Latest / Own More Doors / #87 He Tested 712 AI Tools for Real Estate. Here's What Actually Works

Transcript

#87 He Tested 712 AI Tools for Real Estate. Here's What Actually Works · Own More Doors

0:00
SPEAKER_00: These tools are gonna turn you into Superman.
0:02
SPEAKER_00: I mean, I've seen you draw building codes from a city and it reads it and it spits out like literally the highest and best use.
0:08
SPEAKER_00: If you wanted an appraiser to give you a highest and best use report, you're paying thousands of dollars.
0:14
SPEAKER_01: We are all sitting on a gold mine of data which we are not using, and we need to start thinking about how we can monetize that.
0:20
SPEAKER_01: If you're thinking that you're just gonna sit on your hands and knees and continue doing things the way that you normally do, think again.
0:27
SPEAKER_00: AI is changing commercial real estate faster than people realize.
0:30
SPEAKER_00: Today's guest, Jake Heller, has spent hundreds of hours testing virtually every AI tool on the market, from broker opinions of value to offering memorandums to underwriting acquisitions and analyzing ground up developments.
0:42
SPEAKER_00: Instead of guessing what's useful, we're gonna show you exactly what's working today and what every broker, every investor should be using before they get left behind.
0:50
SPEAKER_00: Jake, it's great having you back, man.
0:53
SPEAKER_00: Thank you for making time for us.
0:54
SPEAKER_00: Round two.
0:55
SPEAKER_00: Happy to be here.
0:56
SPEAKER_00: Man, so I'm following you on X.
0:57
SPEAKER_00: You're killing it.
0:58
SPEAKER_00: Um, you're growing your YouTube channel, folks.
1:00
SPEAKER_00: Make sure that we uh that we support Jake.
1:02
SPEAKER_00: Um, he's doing a lot for us here, helping us, and we want to make sure that we have him back for a third round.
1:08
SPEAKER_00: And so, Jake, you you are constantly putting out videos, you're checking every every system that's coming out.
1:14
SPEAKER_00: I don't even know if it's called a system, but from from a commercial real estate perspective, I don't know anyone else tracking and really diving deep as you are.
1:25
SPEAKER_00: Uh, catch us up from where we're at we where we were at last time, where we're at today.
1:29
SPEAKER_00: What do we need to know?
1:30
SPEAKER_00: And in twofold, because this is a two-part is yes, we have real estate brokers who listen to us, and there's there's ways that we could uh be more efficient with our business, but the our community is based, is really composed of housing providers, people that are looking to acquire, people that are looking to scale, uh trade up, maybe downsize sometimes.
1:50
SPEAKER_00: Hey, we're gonna go into a DST.
1:52
SPEAKER_00: What tools should we be using right now that are that are simple, easy to use?
1:58
SPEAKER_00: I don't care if it's free or not because some of these things shouldn't be free.
2:01
SPEAKER_00: They're really valuable.
2:02
SPEAKER_00: Um, I'll let you take it away.
2:04
SPEAKER_00: Yeah.
2:04
SPEAKER_01: Well, again, appreciate you having me here.
2:06
SPEAKER_01: I think last time I was here was what, four months ago?
2:09
SPEAKER_01: And it and in the uh the age of AI, that's a hundred years.
2:12
SPEAKER_01: Okay.
2:12
SPEAKER_01: So what we talked about last time is I'm not gonna say irrelevant today, but it's changed so much.
2:19
SPEAKER_01: And and the pace of of play, the the trajectory with all this stuff is just it's moving and compounding and shaking at at such a fast rate.
2:29
SPEAKER_01: So uh, but we we are uh we're the AI for CRE collective.
2:32
SPEAKER_01: Uh since we we talked last, our kind of big vision has has really changed.
2:37
SPEAKER_01: What we are is is really media, and we're focused on how technology is changing the built world.
2:43
SPEAKER_01: Okay.
2:44
SPEAKER_01: And uh right now a lot of that's AI, and specifically in commercial real estate.
2:48
SPEAKER_01: That's how we got started, and that's what we're focused on currently.
2:51
SPEAKER_01: We manage or we we, I should say, we track a tool database of about 712 tools as of today.
2:58
SPEAKER_01: Uh, you mean AI tools?
3:00
SPEAKER_01: AI tools, specifically for commercial.
3:02
SPEAKER_01: Well, there's there's traditional prop tech tools, which aren't really, I mean, now a lot of them have AI capabilities, but they were built pre-AI, right?
3:09
SPEAKER_01: So you we call them traditional prop tech.
3:11
SPEAKER_01: But uh a lot of them AI in commercial real estate.
3:14
SPEAKER_01: Again, that's that's sort of our focus.
3:16
SPEAKER_01: And so a lot of the tools represent that.
3:18
SPEAKER_01: Um there's some really good ones, there's some really bad ones.
3:21
SPEAKER_01: What's uh really important to talk about is how the general purpose tools like Claude and ChatGBT and Manus and Gemini and a lot of the ones which I'm sure all your listeners have really heard of, they are so incredibly powerful, right?
3:33
SPEAKER_01: And and specifically for commercial real estate, you can do endless amounts of things, right?
3:38
SPEAKER_01: Just on my screen here, I have uh Claude Design, okay, which last time you and I spoke, this was not here.
3:44
SPEAKER_01: This this did not exist.
3:46
SPEAKER_01: Okay, you had tools like Adobe Design and Figma, et cetera, but you did not have Claude Design.
3:51
SPEAKER_01: And what people don't realize is just how powerful Claude Design is for creating materials, whether it's decks, right?
3:58
SPEAKER_01: If you're an investor, you're you're syndicating a new deal, let's say in Long Beach, shout out Long Beach.
4:03
SPEAKER_01: Um putting together a deck for your investors to see can be done very easily.
4:08
SPEAKER_01: And I have some examples we'll show in just a moment, or creating a website.
4:11
SPEAKER_01: Okay, let's say your website's outdated and you want to create a new, beautiful looking website.
4:15
SPEAKER_01: Easy.
4:16
SPEAKER_01: Claude Design.
4:17
SPEAKER_01: Amenity maps, really interactive type documents you could do with Claude Design.
4:21
SPEAKER_01: It's absolutely incredible.
4:22
SPEAKER_01: And then not to mention you have Claude in Excel.
4:24
SPEAKER_01: Okay, so underwriting, we showcased some uh examples, I think, last time.
4:28
SPEAKER_01: I don't know if we talked about Claude uh in Excel.
4:31
SPEAKER_01: We we maybe showcased uh shortcut AI uh in NDEX, which are great.
4:35
SPEAKER_01: They're purpose-built AI agents that live within Excel, but there is this reckoning moment happening in Excel.
4:41
SPEAKER_01: Excel is one of the most used tools uh worldwide, right?
4:45
SPEAKER_01: Across all industries, but specifically in commercial real estate, we live in Excel, right?
4:49
SPEAKER_01: Uh and so the fact that we can now use this incredible technology to do the things we we normally do within Excel and just do it that much better, it's it's awesome.
4:59
SPEAKER_01: So, you know, underwriting, analyzing data in just different ways, it's all really incredible.
5:04
SPEAKER_01: So we we could showcase some of that as well.
5:06
SPEAKER_01: But I tell anyone who's kind of just starting their journey, don't go chasing all these different AI tools.
5:12
SPEAKER_01: Okay.
5:13
SPEAKER_01: Start with one.
5:14
SPEAKER_01: I recommend Claude or ChatGBT.
5:15
SPEAKER_01: ChatGBT is incredible, um, almost equally as good as Claude.
5:19
SPEAKER_01: In my opinion, they've kind of gone different routes.
5:21
SPEAKER_01: Claude more enterprise, uh, ChatGBT more consumer.
5:25
SPEAKER_01: Uh, but start with one, okay?
5:27
SPEAKER_01: And then what you're gonna realize is that the the depth in in the capabilities of these tools is is it is deep.
5:36
SPEAKER_01: Okay.
5:36
SPEAKER_01: I mean, you have Claude Design, you have Claude in Excel, Claude and Microsoft Word now, right?
5:41
SPEAKER_01: Redlining uh uh LOIs and contracts and whatever it might be, your purchase and sale agreements, you have uh Claude and PowerPoint, there's Claude Desktop, there's Claude uh in browser, there's there's uh Claude Code, there's Claude Cowork.
5:55
SPEAKER_01: There is a lot to cover.
5:56
SPEAKER_01: So I recommend just starting with one Claude or Chat GBT and just going deep and going through the motions.
6:04
SPEAKER_00: Anyone listening to you right now is gonna be like, wow, how do I catch up?
6:07
SPEAKER_00: That was a lot said.
6:09
SPEAKER_00: Give us some basic um tasks that if you're newer to Claude, some simple things.
6:15
SPEAKER_00: Uh let's just say that you're a buyer.
6:17
SPEAKER_00: Let's say you're a buyer in Long Beach and you're in a 1031 and you're looking to buy, and you just like, you know, you're analyzing a few buildings.
6:26
SPEAKER_00: Like, how can someone take that and and instead of waiting for the broker to get back to them?
6:31
SPEAKER_00: Hopefully they're working with us, but they're waiting for us to get back to them.
6:34
SPEAKER_00: And, you know, but they're sitting at home, like, man, which which one of these two or three eight-unit buildings do I buy?
6:40
SPEAKER_00: What are some simple things that they could do?
6:42
SPEAKER_00: We don't necessarily need to do the example, but what are some things that they could prompt it to do that you're like, this is an obvious one where anyone could do this today after listening to our podcast?
6:49
SPEAKER_01: Yeah, I mean, simple analysis, right?
6:52
SPEAKER_01: Um, uh the most basic way to do it would just be dragging and dropping an offering memorandum and putting it in the cloud and saying something like, Hey, uh, I want you to initially screen this deal for me.
7:02
SPEAKER_01: Come up with the 10 questions that I should be asking the broker, for example.
7:08
SPEAKER_01: Come up with 10 things you believe might be a red flag or an issue with this type of deal.
7:16
SPEAKER_01: Analyze the market, right?
7:17
SPEAKER_01: Take it as far as you want, but just basic things like that.
7:20
SPEAKER_01: Um, you can get a lot more uh advanced than that uh with with your underwriting and analysis of it, but just basic stuff like that, right?
7:28
SPEAKER_01: And and you could do that simultaneously.
7:30
SPEAKER_01: What a lot of people don't realize is the ability to multitask with these types of tools, right?
7:34
SPEAKER_01: So instead of just doing that with one offering memorandum, you can queue it up to do it for all three or four deals that you're looking at.
7:40
SPEAKER_01: And and you know, that's in a fraction of the amount of time that it would typically take.
7:44
SPEAKER_01: So uh things like that, and then and then of course, um uh taking it a step further and like I said, analyzing the market.
7:51
SPEAKER_01: I I think with AI, what people don't realize is just uh how how much better you could analyze data, right?
8:00
SPEAKER_01: What's happening in in the city of Long Beach, what's happening at the port of Long Beach, right?
8:04
SPEAKER_01: How are uh new businesses coming into a town affecting the the the different real estate markets and submarkets and different types of assets and hey, this office building, how how is that changing, right?
8:18
SPEAKER_01: Et cetera.
8:19
SPEAKER_01: So um I always tell people like, if you're not sure, just ask AI, right?
8:24
SPEAKER_01: If you don't know how to do something, ask AI.
8:27
SPEAKER_01: It's very simple.
8:29
SPEAKER_00: Always ask AI.
8:30
SPEAKER_00: You know, sometimes people will several months ago.
8:33
SPEAKER_00: I mean, I doubt people are still having a problem, but you know, we have to understand that not everyone, I mean, I would say, Jake, you know, you're you're for my opinion, you're at the advanced level of using Claude and you're talking about things that even I and I use we use Claude and Gemini and perplexity every day for our business.
8:49
SPEAKER_00: But and even when talking to you, I'm like, I realize there's like a big difference between where I'm at and how I'm using it and where you're using it.
8:56
SPEAKER_00: And but even at the basic level, I think when people say, Well, man, I don't really know how to prompt it, and I say, ask it how to prompt it, tell it what you want to do to give you the prompt.
9:05
SPEAKER_00: It's gonna give you the prompt.
9:06
SPEAKER_00: Yeah, it's called meta prompting.
9:08
SPEAKER_00: Okay, that's why I didn't know there was a name for it.
9:11
SPEAKER_01: No, it's called meta prompting.
9:12
SPEAKER_01: Okay, right.
9:12
SPEAKER_01: And so I I I think um prompting is is one of the most basic things you could you could understand um how to do with AI.
9:20
SPEAKER_01: And um the the important part is that you get out what you put in, okay?
9:26
SPEAKER_01: And so what you put in is the prompt, right?
9:28
SPEAKER_01: Uh people oftentimes expect this perfect answer from the AI, but they don't make the time investment on the front end, giving it the context that it needs.
9:37
SPEAKER_01: So you got to think of AI as as somewhat of a junior analyst, right?
9:41
SPEAKER_01: Very capable.
9:42
SPEAKER_01: Um, but let's use underwriting as an example.
9:45
SPEAKER_01: If you're not gonna give the AI um assumptions that you would typically use to underwrite a deal, right?
9:52
SPEAKER_01: Let's say rent growth.
9:53
SPEAKER_01: I don't know what rent growth is in Long Beach.
9:55
SPEAKER_00: I think if I'm doing it, if I'm doing it for a client, I'm gonna keep it very conservative right now because we're not seeing a whole lot of rent growth.
10:02
SPEAKER_00: So I might be like at 2%.
10:03
SPEAKER_00: 2%, right?
10:04
SPEAKER_01: So if you don't give the AI that context, right, what Claude is gonna do, or any AI for that matter, they're gonna go online, they're gonna do re web research, and and let's just say some other broker who just doesn't know the market, he's from Minnesota, is he writes an article online about Long Beach rent growth for apartments, I believe, is 8%.
10:25
SPEAKER_01: Claude's gonna go and look at that, and they might use 8% rent growth instead of 2%, which is which is accurate, right?
10:33
SPEAKER_01: And so that's really important context that you need to give the AI because otherwise it's gonna make assumptions.
10:38
SPEAKER_01: And what happens when the AI makes assumptions, that's what creates these hallucinations, right?
10:44
SPEAKER_01: And so um taking the time to give the AI context is a very worthwhile time investment because that'll get you closer to the result that you actually want.
10:55
SPEAKER_01: But to expect, okay, let's just say in our heads, I mean, you know, you know Long Beach multifamily better than anybody, right?
11:01
SPEAKER_01: Rent growth, uh, what post-renovated units might get as far as a rent premium, uh, one-bedroom apartment, whether it's a light rent, medium rent, full gut rent, how much that's gonna cost.
11:11
SPEAKER_01: These are very nuanced details, right?
11:13
SPEAKER_01: Which if you don't give to the AI, you are allowing it to make those assumptions and you cannot expect it to get those assumptions exactly how you want them every time, right?
11:22
SPEAKER_01: Does that make sense?
11:22
SPEAKER_01: Yeah.
11:23
SPEAKER_01: And so that context is critically important.
11:25
SPEAKER_01: And so when you're prompting, okay, it again, it take the time to really give it this context because the results you're gonna get are are far better.
11:36
SPEAKER_00: Jake, so we talked about analyzing, say, an eight-unit building.
11:39
SPEAKER_00: We talked about underwriting.
11:41
SPEAKER_00: Um talk us through how, you know, I've seen you do things where you're you're doing full-scale developments, and and I'm not, you know, with with the with our community, I would say sh maybe give us some examples on how we could use it for building ADUs on um on an apartment building or hey, we have this excess land.
11:59
SPEAKER_00: Um, I mean, I've seen you drop um building codes from a city and it reads it and it spits out like literally the highest and best use.
12:07
SPEAKER_00: I mean, these are, and I don't know that it's accurate, but these are um if you wanted an appraiser to give you a highest and best use report, you're paying thousands of dollars, right?
12:17
SPEAKER_00: And maybe you should pay, you know, like if you're if it's if it's a very unique property, maybe you should pay that.
12:22
SPEAKER_00: But in the meantime, you could play around with it and really come up with something pretty neat.
12:26
SPEAKER_01: Yeah, no, uh, no doubt.
12:28
SPEAKER_01: Look, I think the general purpose tools like Claude and ChatGPT are are good at a lot of stuff, right?
12:33
SPEAKER_01: But when you come when you talk about feasibility and highest and best use in zoning, zoning is a beast, right?
12:39
SPEAKER_01: It uh each city uh and county, it's all different, right?
12:42
SPEAKER_01: Zoning is one of the most fragmented, most difficult data sets to really uh tackle at scale.
12:49
SPEAKER_01: Okay, and so uh you can't really expect Claude or ChatGBT to do that perfectly each time.
12:54
SPEAKER_01: And so there are purpose-built AI tools which allow you uh to analyze the feasibility or put together highest and best use.
13:02
SPEAKER_01: I mean, there's there's go dirt.ai, there's a lot of these land acquisition platforms like pillar.codes and acres.com and propheticsoftware.ai, there's uh algoma, which is specifically a multifamily feasibility tool uh feasibility tool.
13:15
SPEAKER_01: Um and so they allow you to get really granular with a lot of the nuances that come with zoning and and and the site characteristics and infrastructure, and hey, what does the sewer line and that kind of stuff look like?
13:25
SPEAKER_01: Um but when building uh ADUs uh here in Long Beach, I mean, yeah, you could you could uh you could start with with Claude and ChatGBT um and um and really just ask it um you know what's the zoning of this particular property?
13:41
SPEAKER_01: What's what's been the precedent for approvals for these types of properties?
13:44
SPEAKER_01: How likely is the city to maybe permit this type of project, right?
13:48
SPEAKER_01: And then, hey, give me some concepts of what the project could ultimately look like.
13:51
SPEAKER_01: ChatGBT just released a new family of models 5.6.
13:55
SPEAKER_01: What it could do with image generation is unbelievable.
13:59
SPEAKER_01: And I used to use Gemini for that, by the way, Nano Banana.
14:02
SPEAKER_01: Haven't touched it in a long time, since ChatGBT.
14:05
SPEAKER_00: Uh since since two months ago, maybe.
14:06
SPEAKER_01: Uh honestly last week.
14:09
SPEAKER_01: Okay, got it.
14:10
SPEAKER_01: Uh, just since ChatGBT released their new model, last week being July 8th or whatever that was.
14:14
SPEAKER_01: But um, so preparing these renderings of of ADUs, for example, floor plans, we're getting it's getting better and better at actually producing like accurate, dimensioned floor plans, right?
14:27
SPEAKER_01: And then from there being able to use it and take it a step further and say, hey, what you know, what are some estimated costs associated uh with this ADU, right?
14:36
SPEAKER_01: And benchmark that to industry standards.
14:39
SPEAKER_01: Okay, and then you could ultimately take Claude and have it do that, right?
14:43
SPEAKER_01: And then Chat GBT do that, and then Mannetz do that, Gemini do that, and take a multi-model approach and then kind of compare all of those against one another just to test the accuracy of such.
14:53
SPEAKER_01: But um yeah, there's there's really no shortage of of the things you could do with these tools specifically to uh to commercial real estate.
15:07
SPEAKER_00: Give us some advice for commercial brokers.
15:09
SPEAKER_00: Um so so typically a commercial broker, in most instances, you know, I we're in the business of uh lead generation, and that could be done a multitude of ways, right?
15:20
SPEAKER_00: And so take my business for instance, and so I, you know, I'm I'm able to generate leads in and what I call inbound marketing.
15:28
SPEAKER_00: I provide value through this podcast, I provide value through our YouTube channel, uh, through our conferences, through our monthly meetups.
15:35
SPEAKER_00: I provide value where where I build trust through what I'm providing.
15:40
SPEAKER_00: And then you get inbound marketing.
15:42
SPEAKER_00: A lot of people that are real estate brokers that get into commercial real estate are hammering the phones.
15:46
SPEAKER_00: That's the old school, tried and true.
15:48
SPEAKER_00: All you're gonna do is call.
15:50
SPEAKER_00: Um you work with a lot of brokers, a lot of successful brokers.
15:55
SPEAKER_00: What are some of the tools that they're using right now to help their real estate business?
15:59
SPEAKER_01: So I think there's five really interesting tools specific for brokers.
16:05
SPEAKER_01: Um number one is terracotta.
16:08
SPEAKER_01: And this is in no order of priority, by the way.
16:11
SPEAKER_01: Uh terracotta is really interesting because I'm a firm believer that the tools that kind of meet us where we already are and allow us to do that that much better are really interesting, versus the tools that say, hey, abandon your process of doing things and learn a new way, right?
16:31
SPEAKER_01: You mentioned cold calling.
16:33
SPEAKER_01: That's a tale as old as time.
16:35
SPEAKER_01: It works, right?
16:36
SPEAKER_01: Traditionally, yeah, cold calling works, it'll always work.
16:40
SPEAKER_01: Um, and so in in commercial real estate, uh, that's great, right?
16:44
SPEAKER_01: Terracotta allows you to make more calls and better calls, right?
16:48
SPEAKER_01: And so just a quick little example, and and by the way, they do a lot more than just that.
16:53
SPEAKER_01: Um, skip tracing, prospecting, that sort of thing, they're great.
16:56
SPEAKER_01: Okay, but the they showed me an example which has really stuck with me, and it's it's not a major uh thing that's gonna move the needle ultimately, but let's say you're making 150 calls every single day, right?
17:08
SPEAKER_01: You're gonna get some callbacks from people, people.
17:11
SPEAKER_01: And and it's hard to keep track of the people you've called, their number, et cetera, right?
17:15
SPEAKER_01: So you might get a call back, you don't recognize the number, and oftentimes you're scrambling, like, oh my god, who is this, right?
17:20
SPEAKER_01: So what Terracotta built was sort of this AI summary.
17:23
SPEAKER_01: And so let's say you called Anita at 1234 Main Street and Long Beach was an apartment building, okay?
17:28
SPEAKER_01: She called you back.
17:29
SPEAKER_01: You didn't have her number saved or anything.
17:30
SPEAKER_01: Traditionally, you would have no idea that it's her.
17:32
SPEAKER_01: Terracotta will put a voice or a an AI summary together that you could see while they're calling, and it says Anita from 1234 Main Street in Long Beach is calling you, right?
17:45
SPEAKER_01: So you could answer the phone and say, Hey, Anita, I was just thinking about you and your property at 1234 Main Street in in Long Beach.
17:51
SPEAKER_00: Yeah.
17:52
SPEAKER_01: Um, and so again, while that's not like gonna move the needle and ultimately get you the deal, it's something that that is incredible, right?
18:00
SPEAKER_00: Uh it does move the little things move the needle.
18:03
SPEAKER_00: So yeah.
18:03
SPEAKER_00: Yeah, yeah.
18:04
SPEAKER_00: And all of them together really have a profound effect.
18:06
SPEAKER_01: And so terracotta is really interesting.
18:08
SPEAKER_01: Um, you could also query a natural language, uh, something along the lines of, hey, I'm looking for um uh 50 office owners in in San Francisco that are on two acres less of land uh with a 50,000 square foot office building, right?
18:25
SPEAKER_01: And it'll surface those owners for you with their contact information, right?
18:28
SPEAKER_01: Which is great.
18:29
SPEAKER_00: Um and so just to give some I guess some perspective, like if you if you were to put that prompt in to say Claude, it wouldn't generate the same information, maybe because it's not built specifically for that.
18:42
SPEAKER_00: Is that what we're saying?
18:43
SPEAKER_00: You you're not gonna get the same result.
18:45
SPEAKER_01: Um uh not even close, I don't I don't think you will get some people, and and perhaps there's a couple people on there that are interesting.
18:53
SPEAKER_01: Um, but like Claude for that I wouldn't use Manis, on the other hand, Manis is an orchestration layer.
18:59
SPEAKER_01: It was bought by by Meta.
19:01
SPEAKER_01: It's not actually a large language model like Claude, uh, but it's so incredibly good at research.
19:06
SPEAKER_01: And so if you were to use Manaus for something like that, I think you can get a pretty decent result.
19:10
SPEAKER_01: But Terracotta, uh purpose built for that exact purpose, and so you're just gonna get a better result from it.
19:16
SPEAKER_01: Um, that was number one, right?
19:18
SPEAKER_01: So uh another tool is Deal Ground, they're really interesting as well.
19:21
SPEAKER_01: And uh what they do is really aggregate um all marketing brochures and uh turn that into like one searchable database, right?
19:31
SPEAKER_01: So what people don't really realize is that all these offering memorandums that are out there, it is incredible data.
19:38
SPEAKER_01: Incredible, right?
19:39
SPEAKER_01: Uh and oftentimes they just go in the graveyard.
19:41
SPEAKER_01: Yeah, that data's not used.
19:42
SPEAKER_01: Or like CoStar is gonna pick up that data and they're gonna make money from it.
19:45
SPEAKER_01: But what a lot of brokerages need to realize is that they are the breadwinners, they are the ones that are creating the the valuable data sets, which traditionally goes to CoStar and they make the money off of it.
19:57
SPEAKER_01: But we give it to them for free and then they charge one.
20:01
SPEAKER_00: I mean, that's well, they charge it back to me.
20:03
SPEAKER_00: Exactly.
20:04
SPEAKER_01: Genius for them.
20:04
SPEAKER_01: Yes.
20:05
SPEAKER_01: So something to really think about is like you're creating these data sets.
20:09
SPEAKER_01: And so I think what you need to think about, or anyone else for that matter, is how could we start to leverage this proprietary data that we ultimately create, right?
20:16
SPEAKER_01: And with offering memorandums, think about all the comps that that go into uh those decks, right?
20:22
SPEAKER_01: You're gonna do all that research, you're gonna put those comps in there, and then those are just gonna go to die, right?
20:26
SPEAKER_01: But like what Deal Ground ultimately does is like it starts aggregating all that data, and then you can start to query against it, right?
20:33
SPEAKER_01: So, like, hey, I'm uh looking at comps in um or I'm uh I'm looking for a Chick-fil-A in Long Beach that's lease is expiring in three years.
20:43
SPEAKER_01: Okay, and it's able to get that data just because it has endless amounts of these like offering memorandums and and and data sets that they've consolidated.
20:51
SPEAKER_01: Uh so really interesting uh deal ground.
20:54
SPEAKER_01: Another one um that or another two, I should say.
20:58
SPEAKER_01: Henry AI is is interesting.
21:00
SPEAKER_01: Uh they were purpose-built for for uh creating offering memorandums.
21:04
SPEAKER_01: Now they do much more than that.
21:06
SPEAKER_01: They uh they they pretty much cover the life cycle of the acquisition.
21:10
SPEAKER_01: So uh all they really need is a rent roll and uh some raw financials.
21:14
SPEAKER_01: They'll uh underwrite the deal, they'll go get comps, uh, they'll connect to all of your internal systems as well.
21:20
SPEAKER_01: So they're not just pulling comps from CoStar, they're pulling comps from your internal database.
21:24
SPEAKER_01: And that's what uh just to go on a tangent for a second, that's where AI again becomes really interesting at analyzing data.
21:31
SPEAKER_01: We are all sitting on a gold mine of data which we are not using, and we need to start thinking about how we could monetize that.
21:38
SPEAKER_01: Okay.
21:39
SPEAKER_01: And so I have what's called an obsidian vault, without getting too technical, but uh call it my AI memory layer.
21:46
SPEAKER_01: All right.
21:47
SPEAKER_01: And we should all have one of these AI memory layers because what happens when you use these tools is that you're gonna chat with them and they oftentimes forget, right?
21:53
SPEAKER_01: They they just don't have that context.
21:55
SPEAKER_01: Sure, they have uh memory, but but that's the biggest problem with AI right now.
22:00
SPEAKER_01: Is that that it's missing that context.
22:02
SPEAKER_01: And so oftentimes you'll have to say, Oh, I'm Juan Weezer again.
22:05
SPEAKER_01: I Long Beach multifamily, and it's annoying, you know?
22:08
SPEAKER_01: And so I've effectively created this AI memory layer for myself where it doesn't matter if it's Claude or Chat GBT or whatever it is, but any AI that could access files on my computer could access my AI memory layer.
22:22
SPEAKER_01: And in that memory layer, it knows my entire life, my entire business, all the deals that I'm working on, right?
22:28
SPEAKER_01: All the different comps that are out there, I can query against it and say, hey, who do I know that's a landscaper in Long Beach?
22:33
SPEAKER_01: Or who do I know that's a general contractor that focuses on 50 units in Koreatown?
22:38
SPEAKER_01: Right.
22:39
SPEAKER_01: And because I'm building this proprietary data set of every conversation that I'm having, every email that I've sent, every deal I've underwritten, right?
22:47
SPEAKER_01: Every offering memorandum.
22:49
SPEAKER_01: And so using AI to effectively store this data the right way, of course, and create this knowledge base for yourself allows you to just make better decisions.
22:59
SPEAKER_01: You cannot uh, if you cannot measure it, you ultimately cannot cannot manage it.
23:04
SPEAKER_01: Okay.
23:04
SPEAKER_01: And so just being able to store this data, I just think is one of the most exciting uh things we could do as just real estate professionals.
23:12
SPEAKER_01: It's awesome.
23:13
SPEAKER_00: Yeah.
23:13
SPEAKER_00: Um, you know, I wrote that down that that we are sitting on a gold mine of data.
23:17
SPEAKER_00: I mean, even when people are asking me uh for a broker opinion of value, and a lot of times we do them, and we're doing a lot of research um on the comparables.
23:26
SPEAKER_00: You know, you might think I'm crazy, Jake, but I I'm still pulling my own comparables.
23:30
SPEAKER_00: Like I am hand selecting personally, like not anyone in my team, not an at like I'm the one choosing it.
23:36
SPEAKER_00: And then I feed that into um these tools, and then and then I'm having the broker opinion.
23:43
SPEAKER_00: Like I'm personally saying that is comparable.
23:46
SPEAKER_00: That one's not.
23:46
SPEAKER_00: I'm almost not to the point where I'm trusting, and I know you're saying, hey, you could use some of these um um tools and it will pull the comps for you.
23:54
SPEAKER_00: I guess I'm not to the point where like I believe that those are gonna be the best ones, only because I've been doing it for so long.
24:00
SPEAKER_00: But to your point, is yeah, we have tons of BOVs, tons of OMs, and they're not working together.
24:08
SPEAKER_00: It's almost like when a new lead comes in, I'm starting from scratch.
24:12
SPEAKER_00: It's a big deal.
24:12
SPEAKER_00: Only only because yeah, and I'm probably not, you know, I'm one of many brokers that could probably say the same thing.
24:18
SPEAKER_00: Uh, but but you're right about wouldn't it be nice if if, and I'm just imagining this, if I said a new lead came in, uh, go through our other OMs and BOVs because some of the uh some of the comparables are gonna be the same.
24:31
SPEAKER_00: Yeah, and then it starts assembling it for us.
24:33
SPEAKER_00: Totally.
24:34
SPEAKER_01: I mean, think about it.
24:35
SPEAKER_01: I don't, I'm not gonna assume, but say you've worked on 500 deals in in Long Beach, right?
24:39
SPEAKER_01: You're a multifamily specialist.
24:41
SPEAKER_01: Again, not gonna assume, but let's just say that's the case.
24:44
SPEAKER_01: In Long Beach, you've done 500 deals, all of those you've hand selected uh what the comparables should be, how you should have that data, right?
24:51
SPEAKER_01: The vintage, what's parking look like, et cetera?
24:54
SPEAKER_01: That's all data which you yourself would feel comfortable using, right?
24:57
SPEAKER_01: Because you did it yourself.
24:58
SPEAKER_01: And so if that was organized and structured the right way, then it could ultimately be used.
25:03
SPEAKER_01: And that's that's the bigger problem with adoption of these tools and this technology is that nobody has their house in order.
25:10
SPEAKER_01: Yeah.
25:10
SPEAKER_01: Their data is not structured the right way.
25:12
SPEAKER_01: It's a total mess.
25:13
SPEAKER_01: And so you have in these incredible theories of, oh, I want to automate my business and I want to do this, that, and one another, right?
25:19
SPEAKER_01: But think about the process you yourself would would do to get that result, right?
25:23
SPEAKER_01: So say you wanted to underwrite a deal.
25:25
SPEAKER_01: And I'm not saying this is the case with you, but but oftentimes it is for a lot of people.
25:28
SPEAKER_01: It's okay, you got to go into your Dropbox, right?
25:32
SPEAKER_01: And uh the rent roll lives there, but oh, the P uh the the T12 is organized in another folder, right?
25:38
SPEAKER_01: And then you're gonna make a copy of that folder and save it here.
25:40
SPEAKER_01: And then you go to Mary and the other office and say, oh, Mary, uh, where did you put that thing?
25:45
SPEAKER_01: Or uh, hey, can you tell me what you have for this?
25:48
SPEAKER_01: Right.
25:48
SPEAKER_01: And it's this really nuanced, uh fragmented process, and people are expecting that they could train AI to do that exact process.
25:57
SPEAKER_01: No.
25:58
SPEAKER_01: So you got to get your house in order and you gotta have your data and files clean, crystal clean, right?
26:05
SPEAKER_01: Because that's what's gonna allow this technology to be able to use it.
26:10
SPEAKER_00: And you said that you're that you've created something that would allow us to do that?
26:15
SPEAKER_00: Or you're using it on your own and it's just um I have what's called an obsidian vault.
26:21
SPEAKER_01: Obsidian's just a note taker, okay?
26:23
SPEAKER_01: And without getting too technical, um, obsidian allows you to create this type of knowledge base I'm describing.
26:30
SPEAKER_01: Anytime you save something in obsidian, it's going to save that as what's called a markdown file on your computer.
26:38
SPEAKER_01: So uh maybe I could show you.
26:40
SPEAKER_01: Um if I were to go into this one here, this is actually something that uh I did a demo for this previous previously, okay.
26:50
SPEAKER_01: And so I'm gonna pull up obsidian.
26:54
SPEAKER_01: And what you're looking at here, I actually this is a demo for a broker, so you could literally steal this, okay?
27:00
SPEAKER_01: What you this is what obsidian looks like, okay?
27:02
SPEAKER_01: You have this knowledge graph on the right side of the screen here.
27:05
SPEAKER_01: All you can see all your files are connected.
27:07
SPEAKER_01: Uh we have a handful of different folders, right?
27:10
SPEAKER_01: So there's the company, okay, and that's Jay Keller, and and this was just a demo, but assumed I work at Marcus and MillerChat, it has my license number, my email, my specialty, my focus.
27:21
SPEAKER_01: And then as I'm um saving more information into my obsidian vault, it creates contacts for every individual person.
27:29
SPEAKER_01: So I I basically just all I did was gave it two offering memorandums, two deals, and I ingested it this into the obsidian vault.
27:37
SPEAKER_01: This was just a demo, so it's not comprehensive at all.
27:39
SPEAKER_01: But it creates individual contacts for all these individuals.
27:42
SPEAKER_01: Okay, again, their license number, their specialty, their email, their phone number, et cetera.
27:45
SPEAKER_01: And that alone is really powerful, right?
27:47
SPEAKER_01: Imagine you had this vendor database of every single vendor that you've communicated with, and I'm sure you do, it's in your phone, but getting to that, thinking about it, it it's awful, right?
27:55
SPEAKER_01: So instead, what you could do is use Claude and say, hey, Claude, who do I know that does this?
28:01
SPEAKER_01: Or who is the best person to sell South LA apartment buildings in my vendor database, my obsidian vault, right?
28:06
SPEAKER_01: And so it's gonna track all my listings, it's gonna track all my comps, the lease comps, the sales comps, right?
28:11
SPEAKER_01: So then I can query it and say, hey, uh, in the last two months, what's the average sale price per unit in 900 two six zip code?
28:20
SPEAKER_01: And it's gonna go into my proprietary comp database that only I have access to.
28:25
SPEAKER_01: This is the Jake show.
28:26
SPEAKER_01: Nobody else has access to this database except myself.
28:29
SPEAKER_01: I've done the hard work to build this, right?
28:30
SPEAKER_01: And over time, sure, it takes a lot of work to build this and maintain it, etc.
28:37
SPEAKER_01: But the bigger it becomes, it needs to remain clean, but the bigger it comes, the more comprehensive that it is, the more powerful it becomes to you, right?
28:46
SPEAKER_01: And so what what people don't realize too is that I didn't just hand select and build this vault here in in obsidian.
28:52
SPEAKER_01: I instead used Claude to ingest this information and put it into Obsidian.
28:58
SPEAKER_01: But so what happens is you're looking at Obsidian here, everything that gets stored here then gets saved as a folder on my computer.
29:06
SPEAKER_01: So what you're looking at here is that that vault is saved on my computer, right?
29:11
SPEAKER_01: So I have the company files just like I showed you guys a Jake Heller, Marcus, and Millichap.
29:15
SPEAKER_01: I have the contacts, right?
29:16
SPEAKER_01: These are all markdown files, and markdown files are kind of like Word documents, but they're easy to process for the AI, a little bit cheaper from like a token perspective, but it all gets saved there, right?
29:26
SPEAKER_01: And I also include things like market notes, right?
29:29
SPEAKER_01: And then I have templates and attachments.
29:31
SPEAKER_01: And look, all these offering memorandums are stored in here, right?
29:35
SPEAKER_01: Um as an attachment.
29:37
SPEAKER_01: And so all of this data then gets connected to one another.
29:41
SPEAKER_01: And I could effectively query it with any AI that could access my local computer, which Claude, Chat, GBT, a lot of them can.
29:50
SPEAKER_01: It's gonna be able to tap into a wealth of knowledge that that I have put together over however many years I've been in the business, right?
29:59
SPEAKER_00: Jake, that's amazing.
30:01
SPEAKER_00: Um, let me run an idea by you and just tell me uh so I get a new listing, uh, a 10-unit building in Long Beach.
30:11
SPEAKER_00: And you know, traditionally we would do like circle dialing.
30:13
SPEAKER_00: So then I'm gonna say, well, who else owns similar properties in this area?
30:17
SPEAKER_00: Before this hits the market, I should probably have a conversation with them.
30:20
SPEAKER_00: They might be interested in buying this, or who just purchased in this radius, right?
30:23
SPEAKER_00: Is there something that I could outside of I mean, right now we would go to Rheony, we'd get the telephone and we're calling, right?
30:32
SPEAKER_00: Is there a way that we could utilize these tools that would do that for you?
30:36
SPEAKER_01: I I'm not sure there's there's necessarily a shortcut for that.
30:41
SPEAKER_01: Um but so that particular use case, I I'm not I'm not totally sure.
30:47
SPEAKER_01: Um in a way, yes, but I I'm not sure the um I'm not sure it's worth the time.
30:58
SPEAKER_01: The ROI is not going to be there, in my opinion.
31:00
SPEAKER_01: Right.
31:00
SPEAKER_01: And so that that speaks to a much larger concept for commercial real estate professionals specifically.
31:05
SPEAKER_01: Oftentimes like doing it yourself is just the better thing to do, yeah, versus using AI.
31:10
SPEAKER_01: The biggest mistake people do is they try to AI everything.
31:13
SPEAKER_01: Right.
31:13
SPEAKER_00: Well, it it gives, you know, I I went to this conference and then they were showing this example.
31:19
SPEAKER_00: They're like, you don't have to do your own YouTubes anymore.
31:24
SPEAKER_00: They're, you know, we're gonna take your whatever, your avatar, and we're gonna make the YouTubes, but you've never done them.
31:30
SPEAKER_00: And I'm like, ah, and people are looking at me because I'm in that room, had the biggest channel, and they're like, dude, bro, you're doing it the old school way where you're in front of a camera.
31:39
SPEAKER_00: Like, look, this someone it could use your voice and it could mimic what you're doing and it could tell your story.
31:44
SPEAKER_00: And I'm like, nope, nope, nope, nope, nope.
31:46
SPEAKER_00: I'm like, people are watching me because there's a level of he's sincere, that he's a real person, authentic, authentic.
31:54
SPEAKER_00: Like, I'm like, no, I'm shaking my head.
31:56
SPEAKER_00: Like, I you could do that.
31:58
SPEAKER_00: I'm not gonna take that shortcut.
31:59
SPEAKER_00: And so I'm glad that you brought that up because yes, it's amazing.
32:03
SPEAKER_00: Yes, it's a tool, yes, it could help us elevate.
32:06
SPEAKER_00: You know, I have a boutique brokerage company here, you know, and I'm having to compete with all the big companies and everything else, but now tools like this, we're now um you're dangerous.
32:16
SPEAKER_00: Yeah, yeah.
32:16
SPEAKER_00: You're dangerous.
32:17
SPEAKER_00: I'm gonna take people's lunch with me.
32:19
SPEAKER_01: Okay.
32:20
SPEAKER_01: That's what's so incredible.
32:21
SPEAKER_01: No, the what AI is, is a tool, right?
32:24
SPEAKER_01: And it is going to multiply good fundamentals and A players the same way that it's gonna multiply bad fundamentals and bad players, though.
32:34
SPEAKER_01: And so something uh that not a lot of people are thinking about is that there's this gap forming, right?
32:39
SPEAKER_01: Guys like you've been in the business, you have the experience, you know what you're doing, you're solid, right?
32:44
SPEAKER_01: These tools are gonna turn you into Superman, right?
32:47
SPEAKER_01: And then there's other people in the business who don't necessarily have the experience, they don't know how to underwrite a multifamily deal in in in Long Beach, they don't understand how a pro forma works, they don't understand how valuations work, they're lazy, they're they're scared of rejection, right?
33:01
SPEAKER_01: AI is not just going to make them into a one weezer, it's gonna push them in the opposite direction, right?
33:07
SPEAKER_01: Because it's gonna get them to the wrong result faster.
33:12
SPEAKER_01: So it it there's this multiplication aspect.
33:15
SPEAKER_01: Um, and because it's a tool, it's happening on on both ends, right?
33:20
SPEAKER_01: But I think for a lot of people thinking about, oh, well, when should I use AI versus kind of do it myself?
33:27
SPEAKER_01: It's really important that you establish a measurable baseline.
33:31
SPEAKER_01: Okay, and so take underwriting, for example.
33:33
SPEAKER_01: You know, Juan, that it takes you, let's just say an hour, right, to do a BOV for a new deal.
33:39
SPEAKER_01: Not going to make assumptions, but let's just use that for the time being here on your own, right?
33:44
SPEAKER_01: Without any sort of AI.
33:45
SPEAKER_01: It takes you an hour, right?
33:47
SPEAKER_01: That is a measurable baseline.
33:49
SPEAKER_01: And so now what you could do and understand is that when you're trying to use AI for that exact thing, right, and you cannot forget the time that it takes to review the AI's output and then make any sort of corrections either, because that's time spent.
34:02
SPEAKER_01: Okay.
34:03
SPEAKER_01: If it takes you an hour and a half, that's not ROI, right?
34:10
SPEAKER_01: That that's negative.
34:12
SPEAKER_01: Okay?
34:12
SPEAKER_01: That is 30 extra minutes of time.
34:14
SPEAKER_01: For what?
34:15
SPEAKER_01: Because it was fun or just it was cool to do that?
34:18
SPEAKER_01: No, that makes no sense.
34:19
SPEAKER_01: However, if it takes you 30 minutes to do that with AI, again, including the time that it takes to review and get the final output, that's alpha, right?
34:28
SPEAKER_01: That's ROI.
34:29
SPEAKER_01: That's time back to you, which you could spend doing other things.
34:32
SPEAKER_01: So that's that that I think is really, really important.
34:35
SPEAKER_01: A lot of people just trying to AI everything, it creates a lot of false productivity, and especially in brokerage, too, where time is money.
34:42
SPEAKER_01: Okay.
34:43
SPEAKER_01: And if you're stuck there, uh what I see a lot of too are these uh brokers who are trying to like vibe code and like build their own CRMs and things like that, and build all these incredible tools.
34:54
SPEAKER_01: It is so incredibly hard to build this stuff to a point where it's really production worthy.
35:01
SPEAKER_01: In other words, where you're actually using it, right?
35:03
SPEAKER_01: There is maintenance involved with this stuff.
35:05
SPEAKER_01: There is a learning curve.
35:06
SPEAKER_01: It is hard, and that is taking you away from the things that will ultimately yield you the results you want, right?
35:14
SPEAKER_01: Which are more deals, which turn into more commissions for you.
35:16
SPEAKER_01: So if you're building a vibe coding a CRM, if you're a broker, that's time you're taking away from making cold calls and having touch points with these owners.
35:25
SPEAKER_00: Well, that's like um uh, you know, I shoot videos and other brokers are are now, you know, five years ago, six years ago, people were making fun of me that I was doing these video stuff, and now every commercial broker is out there trying to do it now.
35:36
SPEAKER_00: So, but what they'll say is um, Juan, are you editing your own videos?
35:40
SPEAKER_00: And I look at them like, are you crazy?
35:42
SPEAKER_00: I'm a real estate broker running a brokerage.
35:45
SPEAKER_00: Well, why would I ever edit my own videos?
35:47
SPEAKER_00: And why would I Vivecode to that point?
35:48
SPEAKER_00: It's it kind of the same thing as you're right.
35:50
SPEAKER_00: You got I'm trying to stick in my lane, right?
35:52
SPEAKER_00: I know what I'm good at.
35:53
SPEAKER_00: I'm not a programmer.
35:54
SPEAKER_00: And uh, man, there's a lot of things for us to learn.
35:57
SPEAKER_00: But I'm cognizant of your time, Jake.
35:59
SPEAKER_00: Um, I'm gonna give you the last word.
36:01
SPEAKER_00: Give us any um anything you want to finish with, but please let us know how we could follow you, how we could support you, your YouTube channel.
36:08
SPEAKER_00: And if someone's listening in today and says, hey, um, I want Jake to come into our office.
36:12
SPEAKER_00: I want him to give us a seminar, we want to put him on stage.
36:14
SPEAKER_00: And by the way, I've had Jake on stage.
36:17
SPEAKER_00: He's tremendous, did a killer job, and the feedback was great.
36:20
SPEAKER_00: And hopefully he'll come in next year, February 2027, um, as we're gearing up for that event again.
36:26
SPEAKER_00: And so, final word, and how can we get a hold of you?
36:28
SPEAKER_00: Yeah, well, I I appreciate you having me here.
36:30
SPEAKER_01: Uh, I hope we do a lot more of these.
36:32
SPEAKER_01: Um, no doubt this technology is changing the way we do business, not just in in in real estate, but but across the entire world.
36:42
SPEAKER_01: We haven't even scratched the surface of it.
36:44
SPEAKER_01: So if you're thinking that you're just gonna sit on your hands and knees and continue doing things the way that you normally do, think again.
36:50
SPEAKER_01: Okay, there's guys like Juan Weezer who are gonna blow past you.
36:54
SPEAKER_01: Um, so really something to think about.
36:56
SPEAKER_01: Just just start, okay?
36:58
SPEAKER_01: Start just understanding how this stuff works um before bed, spend an hour on the weekends.
37:04
SPEAKER_01: You got to practice it.
37:05
SPEAKER_01: Well, they should be spending an hour watching your YouTube channel.
37:08
SPEAKER_01: Or that, yeah, exactly.
37:09
SPEAKER_01: So AI for CRE Collective.
37:11
SPEAKER_01: Uh you can check us out, ai for CRE Collective.com.
37:13
SPEAKER_01: Again, we are the independent publication covering how technology is changing the built world that includes uh real estate, and it's changing it very fast.
37:22
SPEAKER_01: There is this reckoning that is happening.
37:25
SPEAKER_01: Uh, we're super excited about it.
37:27
SPEAKER_01: We're on YouTube, AI for CRE Collective.
37:29
SPEAKER_01: We have a lot of videos.
37:30
SPEAKER_01: Uh, I'm on LinkedIn, Jake Heller.
37:32
SPEAKER_01: You can follow me there.
37:33
SPEAKER_01: We have a community of about 800 commercial real estate professionals.
37:36
SPEAKER_01: And if you want to learn how to use AI uh specifically in commercial real estate, get in there.
37:42
SPEAKER_01: We do workshops, we have prompts and workflows and uh events and education and certification and master classes, all that good stuff.
37:50
SPEAKER_01: So you can check that out as well uh on our website.
37:53
SPEAKER_01: But Juan, thank you.
37:55
SPEAKER_01: Uh thanks for having me, man.
37:56
SPEAKER_01: Really appreciate it.
37:57
SPEAKER_00: We'll do it again until next time.