[00:00:41] Zach Hammer: Welcome, welcome back to another episode of the Real Estate Growth Hacker Show. On today’s episode, we’re going to be talking about [00:01:00] creating AI assistants with GPT’s. So we’re going to be diving into that, giving people all the context that they need in order to have success, understand how to use these things and get into action, leveraging them for success.

[00:01:12] Zach Hammer: Hopefully on today’s episode, again, back with us, we have Charlie Madison, founder, developer, realtor, amazing Christmas holiday shirt, sweater wearing.

[00:01:25] Zach Hammer: Madison.

[00:01:26] Charlie Madison: Ho, ho. Roll Tide.

[00:01:30] Zach Hammer: Indeed. Welcome back, Charlie. What do you think about this topic today about creating AI assistance with ChatGPT and GPT’s.

[00:01:38] Charlie Madison: GPT’s. It sounds like something I might need a penicillin shot for. Is that true?

[00:01:41] Zach Hammer: You know, it depends on if you are allergic. They might have to give you an alternative if you can’t deal with the seals.

[00:01:47] Charlie Madison: Yeah. No, what I like about this is I’ve played around with the GPT’s and I’ve asked you questions about it and I love kind of the promise and the ability. And so you’ve dove into [00:02:00] it and I’ve kept me from some pain and, you know, I haven’t dove further into it probably because I guess I was waiting for this episode.

[00:02:09] Charlie Madison: So, this may help me on my next step. But I’d love for you to kind of share, like, what are the benefits of a really good GPT?

[00:02:15] Zach Hammer: Yeah. So, and that’s the key here. So, the ideal outcome from this is that you’re able to leverage some of that promise of AI to get a little bit further than we had been able to get before. Right? So, before GPT’s you know, we had a really great tool with ChatGPT or GPT for that model from open AI, or even some of the other models from Claude and all the other things that people may or may not have played with already. But you sort of had to know what you were doing maybe a little bit more in order to get a good result.

[00:02:49] Zach Hammer: And so this is hopefully going to help sort of steer the tool in more of a specific direction for us rather than kind of having everything open to [00:03:00] it. Lets us be a little bit more specific with our end results. So, in terms of what we want to cover, what I’m going to dive into, I’m going to talk about, give people the context, what these things are.

[00:03:10] Zach Hammer: A little bit about why it matters and why it matters to talk about it specifically. And then we’ll dive into how to actually, you know, how to leverage them, how to think about this, how to put into work. How’s that sound?

[00:03:20] Charlie Madison: Let’s do it.

[00:03:21] Zach Hammer: Perfect. So in terms of the context, so when we’re talking about GPT’s, I actually, I forget right now what GPT stands for.

[00:03:28] Zach Hammer: It’s like general something

[00:03:30] Charlie Madison: Transformer. All right. Yeah.

[00:03:33] Zach Hammer: Yeah, transformer is the last word. I don’t remember what the GPT part stands for. I can look it up, but again, honestly, it doesn’t matter in terms of what we’re talking about this. What we’re talking about is we’re talking about an instance of ChatGPT, that’s essentially, not exactly fine tuned, but it’s specified to be a specific thing, right?

[00:03:51] Zach Hammer: Where if ChatGPT is designed to be kind of this overall assistant, that’s designed to help you with tasks, and it’s kind of got everything available. These [00:04:00] GPT’s are like, you tell it, no, I don’t want you to be a general assistant, I want you to be very specifically focused on this, I want you to talk it this way, I want you to think about things this way. Right?

[00:04:11] Zach Hammer: Like you’re sort of specifying and narrowing the scope so that it’s better at doing that job hopefully, that’s the idea here. It’s that same sort of interface of like chatting with ChatGPT, but instead of being sort of, generally good at a lot of things. You’re telling it, no, I want you to be an expert at this thing. Does that make sense?

[00:04:31] Charlie Madison: Yeah.

[00:04:31] Zach Hammer: And if you followed some of our other episodes, you may realize, that idea is actually part of what we deploy in our mega prompt framework of saying that, when you want to get a good result from a prompt, the first thing that you start with is you start with that persona.

[00:04:46] Zach Hammer: And you say, who do I want you to act as? Because that ultimately gets you a better result in the answers that you get. So, part of what you should be thinking about for this is that, it’s like having that persona, part of your prompt sort of built in, [00:05:00] right? Where you’re telling it right from the get go, how you want it to act in order to get to that better end result.

[00:05:05] Zach Hammer: So, this as a feature is currently only available to those who are paying for ChatGPT plus it’s really easy to work with and play with in there. There’s a little context window up toward the top where you can go into the GPT’s and start a new one. And then the way that you set them up is you sort of have a conversation with this GPT builder and it starts putting it together for you, suggesting a name, starting to build out the descriptions for you.

[00:05:28] Zach Hammer: You can either sort of use that ChatJAT based interface to, you know, to build it out and have it do most of the work for you. And it does a pretty good job. Or you could go in and you could sort of manually fiddle with the settings either like I find myself needing to do both, but they’re fairly easy to get started and play with.

[00:05:43] Zach Hammer: And then you can actually publish these things, right? So you could publish them. You can make them available either, you know, just privately for your own account, you could publish them and make them so that anybody with the link can use it, or you can even publish it fully publicly where anybody can access it through, you know, this GBT store and [00:06:00] people can use it, can play with them themselves and and use them within their own account.

[00:06:04] Zach Hammer: So there’s kind of a lot of variety here in terms of how you can actually deploy these, but and that’s really where part of the opportunity here is that. At a baseline level, you could set up these GPT’s to be useful for you, right? You could set them up to be useful to accomplish things that you’re doing day in, day out in more of that specific way, right?

[00:06:22] Zach Hammer: Or for those of us that have people that we’re looking to equip, people on our team, people, agents, maybe that are part of our team, this could be ways that we put effort in on the front end in order to create a system that helps equip people on our team to do something that we’ve already prethought through, right?

[00:06:37] Zach Hammer: So, whereas in the past, we might’ve had to develop prompts to help people, right? We’re going to start to be able to leverage GPT’s to maybe take care of some of that legwork, make that a little bit easier. Right? So that’s part of where we’re looking at doing this. And further, if you do this right, shoot, GPT’s might even be part of the way that you help to recruit, right? It could be part of the way that you say, Hey, this is a tool that I’ve built out. Would you come to our team? This is one of the tools that we [00:07:00] give you access to where we’re continually training this thing and making it better and making your life easier as well. Right?

[00:07:04] Zach Hammer: So there’s lots of opportunity for this in terms of the basic concept, it can be similar to having like your own app, right? Like that’s literally part of the the scenario and the framework that open AI is thinking through this as a concept is that they’re seeing people build and develop these GPT’s almost as if they’re apps of the app store, right? The GPT store is like the AI equivalent of the Apple app store, the Google play store, right?

[00:07:29] Zach Hammer: Where you can actually get these specific use cases of this larger AI power at your disposal. And so, yeah, those are kind of some of the foundational ideas about them. Any in terms of just conceptually, anything there that you feel like I should dive into further or that you would you feel is confusing or maybe unclear.

[00:07:47] Charlie Madison: Yeah. I’d love to share kind of what I heard and some questions. And so, kind of, I almost think of this. It’s almost like taking. Like if I just go into ChatGPT, I can give it a super [00:08:00] prompt. I can upload documents to it, it’s got a shorter memory, whereas with these GPT’s, it’s almost like putting the bumpers on a bowling alley, like I’ve got like my vertical now and I can give it a bunch of content, so you know, if I’ve got, let’s use Keller Williams, 3030 Ignite, I can upload all of Ignite, I could maybe upload the teacher workbooks, I could upload maybe some teachings we’ve had in the past. And so now that all of this is in my Ignite GPT’s, now instead of me, Telling the people, hey, go into ChatGPT, upload these documents and create a prompt, they can kind of use any prompt they want because we’ve already loaded the content and that way they can just use the prompt and say, tell me which one talks about Fizzbos.

[00:08:52] Charlie Madison: Is that a good description?

[00:08:56] Zach Hammer: Yeah, potentially. So there’s a couple of different use cases and we’ll dive [00:09:00] into how I would think through right sizing, you know, what you expect from the GPT’s, but yeah, at a basic level it is. There’s a combination of giving ChatGPT some superpowers. Part of it is about really focusing it’s powers into one way.

[00:09:16] Zach Hammer: And so, yeah, there’s a blend of ideas, that definitely lets me know kind of what to cover here. So in terms of why I think this is important to talk about is there is potentially a little bit of confusion on what you just said that I think is worth addressing in terms of what can we reasonably expect right now versus what’s on the horizon.

[00:09:33] Zach Hammer: So what people are seeing when they’re looking at the GPT’s, is there starting to think about them as if they can become like a fully trained assistant that understands these things completely. And unfortunately, the reality right now is that, these GPT’s are able to be made really effective in a narrow vein.

[00:09:55] Zach Hammer: So if you think about what it is that this thing does, needs to be somewhat [00:10:00] of a narrow field, a narrow view in order to really be able to do it well and do it consistently and so, you know, there’s a lot of opportunity in building something that becomes a lot easier for somebody to use, but whereas people are typically expecting that this is going to be like the equivalent of training my own specialized HAL, or my own specialized C3PO, right? That’s going to be able to really adapt to me. It’s not quite that, it’s a little bit more like, I don’t know, have you ever watched Rick and Morty?

[00:10:36] Charlie Madison: Yes.

[00:10:36] Zach Hammer: Yes. Okay. So, it’s a little bit closer to building the robot whose job it is to deliver the butter, from there to here, right? Where it’s going to be able to do that effectively and adapt to that scenario, but it’s really designed to just deliver the butter and you more so want to think about building yourself multiple of these GPT’s that are designed to do specialized [00:11:00] tasks. Now it’s a little bit more adaptable than that, but that’s the thinking that we want to get down into is you’re building pieces of a bigger puzzle rather than building something that’s able to solve the whole puzzle in one go. Does that make sense?

[00:11:10] Charlie Madison: Yes.

[00:11:10] Zach Hammer: Just at least conceptually.

[00:11:12] Charlie Madison: Yeah.

[00:11:12] Zach Hammer: All right. Perfect.

[00:11:13] Charlie Madison: So.

[00:11:14] Zach Hammer: And,so go ahead.

[00:11:15] Charlie Madison: So my question is, thinking about training my real estate team. So is answer questions about, you know, let’s say the ignite course, you know, it’s I can’t remember, it’s like 16 weeks is that narrow enough versus, you know, tell my team about any of our education sources on, you know, is more broad or tell only about the FSBO course is, like, is that a good middle ground? Like what, how do you decide that?

[00:11:48] Zach Hammer: Yeah. So here’s how I think through this. So first off in terms of the context around this, what I see people missing is they get too excited about what might be possible. And so [00:12:00] they get disappointed when it can’t do that big vision. And we will get there, we will get to a point where I think a lot of these things come together to be sort of this super powered thing that thinks even smarter than us on some of this, right.

[00:12:12] Zach Hammer: And being able to pull together a lot more, a lot quicker than we can do. But the other side is not knowing when to leverage this, right? So at the, you know, the smaller end that maybe you’re using older ways of doing this, where you can actually make your life easier by deploying as a GPT rather than what you have been doing.

[00:12:29] Zach Hammer: So that’s what we’re going to be diving into is how to right size this, that you’re getting the maximum benefit by knowing what it is capable of doing without being let down by expecting too much of it. So in terms of how we do that, really the foundational sort of framework that I like to think about this is actually really similar to how I think through setting up an SOP in my business, which is, if I’m going to put together a training on a concept. I have a mental framework of saying this process should be able to [00:13:00] be clearly explained in 15 minutes or less, right?

[00:13:03] Zach Hammer: If it can’t be explained in 15 minutes or less, then it’s too complicated of a process and I need to break it up into multiple processes.

[00:13:11] Charlie Madison: Yeah.

[00:13:12] Zach Hammer: And so with that being said, I can explain how to do a process within 15 minutes that maybe the process itself can’t be completed in 15 minutes, but it could be understood effectively in 15 minutes.

[00:13:27] Zach Hammer: So a good example of that would be, Hey, I’ve got this library of every training imaginable. And my process is on how to find the right training for the right topic at the right time. And I might be able to explain that clearly in 15 minutes. But the library itself is much more vast than that, it may not be able to understand, like, all of the content in that library, but if I point it to the documentation that says, Hey here’s this course, here’s what it’s [00:14:00] about, here’s kind of the ideal outcome, then yes, like your sort of scenario of saying, Hey, let me upload all of these trainings from ignite and be able to tell people, I want to learn this, where should I go to dive in, that’s going to work well, but it may not be able to actually do the teaching, right?

[00:14:17] Charlie Madison: Right.

[00:14:17] Zach Hammer: May not be able to say, here’s what you need to know about this topic from step A to step B and how to implement it and now build an email template off of it, right? Like it may not be able to do that well and it’s going to be having to make a lot of guesses in that process.

[00:14:30] Charlie Madison: But it may like point to a video of someone already teaching it or something.

[00:14:35] Zach Hammer: Exactly. And so there’s a little bit of a fluidity here because it is adaptable and it will try to do things on your behalf. It’s just really where you would question how good of a job is it going to be able to do. So that’s the way that I like to think about this is I like to think through it as if I can give it. You know, a series of templates or like a document that sort of has broken down chunks and the process itself could be fairly clearly [00:15:00] explained. It’s then going to be able to take that, adapt it, remix it and give you some really solid stuff.

[00:15:04] Zach Hammer: So like some good examples. I can potentially, whereas in a you know, historically, I would have had to build a prompt that you’re like one prompt for generating images and one prompt for developing ad copy and one prompt for doing that. I could potentially build a GPT that can help me do a lot more of those things. As long as I do a really good job of structuring what I upload into it so that it’s able to pull on demand different concepts and sort of a bigger prethought through process.

[00:15:36] Zach Hammer: But if I don’t do that, if I don’t do the work of giving it those different steps, the ways to think about that, and I’m not able to communicate that process clearly or upload all the pieces in a solid framework, then it’s not going to be able to do a good job of implementing on that. Does that make sense?

[00:15:51] Charlie Madison: Yeah.

[00:15:51] Zach Hammer: Like, so it’s really, it’s all about. Like, because that’s the key, essentially what these GPT’s are, if you’ve [00:16:00] played with ChatGPT so far and you’ve played with the system instructions and you’ve played with the custom instructions and the context, this is the, it’s like a more advanced version of that.

[00:16:11] Zach Hammer: It’s you being able to tell it kind of how it should act. And giving it a knowledge base to be able to operate out of as well. And so the limitations currently is that you could upload about 20 documents. I forget exactly how big they could be, but if you’ve got a massive library that you know, it’s going to be too much to fit in inside those 20 documents and you’re trying to accomplish, you’re probably trying to accomplish too much, but if you could really cool down, like, what are the key elements of this and what’s the documentation that’s necessary, right?

[00:16:41] Zach Hammer: Where it’s like, if I want something, that’s going to help me write follow up sequences. Well, hey, that might be something that I could do right? Because I could say like, hey here’s a document that shows some examples of the types of follow up sequences that work or the types of sequencing, or you know, explaining like, we build [00:17:00] our follow up sequences in a Fibonacci Sequence for these reasons. Right?

[00:17:04] Zach Hammer: And so it’s going to give you the timing, it’s going to give you what style of messaging works the way, when you go out further maybe you’ve got documentation that explains this is you know, a cold outreach followup sequence versus this one is a long term nurture followup out you know, sequence and what sort of messaging makes sense depending on those, right?

[00:17:24] Zach Hammer: Like that sort of documentation you might be able to do, so that when I go into the GPT and I ask, I’d like a long term follow up sequence for people that already know me, it’s going to have the context in its database to actually be able to pull from that knowledge base and then build based off of that context that you’ve already given it to prove, right?

[00:17:45] Zach Hammer: So, that’s sort of the idea there is the same way, like, if I was going to teach somebody, Hey, I want you to be great at writing follow up sequences, here’s what your process looks like. Your process is that you go to our library of follow up sequences, [00:18:00] and you ask yourself, what type of follow up sequence am I wanting to write? You go and you reference our templates, and then you start writing based off the templates of what’s already been proven. I could teach you that process in 15 minutes.

[00:18:13] Charlie Madison: Right.

[00:18:13] Zach Hammer: Right? You might still have a bit of review in terms of being able to leverage that. But I could teach you that process of going to the library, pulling out the relevant examples, and then writing based on the inspiration of those examples.

[00:18:24] Zach Hammer: That’s sort of the idea here as well is, if the concept that you’re looking to achieve could be trained effectively in 15 minutes, then you’re probably going to be able to get a good result from a GPT as well. Does that make sense for sort of how to think through the level of a task that you might be able to get it to do?

[00:18:40] Charlie Madison: Yeah, can you describe it in 15 minutes? It’s kind of a good over.

[00:18:43] Zach Hammer: Exactly.

[00:18:45] Zach Hammer: And that’s a good rule of thumb, right? Where typically if you’re able to do it in that timeframe, that’s probably the right size. So as a, for instance, I’m probably not going to be able to put together a good GPT that is going to be able to do great [00:19:00] work. Writing the ad, creating the follow-up sequence, creating the landing page, creating the phone scripts, or because now I’m starting to get into a cons into a level of training. I’m probably not gonna be able to adequately explain it well in 15 minutes, right?

[00:19:15] Zach Hammer: So I probably need a separate GPT for generating the ads, and I need a separate GPT for generating a good landing page and a separate GPT for putting together that follow-up sequence. Right?

[00:19:27] Charlie Madison: Right?

[00:19:27] Zach Hammer: But. Overall, I can know, okay, here’s the steps in this process that I need to achieve. Let me put together separate GPT’s for each of these major components and be able to make that easy. And so that’s one of the key things for sure. The other thing that a GPT is going to be really good for is if you’ve already been using a Megaprompt framework. GPT’s are going to be really great for basically taking that megaprompt framework and sort of putting it behind the scenes so that somebody just shows up and interacts [00:20:00] with the GPT as sort of a pre trained model to follow, right? So I don’t have to train somebody on doing a megaprompt.

[00:20:07] Charlie Madison: Right?

[00:20:08] Zach Hammer: I don’t have to train my agents on creating their own megaprompts. I can say, hey, here’s this GPT that’s going to help you do this thing. Come to it with this sort of information, these sorts of questions, and it’s going to help you be able to deliver on that effectively.

[00:20:19] Zach Hammer: Because it’s already going to have the persona, the task, the goal, the steps and what you’re coming to it with is just the here’s the inputs that you need. You need you know, what’s the topic that you want to talk about and what’s the context, right? Like that. And you tell people that you say, come do it with this and it’s going to be able to give you good results based on what it’s already trained on.

[00:20:41] Zach Hammer: And so it’s like the equivalent of being able to give somebody the end result without them having to copy and paste to understand make modifications. It’s able to adapt more on the fly for you on behalf in a way that you could share without sharing the prompt itself. Does that make sense?

[00:20:57] Charlie Madison: Yeah, so you could possibly like [00:21:00] have a one of your Realtors upload like a Zillow link and it could create a different description of the house, you know, like this is prelisting, just to have something nice, create a different description, go through each of the images and create a description for each image.

[00:21:22] Charlie Madison: So you can have like a listing marketing GPT’s. Like it’s not going to do everything, but it might could like create a flyer, create the open house flyer, maybe create an ad for Facebook, like just all of this in this one little vertical, but it’s not going to help you auction a property.

[00:21:40] Zach Hammer: Right. Exactly. And it may not help you get the property, right? Like it may not also be able to put together the listing presentation and also do these things that are sort of a different frame of mind and the different set of templates that you would need. And so like that’s actually a really good example. Some of those things where you’re looking to get it to interact with the internet, you’ll have to actually [00:22:00] test that out and see if it works like you’re expecting or not.

[00:22:03] Zach Hammer: Cause I find that it’s a little bit hit or miss where it can interact with the internet, but the way that it tends to do it seems to be less visiting a specific link or some links work and some links don’t. It sort of depends on how that website structured. I mean, you as a developer are more familiar with that sort of concept. I don’t know exactly how it’s always trying to load. I know, half the time, it seems like the way that it wants to do it is it wants to do a Bing search first around the concept and then try and access things from the search rather than directly going and visiting links, if that makes sense?

[00:22:40] Zach Hammer: So, and again, all of this is changing over time. So in terms of the specifics, this is one of those areas where you don’t want to assume that the idea is going to work. You want to get in and you want to say, I think I might be able to do it this way. Let me put together a test and see if it does what I would expect it to.

[00:22:59] Zach Hammer: [00:23:00] And just like with other things right now, the key is to try and test and figure out how can I prove that it’s wrong? If it was wrong, right? Like go to this page and first, I want you to tell me the most recent details about beds, baths, this sort of thing, and you want it to be very clear, very exact, because if it can’t do that, then it’s probably not seeing the same page that you are, it’s not accessing it in the same way that you are, and so you won’t be able to rely on it to do the listing description as well. Does that make sense?

[00:23:33] Zach Hammer: So you want to start with some of those tests first to make sure that it’s performing as expected. But yeah, those sorts of ideas are fantastic in terms of what can be done GPT’s can be built, having a knowledge base of stuff that you pre upload that could be a text documents, PDF’s even images.

[00:23:49] Zach Hammer: You could pre upload images and it could reference those images and then seek to understand them because it can understand images at this point. So you could do that, it can you can leverage Dolly in it. So it can [00:24:00] actually create visuals for you. Now it’s going to have the same limitations as Dolly already does.

[00:24:06] Zach Hammer: So if you’re not familiar with how you like make images with Dolly, what can do well, what it can’t do well, it’s going to have those same sort of limitations. But you can have it, be designed to create images for you through what already works well, it’s also going to have I forget what they call it now, it’s the code interpreter though, which is basically, it can execute Python code and interact with data through that code.

[00:24:29] Zach Hammer: So if it could be deployed through Python, a lot of those things can actually be done within there as well. That could be really useful for things like data analysis where maybe somebody uploads a CSV that has some data and it could take that data and then process it. It could extract some meaningful data correlations run numbers to say what average price points were average days on market, it could do things like that.

[00:24:50] Zach Hammer: So you’d want to experiment and see what’s possible, but it can do some of those things as well. So not only do you have the large language model where it’s able to understand the language, you can actually have it process [00:25:00] data and give you interpreted results based on that.

[00:25:02] Zach Hammer: And you can have it generate images and you can have it access the internet. So like you could one of the tools that I’ve seen somebody do for instance, is build a GPT that’s designed to help you create great real estate videos. And one of the things that it’s trained to do is actually be able to go out and look for updated market stats to give you a script, to create relevant to your area, current market data about what’s going on in your area and put together a script and a template for you to go through and talk about what’s going on in your market based on data that it’s actually able to research on your behalf, right?

[00:25:34] Charlie Madison: That’s pretty cool.

[00:25:34] Zach Hammer: So, like, yeah. So there are things like that, that you could build out and, you know, if you’ve got a flow, you know, you could start to think through what is that flow? Let me document that into a GPT. And then I could actually guide somebody through that process you know, pretty readily where they just show up with their unique aspect, their unique spin, their human approach their opinion, where that’s relevant matters.

[00:25:53] Zach Hammer: But again, narrow that task down to that 15 minute explanation and that’s about what a GPT could be expected to [00:26:00] do. Does that make sense?

[00:26:01] Charlie Madison: Yeah. I like that.

[00:26:02] Zach Hammer: And so where we go from there is that right now, that’s what GPT’s are. And so, to sort of give people a little bit of maybe what’s on the horizon, what we could start to expect, who knows the timeframe, some of this stuff, like I would imagine.

[00:26:16] Zach Hammer: We’re seeing some of this at least by next year, and maybe sooner, right? Cause this stuff goes really quick. But part of what I would expect to see is that as these GPT’s are getting trained, we’ll probably start to see what they’re calling GPT swarms, where you have a centralized thing that knows that these different sub-processes exist, these different GPT’s, and then it intelligently leverages those GPT’s in order to create, you know, in order to go after a bigger task, where maybe you are looking to put together a full marketing strategy to get the listing, market the listing, etc. We have a specific tool that’s designed to go out, leverage those smaller tools to do the [00:27:00] sub-processes and then go out and do bigger things. Right?

[00:27:03] Zach Hammer: But, we’re not there yet, that kind of stuff isn’t ready or easy to do, but GPT’s are still really powerful right now. As long as you’re right sizing your expectations for what they could be doing. And again that’s back to that. Can I train somebody on how to do this process in 15 minutes?

[00:27:19] Zach Hammer: Where they at least know where to go and get the information and understand how to use that information in a 15 minute explanation, and if not, then you’re probably needing to break that up into multiple processes.

[00:27:31] Charlie Madison: Yeah, you know, I imagined wanting to empower my realtors to use ChatGPT. I mean, right now it’s a multi step process.

[00:27:39] Zach Hammer: All right.

[00:27:40] Charlie Madison: Like here’s your master prompt and here’s the files you need to upload. Here’s it’s resources and now you do all that.

[00:27:51] Charlie Madison: What I imagine with this is now you can create that same thing. This is just a super prompt, it’s a vertical, but you can have multiple, you know, Hey, [00:28:00] when you’re at this part of the listing or when you’re at this part of the buyer process, like they can have them and all they got to do is talk in their normal language and I think one of the things too is when you create your own GPT’s don’t you get to see a log of all the questions and responses that people have used. Don’t you see a history?

[00:28:21] Zach Hammer: Yeah, I forget exactly what all is accessible from when other people are using it or not. What you might be thinking of is there’s a very similar thing where it’s the same technology, but you could build them to be leveraged by the API. And if people are interacting with that, definitely you could see the log of what’s happening. I’m not sure if it’s available for the GPTs’ or not.

[00:28:41] Zach Hammer: I don’t know if you get access to all those questions for the GPT’s themselves. Conceptually, yes that idea is available, but I think you have to do that through the API rather than, through just the standard GPT builder.

[00:28:51] Charlie Madison: Gotcha.

[00:28:52] Zach Hammer: Anyway but yeah, so leveraging that though, you would be able to start to refine it, get context and that sort of thing.

[00:28:58] Zach Hammer: And I could be wrong. I need to double check [00:29:00] to be sure. But I think that’s just through the API that you could get that level of seeing what’s happening.

[00:29:05] Charlie Madison: Well, it’ll be nice if they, once they allow everyone to see that would be pretty cool stuff.

[00:29:11] Zach Hammer: Yeah. It’d be really useful to be able to refine it, make sure of like, what should I make sure that there are answers for and refine that further.

[00:29:18] Charlie Madison: Yeah, absolutely.

[00:29:19] Zach Hammer: But yeah, so in terms of like next steps for what people should do to leverage this, I mean, this kind of information, I recommend going out thinking through what is something right now that is a 15 minute explanation process that I can make easier for somebody on my team, make easier for the agents, that are part of my team, make easier for myself, right?

[00:29:37] Zach Hammer: What’s something that you go to that maybe, it’d be useful to have somebody that’s sort of pre-trained on your templates, your processes, so you don’t have to go and look them up to find them. But it’s just available and aware and knows how to do that. What are those sorts of tasks that you could start replacing with something like a GPT?

[00:29:54] Zach Hammer: I know that you’re building the skillset and building the processes into your business, that over time, it should just start to get [00:30:00] easier. The more that you’ve built these out, you’ve built out this custom library for yourself. Now, if that’s something that you want to make easier and you want to do this alongside people that are in the trenches doing this as well.

[00:30:10] Zach Hammer: That’s exactly why we put together our mastermind for teams that are looking to implement these AI tactics in their business. We’re already at work building out these GPT’s for things that we see commonly being necessary. Whether that be the listing description, the listing marketing whether that be the Facebook ad generation or the listing presentation adaptation, the scripts around markets, and those sorts of things.

[00:30:34] Zach Hammer: We are in the trenches building these things out ready already so that you don’t have to build them and you get to just leverage them already, that you get to just have the GPT’s available as we’re building them out and further, we’re listening, we’re hearing, what do you need? What do you think? What do you see as an ideal opportunity for something that I feel like I’ve got a process that it only takes 15 minutes to train on.

[00:30:55] Zach Hammer: But man, if we could have that be easy for people, it would make that process so much easier where they could [00:31:00] just have it ready to go. We’re building these things together in the trenches each week, getting down into work, making sure that we’re implementing this stuff together.

[00:31:07] Zach Hammer: If that sounds like something that you want to do, where you want to be on the forefront of what’s going on, what’s working right now in AI, and Getting to have me in your corner, helping you make those things happen, as well as others, you know, top successful Real Estate teams that are building this stuff out, that’s what we’re doing in our AI mastermind.

[00:31:23] Zach Hammer: If you want more information about that, you can find out about that by going to RealEstateGrowthHackers.com/Contact, let us know that you’re interested. We’ll get you more details and and see if it happens to be a good fit for you as well. But yeah, that’s the key next steps that any final thoughts on GPT’s, anything that’s clearer now than it was before when, you know, in how you’ve played with them in the past.

[00:31:42] Charlie Madison: Yeah, I like the thought of just the 15 minutes. Like, I mean, you know, imagine if you could take your SOP’s, and, you know, for all of your agents, like, Instead of them coming to you or, you know, the office admin, like, how can you make it simpler where they’ve got [00:32:00] one place to go for each thing?

[00:32:02] Charlie Madison: And like you said, you know, right now they’re individual, but, you know, we’re not far from a place where, you know, it’s almost like a phone directory. Like, you know, where, you know, you just choose which one you want. And what’s cool is this isn’t just like a Wikipedia that’s regurgitating stuff. Like this responds creatively, like this thinks in a way, you know, like it’s unique in that way, which is, you know, one of the things I always wanted in real estate is I don’t want each of my team members to just have exactly the cookie cutter thing.

[00:32:37] Charlie Madison: I want it to be the same thing, but I want it to be unique. And now, you know, you can do that, which is really cool. And it’s just amazing how quick Each of these happens and like, I just imagine, you know, all the folks in your mastermind, like, you know, like they’re going to have different things that they’re working on, you know, at the end of this mastermind, everybody comes together.

[00:32:57] Charlie Madison: There might be like 30 to, you [00:33:00] know, each person had three things. Now you got 30 of them. Like, that’s really cool. It would take how many years to do that one on one.

[00:33:08] Zach Hammer: Right. Exactly. And that’s the key is that like all of this stuff is stuff where we’re looking at not just, Hey, this is cool and fun to play with, but really looking at it from the perspective of what is the day by day work that you were doing in real estate? How can we make that process easier? Right?

[00:33:24] Zach Hammer: Most people seem to be leaning on fun, quirky tricks that are interesting and sure, they’re fun to play with, but they’re not solving real problems. They’re not looking at what are you doing day by day? How do we make replace parts of that process with AI to either make it go further and make it more leveraged or make it take less of your mental creativity each day to make that work happen.

[00:33:44] Zach Hammer: And we’re putting that stuff into practice. In that team environment, either on behalf of the agents in your team, on behalf of your staff, making sure that processes that you’re putting into place can be done more at scale or done at a higher level, more consistently wherever it makes [00:34:00] sense to apply that stuff.

[00:34:01] Zach Hammer: AI is definitely right now it’s able to take things that previously took a lot of human creativity and you still need a human at the helm, kind of controlling this thing and making sure that both the input and the output is thought through, but that middle work, that sort of just taking the process and running with it.

[00:34:19] Zach Hammer: AI is really good at doing that and then getting a little bit of final polish and, you know, actually being able to go out into the world for you. But yeah, so that’s what we’re doing. If that’s interesting to you again, RealEstateGrowthHackers.com/Contact, reach out to us, let us know you’re interested and we’d love to connect you with information on how to get started.

[00:34:36] Zach Hammer: But in the meantime, definitely get out there. Start playing with GPT’s. I mean shoot, hopefully this episode is relevant for more than five minutes and and people will get value from it get to learn from it. But man, I’ll tell you what, that’s the key. This AI stuff is moving so quick.

[00:34:51] Zach Hammer: You know, we, you know it, the principles keep being relevant but the specifics and the implementation are often changing just [00:35:00] slightly as you go. And it’s not a matter of the right tool, it’s not a matter of the right, like, you know, what’s coming out after ChatGPT, it’s so far, that’s not what it’s been about.

[00:35:09] Zach Hammer: It’s more so been about wrapping your head around. What are the opportunities? What are the key things to actually tackle what’s working right now? And being able to sort of have that group of people to rely on that are in the trenches, actually figuring this stuff out and figuring out what works really right now versus what’s on the horizon, because none of that’s been especially clear, right?

[00:35:31] Zach Hammer: There’s lots of things that we don’t know what’s possible because we haven’t tried it, but then we try it and we see what’s possible. And there’s lots of opportunity, but some things become a lot easier, some things still really require the expertise and it’s getting into the trenches and figuring out what’s working that seems to be making the difference for those who are actually getting the biggest leverage out of AI.

[00:35:51] Zach Hammer: So, yeah, there you go. Well, with that, I am Zach Hammer. This has been the Real Estate Growth Hacker Show. With me, I have Charlie Madison. If you are looking [00:36:00] for awesome strategies, tactics and tools to help you build a waiting list of realtors or really get in front of your sphere in a way that’s super powerful.

[00:36:09] Zach Hammer: You should definitely check out what he’s up to at RealtorWaitingList.com or ReferralsWhileYouSleep.com. If you want to go and connect with him and make sure that it’s clear that he came through The Real Estate Growth Hacker Show, you can go to RealeEtateGrowthHackers.com/ReferralsWhileYouSleep.

[00:36:25] Zach Hammer: I believe is what we have set up currently. So if you go there, then that’ll be a special friends of Zach link and then, you know, he’ll make sure to to say, Hey,

[00:36:33] Charlie Madison: I’ll take care of ya.

[00:36:35] Zach Hammer: Yeah,

[00:36:36] Charlie Madison: Mention the real hammer.

[00:36:39] Zach Hammer: Indeed. So there you go. Yep. That’s that’s GPT’s a good overview for you. And yeah. If you guys like this content, be sure, let us know if this was helpful to you, let us know honestly, I thrive on that feedback of knowing if this is helpful.

[00:36:53] Zach Hammer: If this is useful reply, drop comments, send emails, let us know that this stuff’s helpful for you, cause [00:37:00] honestly, if we go long enough without hearing people tell us, Hey, this was useful to me we’ll probably just stop creating that kind of content.

[00:37:05] Zach Hammer: So if you found value in it, reach out, right? I need to know that it’s actually helping people to know that it’s worth to continue doing. But yeah, there you go.

[00:37:12] Zach Hammer: Thanks so much. We’ll catch you on the next one.

[00:37:14] Charlie Madison: Bye.