I Replaced 100 Hours of Real Estate Work per week With This AI Strategy

Hello and welcome back to the Owner Meeting Podcast hosted by Multifamily Strategy. I'm Christian Osgood, your channel host. Guys, we're going to have a ton of new episodes that are in production right now coming out. I'm super excited. However, today we have one of my favorite topics with one of our favorite guests, Caleb Hommel. Caleb, welcome back to the channel.

Thanks for having me.

Dude, this is going to be a fun one. So AI in real estate, I've done some videos on this. We use it a ton, but this is going beyond the basics. You're using this at a very high level. I've developed some AI projects that are like real-time working in the mentorship group, making huge difference in the way that we look at real estate. I think that you are at the front end of some of the things that you can do in real estate. We're talking capital raising, relationship management, some really cool tools that you are using right now that I want to share with other people just so they even just start thinking about how to use AI differently. Caleb, what projects are you currently working on that you were most excited about?

Yeah, gosh, I mean, they all have to do with AI. Literally everything I'm excited about when my head hits the pillow at night, last thing I think about is AI. First thing right now is first thing I wake up is AI. It's— I didn't realize a lot of the monotonous stuff we do as real estate investors. Real estate's kind of obviously an older industry, especially property management. There hasn't been much integration for a long time besides certain softwares. So being able to come in even to PM, something like expansion, hey, how do you go find old properties that maybe need new management. Hey, how do you go find a property management company? Bad website, maybe looking to sell, that kind of thing. And really organizing capital lists, going back through old capital leads that may have slipped through the cracks, generating follow-up messages for these kinds of people automatically with AI are some of the main things we're working on on the real estate side of this.

And I'm absolutely loving it. How are you finding, as real estate has been a relationship game and both of our model has been build relationships with brokers, with owners, How is AI changing the relationship game for you? And is real estate still a relationship-based game?

Yeah, I don't think that's ever going to change. I think no matter how much AI gets involved, whatever, it's just going to increase the load of scamminess and really, I think, increase how much relationships really matter. Like, you see the guys sending out mailers and all this stuff. Okay, now they can automate it where they don't even have to go to the mailbox. Somebody's already had that down, but now they don't even have to do prospecting. Everything's dialed in for them. And at the same time, all these old owners who you actually want to build relationships with are now going to be even more turned off to all the mass marketing and really wanting to go on the relationship base. So AI is a great way to get your foot in the door. It doesn't replace building a relationship by any stretch of the imagination.

So talk to me about the projects that you are using right now that are on the, you know, the cutting edge of AI and real estate. What things, you know, so we're going to assume like, okay, we're still meeting with brokers, still meeting with owners, we're still finding deals the old-fashioned way, right? What things does AI replace? That has significant time savings or competitive advantage?

Oh, I think it's the, like, the grunt work and the prospecting are the number one thing. Like, even as something on, like, CRM, like, maintenance or going through, like, projects, like, making leads fall through, like, leads that have fallen through the cracks in the past to use AI instead of praying that ChatGPT, instead of being a yes man, will give you a good answer. You can actually use AIs to sync with whether it's your email, your SMS, whatever, and actually go back through your messages, pull these leads out, extract these leads out that you may have dropped the ball on, put them in a spreadsheet. You could even program it to send automated emails out for you. If you have certain SMS software, you could send out automated messages that are actually tailored to the person. So it's not, hey, Caleb, just following up on this. It's, hey, Caleb, last time we spoke, you were working on XYZ. Did you end up finishing that project or how is that going? Way more powerful on a follow-up basis as well.

This is really interesting. So you said instead of ChatGPT, what are you using instead of ChatGPT that you—

I only use pretty much just Claude and Anthropic at this point. And I'm not some AI savant by any stretch of the imagination, but how much progress I've even been able to make with that stuff, it's been night and day versus any other product we've used.

This is what I love about AI. AI will tell you how to prompt it. You can start with the end. So you go, here's my end goal. What questions should I be asking? Or how would I build this? And it'll actually teach you how to use it better. It's the beautiful thing about AI is that you actually don't need to be an AI genius. However, some of the stuff that you're doing does feel pretty genius. Can you gimme some depth into, so you mentioned you can look for messages that you've missed or people who, you know, talking like investor capital. Let's use that as our case study.

Yeah.

So we have a lot of people over the last, what, 6 years now who have reached out to me, reached out to you. Hey, I want to invest in a deal. I want to learn more. I want to participate. We now have a deal come up instead of us looking through every single past lead or every single past form beyond just copying and pasting a list into AI, how is it actually going through and helping you sort this and automate it?

That's really a question. So it depends on the platform. It depends if you're in your CRM, your iMessages, or your email. If it's on your iMessages on Anthropic or Claude, you can literally go into a certain thing on into the cowork section, enable something for SMS. You give it the right permissions, this thing can literally go through. Like today I did this on my Mac for some stuff, looking for some certain leads in the past that fall through the cracks, prompted it correctly, said, hey, at the end, create me a spreadsheet with all the information, X, Y, and Z. And 5 minutes later it has 597 people in there already in that spreadsheet pulled up with context on each and every single person ready to go. And I can follow up with those people. Now I can either have it go in and actually send these messages out for me Or I can go plug this into a CRM or send those, hey, like if they have their emails or we have their emails anywhere, hey, email this person based off our last text thread and follow up. Hey, I have an opportunity because you said X, Y, and Z.

Oh, this is really interesting. So if you're doing this as like a 506C, so if you're doing like a syndicated thing where you're allowed to market, you can literally just write an awesome marketing campaign and send it to everyone who's expressed direct interest in investing with you.

And you've got 500 people.

How long does that take AI to pull? I'm assuming this is pretty much instant.

Like, Going through all the messages, all the approvals, 5 minutes tops.

So in 5 minutes. So you're getting an average of like 100 leads a minute going through. All of them that we've not found.

If I had all the permissions enabled, it would have been like a minute or less. So yeah, it's pretty much instantaneous on actually going through this stuff, let alone even like talk about underwriting on a deal or talking about, hey, broker sends you a deal, you create something with Claude Code that actually plugs straight into your calculator. It runs all the things for you. It spits out underwriting that would have taken you hours. Now it takes the AI, what, 30 minutes? Minutes or 30 seconds to a minute, maybe tops, and just breezes through everything.

And that is absolutely amazing. It's a really fun way to do this. So, you know, again, you can use that to market. The other thing you can do is you can take that list, everyone who's expressed interest in investing. Let's assume that you're not allowed to market directly to them. You can host events, you can rekindle relationships. You can put together a list of like, hey, who do I have an existing relationship with? Who I've had multiple points of contact, who I could reasonably reach out to. And you can put together an actual hit list and you can talk to these people and you can go through the list and say, hey, do they qualify? Do they not? You can make this instead of looking through, through the years, roughly how many contacts do we have in our backend system for people who've reached out?

It's like 85 or 90K.

Okay. So around 100,000 people, we can bring that list down to the top 500 people that we need to speak to. And now I can look through a list of 500 people and say, hey, who's appropriate with the regulations of this deal and who's not? I can and make informed decisions on a list of 500 people instead of informed decisions on a list of 90,000 to 100,000 people. That's an insane difference.

Yeah. And maybe you spend $10, like you could use— there's different ones with APIs because that integrates with this too. Grok, you could use OpenAI, Anthropic, whatever in the backend. Like the cheapest one, Grok, probably costs for that many people. I don't have a ballpark, but I assume anywhere from under $100 to screen 90,000 contacts. Like it's like every time, like I was looking at something that I had ran recently, it was 1.1 million tokens and that equated to $0.28. So it's like, if you could actually get this down where you have your system down, I don't have a finite number of how much it would cost. Depends how much data and like how much into the weeds you're going with that, so to speak. But I mean, you're talking like, okay, you spend, let's say even $100 in API tokens and it saves you what, 3 weeks worth of work?

Well, and remember what we used to do 1 or 2 years ago together when we were, you and I would, but we, we've gone through a few, but usually we'd have 2 VAs on staff. You'd have one, I've had one for our various tasks and it's usually like $800 to $1,000 a month. So you're paying them like $200 a week. All of these functions now take way less time and are way better. I mean, it's kind of a bummer if you're a virtual assistant, that's, uh, it's really kind of probably going to make that job a lot harder. But as an entrepreneur doing massive amounts of tasks for 28 cents, it's crazy.

Why wouldn't you? It's like the time savings and efficiencies there, or you can literally build something like I'm working on this right now. Like, literally goes through my email, goes through, like, I mean, you get more emails than I do. You can literally build a bot that will just go through your entire email and put the top 10 most important or urgent emails that are actually from people and not just spams at the top of your to-do list. And then instead of you having to sort through hundreds of thousands of messages, like, okay, cool, I need to bang these out this evening. Awesome, I'll get it done. And you can literally just have exactly what you need to do. You could have it respond. Or another thing we do is on the property management front, on the leasing front, we have a lot of stuff on Facebook Marketplace. You can literally create a bot even in Cloud Cowork. You just program it, tell to do, it'll go on Facebook Marketplace and just respond to the leads.

I have a project I need you to help me with after this call. That is so smart. I need some help on some Washington State properties. If you can build that bot for me on that, that is, that, that, that would, I would love to see how you're doing that. But these are the ideas. Like if you think it for the most part, you can prompt AI on like, hey, how would I build this? And what have other people done similar projects? What should I be asking? Like you can write down your questions, your end goal, and have it tell you what to do to try to optimize it. You follow these strategies, this works really well. One thing that, I mean, this is insane. Mentorship. My goal, every single person who signs up for the mentorship buys property. Nothing else really at the end of the day matters. You can have the best courses, the funnest community, the biggest community. If you don't have active investors who are getting their desired result, like if you sign up, you need to buy real estate. So I want you to feel supported. Absolutely. For us to better support people, we've actually rolled out some AI things that have been awesome. A custom GPT. Which I just asked ChatGPT to help me make. Took me 2 days of just like nonstop prompting, 8,000 lines of prompt going through our methods, everything I've filmed on YouTube, everything that I've filmed in the course. It downloaded the transcripts of all of this based its personality off of that, asked a ton of questions and situational things. It now recommends deals, analyze them exactly the way I would, except it has built-in calculators. So if you're a student, Of course you have one-on-one support. Of course you have coaching. In addition to that, you can instantly, you may have this big math problem of trying to figure this out. You can take that while you're waiting to book a call with me, or you're just like, hey, I want the answer in 2 seconds. You go to GPT, you can upload the actual OM for your deal and prompt it, and it'll actually tell you what to prompt. It'll say, hey, Christian recommends these parameters. A bank's probably going to look at the deal like this. Based on your books, your offer price range is here to here. Now you have inputs and you can test all of these theories in the calculator that's provided in the course. By the time my students hop on a call with me, the difference is not that they get less coaching. It's when they hop on the same call they were scheduled for, they're 10 questions farther. Their deal is almost ready to go. And instead of me filling in like, hey, this is what you need to factor for, for landscaping, or hey, you forgot taxes are going to reassess after you close. They're coming in with a fully underwritten deal and I'm checking their assumptions and we're working on writing the offer. The difference in just, you just need the coaching space. It's, yes, it's saving me time, but more importantly, the same amount of actions with students have 10 times the output. We just came out with an app. It's my virtual mind. It literally, it's like a, you can call it, it actually has my voice. And you can go through, downloaded, I think it's 150 million words that I've said on YouTube. Wow. On the course. That's a few. On other people's podcasts actually plugged in like other podcasts too. It's everything I've said over 5 years. And same thing, multiple days of interviewing me. It got down to details of like, hey, when you bought the Robinhood Village Resort, you mentioned that you have a leasing strategy of 95/95 in property management for multifamily. Are there similar parameters that you look for in hospitality? Like, I mean, it went down to like the level of like, yeah, very precisely specific questions. My mentees, if I'm on a call with another mentee, you can call basically my mind and ask it any operating question.

Yeah, that's powerful.

Those both took 2 full days. I mean, I had to take the day off, so we're talking 4 days of my life. My mind is now in ChatGPT on a math and analytics perspective. You can ask technical questions and combine it with the search power of GPT, or you can ask it like functional or operating questions that only I would have the answer to. And those are also downloaded in another AI. They're both available in our course. You can talk to me, you can message, you can calculate, you can get deal advice. It's life-changing. 4 days. 4 days of work. That would take years to train an employee to be able to do that. Oh, absolutely. Unbelievable. The time savings and value that it's able to create. By the way, this isn't an ad for the mentorship. This is just a straight up plug for like, hey, if you're not thinking about these things for how you're communicating with your clients, how you're building your real estate and business, I'm looking at building a very similar AI for my employees of multiple companies where they can go in and like, hey, when you're stuck on this, sometimes I'm coaching, sometimes I'm at a meeting, sometimes Caleb and I are in a closing, we're signing docs. How cool to be as a property management company to have all of the data of our operations, our SOPs, our best practices, and you can literally come in and just talk to your boss when they're busy.

Well, talk about saving like employee costs as well. Like you have all these employees, like, like executive assistants. I'm like, how do you guys any future in the jobs you guys are at the moment. Not saying our own people internally, just generally. It's like, okay, you could build an AI bot that does absolutely all of this. And granted, you want to put all this stuff out there everywhere on the internet with all your passkeys to your backend, your root passwords. But at the same time, you can create something that runs locally on your machine that is password protected and everything that can literally just do all the tasks you want it to. It's custom built for you. You don't need to go try to, hey, let me grab this software for $20 a month. Let me grab this one for $200. Let me grab this one for $300. Hopefully we can spice it together and be an employee. You could just build it. Granted, it takes some finite work and there's gonna be some errors and some troubleshooting, but you could basically build an employee for the most part.

Yeah. Well, and you can build a lot of employees and what I've done is, you know, for like a media editing, for example, AI has a lot of software that can do it really quick, but it still doesn't quite replace having a person and a creative director who can come in and organize it, come up with the strategy, implement the strategy, get high quality edits, or just the person to interface with me of like, hey, can you film this, this, and this? I need this clip. It doesn't quite replace a media coordinator. What it does do is it replaces every single employee underneath them. I can build out an entire media company with one person. Our production, I can get, now I don't want to blow up my audience, so I don't do this, but I can easily produce 10 short-form pieces of content for every platform every single day, rip it out there. Do long-form YouTube. We can streamline all of these things, but your entire production company, you need someone who is a skilled photographer and a good creative mind. And if you have those two things and a decent editor, if they have those three skills, AI can do every other piece of that thing. You can have a 10-person company with one employee.

Yeah, it's life-changing as far as expansion, as far as systems tracking. I mean, even like, I think there's a Claude extension or I don't know, I think it's Claude. I haven't used it, but it already just plugs straight into your Excel sheets. Like you can actually use something where it integrates with Excel and like whatever you need it to do, run whatever data you can pull. Like you can create custom webhooks. Like literally you could create your own webhook based off your own domain and use a webhook function, whether it's Zapier or something else, to actually integrate with your own website and send webhooks. It's like there's so much you can do as far as actually with AI and what that looks like building it out.

Oh, so cool. Well, and so one thing that I don't quite understand yet, you're farther than me on this AI stuff. You mentioned having like a coworker, like coworking in Cloud, and then I've been playing with AI Agent. Is Coworking and Agents, are those similar? Are those completely different? Uh, it depends. What are those two things and how are you using them?

Yeah, so Coworking is like tasks. So like Coworking on Cloud, like let's say for that Facebook thing this morning with all the, on all the Facebook leads, you're like, hey, I need you to go on Facebook and I need you to respond to these people in this way with all the unread messages. And if there's already a conversation going, pick up where it left off. You could basically do that. And there's an integration where it can go on Google Chrome and from Google Chrome, it can then log, if you have the Facebook login up, it'll log into your Facebook. Go on your messages tab on Facebook Marketplace and tackle all those. So Coworking is more of an assistant and you have to constantly prompt it. Where Claude Code is different is you're actually building websites, you're actually building domains, you're actually building backend infrastructure, and it's going through your code and optimizing it and fixing things. So there's Chat with Claude, there's Claude Cowork, and there's Claude Code are the three different ones. It basically wipes out the whole OpenClaw thing a few weeks ago. That basically has taken a backseat because Claude basically came out with its own better version.

Interesting. So does that completely replace the AI agents that, that do the thing? You create your own agents who are doing all these things. Is the coworking basically its own agent?

Yes and no. It depends. Like if you want to have an agent doing it like automatically on their own, maybe you create it like with NotebookLM and you actually create an agent that way. But if you want to like, if you just say, I need you to go do this task, like XYZ, update these people's data in XYZ spreadsheet, it can just go do that for you. And you can actually generate another coworking session where you, let's say you have, hey, this person's doing this, this person's on Facebook, this person's on data, this person's on expansion. You can have 3 things doing different prompts. It'll let you know when you need to go access permissions for something, but you can have 3 different things firing at the same exact time, all doing different tasks.

So you can basically build your architecture. So if I'm building this as a company, and by the way, you people who are really good at AI are probably like, Christian, you're an idiot. We know all this stuff.

Oh, same here. Like there's some people who are so much like so far.

I'm super curious though. So can, if I'm doing this the way that I build a company is I'd actually have an org chart. Can I create an agent who manages or a coworker, however this works, but can I create a bot essentially create an employee who I instruct and then they instruct the employees underneath them? Like, can I interface with one?

Feasibly, yes, it should be doable. I haven't gone that far yet. It's actually built like a site that'd probably be on how I would think about it, and I'm probably wrong. There's probably some AI nerd in this like, Caleb, such an idiot. But you'd probably create like a head bot actually with NotebookLM is what I've heard is the best one to use and then actually create sub ones under them. And then obviously you'd make your own domain so you can manage and actually see progress through all of them is probably how you'd feasibly do that.

So you have a high school degree, you have like 400 rentals today. Are you 400 or 500? I forget. Did the 80 units count? 425?

Yeah, I'm 413.

413. Okay. 413. You have 413 rental units. You to run the property management company. Well, you have been helping with Facebook, but it actually might not even be Caleb anymore. And by the end of next week, how long did it take you to learn this to the level that you're at? Like from the time that you're like, hey, I'm going to learn how to implement AI in my business to now. You're 23 years old. Again, high school education, very busy individual, insanely busy individual. You have a full-time job and you run multiple companies. How long did this take you with all of that on your plate?

I would be lying if I said it was instant. There's definitely like the premise of AI was fairly quick, but actually learning how to build these things efficiently has taken time and I'm still not where I want to be. Obviously, like there's some things where today and yesterday has pretty much been troubleshooting this one domain I built to analyze old conversations, score them on a scale of 1 to 10, then send a custom follow-up message. Like for something like that, it takes extended time, but overall it didn't take that long at all.

Like, uh, best guess, when did you— we obviously have been using AI for, I mean, it's been about like a year. It hasn't even been that long. When did you really start building in AI beyond, you know, beyond, hey, I'm using ChatGPT to ask questions. Like, when did we actually start implementing AI in the business?

Maybe a week and a half, 2 weeks or so, like to actually going in depth on it.

I just want to point out how crazy that is. You're talking about building basically an AI company of AI employees. You've done coding, you've built apps for me. Uh, we've done multiple apps. For my company for integrations. And this is like a few-week project. This is amazing how fast learning happens. Really fun to see where it's going. But this is the goal of today's podcast is talking like it's the advanced part of AI in real estate. Can you create agents to take out tasks? What things should you outsource and what things should stay as relations? The way that I have been doing this, and I think this works really well, this is a friend of mine actually back when we were doing the VA strategy before all this AI stuff was available. One thing they would do every day, and I think this is a practical drill, write down every single task you do for a few days. And it is a pain. Every time you do stuff, just write it down. Then you're going to color code it as things that absolutely do not need to be you, things that sometimes you should have some input on, and things that always need to just be you. If you have a color-coded list, everything we'll call red, our magic color for like, hey, I shouldn't be doing this task. And if you're on the fence, ask yourself, does this task actually make money? If you're an entrepreneur, that's pretty much your main thing.

Yeah, that's pretty much for me.

Does me doing this task make more money now? Does it make more money later? And I probably should have input, or is this a, just like, is this just a non-money task? Those non-money tasks, those things that you have to get done but probably don't need your input, every single one of those should be offloaded. Either to an employee or an AI bot. And now the question, in the past it was like, does it need to be a US employee or can this be a virtual assistant? Today, the question is, do I need an employee or can we build an AI bot to do this with this set objective by the end of today? The biggest thing that Caleb does, and this is, I think this is most of the success that Caleb and I have had in our business, instant implementation of ideas. Someone has an idea on a call by the end of the day, It's a thing. Hey, we should build an AI bot for this. By around 9:30 tonight, we'll probably have an AI bot for it. Hey, I have a great marketing idea. We should try this ad campaign. It'll be filmed, edited, in production, and we'll have beta tested 4 different hooks by the end of the day. You can instantly analyze this. You could listen to this podcast and you can put this in your business by the end of the week in multiple levels.

Absolutely.

Super exciting stuff. Caleb, any closing thoughts on AI and real estate and technology? Anything that you have learned that people absolutely need to be doing right now?

There's some part of it actually building this stuff. It's kind of like real estate usually takes a lot, like not more expensive because I mean, it's very minimal on like the actual costs of this. Closing thoughts, I've started doing this, making sure you actually ask like, hey, does this cost tokens? Does this cost money to your AI? Cause there's some people I've seen on the internet that rack up, like, if you're using Claude without a paid monthly plan, you're going to rack up a very hefty bill if you're using paper, like paper use. Just get a regular plan, but actually make sure you're tracking this stuff. Cause there's a lot, like, you're like, how does somebody make money off this? How does this actually run? How does this fire? Like, why does this make sense? And actually looking at that stuff's pretty damn important. So that's the big one. And then building this out, you're, it's probably going to take a lot of debugging eventually. You're going to tell it to do something. You're going to get it there. And then you're on the finish line, you're like, I completely missed how to do that part. I should need— and you have to go back and rebug and redo the whole thing. Doesn't take forever, but it's going to take longer, especially at the start.

Oh, that's great advice. Fantastic ending point on this too. Yeah, it's like, hey, if you're doing this, you know what can tell you the most cost-effective way to run it? AI. Verify once you build the system.

Absolutely.

Did I build this optimally? Optimally? Was there a better way to run this? I actually ran an app recently where in that app I was like, I'd finished the whole thing and I'm like, this is actually pretty cool. Like, is there a better way to run this? And it's like, wait a second, this code is really inefficient. You should add this and this. And it literally, it tried to patch it. It kept crashing itself. And then it said, hey, I'm going to run my, leave your computer on. I'm going to recreate this entire thing in an optimal way over the next couple hours. And when I woke up in the morning, it had already not just built,, but it had tested everything that I'd previously tested, found the errors, restarted, and it's like, hey, this is your optimal amount of code for the most efficient version of the app that delivers everything that you wanted. And I've already tested every feature to make sure that you didn't have the app crashing the way that it was prior. Literally wrote itself overnight. It's a beautiful thing. So even when you're checking these things, all you have to do is just ask your app you designed it on the question, did I build this right? Literally go in and tell you, you did a good job. However, I'm going to build it for you again one more time. I'm going to build it better.

Or you can ask, hey, can I build it cheaper? Can I do this? You can just ask it questions. It'll spit back at you. It makes it a lot easier.

Yeah. Well, I hope this was helpful. If you guys are listening to this and you're bored out of your mind, then pick literally any other episode about real estate on this channel or business updates in the field. Again, we are going to bring a ton of new speakers onto the podcast. We took a little hiatus as we opened up Abilene. We've been building out the PM company. We started this year crazy, crazy, crazy heavy on acquisition. So buckle in, this podcast is about to get a ton of new episodes, ton of content. So save this, comment below, give us a 5-star rating on whatever app that you're on, help us spread the word, and we appreciate you listening. This is the Owner Meeting Podcast hosted by Multifamily Strategy. I'm Christian Osgood. This is Cael Hommel. I'll see you on the next episode.