[00:00:00] Zach Hammer: Welcome back to the Real Estate Growth Hackers Show, on today’s episode. We’re going to be walking through my process for creating the perfect prompts. The whole idea here is, how do I develop my prompts for AI that get me the result that I’m looking for? I know a lot of people struggle to, you know, put together prompts that actually get them the result that they’re looking for.

[00:00:20] Zach Hammer: That maybe output consistently or get that nuance that they’re looking for. And they feel like, the problem is with AI. But really, the problem might be with how you’re structuring your prompt? That if you’re not including enough detail, maybe in some cases, including too much detail. Regardless, that there might be a way to massage what you’re doing in order to actually get a prompt that gets you the result that you’re looking for time after time.

[00:00:44] Zach Hammer: So that’s what we’re going to walk through. We’re going to walk through what my process looks like. I’m going to open the kimono and reveal the secrets to, what I’m doing to make my prompts. Hopefully it helps you get clarity on what that process looks like for yourself as well. With me today, I have Charlie, as well as Charlie’s kids in the background, [00:01:00] what’s up? Coming to you today to talk about this, Charlie. I see, you’re back in the Hawaiian shirt today. I like the yellow and the sailboats, it looks great.

[00:01:06] Charlie Madison: Yeah, I’m visiting my brother down here in Jasper, Alabama. So, since I’m feeling vacationy, what am I kidding? I try to wear a Hawaiian shirt every day but it seemed appropriate.

[00:01:17] Zach Hammer: Every day’s a vacation, I love this. Before we started recording, you had mentioned a couple of questions about just thinking through, like this process. Can you help pull us into this process? What were your questions in terms of what comes to mind when I bring up this topic?

[00:01:29] Charlie Madison: Yeah, you’ve got really amazing prompts for everything I can think of. Like, the way that I write a prompt, if you guys don’t know, is I’m at Zach Hammer, what is the prompt you have for X? That’s the way that I write the perfect prompt. Part of it because, it takes a skill and patience and mindset, because I can get like, 60% of the way there or 90% of the way there and then it’ll go completely sideways and now we’re back at square [00:02:00] one. So, you’ve got a thorough process that you’re going to go through and explain here. What were the hurdles? What were the things that you had to jump through to actually take the time to develop this process?

[00:02:15] Zach Hammer: Yeah. It doesn’t actually feel like hurdles. It feels like a fairly natural part of my process. But maybe understanding why that is could be helpful. So we’ve mentioned it before, I forget if it was an episode of my show or yours. But we talked about the idea in Ender’s Game, where Ender is in the battle school and he takes out this kid, Bonzo Madrid. And what he described is he described, I didn’t want to win this battle, I didn’t want to win this fight.

[00:02:39] Zach Hammer: I wanted to win all the battles to come. And I found that’s a natural element of who I am, that I actually suck at doing something consistently, right? If I have to do the same thing over and over again, I don’t do a good job with that. But I’m really good at diving into something and solving it now, so I never have to solve it again. [00:03:00] I am aggressively lazy.

[00:03:03] Zach Hammer: I will work incredibly hard so that, I don’t have to keep doing the same thing over and over again. And it’s not that I’m afraid of hard work. There’s something about the way that I show up in something. That if I could do the work now and then not have to think about it again in the future, I’m going to do that every day.

[00:03:18] Zach Hammer: Cause that’s really what works for me. I’m really good at building a machine that takes care of the rest of the process for me. Because I’m not good at just continually showing up and doing that thing. And so, that’s really where I see AI coming into play is that, it’s a really useful tool for somebody just like me, where I know how I want something done systematically. I know how I want it done every day. I’m just not going to be that person to do it. That sort of created in me this process to be willing to do the hard work now, so that I don’t have to do the hard work later. And so I put in that extra effort knowing that it’s going to pay off, even if it’s not for me, right?

[00:03:55] Zach Hammer: I think, this came out of my own. I’ll do the network now [00:04:00] so that I don’t have to do that aspect of the work later. But really for me, it’s also paid off where I seek to create things in a way that my team could use, right? So that I could do the work now, so that they don’t have to do the work later. So that, the work that they have to do is lower or less or easier or more consistent or more scalable, right?

[00:04:17] Zach Hammer: Like all of those ideas that I could put a lot of work in on the front end in order to get that process on the backend. Does that make sense?

[00:04:23] Charlie Madison: That does make sense, aggressively lazy.

[00:04:26] Zach Hammer: Indeed. And so, really that’s where this process comes from is, like, consistently show up and do the same thing over and over. I have had to get good at how do I build systems that do the work for me and AI has just been a really good fit. And setting up and learning how to do prompt engineerings that I can make a prompt once that takes care of a task indefinitely for the future. And that’s really where this comes from. Over time, I’m also very voracious researcher. So I do a lot of research into what’s working, what other people are getting success with. And then I’ve also learned to [00:05:00] test things pretty aggressively as well.

[00:05:01] Zach Hammer: And really, that’s where this process comes from. It’s a combination of me researching a lot of things but more importantly, testing a lot of things. Taking the ideas that people have said, trying them out, seeing what actually works in the real world. And that’s where I’ve come down to this system of what works for me to develop prompts that consistently get a good, high quality result time after time. And really knowing what I can expect of my prompts of AI and how to string things together and what I can’t. And continually growing and learning and improving over time. That’s where this process comes from. Anything there that you think I should dive into more before we go into the system itself?

[00:05:38] Charlie Madison: So, one thing I got is that, you are aggressive, aggressively lazy, aggressively searching and aggressively testing. What I find fascinating about your prompts, is that AI as we know it, is not built to give consistent results. AI as it is known, [00:06:00] like we’ve got the temperature, where 0 is completely random and 1.0 is the same over and over. And it is dialed to be like a .8, it fills random, so it fills alive. And the fact to me, what makes your prompts amazing, is that they are consistent even though not identical. It’s got the life of AI without the craziness of AI.

[00:06:25] Zach Hammer: Yeah, exactly. And there’s some strategy to that, right? Like sometimes, the right way to say something for a social post or an email might be to intentionally use repetition, right? And so AI naturally wants to not do that. And yeah, you do sometimes have to force it into submission. And yeah, so let’s go in and dive into some of, really what this looks like? So first and foremost, I’ve done a couple of episodes at this point on my mega prompt framework. At this point, it’s changed and adapted over time.

[00:06:54] Zach Hammer: I’ve expanded some of the ideas, I’ve clarified some of the ideas. But that’s the foundation [00:07:00] here is that, I start with understanding what I would call like, the core components of a prompt, right? And those core components are role, task, goal, steps, context, constraints, format, tone, style, templates and examples and user fields.

[00:07:19] Charlie Madison: Just those, just those things.

[00:07:21] Zach Hammer: So all of them serve a purpose. They may not always be needed but when they’re needed, they’re very necessary. And it’s useful to just understand where they come into play depending on what you’re looking to achieve. Role and task are probably the two most definite, you need them most of the time, right?

[00:07:41] Zach Hammer: So role is where you’re telling the AI, here’s what I want you to act as, right? And you get provably better results when you say, I want you to act as a copywriter rather than just a general assistant, I want you to act as an expert videographer rather than just general AI. That when it narrows its view down toward somebody who would have an [00:08:00] expertise in what you’re looking at it to achieve. The results that you get are better, more consistent and more in line with what you’re likely looking for. Task is just really pretty simply, it’s what do you want it to do, right? So you’re acting as this person and I want you to do this thing. This is the thing that I want you to do for me, right?

[00:08:15] Zach Hammer: Pretty much, every prompt is going to have those in some way. A lot of people are just doing the task, at least include the role and you’re going to get better results. The next one is goal. And goal is really, it’s one of our first attempts at narrowing and giving clarity to the AI on what success looks like.

[00:08:34] Zach Hammer: And letting it know, here’s what we’re trying to achieve, right? It’s not only, here’s the task that I want you to do but here’s why. And that sort of helps it to know when it does it right, it’s going to meet those needs. It’s going to achieve that goal. And that gives clarity of what success looks like and it’s output that gives you a better result.

[00:08:52] Zach Hammer: Steps is us taking tasks and really defining a process for what it should do. Defining that process gives [00:09:00] a better, more consistent result as well. Step one, I want you to understand this context that I’m giving you. Step two, I want you to take that context and I want you to create an outline.

[00:09:07] Zach Hammer: Step three, I want you to fill out that outline. Step four, I want you to take what you filled out and I want you to massage it a bit, right? Like that sort of clarity of walking it through a process, getting it to think step by step, get some more consistent results in the end run. Those are probably foundationally, some of the most important aspects. Some of the other stuff is really how you start to massage it.

[00:09:25] Zach Hammer: Any questions on what we’ve covered so far?

[00:09:28] Charlie Madison: Nope, that all makes sense.

[00:09:29] Zach Hammer: Perfect. So continuing on, we’ve got context. So context, this sort of sits outside of the specific tasks, right? Where it’s like, context might be telling AI a little bit more about you. It might be telling it a little bit more about the nature of the thing that you’re doing.

[00:09:44] Zach Hammer: As for instance, I’ve been working on prompts to create social content for my book club. And part of what’s important for it to know is to understand some context about the book club. That we meet by weekly and who it’s geared toward, what sort of things we cover, what kind of books we read, right? Like, that’s not [00:10:00] necessary for the task itself. But it’s necessary for it to properly understand how to word things, how to word like a call to action, that sort of thing. It might also be necessary for AI to understand who you serve, right? Who is this geared toward? Is it geared toward moms?

[00:10:14] Zach Hammer: Is it geared toward real estate professionals? Is it geared toward amputees who also like dressing up as clowns, right? Whoever it’s geared toward is important for it to understand how to craft messages or run processes. It’s not always necessary but when it is it’s incredibly helpful.

[00:10:31] Zach Hammer: So any context like that is useful where you just giving an understanding about things surrounding whatever you’re doing. The next one is constraints. So constraints are going to be things like, in real estate, adding information, adding listings to the MLS is incredibly specific.

[00:10:46] Zach Hammer: Sometimes there’s really clear requirements. Some of them are technical in terms of things like character counts. Some of them might be like, I don’t know if legal is the right word, but the rules, right? Like you can mention this or you can’t mention that. [00:11:00] And you could be facing violations or issues in that way. And so constraints are anywhere that you’re very clearly laying out and defining. These are the are the boundaries that you need to exist between, right? I want this many sentences. I want the sentences to be no longer than this. I need character count to be less than this, et cetera.

[00:11:18] Zach Hammer: And do keep in mind. What you just talked about earlier is important here. AI is good at trying to comply with what you described there but it may not be perfect. So do note, if something is like mission critical, you might need to have a human double check it. But it’ll at least get you closer to that zone. If you say, I want it to be 500 characters or less, it might give you something that’s 510 characters, right? So you might need to massage it a little bit further. It might give you something that’s 400 characters and you’re really trying to maximize the 500. So far, these models aren’t perfect at that granular specificity, which is surprising to some people but it’s important to note, does that make sense?

[00:11:56] Charlie Madison: Yeah. I was actually going to ask, like, how do you decide [00:12:00] when the constraints are followed good enough? Because to me, it’s like dealing with my three year old. They are going everywhere except where I want them to go.

[00:12:09] Zach Hammer: That’s a good question. Part of that is, what we’ll get to after we go through literally just like, how I think through the sort of the foundational structure of the prompt? And that’s going to be in like, how we test and how we iterate, right? Because there’s a certain level of just understanding what’s reasonable to expect based on what you’ve tested and what you’ve tried from AI.

[00:12:29] Zach Hammer: Like, one of the things that I’ve realized over time is that, some of your constraints, they need to not be constraints in the prompt. You break a prompt into multiple steps, right? Because then, it has less to try and keep in head at one time and it could do it well, right? It’s almost like you want to think of, this changes over time and some models get better than others. Claude 3 for instance, has been able to handle a lot more complicated stuff than GPT-4 was. So like, this changes and adapts and that’s part of where experimentation comes into play here.

[00:12:57] Zach Hammer: But there’s a lot of times where AI sort of starts to [00:13:00] feel like, you’re dealing with Dory from finding Nemo. Doing what you asked it to, it forgets key components that you told it you wanted to do. And so sometimes you have to really simplify it. I need you to go and do this thing and then once it’s done that, okay, now I need you to go and do this thing. And so breaking things into separate prompts, sometimes you’ll discover through the process of building out these prompts. So constraints again, anywhere where we’re detailing out the specifics of what it needs to comply within.

[00:13:24] Zach Hammer: Something related but different is the next one, which is format. So format is anywhere where that I’m describing, was it using markdown? Is it plain text? Is it using bolding in Italics and emojis or not? Is it outputting this as JSON or code or a text? Like anywhere, where I’m describing those sorts of ideas. That one’s pretty easy to understand but it’s just where I’m detailing out, if I need it in a specific format for whatever my purpose is, I’m trying to convey that in the format section.

[00:13:53] Zach Hammer: The next one is tone, which also relates to style. So I’ll talk about both. I describe this [00:14:00] as how should it feel, right? Like, how should sound? Should it sound friendly and uplifting or should it sound sarcastic and snide? Should it sound, maybe just a hint of cynicism, right? Knowing how to use those elements like in a recipe, can create good results. Some of my favorite authors write things in a way where they’ve got just a hint of sarcasm in their tone, right? Like they may be coming at it from a place of service and giving but you can tell that like they’re tone.

[00:14:27] Zach Hammer: It’s not just friendly neighbor down the street. It’s pissed off, grumpy guy telling you to get off his lawn. And knowing the tone that you want to convey can make difference in getting the result that you’re looking for so, that’s tone. Style is similar, but style to me is referencing things like, does it use emojis? Does it not use emojis? Is there persuasion used and how is persuasion used? Right? So it’s related, but for me, I find it helpful to mentally break it out where tone is, how does it feel? Style is effectively, how is that conveyed? It’s more of the tactics of how [00:15:00] something is conveyed. So again, it’s not always used but it can be helpful when it’s necessary.

[00:15:04] Zach Hammer: The next one that I use is, how I can create consistency time after time? And that’s leveraging templates and examples. Templates and examples come into play when I know exactly how I want something to be structured. Good examples of this would be something like, if there’s a really clear style for writing a social media post.

[00:15:22] Zach Hammer: Where it’s like, three lines that cover these ideas. Then it flows into a step by step how to ends with a call to action, that sort of idea, right? If I know that I want it to follow a structure, then providing a template makes it a lot more likely that structure is followed rather than trying to describe a structure and then hope that it follows it, right? A template where I say, this is the template that you should follow, gives really good results. And then if I really want that clear, that I provide an example or two as well. Where I say this is the template, here’s an example of that template actually being leveraged to get the end result that we’re looking for.

[00:15:56] Zach Hammer: It gives that context to say, all right, here’s how we [00:16:00] adapt these ideas. Like I don’t need to follow it word for word. I can adapt the words a little bit and here’s how those ideas relate.

[00:16:05] Charlie Madison: And it seems to be really good at understanding. This is a template, here’s examples and actually. From what I can tell, looking at your prompts, even better than a human. Being able to picture the template, seeing the examples, thinking about it now, probably like, when I think about creating prompts like you’ve done, that’s for the part that just drives me bonkers.

[00:16:28] Charlie Madison: Like to be able to think, all right, how do I templatize what I want? How do I actually go find real examples? Which is why you’re my prompt. The reason I want to go to your prompt. Because you’re really good at finding both the template and the examples so that you’ve just got it there. Because it seems really good at following that.

[00:16:47] Zach Hammer: Yeah. And it’s amazing how often that part is necessary, right? If you want consistency, if you want it to do the way that you want it to, then the templates are almost always necessary. There are times where what you want AI to do, isn’t really a [00:17:00] template sort of thing, right?

[00:17:01] Zach Hammer: If you’re looking to leverage it as an AI role play partner that’s going to give you some back and forth. That’s not a thing where you need to include a template. That’s more of a assume this role and kind of follow this general process and it’ll do a good job. But if you’re looking to leverage AI to help you create assets that go out into the world on your behalf and follow a specific structure, that’s proven that you’ve tested, that what you want it to do. Then yeah, templates and examples are really necessary.

[00:17:26] Zach Hammer: And honestly, having both really makes it come together. A template alone can do a good job but template plus example is almost always what’s necessary to really have an understand. And an example or two can be powerful in the right context. And I think you’re right, I think a part of this is because it’s so important to me, where it’s I know what I want it to do. And I know that if I do the research and if I take the time to abstract what’s going on in something that I like, that I can have that consistent repeatability.

[00:17:54] Zach Hammer: And so I do, I go through and I do the work, I find the examples, I break them down. And I’m just continually [00:18:00] building my library of those examples, of those templates and putting them together with prompts in order to really continually improve that machine of what I’m able to do with AI.

[00:18:11] Charlie Madison: That’s probably where your research of copywriting for so many years, marketing and just your hunger. Like, you’ve got more writing examples in your swipe file or head. Like even if you don’t know exactly what it is, you know where to look. You’re decades of experience of studying that, you’ve got them a lot more than even I do.

[00:18:34] Zach Hammer: Thank you. Yeah, I think it does pay off that I really do have a working knowledge of what I want to achieve. And what sort of results I could expect and names of people that I think this is somebody that I’d want to mimic that I’d want to evoke, or this is somebody who talks about this concept. So I could take their formula and bring it into this world for sure. And yeah, so that brings us to the last part, which is, if you’ve put all of these things [00:19:00] together, then very often, you’re leveraging AI because you need it to create something from some level of your changing context, right? You want it to be an adaptable thing that your unique aspect in order to get back out.

[00:19:12] Zach Hammer: Something that’s running that process consistently. And that’s where, what I call user fields comes into play. Where it’s, what are the things that you can’t store on the prompt, right? If you’re wanting it to write an article, that’s SEO optimized, you probably need to provide it. What keywords you want it to optimize for? So that would be a user field that, when you use the prompt, you provide the keywords, you provide the audience, you provide those sorts of things. And the user fields really, they’re just there as a reference to remind you. What do I need to provide in order to get the desired result out of this prompt?

[00:19:45] Zach Hammer: For an SOP prompt, it might be the transcript of the video that you’re leveraging to turn into an SOP. For a social post, it might be the transcript of a short form video that you’re using in order to then turn that into the caption that goes along with it. So it’s all just about understanding, [00:20:00] what are the things that are required for AI to be equipped, to do what it needs to that I expect are going to change time over time when I use this prompt. And having those clearly laid out is a big part of the process.

[00:20:11] Zach Hammer: And so, what you end up with all of that is that, those pieces aren’t always used. But they come in when they’re necessary, depending on what you need. And you may not use every single one of them all the time. But depending on how specific you need your prompts to be and what you need to achieve, you put together those pieces to ultimately structure your prompts. But that’s just the foundation.

[00:20:31] Charlie Madison: That’s the foundation.

[00:20:32] Zach Hammer: That’s the underlying code that gives you the pieces to know what to work with, the parts of the recipe that you bring in or out as you need to. And so, really, my process starts there. And then the next part of the process is where you really make sure that these things actually work, which is in testing and iteration. And so, what I do is I actually come up with an idea of what I want to prompt to do. And then I’ll go through and [00:21:00] I’ll do multiple rounds of trying a prompt in different ways and seeing, do I get the output that I’m looking for?

[00:21:05] Zach Hammer: And then wherever it’s off, I go back and I add context or I add templates or I add format or structure. And I build that out over time where I say, what’s my baseline rough draft idea of what I want this thing to achieve? And then I go through and I try and use it and I expect that the first time I get it back, it’s not going to be right.

[00:21:27] Zach Hammer: And so I’m going to have to iterate. So I test and iterate and part of what I’m looking for is, I’m not only looking to get what I want back. I’m looking to try and test to see if I could figure out where does this go wrong? What’s the thing that I’m missing? Where, if it goes slightly differently, it would go wrong. And so I also try, they call the concept in like, network security red teaming. Where, I try and come at this from a way where, I’m saying if this is going to fail, where is it going to fail?

[00:21:56] Zach Hammer: And I try and test it that way to see if I could still get the results that I’m looking [00:22:00] for in a slightly different context, with something like a different transcript, a longer transcript, a shorter transcript, something that might be a little bit confusing. And I iterate over it to try and figure out where is it going to fail until I ultimately end up with a prompt that, even with this, it’s still not going to work 100% of the time, but it’s getting us there most of the time.

[00:22:18] Zach Hammer: And I’ve covered most of the likely edge cases. And what I end up with, is I end up with a prompt that is tested, proven that I’ve seen in this system, it works this way. And that is an important aspect of this too. When I’m testing these, I am not just testing in ChatGPT, right? Like, I’ll test it in the free version of ChatGPT, I’ll test it in GPT-4, I’ll test it in Claude, I’ll test it in the different levels of Claude 3, right? So there’s Opus, Sonnet and Haiku, I think are what they’re called right now. So, I’ll test it in all of them and I’ll see what level do I actually need order to achieve this result?

[00:22:55] Zach Hammer: Because as I’m building these out, if I could use [00:23:00] 3.5 to get the result, I’ll use 3.5. Because, tenth to a hundredth of the cost, right? And so, I’ll use that to get the result. But if I need GPT-4, then I’m sure that I know where and why. And over time, I’ve also developed a little bit of an intuitive knowledge for where those models are best deployed. What sorts of tasks I can expect, 3.5 to do well versus where I potentially need to bring in the bigger guns of something like GPT-4 or Claude 3, Opus and how to deploy those. And so, not only do I have an understanding of the prompt itself. But I understand where that prompt can be deployed. And I have an expectation of this needs to be used with GPT-4, this needs to be used with Claude. There’s times where I have a prompt that uses Claude but I have a prompt that follows up in GPT-4 to make sure that it’s structured properly.

[00:23:47] Zach Hammer: I’m even pulling together prompts in different ways to do different parts of the process, where they’re best suited for that aspect of the process. And really, there’s no way that I found to do [00:24:00] that other than doing the testing, experiencing it and seeing what the results are. And really, the only other shortcut is leveraging the prompts that somebody else has already done that testing on.

[00:24:11] Zach Hammer: If you’re building your own prompts or if you’re going through that process, if you want to know that they’re going to work consistently, I haven’t seen a way to shortcut doing this process, right? Where you test it, you iterate, you see what happens. Cause it’s still a little bit of the Wild West and it doesn’t always perform exactly how we would want it to. The only shortcut that I’ve seen that does work pretty consistently is, when you’ve got something that has been tested and has been proven to work in your use case, you can expect that it tends to keep working fairly consistently in that environment for that expected.

[00:24:41] Zach Hammer: And that’s part of what I do, right? Is I take these prompts that I’ve worked on and I put them into a library so that people can leverage them, that people can use them over time to get those consistent results. And yeah, that’s what we’ve put together in our real estate Mastermind. And other ways that we make those prompts available to people.

[00:24:58] Zach Hammer: So I know, you are part of the [00:25:00] Mastermind, you’ve seen this prompt library. What are some of your thoughts on just like the library of stuff that we’ve put together?

[00:25:06] Charlie Madison: I think just the iteration and testing is probably the part that I’ve got the least patience for. Because like, it’ll improve and then it’ll go far sideways. I’m like, what do I do now? Do I go back to the drawing board? And it just takes so much knowledge to be able to know where to guide it and where not to.

[00:25:31] Charlie Madison: I think that’s probably what it is that, just the iteration of getting it to where it actually works consistently. Cause only have to have the patience but you gotta be able to know where to guide it. And so the fact that it’s just done for me, that I can part of the Mastermind. I’ve seen it, like, people say, all right, here’s my problem.

[00:25:51] Charlie Madison: And the next week you’ve got a prompt that solves it consistently. Everyone can say, here’s my problem. And wait a [00:26:00] week to get a prompt or even better. The last I looked, I don’t know how many prompts you have, but I just want to hug them all because they just work.

[00:26:08] Zach Hammer: That’s the thing that’s interesting is, there are people who have more prompts than I do, right? Like in terms of the overall quantity, I think I’m up to somewhere around, at least, in the library itself of prompts specifically. Now keep in mind, I keep prompts mentally separate from the templates and from the processes and from the playbooks, right?

[00:26:28] Zach Hammer: But the prompts specifically, I think I’m up to 47, is what I’m up to right now. But those 47 prompts are dialed, tested, consistently achieving a powerful end result in your business kind of prompts. They’re not mass generated things that somebody said, you know, generate a hundred prompts about Facebook and then told you, I’ve got a hundred prompts, right?

[00:26:51] Zach Hammer: Like it’s not that. So there are things that I’ve tested every single one gone through and I’ve actually put them to use and tried to leverage them in my business. And [00:27:00] yeah, that’s the key idea. And regardless if you want to leverage this process and go and create your own prompts, I encourage you to do so, right? Like, this as a skillset is very learnable. What I’ve found is that I get better at it. The more that I do it. And so I’m potentially ahead of a lot of people just because I’ve done it a lot. I don’t think there is a shortcut for that knowledge, for that experiential knowledge of seeing what can this model do?

[00:27:26] Zach Hammer: What can’t it do? And knowing how to deploy those and how much context is too much context? How much context is the right amount of context? I learned that, I get quicker at that over time as I test it more but further I’m willing to do that testing. So even if I’m wrong, I test it and I see what’s actually working, what’s getting me the end result and I know how to go back at it and attack it again in order to get those results. If you want to leverage that process, really I’m not holding anything back there. That is what my process looks like, those are the thoughts for how I go through and I work on perfecting prompts to get an end result that I’m looking [00:28:00] for.

[00:28:00] Zach Hammer: Everything else is literally just the work of doing it and the experience of doing it and getting better at that over time, learning how the tools work. It’s something that you can achieve if you want the shortcut for it. And you’d rather be able to show up and essentially leverage ZachGPT and be able to put forth your problems and get back AI driven solutions that could hopefully help you do that at scale.

[00:28:22] Zach Hammer: That’s what we look to achieve in our cohorts where we go through and create some very specific solutions to very specific problems. Like, making sure that you have an unending content strategy to leverage easy ways to show up and have all the marketing content that you need all the way down through to our Mastermind. Where we’re continually taking the opportunity to take in what are your problems that you’re looking to solve?

[00:28:43] Zach Hammer: And let’s build those prompts together so that you can leverage AI to be scalable, consistent, keeping those costs down, being able to achieve more results with either the same staff that you have now or less cost as you continue to grow. If that sounds interesting to you and you want to learn more about what we’re up to, feel free to reach out to us at [00:29:00] RealEstateGrowthHackers.com/contact, where you can find out more about those offerings that we have available. And get onto the newsletter where we’re continually talking about this information. And stay tuned as we put out more information and more offers to help make the world of AI easy for you but otherwise, any final thoughts on this, Charlie, before we call this one a wrap?

[00:29:18] Charlie Madison: I love the idea that I can just do at Zach Hammer GPT and I’ve got the prompts and they work consistently. Being able to repurpose content, just being able to do a little video and it turns into a SOP.

[00:29:32] Charlie Madison: Not to mention, to be able to say, Hey, here’s what I’m looking at, what can you do? And I know a little bit of the behind the scenes stuff of what you’re working on right now. We’re messaging this week and I’m pretty techy and I was like, I’m blown away. It’s pretty cool what’s happening now. And I see the stuff that’s coming behind the scenes, which is really cool.

[00:29:54] Zach Hammer: Yeah, absolutely. What we’re working on these days is, we’re trying to make everything as [00:30:00] easy as possible and building systems to take care of as much of this stuff for you as we can. I’m working on ways to make it one click simple that I could take it a transcript and output like all the social posts that I need based off that transcript.

[00:30:13] Zach Hammer: And yeah, we’re building out a lot of cool systems. I definitely encourage you if you want access to that, you want the golden ticket. Reach out to us at RealEstateGrowthHackers.com to learn more about what we’re up to and how we might be able to help you implement AI in your real estate business or for your company. So there you go, thanks again. I’m Zach Hammer, Charlie Madison here with me co hosted today. Thanks for joining us for another episode of the Real Estate Growth Hackers show. Until the next time, we’ll see you on the next one.