healthy relationship. Right? They're like, oh, my gosh, they’re calling again, something must be broken or they need a check. But to now change it where your Property Managers call, Hey, how's it going? You know, they can look at the notes to see how the past conversation went and just to check in and see how life is going. We've been able to get listings from that, new sales, lots of different things, because now we have more time to focus on higher value tasks because now we've unloaded things to the computer. For example, I brought up the Property descriptions, we went around our office, we figured out that people were spending between 30 and 45 minutes to write what they felt was a really good Property description. We've now saved that time, right? Because what the computer spits out, it's good enough and sometimes it's better. I think it's how do we use these tools? Not so much to have bells and whistles, but to create time that we can use and what we feel are in higher value ways.
Ray Hespen
So one of the things that I think in a delineation, is AI is a how. I think identifying what you're going to move is really important. So I released a LinkedIn post about what we're calling our ladder of maintenance. It's kind of a play on Maslow's ladder of the hierarchy of needs but its maintenance. And so the idea behind it is, I think starting out with saying what is the metric in your business you need to move the most to distill the most prudent conversation, not how do I inject AI into my business. And if you're sitting there saying, I want to move this particular metric or this particular number like, how can I do that now? Now, if your metric to move is the amount of time you spend on mundane tasks, like that's a great one. Wolfgang gave some great examples of that. Then you can sit there and say AI makes a lot of sense. I still think we're at a point where businesses getting more crystal clear on their KPIs, what metrics they want to move, the how in the AI being able to move that, is still a little ways off, in my humble opinion. I still think that that's the direction of the decision making of how to inject AI is really still starting with the what, what you're going to move.
Andrew Smallwood
That's great. Well, guys, here's what I want to do just with timing that we've got left. Call it like a parting shot here or if there's anything you feel like man, it was it wasn't said and I've got 30 seconds. You know, here's some tactical practical value to end with it. Or just a thought, you know, to leave people with, you know, before we have a couple exciting announcements to make. And we'll ask you in the audience to reflect back some of their top takeaways in the chat. Hear some good feedback as well. But why don't we do this? Let's do Tom, Ray, Wolf. Quick parting shot.
Tom McGarry
No, sure. I was just going to jump in and, you know, talk about what Ray just said. Like, I've been doing that the entire time. But this is an important thing to note is that, as I said before, the application and figuring out what you're going to do with it is one of the aspects, but obviously what the really good thing about ChatGPT besides it, writing content for you and being an endless source of entertainment, right? It's starting to actually bring AI to the forefront and create this momentum that wasn't totally there before. And as Ray said it, AI in its current state, isn't totally accessible by everybody. Nice thing about ChatGPT, is it's accessible. You can go there, you can start playing with it. The end. There's no math involved. There's nothing. So other parts of AI, there is, right? We’re not totally there for it to be completely consumable by the population. But with things like GPT, momentum is going to start pushing and you're going to start seeing is this stuff being commoditized, productized to be consumable by the person who doesn't need, don't need an engineering, don't need a data scientist, you don't need to know statistics, but you do need to know your KPIs and you want to know where to get to in your business. But you have no idea how to use AI. Very soon. In the next two years, you will those products emerging and enabling us to do that. So I'm very excited about that. So that's my parting shot.
Andrew Smallwood
Well said, Tom.
Ray Hespen
I think there's kind of stages this is ultimately going to go. First thing is data collection, like, real world data collection. And once you have data, you can start doing measurements and then once you have measurements, you can start implementing automation to figure out what's moving the metric. And then we're going to move to optimization and then we're going to elimination. So I just wrote that down it’s probably really stupid, but the idea is I think we're still in data and measurement as an industry, and I think it'd probably be a little obtuse to say we're going to go to elimination out the gate. We're still learning, Property Meld has 450,000 rental units on our platform. We process around $550 million in invoices annually. We're still getting the measurements and the data to be able to even start thinking about these things. And so the idea of jumping to the end, I think we're still a ways off.
Wolfgang Croskey
I think for me, I've used this image a couple of times, but the Roomba that's spreading the dog crap all over the house. Don't worry about the AI and the zaps and all the those sexy bells and whistles. It's policies and procedures and your processes that, you know, if you don't have rock solid policies that are then put in action with your processes, all AI is going to do is screw it up even more and frustrate people. So going back to the KPIs, get your policies nailed down on all your situations and how do you do things, What's our policy on waiving a late rent fee? What's our policy on adding/removing a roommate? Nail that down first? Because if you don't have that, it doesn't matter what GPT or super neutron you have or whatever, it's not going to matter. And then from there roll up your sleeves, do the dirty work and make your process. You cannot outsource that. That is your business. Your business isn’t your lead simple template. Your business isn't your zap or your cool graphic you have on your website, your business is your processes. Without that, you have nothing. So nail those down, once you have that, then start looking at, you know, these other items. But to stay on top of where I learn about AI or all this for the us common folk down here. Zapier has a great email newsletter that isn't just a sales page. They're sharing all kinds of different things. I highly recommend you subscribe to that and start looking for people on Twitter. There's a wealth of knowledge that people are sharing and then LinkedIn, those would be the three areas. You know, you find people that seem smart. Follow them and then, you know, start going from there. But the knowledge that you need as a Property Manager is not going to come from a keynote. It's going to come from your peers and sessions like this where people are talking. But if you're feeling a little anxious right now by oh my gosh, one more thing I got to figure out. Look back. Do you have your policies and your process nailed down? If not, start there. I can tell you personally that when you don't have those, things get screwed up. So focus there first and then start looking at… and sometimes you have to do it simultaneously. You have to dual track it, but then look at the tech, the AI, and all those other things that you can start bringing in. But the reality is that you need to be able to learn, unlearn and relearn at the speed of the internet to stay successful.
Andrew Smallwood
Awesome! Guys some great stuff there. I did promise people there was going to be something we had to announce at the end today. You know, we'd love to, before you all leave, feel free to drop in the chat what you're taking away. I heard a lot of interesting things that might shift how you're thinking or your perspective today. I heard a lot of interesting things about practical applications to be put into place and things that maybe you were considering rushing into that maybe you're now feeling better about putting off and knowing what to look for. Lots of great stuff here. Really appreciate the panal. I want to express my gratitude. Thanks to all three of you for being here today and generously giving up your time and your expertise here. And the announcement that I promised to folks. In previous years we've had TWLX. If you've attended TWLX, you know, it's really like the premiere online digital event for professional Property Managers and we do things a little differently. It's not just like YouTube video after YouTube video, sitting through it, we get you actually, unlike this, you know, really into small group interaction with your peers to talk about, okay, how might we actually get some momentum in putting this in place? So we weren't prepared to facilitate that today, but we've got TWLX coming and we're actually, based on feedback, going to have a broker owner only option. It's going to be May 16th. We are going to have an option later in the year where the entire team can come in as well, but we just heard so much from people on broker owner day of the two day event, how being able to have conversation with their peers at that level was great as it was great also talking to people from diverse parts of an organization for different kinds of conversations. So, really excited, May 16th, that's going to be happening. This will be a fundraiser, 100% of proceeds going to NARPM’s charity, Communities in Schools is the name of the charity, if I remember it correctly. We're so excited to be a part and grateful to be a part of this amazing Property Management community and share it with you. It's really cool how this community coming together, learning together, growing together, can also make a great impact in your local communities on people, people who need it. So our goal is to raise some five figure amount. So just now every dollar and a ticket, you know, above costs is going to that excited to announce keynote speakers great panels like this one you know we'll be a part of that. Really excited to announce keynote speakers, great panels like this one. We’ll be a part of great interaction, collaboration, conversation with peers, is a part of that. You won't regret it. There's going to be a link down in the chat that opens up today and more information to come. I encourage you to sign up. Sign up today. So with that, we're right on time. Laura Mac, I think we have one more bonus free event, the modern Property Management tech stack, which is coming next month. We've already got one or two panalists for that, we are looking for one or two other panelists. So if you're here, my guess is you're probably interested in Property Management technology, maybe some expertise there you'd like to offer from the Zoom stage similar to how we did today. Feel free to just email Laura Mac about that or let us know, whatever way Laura is saying to let you know, let us know in the chat and we've got a bonus free event for everybody as well next month. Thanks again to all of our panelists. Seen a lot of appreciation in the chat with that. Happy Tuesday, keep stacking your triple wins, and we will see you next time. Take care, everybody.
That's all for today's Triple Win Property Management podcast. Thank you so much for listening. Thank you so much for sharing a piece of your life with us. We do not take it for granted. I also want to give a shout out to Carol Housel for everything she and her team does to make these possible. It's crazy to think about, over 5000 professional Property Managers have pressed play on episodes and season one and season two now, and we really want to encourage you to keep giving feedback because more and more people are listening. It's getting better and better and better thanks to everything that you're sharing with us. If you liked this enough to listen, I want to encourage you to share it with other people. You can give us feedback directly on the social media channels, Facebook, LinkedIn, wherever you're hanging out. You can also send us an email at TripleWin@SecondNature.com. And we just want to give more! There's no sales pitch here, just want to offer more resources that help you find and stack your next Triple Win and become a Triple Win driven Property Manager. So where can you find that? You can find a private Facebook group. You can find our blog, you can find our newsletters to find more resources all at RBP.SecondNature.com to search for what you're looking for there. And every time we see you, we want to see a better version of you and your business. To that end, keep it going, feel inspired. Take our encouragement and we'll see you next time.