7 AI Tools To Help You Buy and Operate Hotels (With No Staff and No Time) | Jake Heller E60

Click here to start your free 7-day trial of the AI for CRE Collective.

Most boutique hotel investors are running lean. No analyst team. No data department. No six-figure tech budget. Yet the demands of buying and operating hotels keep getting heavier.

In this episode, I sit down with Jake Heller, founder of the AI for CRE Collective, to break down the most effective AI tools that give small operators the firepower of a full investment team without hiring more staff or burning more hours.

We workshop real tools you can start using today to:

• Research any hotel market in minutes
• Underwrite deals 5 to 10 times faster
• Build pro-grade investor decks and OMs
• Create renovation concepts and renderings
• Cut legal costs with AI-assisted contract review
• Streamline asset management and portfolio reporting
• Automate day-to-day admin and communication

The 7 AI Tools We Cover in This Episode

1. Claude (Deep Research) – Market feasibility, multi-source synthesis
2. Perplexity (Market Intel) – Accurate analytical research with citations
3. Shortcut AI (Underwriting) – Flexible financial modeling inside Excel
4. Google Gemini (Nano Banana Image & Video Generation) – Property renderings, façade upgrades, interior redesigns, branding, teaser videos
5. Wilson AI (Legal & Due Diligence) – Contract review, redlines, title exceptions
6. Gamma (Decks & OMs) – Investor decks, offering memorandums, pitch materials
7. Wspr Flow (Voice-to-Text & Admin) – Ultra-fast transcription, notes, emails, and communication

Whether you’re trying to scale your hospitality business, underwrite more deals, or operate your hotel with fewer people, these tools will change the way you work.

If you want to buy and operate hotels without a big team or a ton of time, this episode is your new playbook.

Get One Week Free Membership to Jake’s AI for CRE Collective

If you want help implementing everything we discuss in this episode - the prompts, the workflows, the tools, the systems - Jake’s community is the best place to start. 

Listeners of The Hotel Investor Playbook can get full access for FREE.

Click here to start your free 7-day trial of the AI for CRE Collective.

About Jake

Jake Heller is a third-generation homebuilder and entrepreneur who co-founded the AI for CRE Collective, a central hub where AI and technology meet real estate. As an expert focused on practical, implementation-oriented solutions in AI in Commercial Real Estate, he built this 400+ member community to provide battle-tested workflows, automation templates, and a database of 300+ vetted AI tools for professionals. Drawing on his experience in development and consulting, including cofounding JDJ Consulting Group, Jake specializes in helping CRE firms and individuals scale and streamline processes to accelerate deal speed across critical tasks such as underwriting and deal analysis.

Connect with Jake

Website: https://aifor

Connect with Michael on Instagram or LinkedIn.

Email Us at info@hotelinvestorplaybook.com

Visit the Hotel Investor Playbook Instagram

Invest with Malama Capital

Submit a deal

Transcript

Michael Russell 0:03

The Hotel Investor Playbook, your guide to building wealth and freedom through hotel and hospitality ownership. Welcome back to the Hotel Investor Playbook. I'm Michael Russell, co-founder of Malama Capital, and your host. Today we're diving into one of the most exciting topics in hospitality right now. AI. Specifically, how small boutique operators like us can use it to punch way above our weight. Now, let's be real, most of us, we don't have a team of analysts, a data department, a six-figure tech budget. I mean, we're just we're running lean, we're wearing all these different hats, we're trying to buy and operate hotels and without going crazy in the process. So that's why I'm super fired up about this conversation. My guest today is Jake Heller. He's the founder of AI for CRE Collective, and he's built a community of real estate operators and investors who are learning to use AI every day to underwrite deals, manage assets, and scale faster, and all without hiring big teams. So Jake and I are gonna workshop live some of the most powerful ways you can use AI to streamline your hotel investing business. And look, I'll be the first to say it. I'm no AI expert. You know, if this were, I guess, human evolution, I'm somewhere between a caveman and a guy who just discovered fire. But Jake, on the other hand, Jake is building rocket ships. So let's get into it. Jake, welcome to the show, my man. Thanks for having me, man. I want to start with deep research and market feasibility. Because before we ever model a deal, I think the real work starts with research, right? Figuring out whether a market even makes sense. Most investors, myself included, waste days collecting data and we don't even know how to interpret it. So let's start here. When you come across, let's say, a deal in a market that you've never looked at before, what AI tools or workflows would you use to quickly research that market before you even open an Excel spreadsheet?

Jake Heller 2:15

Well, I think there's a lot of different ways to go about this. I'm fortunate enough and unfortunate uh to be subscribed to all five large language models: ChatGPT, Claude, Grok, Perplexity, and Gemini. And there's others as well, but those are kind of the core five for me. And so what that allows me to do is be effectively multi-model uh versus just a singular source of information. And each one of them is better at the other for specific purposes. Grok, for example, is pretty integrated with Twitter, and Twitter isn't a is a wealth of knowledge, right? Sometimes good, sometimes bad, but real-time knowledge from the average Joe. Now, it's important to verify some of the information that you get from Grok and from Twitter. So I wouldn't rely on it entirely. And that's why I think a multi-model approach for collecting research, for collecting data is so critical. So what I do, and just bear with me, it's pretty tactical. I use all five, the deep research capabilities with all five. Now, it's likely that you or your listeners in this case probably aren't subscribing to all five. In fact, I wouldn't even recommend doing that. Perhaps start with one, try as many as you can, and you'll figure out which one of them works best for you, which with for what specific purposes. But what I might do, let's say we're looking at a new market, I might start with perplexity. There's a big difference between generative and analytical AI. Generative is creating what doesn't already exist in a way. Analytical is what already exists. So Google, you go online and you're looking for information that already exists versus ChatGBT, you might build off of that or create something new entirely, right? So generative versus analytical, perplexity for the most part's analytical. It's gonna function in a similar way that Google does. So it's gonna pull information, sources, whatever's out there, it's gonna cite all that. And you could be pretty confident that the information you're getting from perplexity is pretty accurate. So I always like to start with perplexity. I'll say it's something along the lines of, hey, we're investing in a in a hotel in Long Beach, California, right? Like help me pull together a market report for Long Beach. Then I'll do the same thing with Grok, and Grok's gonna just dig into Twitter deep, right? What are people saying about it? Keywords, as much as it can possibly grab from Twitter. So now I have grok, I have perplexy. Then I'm gonna go into the three big boys, Gemini, Claude, and in ChatGBT. And I'm gonna do deep research using each of those. Deep research is different than just using it ordinarily. You'd actually go in and where it says tools, you would click deep research. And so I would do the same with Gemini, I would do the same with Claude, and I would do the same with ChatGBT. At this point, I have five different results from five five different platforms, and that's okay. What you'll find is that there's a lot of overlap, there's a lot of synergy, there's a lot of the same. There's also a little bit that's different. And so what I'm doing at this point is I'm consolidating a lot of the data that I'm getting from these different tools, and I'm synthesizing it. And when all is said and done, what I'll usually do is I'll go to Claude and I'll give it a prompt like, hey, here's all of this different market data, all of this different research that we've done. And I'm gonna tell it to effectively pick apart the stuff where there's overlap, right? Because at that point you can have higher confidence that it's correct if you're getting the same thing from multiple locations. And I'm having it synthesize it, put it all together, and then come up with this really robust and comprehensive market research report. Now, again, the average person's probably not going to subscribe to all five. My personal favorite, the tool that I am using at the most, the most at the moment is Claude. I think it's far superior than a lot of these other tools. ChatGBT, in my opinion, has been on the decline since they released GBT5. Now, their deep research capabilities are great. I wouldn't trust it wholeheartedly and in 100%. I'd always fact check. And again, AI, you never want to trust it 100%. But Claude, uh, if I were to just pick one, Claude is at the top of my list. Claude Deep Research is on a whole nother level. I actually was just doing a similar type example, and I think Claude pulled like 900 different sources as part of its deep research for a market report that I was doing.

Michael Russell 7:03

Yeah, I mean, like the amount of work that it would take someone to be able to review 900, but then synthesizing all that information. Obviously, we understand Chat GPT is extremely valuable, but you're describing these other LLMs. I had not, I've not used Grok before. I have not used Cloud before, but I've started using Perplexity. And what you said there, I think is an important distinction that perplexity is not a large language model. It actually just uses what's available on Google or really what's on the internet. It'll validate by by providing sources to actual websites that you can go and you can review. The other thing I noticed about perplexity is that they have this labs feature for when performing all this research, at least in the presentation and synthesizing the data. Sometimes when I use ChatGPT and I do deep research, I get all this information, but it's still visually I'm having a hard time interpreting what all the information means. What I like about Perplexity Labs is they will create charts, they will create graphic images, they will create dashboards, which are just ways to be able to interact with the data. You can filter, you can ask it certain questions. I think to your point, each of these tools are valuable in their own right. But I definitely would like to check out Claude because from what you shared, it seems like there's a little bit more accuracy compared to Chat GPT. I will say that what I've noticed sometimes is that ChatGPT's goal is to give you an answer. And sometimes those answers are what do you call it? It makes stuff up, right? It hallucinates. I've had the situation where I'm using ChatGPT and I'm doing deep research and I'm asking for comps, and it's created a comp that doesn't actually exist. It just made one up. It said, Oh, there's this building here, this property here. And then when you go to try to find that property, last ChatGPT, and it says, Oh, I'm sorry, you're right. I did hallucinate. Uh, forgive me. And so there is a bit of skepticism. I guess I'm curious when going through that, you've talked about using all the different sources to kind of like cross-check each other. Is there really any like specific prompt that you use to validate the accuracy? Is there some way that you're doing so beyond just checking the individual resources to make sure they exist?

Jake Heller 9:15

There's no like secret sauce or magic bullet with that. I think as humans, right, quality control is is important. My my goal with going multi-model for something like market research is to just gain a lot more confidence in in the result than if I were to just do it with Chat GBT, right? ChatGBT, I'm hearing time and time again, is is just on the decline. But it's important when you're doing a lot of this stuff that you give it the right prompts, that you give it the right guardrails, that you give it the right direction, because you're gonna get out what you put in. And so when prompting, there's there's something called the RTCF framework. And it it stands for R is role, T is Task, C is context, and then F is format. And so when you're asking these LLMs or any AI for that matter, it's important that you follow the RTCF framework and you want to tell it like what's your role. So, for example, in this case, you might say you're a you're a 30-year market research expert at Blackstone, for example, right? You've been putting together market research reports for 30 years at Blackstone, the biggest and baddest in the industry. And your task is now to do X, Y, and Z, right? And that might be, okay, we're buying a hotel in Long Beach. It might be to just do market research for someone who's purchasing a hotel in Long Beach. And then you want to give it as much context as possible because what will happen if you don't do these things is that it's gonna start assuming, right? Like, okay, well, he didn't give me the context that uh here's our specific value add strategy for this hotel, right? And I might start going on a wild goose chase with more of like a core plus type strategy where it's not value add. And so the the result will be kind of more catered to that specific context. So the context in all this is really important as well. And then the format for how you get it, okay, do we want it in an Excel file or markdown or PDF or whatever it might be? And then it should look like executive summary and this, that, and one another. So it's so important that when you're asking and prompting these tools, you give it the right framework, the right prompt. Garbage in, garbage out. And so I've I've become quite good at at prompting these different tools. Some something you could also do as well is you'll give it some context, then you'll ask it to give you the prompt. You'll take that prompt and then you'll prompt it. Say, hey, we want to do a market research report. Give me a prompt for this, and then it'll give you a prompt. You'll then take that prompt, you can start a new chat, or you could throw it back into that existing chat. And next thing you know, you have this great prompt, which is ideally most likely going to follow the RTCF framework, or at least be pretty darn close to it. But you're gonna get an entirely different result, whether good or bad, if you were to say, Hey, I want a market research report for hotels in Long Beach, versus you're a 30-year Blackstone hotel executive who's been putting together market research reports for an absurd amount of time. And you also have a friend who's an even bigger uh hotel investor at Blackstone and knows this better than you. Here's what I want you to do. Here's all the context with it, et cetera. So you'll get an entirely different result.

Michael Russell 12:38

Hey guys, quick favor if this episode is helping you, text it to one friend who is either trying to buy a hotel or owns and operates one already. Sharing the show is the best way you can support what we're building here. Now, back to the episode. Yeah, I mean, look, the the prompt engineering that you're describing, in my mind, I get a little anxious about it because I'm like, oh man, like I don't know what to say or exactly how to say it. But what you're describing is if you can speak a normal language, you have a basic premise of like, okay, well, what is the framework that I need to ask? And then you ask the LLM, let's say ChatGPT or whatever you're using to create the prompt for you, then that takes a huge burden. And the results that I've found are like I could never create a prompt to the same degree. What I do find though is I'm I'm familiar with ChatGPT, is I use the thinking model to provide a better prompt than just using ChatGPT auto. And usually I find better results. On that note, though, one other thing that I've found, I've learned, is that sometimes I will have to iterate two, three, four times to finally get what I'm looking for in the results. I may have the prompt, but then ultimately I might have to filter a little bit more. And one thing I've found that I'm doing now is I will ask the LLM if you were to reverse engineer this, what would have the prompt needed to be to get this result with all without all of that iteration to finally get here? And it'll spit it out. And then you can save it. You can save that prompt if you're doing constant due diligence and research. Then that prompt can go into whatever you could save that prompt to whatever LLM you're using for future use.

Jake Heller 14:15

Yeah, I I think the reverse engineering is an interesting concept. And I first learned about it when I made a LinkedIn post that went viral and it was all about utilizing ChatGBT's new agent mode, which for those of you that don't know, the big difference between agent mode and just kind of its standard feature is that agents could could execute tasks, right? So, like think of just booking a haircut, for example. You could have an agent effectively do that. But I use Chat GBT agent mode to prepare a feasibility study, just generally, right, for a piece of property. Hey, you have X, Y, and Z Main Street. Please go look at zoning and help me find highest and best use. And you at that point, you have to jump through all of these different interfaces, right? To first find zoning and then the development standards. Hey, it's R3, you could build 50 feet of height or three FAR, or your setbacks need to be a certain thing here. And so I I did this, I did this post, and at the end of it, I I said, comment and I'll I'll send you the prompt. Right. And what someone did, sneaky little guy, was he took that, all that information there and he reverse engineered it. And then on that same post, he's like, hey, for anyone who wants the prompt, I reverse engineered his output. Here's the prompt that he used for it. And I was like, oh, that's so interesting. And so that was the first time I kind of learned of this like reverse engineering concept with prompting. But look, as it relates to prompt, I think it's important for everybody to just create a prompt library for yourself. And think of all the different scenarios within your business, all the different roles that you play and make that time investment, right? To just have these prompts. And as you delegate, build out your team, et cetera, it's then easy to to for anyone to have access to these different prompts. But I think it's a really good time investment just because of how important prompting is today to just create a prompt library for yourself.

Michael Russell 16:11

Yeah, absolutely. And you know, I'm starting to use it now. I've got a an executive assistant, and I created uh a series of prompts, and she's using these prompts to perform research. It's incredible how much um information she can gain, some valuable information. And look, she's an executive assistant, she's not a trained analyst, she's not an expert real estate person, she has no experience in real estate, but she's able to get the information I need just by using the prompts that I spent time to create and to share with her. And and now it's really uh expedited the process. So it's been very valuable. I want to introduce the opportunity to discuss maybe some other AI software programs or or products that are out there. So let's let's jump into underwriting and deal analysis. And this is kind of where most deals they they live or die based on underwriting. And I'm sure every listener here has stared at a broker OM, wondering, okay, are these numbers real? So what I want to know is what's an AI tool that's a total game changer for underwriting and deal analysis?

Jake Heller 17:11

A couple that I I really love at the moment aren't necessarily purpose built for real estate at all. But what's so great about them is that they live within Excel. They have created these Excel AI agents that are native to Excel, their plugins in Excel, shortcut AI, uh, and index AI are brilliant tools uh growing like crazy, purpose-built for financial modeling and underwriting and analysis, et cetera. And so you probably have shops out there, listeners, uh, individuals who have been on iterating their underwriting model for years, five years, 10 years, 20 years, 30 years in some cases, right? And so a big mistake I think a lot of these underwriting softwares and tools make is that they ask you to essentially abandon your way of underwriting deals, and just hey, come on to my platform, learn this interface, we'll onboard you, all we'll be great, right? But that'll never happen. I shouldn't say never. It's happening, just not necessarily at scale, I don't think. And so what's so brilliant about shortcut and index is is they meet you where you're already working. And instead of replacing anything, they allow you to do things more efficiently, multiply your capabilities, right? Instead of just underwriting one deal, underwrite four, five, ten in the same amount of time. And they're only getting better. Okay, but like to help you guys kind of visualize this, think of just any Excel workbook. Top right in your kind of ribbon bar, you hit a button, and they're shortcut and index, and then a chat interface will pop up the same way that you communicate with Chat GBT or Claude. And now you could do a lot with this. Okay, you could take an OM, drag and drop, and say, hey, help me underwrite this deal. Now, again, this is prompting is very, very important in this particular case. You want to be very specific. How you're thinking about this deal, is it value added? Is this that, one or another? So again, garbage in, garbage out, but you could analyze that deal and essentially do nothing. Walk away from the computer, come back, and you'll have a recommendation of some sort or full type pro forma or analysis on that particular deal. Yeah, and as humans, we use judgment, quality control, et cetera. So you want to just be mindful that you probably don't want to rely on it 100% beyond just that scenario I just described. What they allow you to do is analyze data and scenarios in ways you wouldn't do previously. Our models that we've built, some some are are robust, but oftentimes they're not flexible where maybe there's something about the deal that you want to understand, but you can't because your model's not like specifically built to be able to do that. These tools. You could just say, hey, like, here's what's going on with this deal. Like, help me analyze it or help me build a model that I could analyze it for. So, some examples, you have a residential property and you want to see if an ADU makes sense. You can go to index or shortcut and just say, hey, I want to understand if building this additional ADU on my property makes sense. Traditionally, you you may have to hire someone to do that to build that out for you, or you may spend the time thinking about it and trying to build it yourself. The other scenario is you have your Excel model and you pull up these tools and you could map this deal to your exact model. Right? So, hey, underwrite this deal for me using my exact model. And it's not perfect yet. I've been in close conversation with the executive teams at both of these companies, and we were talking about this perfect scenario where we have our models and we want to underwrite 10 deals at one time. Like, how can we take 10 OMs and underwrite 10 deals at one time? I was like, how far away are we from something like that? And they said probably 12 to 18 months, but it'll be able to do it perfectly. There's another tool called quickdata.ai. That one's purpose built for data extraction for rent roles and PLs. And then as far as like commercial real estate specific underwriting tools, that is probably the most crowded space in Proptech, right? The list is endless. I mean, I could I could probably name 15 off the top of my head, but some like SIFT AI, APERS, ClickDot AI, Intel CRE, Optium. Some are maybe a little bit more specific to a particular asset class, but they're all trying to accomplish the same thing. And that's just underwrite more deals and close more deals. I think the problem that a lot of those tools have is that they're another software. They're another interface. They're another tool that you get onboarded to. And I think as real estate professionals, we're we're not sick and tired of software. We might be if if more and more of this stuff starts to come around. But we want to be more efficient at what it is that we already do. And I think a lot of those tools that are that are building in the space are understanding of that at the moment. But like the perfect scenario for me, I'm going to describe like a futuristic scenario as it relates to underwriting. Like you get all these deals from brokers. Maybe you get 20 deals a day from brokers. Think of this world where you have this like AI underwriting agent. It lives within your inbox. It's going to scan your emails, it's going to take those OMs. It's automatically going to underwrite it for you, maybe all while you're asleep, and it's going to come back to you. It's going to understand your specific criteria, right? I want a levered 15% IRR. I need a 7% untrended yield on cost. It's going to know all this stuff. It's going to be very intelligent on you, effectively like a digital twin. But it's going to go through your email. It's going to underwrite all these deals for you. It's going to come back to you. And for any deal, let's say it's two and a half million for the deal. This AI agent to get a 15% levered IRR and 7% yield on cost says, well, you need to pay $2 million for that deal. It's going to get all that information to you with the underwriting model. You'll wake up in the morning, you'll see that. And it's going to say, okay, do you want me to proceed? And you say, sure, write up an LOI. Then it's going to take that LOI, it's going to create it. It'll send it on your behalf. Next thing you know, it's liaisoning with the broker on your behalf. And all this is done agentically, right? Like with an AI agent. Like that, that, in my opinion, is great because we live within email, we live within Excel. Like nothing really changes. You're just being more efficient. And so these tools that kind of meet you where you already are are the ones I'm most excited about.

Michael Russell 23:37

One of the things that I enjoy about being in your community is you just rattled off all these different software options. And it becomes overwhelming to the point where I'm like, I don't even know where to start. And what I like that you've done is you've got kind of a master list that if you want to check out pretty much any software that's out there, there's a link to it and there's a quick little note about well, what does it do? But your community, what you're doing is you're you provide these little workshops where you have you know 10, 15 minute videos. Sometimes they're they're shorter, and they basically just demo the product and you can get a quick idea of well, is this something that's relevant for me? And so you demoed shortcut a few weeks ago. And so I hopped on that. I've been playing around with shortcut for the last three or four weeks. Two scenarios in which you can use shortcut. If you want to be nimble and create analysis that maybe doesn't fit into the parameters of the existing model you're currently using, you mentioned the ADU. So I've got a property right now that's we're moving from the LO LOI stage into the purchase and sale agreement. So I'm really pumped, I'm excited. This property is gonna be incredible. It's a 43-key boutique hotel, but it has, in addition to the hotel property, it's got two adjacent parcels. One that's about an acre, and another one next to that that's just though just under six acres. The one-acre property uh has a small home on it. And so the reason I bring this up is because my existing model, like you said, it doesn't really account for well what these additional properties are and how that could fit. I I could probably hire someone to help modify the financial model, but that could take weeks to find someone to pay for it, cost thousands of dollars. What I've done with a shortcut is I put in a prompt and I explained the details and it created a brand new model within like 10 minutes. And it wasn't perfect. I continue to iterate, but what I'm able to do now is I'm able to right then and there, without having to rely on some third party and spend thousands of dollars, I can say, hey, let's pretend we take that six-acre parcel and let's turn it into a campground. Let's say we get RVs on there and we kind of model what that'll be like. And the shortcut prompt system. So there's an Excel spreadsheet, and then on in the corner of it, let's say you can just write into it and prompt it. And I think what it's doing is it's using other large language models. It's not just providing you financials, it's also thinking through the scenarios by doing research. So it'll go and say, like, okay, you want to build an RV park. Here's what you could potentially make by by creating an RV park. This is what it will cost to build out. These are the things you need to think about. So this is all operating within this Excel spreadsheet that is is populating the numbers. And then maybe I want to make a pivot. Maybe I say, you know what, let's not do the RV route. Let's make like a micro resort on that six-acre parcel. What is it going to cost to build out, let's say 12 units? It'll research the cost of construction, it'll research all of the steps that you need to perform, and then it'll start doing the research on well, what are the average daily rates? What's the potential occupancy, all within this Excel model? And within about 15 or 20 minutes of iteration, I can have three or four different variations of how I want to go about through this acquisition. This is real. I'm doing this right now. And so, although I don't necessarily have it figured out right now what I want to do with those extra parcels, I know that my existing model wouldn't work. It just doesn't. It doesn't give me the flexibility. And so I'm really loving shortcut because it gives me like PhD level analysis without having to spend tens of thousands of dollars to get these changes or weeks of time to make each iteration it's instantaneous. So, scenario one, absolute flexibility at your fingertips. I love that. Scenario two really has me interested because if you have an existing model and it's cookie cutter, you know exactly what you're looking for. There's not all these variations, and you say, you know what, I need speed, I need to analyze in bulk, I need to just underwrite, underwrite, underwrite. You can use the model that you already have. And like Jake said, just plug in an OM. And once you have a prompt, you can actually save prompts within the system so that you don't have to keep typing the prompt in. You could press a couple of buttons, it's gonna underwrite it. And again, it might not be perfect, but it's gonna get you close enough to know is this within the strike zone? Is this, does this pass the sniff test, so to speak, so that maybe you can take a closer look? Because if volume is the key to success, underwriting a hundred deals is the key to success. It takes you so much time and analysis and bandwidth. But the AI software that we're describing, in this case, shortcut AI, it just expedites the process, number one. And number two, you can have a junior associate now do this work without having to be involved yourself. A lot of people, it's a little taboo to say this, they're afraid to say it. They're like, AI is not replacing humans, it's just enhancing them. I call baloney, it is replacing humans. Like it is. And sorry for a lot of folks that are going to be affected by this, but here's the reality: we were paying $60,000 a year for a junior associate who now I can just pay $20 a month and have software that does not quite, but almost the same amount of work at my fingertips, with the help, of course, of maybe a VA in the Philippines. Like this is so mind-boggling crazy. I'm so excited about it. It's incredible how you can leverage these resources to just get to the next level without having to be an institutional grade investor in a big team and have a huge budget. You can do this as a small outfit.

Jake Heller 29:13

Yeah. Well, I echo a lot of what you said. And I just to touch on the whole VA piece. These tools level the playing field, right? They they lower the barrier to entry. You don't need to have an experienced, Harvard-educated analyst to do the initial data entry and underwriting. And look, a lot of the time these guys spent was on data entry, creating the pivot tables and extracting documents. And it's like that that's a joke. That's that's easy stuff. Now a lot of that can be automated with these tools. But to your point, like you could have a VA in the Philippines use shortcut or index and build institutional grade models. Or if you have a a hotel and you have this excess land, and maybe you want to do a glamping resort, but the underwriting and analysis is a little bit different. The VA just prompts shortcut and it could go ahead and do that. So I think the barrier to entry becomes a lot lower with these tools. I think people are certainly getting replaced. In my opinion, it's it's the Harvard educated analyst who you would normally pay $200,000 a year to. It's the individuals who aren't using these tools. And so I'm super fired up that you could essentially have someone with effectively like no experience whatsoever in Excel andor real estate come in. And as long as you have your prompt library, which they have these within these tools, that person can copy and paste and just do a lot of the grunt work, data entry, first pass at these models for you. And it's it's super easy. So I'm fired up about that too.

Michael Russell 30:53

Quick break. If you're not using AI for your business, you're going to get left behind. And if you're overwhelmed by it or don't know where to start, you need to check out Jake Heller's AI for CRE collective community. I'm a paying member and it has been a game changer for me. Jake shows you exactly which tools matter for commercial real estate and how to use them to underwrite faster, tighten up operations, automate the busy work, and make much better decisions. The trainings are simple. The community is active and you can bring your real questions to the weekly roundtable and get direct feedback. There's a link in the show notes that gets you your first week free. Full transparency, it's an affiliate link. So I do get a small kickback if you join, which helps me keep creating episodes like this. Check it out after the show. So let's move on to another area of pain traditionally, which is legal and due diligence. I think that we all know attorney fees can crush a deal budget. So I've heard you talk about tools like Wilson AI that can analyze contracts and prep redline summaries before you even send them to your lawyer. So maybe you can touch upon how that actually works.

Jake Heller 32:11

Yeah, legal, the legal AI space is interesting. I think it's just starting to get traction. We all know that attorneys are bottlenecks oftentimes to deals. So these legal AI agents are really, really interesting. Wilson specifically, I haven't used many beyond Wilson. I think there's others like Harvey. I just saw something, another tool that that did a big fundraising round as well. And so starting to learn a lot more about these tools. But with Wilson specifically, I actually did an episode with the CEO last week. His name's Gus. They're out of London, but moving HQ over to San Francisco. It's brilliant, these tools. Wilson, again, not specific, built for commercial real estate, but some of the use cases I was able to do with Wilson. One, redline uh LOIs and purchase and sale agreements. You literally sit there, you prompt it, next thing you know, you have this AI agent that's just redlining, beating up this document in ways you you haven't seen before. It happens in minutes. I deal with quite a bit of land, and on these title records, you just have absurd things. You could have Wilson get in there and prepare like a title exception letter, which you could just essentially pass along to escrow. It's brilliant. It's gonna go online, it's gonna do its research and reference this, that, one another. And what these tools are are hyper focused on at the moment are citations uh as well, just because big decisions are made off of attorney's judgment, judgment from the legal side, right? So they're making it a point that they provide you with as many of these citations and sources, et cetera. So at the end of the day, you you could make that decision and ultimately they just cover their, you know what. But I'm pumped about that. I've created contracts with Wilson AI. There's just so much that you could do with it. And I think uh how how that translates into savings both with money and time. I I don't have these specific data by any means, but a tool like Wilson probably costs you a couple hundred bucks a month. If you're doing enough deals, I mean that's justifiable all day long, right? And time is money as well. So being able to turn stuff around within a matter of minutes versus in many cases days with these attorneys is such a game changer. I I do a lot with larger publicly traded home builders. And as you would imagine, the administrative work and contracts are are absurd. I mean, you need an NDA to sign an NDA stuff. You don't need a real estate attorney for the basic stuff. And so having a tool like Wilson is so incredible in that case.

Michael Russell 34:58

So I'm at this point when when I reflect on the last year, there's been three purchase agreements that I've entered into, and uh, they've been painful. Like things have fallen out, like didn't work out. But the attorney fees, like they all get paid up front, right? So we spent, golly, on the first deal, I think we spent $15,000 in attorney fees. And then the second one was easily over 10 grand. It just adds up, right? And so now I'm at this stage where I'm I'm kind of sensitive to the pain of that, and I'm entering to this purchase proposal stage to buy this hotel. And so I'm like, okay, this time around, I'll I'm gonna check out, like you just said, Wilson AI, or maybe I'll check out Harvey for a couple hundred bucks. I mean, these attorneys charge $400 an hour and it takes weeks and it's painful. And then you get something, and all it does is it pisses off the seller's like, what the hell is this? And so using Chat GPT, I guess you could use that, but I really want I'm gonna check this out. I haven't used Wilson AI, I appreciate that. I'm curious to know though, in your experience, what would you not rely on AI for in legal review?

Jake Heller 36:06

A lot at the moment. I think that'll change. AI is not gonna replace humans. Judgment, relationship building, problem solving, strategic thinking, these are things that just humans do very well and and specific to businesses. I I would be quite cautious on the legal AI stuff at the moment. I think there's specific parts within that whole process, which you could use tools like Wilson on and feel very confident that it's gonna do it correctly. Like my title exception letters, those I feel really good about. Right? Redlining LOIs, those I feel really, really good about. First pass reviews for purchase and sale contracts, I feel really good about it. Now, I uh you know, I talked in the beginning of this call about multi-model versus singular with with all the different large language models. And I think that mentality is really important and interesting with all of this, and just being able to know which tools to use for what specific purpose, how to best use them, when not to use them. Those individuals who understand that at the moment, I think are the most dangerous, right? Because what I might do for um a purchase and sale contract is take a first pass with Wilson and beat it up, and then have my attorney look at it with all of that and just kind of do a lot of the initial work up front for them. And maybe they're able to do it in half the amount of time, half the amount of money. And so there's kind of this like strategic combination of some sort. Like as it relates to legal, I think that's an interesting and good approach, and an approach that I'm doing currently. Now, with how fast these tools are changing, how how quick they're just becoming better and better, maybe in two weeks' time you'll have like a purpose-built real estate legal AI agent that you could feel like 100% confident. Uh, well, not 100%, but very confident in. And so I I think just this collaboration essentially with like these AI tools and then humans, like that's where the magic happens at the moment.

unknown 38:22

Yeah.

Michael Russell 38:22

I mean, so basically, look, at this point in time, we're not replacing our attorneys. We're just showing up with a good amount of the work already done, which results in real savings.

Jake Heller 38:33

Mike, one just quick thought. Yeah. So paralegals, junior attorneys. I I don't like to say the word replace, but I think tools like Wilson AI at the moment are affecting them and will continue to affect, right? Just being able to do a lot of the stuff that they do. Maybe your junior attorney for whatever reason, or an first-level associate, they bill at 300 bucks an hour. I think Wilson is gonna compete with them if it's not happening now in a very short amount of time. But your 30-year partner who's been doing real estate contracts since he was a little kid or whatever, he graduated locksburg and just knows it better than anybody. Like, it's gonna be hard to replace someone like that.

Michael Russell 39:17

Makes sense. All right. So jumping into the marketing side. So whether it's raising capital, just listing the property, you know, good design, it sells. So hiring a designer for every investor deck or for every OM, I mean, this is gonna cost thousands. So, you know, what AI tools can help create professional grade decks or investor memos or property listings without spending a ton of money.

Jake Heller 39:44

I love Gamma. Gamma is such an incredible company. The founders are incredible. The growth they've seen has been incredible. And I am so obsessed with Gamma. The use cases are in. Infinite, but essentially it's the PowerPoint killer. You prompt it the same way you would with these other tools, and it builds decks for you instantly. Now, gamma, what they prioritize is within 30 seconds, having the user be able to see it. So in real time with gamma, like it's gonna build this deck for you in instantly. The beautiful decks, and it's only getting better. OMs, investor presentations, BOVs, market research reports, whatever it is, you could build with gamma and do a hell of a job. So I I I love gamma. They have an agent mode that lives within it. So you could take a first pass at creating a deck, and then you'll hit agent mode, and you might say something like, Hey, we're presenting to Chinese investors. Please translate all of this to Chinese. It'll go through and do all of that. Or you well might you might say, Hey, I didn't see any comparables in this deck. So go online and find me sales comps, and it'll do that. Or change the whole layout of this slide. And it'll do that in real time. And so gamma is just a powerhouse, an absolute monster in generating presentations and decks. What I've really enjoyed about it is when I'm pitching investors. Traditionally, I wouldn't I wouldn't spend the time nor money to to build a deck like specific to that investor and the type of collaboration we might have. But what I've done with Gamma is we'll have a meeting with these investors, I'll have my AI note taker come in there, and I'll then take the notes from that AI note taker. I'll take all the kind of correspondence I've had with that specific investor, my email, and I'll give it some context, and then I'll build like a deck where it's like Jake slash Michael J V and like this beautiful deck that's like specific to that investor. And now, is that gonna move the needle much? I I don't know, but I think someone would appreciate oh 100%, dude.

Michael Russell 42:05

That's gonna move the needle. Look, the property we entered into contract earlier this year was in San Diego, and it's it was an eight and a half, nine million dollar property. Actually, it was listed for $11 million or something crazy like that originally, but we ended up getting it under contract for less. My team and I spent hours preparing a deck with Google Slide. We used a Loom video. It was actually pretty decent, but it was like hours and hours, days of effort it took us to prepare this deck, which we presented to the seller and basically put our case. We're like, look, you're you want this much money, this is why we think it's worth this. And we put in the different comps. We had all the visuals that showed the different properties and why basically it was worth the amount that we were suggesting. It was effective. The seller appreciated the effort, but holy heck, the amount of time and hours it takes to build that. If you're saying that that can be done with gamma within a few minutes, that's a game changer. But the concern I have is sometimes this AI software, it can look a little like just a template project. Similar to how someone can read now with the little M dashes, and when someone posts on LinkedIn, they can quickly and easily identify, oh, it looks like that person used AI. I'm wondering, are these decks, are they gonna look like just a generic template? Or does it really look like someone put passion and care into it?

Jake Heller 43:24

So a tool like gamma is not purpose-built for real estate. So I think there will continue to be that scenario you you described. It could do a hell of a job. But like if you're pitching to you know a Blackstone as your investor for this hotel purchase, like you're probably not going to use a tool like Gamma to create that deck. But there are companies in the space which are doing really well. In fact, the fastest growing prop tech company in the world right now is a company called Henry AI. They build decks. And their focus at the moment is with brokers and creating offering memorandums and broker opinion of valleys. Like that, that is their focus. And I think why they're doing such a great job is because they're just belentlessly focused on perfecting that. They're going after whales, but they're so focused on building these institutional caliber decks, right? And they spend a lot of time working with you to just on the front end to make sure your branding requirements, your style, your tone. It's not like YAML where it happens in a second. You'll get your deck in three, four hours. But I with AI generally, like it's creating efficiencies and sort of compressing the timeline for a lot of tasks that we do.

Michael Russell 44:40

Well, I think a big part of the presentation comes down to design and visualization. So investors are constantly struggling with visualizing visualizing this finished product, like especially before committing to a renovation budget. So, what software would you recommend for quickly generating concept renderings or design mock-ups to communicate the vision to investors or or to lenders?

Jake Heller 45:05

Gemini is a winner for all that. I mean, if you go to my LinkedIn, look at my profile photo. Oh, it's it's Gemini. You wouldn't, you wouldn't tell. It's insane. Mine too. I put Hawaiian Islands in the background. Uh, those are not real. Oh my god. Gemini's image generating and video generating capabilities are are unbelievable. I think in the AI race, Google is up there between Google and Claude. That's my arsenal. But Gemini could produce renderings, it could produce logos, whatever, whatever it is. I've taken old kitchens, outdated, and I've given Gemini great prompts to create upgraded, renovated kitchens. And you can get so funky and creative with it, right? Like I've I've battle tested, okay, I want an industrial, industrialist kitchen, I want minimalist kitchen, I want modern farmhouse, contemporary, and it's just gonna keep doing it. And the exact dimensions and layout and structure of this kitchen, it doesn't change. Assuming you give it the right prompt, you don't you want to make it clear, hey, do not change the dimensions, do not hallucinate, just upgrade this, right? You could say, hey, I want LED lights underneath the cabinets, or I want subway tile backsplash, or I want you know these type of appliances, stainless steel, I want this type of fridge, etc. And it'll do all that. And you wouldn't even tell that it's AI generated. So that's insane. Like Mike, you buy assets in in other states and stuff. You can't just hop on a flight and get there within hours of the next day. Go on Google Images, take your little snippet tool and and screenshot that that asset from the from the exterior, throw it in a Gemini and say, hey, I I want to upgrade the the exterior, build out balconies. I want a living green wall on the exterior with metal siding, whatever it is. And it'll generate it. And it's so I'm not gonna say perfect.

Michael Russell 47:08

I hate saying perfect, but it's so close. Yeah, I've been toying around with Google Gemini. It's we're using it even for mocking up t-shirts and things. So it goes for everything. You talk about branding, like if you just want to get something quick and dirty out there to test this out and see, hey, is there interest? It is an incredible way at a very, I mean, I don't know what Google Gemini costs, but a reasonable cost because look, it cost us four grand to create a pitch deck.

Jake Heller 47:33

Have you tried video generation? It's amazing. I mean, again, that exact example. Take take the exterior photo from Google, updated rendering using Gemini. It's called Nano Banana, by the way. Oh, yeah, yeah. Image generator. Funny name. Take that image and then use their video generating software, and you can create like a teaser video. So whether it's for investors or if you're a broker selling an asset, whatever it is, you can create a teaser video, and it's like HD level shit. It's insane. Excuse my legit. So, oh, it's it's legit. I I love it, it's amazing.

Michael Russell 48:11

Yeah, I've seen all these like funny things where people are like self-imposing their photo onto like a superhero and they're flying around like maybe Mars or something. And yeah, wow, that's real realistic. And that would have been 10 years ago completely impossible unless you were like a studio production in Hollywood. And now with a couple of clicks on the mouse, you can generate these incredible videos, which is really relevant, of course, for investing. Because, like we're talking about, a video is just the next level up from a photo, really puts people in the picture. So I'm gonna be checking all that out. I'm excited because I'm going through this process now with fingers crossed, everything works here, but I'm I'm optimistic that I'm gonna start now to be able to implement a lot of these tools. So everything that we've talked about from feasibility to underwriting to preparing in the deck and modeling out future renovations. So it's really gonna be a great opportunity to put these tools to use. The next step I feel like that's really important for folks is once you purchase a hotel, maybe some of our listeners already have a hotel and so they're experiencing some of the pain of asset management and operations. I think that sometimes operations things can fall apart for a lot of owners. You know, you have multiple properties trying to track all this performance and communication, it can get overwhelming fast. So I want to know what AI tools help operators manage their hotels more effectively without having to hire an expensive asset manager.

Jake Heller 49:34

There's a tool called Lenny. I don't know if it's capable of hotels at the moment, but for multifamily, they're they're insane. Just portfolio level insight in ways you you could only imagine. Right. Like I think as as it relates to asset management, being able to communicate with the whole portfolios is critical, right? Like at any point in time, okay, this asset's performing in a way that it needs our immediate attention. And like why? And like just getting that type of real-time insights amazing. And Lenny is a tool that does that. But there's a lot of tools out there for both property and asset management. Tools like Wayline, I think it's called brightwise.ai. And a lot of these tools, what they're trying to solve for is kind of the front end like reception administrative type work. But it's a space, asset management and property management is a space like I'm keeping a close eye on. I I think that there's some players in there trying to crack that nut. It's not an easy nut to crack.

Michael Russell 50:33

Yeah. Well, I think it's one step at a time, right? We've talked about the evolution and the progression of this. And, you know, at minimum, just getting started with some of the things that we've described and then keeping your eye on well, what's the next iteration of AI software as it relates to us and moving forward as things get rolled out and as we all get more comfortable with it? But one thing that I will I want to end with this last tool here, or a suggestion for tools, you know, communication and admin. I mean, every day there's admin chaos. Like there's emails, there's scheduling, there's investor communication. Sometimes I'm just drowning in, like you said, the BS. What tools do you use personally to keep up with communication without having to hire, let's say, an assistant?

Jake Heller 51:14

Yeah, it's interesting. I I was building like AI agents and automations for a lot of that stuff you described using tools like Lindy and N8N, Zapier. They're difficult because it's like a band-aid, more permanent type of problem. There's tools out there that solve like very specific things. Like there's a new one called Iris, which is effectively an AI personal assistant. They're all about calendar and invites and emails, etc. So you you might be able to just say, let Michael know our scheduled podcast at 11. Uh something came up and it'll handle like the emails to you, rescheduling. It could send new calendar invites. So there's tools like that. I think the whole personal assistance space is a really interesting one that I'm also keeping my eyes on.

Michael Russell 52:06

Have you tried the voice to text? Like you have. I think you did it, you demoed this actually. Um whisperflow. Whisper flow. Oh my gosh, so simple, but so helpful. So for our listeners who don't who don't know what Whisperflow is, maybe you can describe this, but I'll take a stab at it. It's basically like a little voice to text automation thing. But in instead of like the the one that you have in your phone, like if you try to use Siri or whatever, sometimes it's garbage. Whisperflow, like you just hit the little button, it's lightning fast, and it just nails it. And it like knows how to space out the text, it adds commas in the right spots. And so whether it's for your own personal use, like just creating notes or bullet points or whatever, that's helpful. But like sending text messages these days, or I mean, you can do full on emails, you just talk into the app and it just instantaneously like spits out a perfectly like grammatically correct, just organized, easy to read. It's awesome. So I I've been impressed with that. I I can't even believe I didn't mention that.

Jake Heller 53:00

That that tool has changed my life, my way of doing things more than any other tool. I have it on my computer and on my phone. And I think there's sort of like a mental block for a lot of people when using a tool like Whisperflow. Once you get past that, you'll never go back. And then you said something about like investor communication. There's three different tools. There's Juniper Square, which is a big one. There's Agora, A G O R A, another big one. And then there's a new company as well. I think it's called Home Base. That's also a pretty crowded space, but a lot of people like those tools. So if you have a lot of investors, a lot of different files, deals, they help streamline that whole process for you.

Michael Russell 53:42

Cool. All right. So listen, before we close, I just want to zoom out for a second here. If someone's listening who feels overwhelmed by this discussion or AI in general, they don't know where to start, they're worried. Hey, it's too complicated. What would you tell them?

Jake Heller 53:56

It's a very real concern. And what's crazy is that I spend all day long on this stuff, and even I'm overwhelmed. And that's okay. I've accepted it. This stuff is moving very fast. There's a new product, tool, software, announcement, feature, whatever, every single second, every single week. So it's very hard to stay on top of this stuff. But there's no doubt that these tools and softwares help save time, help save you money. And I think the most effective people today and heading into the future are the ones that just know these tools exist, know which tool to use for what specific purpose, know how to effectively use it, know when to draw the line, right? Know when, okay, humans are better at this, AI is better at this. And so you gotta just start. You gotta start. You gotta start messing around with these tools, you gotta start asking questions, you gotta get involved because the gap between those that are using these tools and are not using the these tools is getting wider. And I hate when people aren't open-minded about this stuff. Like I've I've gone to battle on LinkedIn with just individuals who are you know, they're prideful in their work and they want to write emails their way and not use AI for this, that, and one, another. But look at the internet when when that all that came around, dot com. Can you imagine just back then just being so resistant to like, oh, the internet? Those people are not around today, right? And in this AI revolution, this new world we're heading into is is arguably bigger than that, bigger than anything we've ever experienced in our entire life. We haven't even scratched the surface of this stuff. We're not even in the first inning, we're not on first base, we're not even close. And so I think in this new world, those that are are embracing this, those that are being open-minded, those that are asking questions and trying these tools, I I think they will do a lot better on the on the other side of this. I think there will be massive advantages. I think there will be huge efficiencies. I think there will be real estate teams that manage billions of dollars and it's just one person. Myself, you as well, Mike. I mean, we are ahead of the curve as it relates to all this stuff. We know enough to be dangerous. We we know that we can create very good decks in 10 to 15 minutes and someone else could spend 10 hours on it. We know that we could underwrite deals in in ways that then others cannot. We know that soon we'll be able to underwrite 10 times the amount of deals someone else could just do what. We know how to search for comps and do market research and utilize AI in in efficient ways that others do not. So I would tell anyone like, don't fight it. Keep an open mind. There's no reason to not have an open mind with this stuff. This is some of the most revolutionary things in our lifetime ever. And so ask questions, get involved, try this stuff. It doesn't hurt you before bed, get online, join the community, try that tool. You won't believe how amazing some of this stuff is.

Michael Russell 57:08

Yeah. Jake, dude, this has been awesome. Thank you so much for taking the time to share some of your insights. For anyone who wants to go deeper into this world, where can they follow you and learn more about the AI for CRE collab?

Jake Heller 57:23

I'm all over LinkedIn. I think I'm Jake Heller1. I'm all over Twitter. We have a YouTube channel, AI for CRE Collective. We do podcasts with founders building product in our space. The AI for CRE Collective, though, is a community on the school platform SKOL. It is amazing. Glad that you're a part of it, Michael. We post daily, we share ideas, workflows, automations. There's courses and roadmaps and role-specific implementations, and we're constantly just releasing new stuff on a daily basis. Yeah, awesome.

Michael Russell 57:58

Of course, we'll put all the links in the show notes. So, listeners, if this episode gave you a few light bulb moments, share this episode with someone who needs to hear it. An operator, an investor, somebody who wants to buy a hotel, maybe someone who already owns one and wants to run it leaner and scale smarter. So do it right now. Come on, press that share button. Use those little numbers, nimble thumbs of yours, type in someone's name who could benefit and send them this episode. That's how we grow the show, and that's how we attract amazing guests like Jake and keep bringing you valuable content every week. So if you're enjoying the Hotel Investor Playbook, share this with five people right now. And all right, that's it for today. Thanks, Jake. Thanks again for joining us, and we'll catch you guys again next week. Aloha.