Episode Transcript
[00:00:00] Speaker A: If you're sitting down waiting for this to get perfected, to make sure, oh, I don't want to see any mistakes, any hallucinations, because that happens, right? Like you come up with the wrong number or just makes make some stuff up and that's legit. But if you're not using it for what it is and you're just saying, oh, I'll wait until it's perfect, I'll use it then. Do you know how far behind the rest of the world, every other competitor you're going to be if you wait three to five years? Do you think you're the leader of the pack? Are you kidding? Like this technology can do stuff right now that is just mind blowing.
[00:00:32] Speaker B: What if you could triple revenue without hiring anyone new? Most local businesses already have the tools to do it, they just don't know it. Join us as we explore how AI automation and smart advertising can help you work less, earn more, and build systems that keep paying off. The local AI show is where Main street gets street smart one simple building block at a time.
So today we have a bit of a 101 on AI for the local business owner, Rhys. So you have prepared a little doc here. You want to tell us what we're going to go through here today?
[00:01:08] Speaker A: Yeah, so last episode, first episode, right. Was introduction to us with Kyle and I and our kind of a background. And then that just kind of makes sense that the next episode would be an introduction to our topic, part of the entire reason that we're doing the show. And so the basics behind AI. And so some of those basics obviously is a prompt, but other things like understanding, you know, what an LLM is and what ChatGPT is and the limitations of ChatGPT and what is a, an AI brain and what is an AI agent. So all these terminologies being thrown around.
So we're just going to break down kind of the, just the basics like that, those answers to those questions, just so you give the audience a general understanding and a, a general idea of the differences and the nuances of AI and LLMs language. Large language models. Yeah. So a prompt, it kind of starts with that. Kyle, do you want to start with the prompt? And I mean, just the basic. What is a prompt? Just on that note, as we're going through this, because this is about AI, obviously the easiest way to, to teach you about AI is AI itself. Right. So we just asked AI the same questions. And so I think at the base of AI is just to understand that it's just an assistant Right. So just treat it like, like a, like an assistant, like a, like a human being, your own little helper. And then just ask it questions like you would ask anybody else. Right. So we'll try to get into it and to keep it very plain. Jane. English and not too jargony.
[00:02:46] Speaker B: Yeah, I think I, I think a good, like a good can, like overview would be.
Let me just. Okay. I think a great spot to start this would be just talking about what an LLM or a large language model actually is.
And as far as I understand it, a large language model is basically this thing that has.
It's. It's absorbed all of the information that has possibly been fed into it. And at any given time it is able to tap into it based on what the user asks it or prompts it. But I think what a. I think an important distinction is that an LLM by itself cannot do much of anything. It's basically just. It's just this thing that is just sucked up all kinds of information, but it is able to access it like in no time flat. And that's kind of the power of it. Like, humans are really. We're not good at processing data. Tons of tons and tons of data. But that is what LLMs are amazing at. And so that is why prompting is one of the first things we're going to talk about, because that is how you. That is like the first step to tapping into this thing called an LLM or a large language model. Is that your understanding? Have I that up? Or does that seem like it's.
[00:04:16] Speaker A: No, that's. Yeah, that's legit for sure. So I think, I mean, the terminology of LLM you might not even hear, right. The average user might not even hear that terminology. But think about the ChatGPT is obviously the most popular, right? So that's a large language model. And a really good analogy to kind of wrap your head around what it is is to compare it to a library, right? If you walked into a giant library and the librarian has read every single book that there is on that's ever been printed, right? Every. Every book that's ever been released. And this library and this one librarian has read them all. So books about everything that you could possibly imagine. And then you might ask if you have a dog grooming business, you might say, hey, can you write me some Kathy slogans for my dog grooming business? And so you're asking the librarian or the large language model to access all of the knowledge that it has on marketing, on dogs, on small businesses, and then it grabs all that information and puts it all together and says, okay, based on all of this information that I have, I'm going to now create this wonderfully patchy slogan for a pet grooming business.
And so it is trained on a subset, like a set of data that is frozen in space. Right. So it is learning from the entire Internet. So ChatGPT knows everything on the Internet and then at a specific time. So I'm not even sure where we are now. October 28th, I think was one, and then we updated this year again, so it might be like three months or six months behind. So it's not good for real world data, it's not good for news. It's not something that you're going to, you're going to ask it, hey, oh, I heard Hulk Hogan died today, for example. That's not in there because that just happened. And it's not programmed into the Data set for ChatGPT because Hulk Hogan died. Yeah.
[00:06:11] Speaker B: Oh, my goodness, I had no idea.
[00:06:13] Speaker A: Yeah, so let's leave large language model for a bit and we'll get back into prompting then. So. Yes, sir. So basically a prompt is just something that you're instructing the LLM or CHAT GBT to do, right? So you're, let's just call it LLM doesn't really roll off the tongue very much like gbp, Google Business profile. So let's just call it CHAT gbt. It's just very easy. Everybody knows it, everybody's familiar with it. So when you're trying to use CHAT gbt, how do you use it? What do you do? So you prompt it or you ask it a prompt, verb or a noun. So a prompt is just a question or an instruction that you're giving the AI, right? And you want from that instruction you want the answer, a very useful answer from whatever you're asking it. The. There's a lot of nuances to prompting and stuff, but at the very basics, the basic stage of beginning, just begin, just ask, just so it's just like talking to a human being. Try not to think of it like, like you have to talk to a machine or you have to know exactly what to say and then what we'll learn later on, you can actually ask CHAT GPT to come up with the prompt. Right? So if you don't know how to prompt, you're like, oh, okay, I need a prompt for writing a catchy slogan for my dog grooming business. And then it will come. Create the prompt and then you feed the prompt back into the machine. Right.
[00:07:36] Speaker B: Yeah, sometimes I find those are some of the most powerful prompts, are just the ones where it's like, hey, based on what we've discussed here so far, what, how would you have prompt. Prompted yourself? Or how would you prompt yourself going forward? Or what did I miss when I asked you what I asked you? And I think like what most people get wrong about prompting, it seems so simple, right? You just spit something into chatgpt or cloud or whatever and it gives you an answer. But I think that it's like, it's kind of like garbage in, garbage out. And so the more clear you can be and the more context, I think the biggest thing is context, like telling it exactly what is going on. I'll often just be like, hey, the date today is July 24, 2025. Here's X, Y and Z. About this scenario. Here is what I want the outcome or the goal to be. That's why I'm asking you this. Here's as much specificity as I can give you, basically. And then the, the more clarity, the more and the more context I give it and the like rooted in an outcome or a goal. That's when I find I tend to get the best results.
[00:08:49] Speaker A: So add to that throwing an example in there. So if you know what you want as an output, then throw that in there, right? You're like, I'm looking for table. And the table is going to have five columns and each column it will have this header and then you just tell it exactly what the output is that you're looking for so that it has an idea, okay? And then when it does, then the next question that you prompted to it will know that it's going to have that table with those five exact columns, right? So there's no. Sometimes I give it the examples and sometimes I want it to come up with something first. I want to see without me saying anything. Like maybe it gives me seven columns and I didn't think of those other two. And then I'm like, so I don't want to say, hey, we stick to five columns because I want to see what it can come up with first. And then I. And the point of prompting isn't that it's the one time end all, be all prompt. You're not trying to, oh my God, like think of this big one pager that's going to cover every aspect. So at the beginning, just ask one little thing and then you get the answer. Read the answer or look at the answer. Did you like it? Is it what you were looking for? And then ask something else. Oh, I kind of like that. But it wasn't as catchy as I was hoping or I really wanted to focus on how we're community driven and we're local, whatever the thing is. And then it'll change, right? It'll fix it, it'll edit itself and then come up with another one and do you like that one? And then just try to work with it and keep on asking it questions until it, it comes up with the output that you're looking for.
[00:10:13] Speaker B: Yeah, absolutely. And, and I also find like voice and tone is important and being clear in terms of what you want it to sound like. I. This is a. I feel like I've wasted so much time saying the same things again and again, which is like all these LLM platforms have solved for this with Things like custom GPTs or in ChatGPT or projects in Claude and you can actually.
[00:10:36] Speaker A: Hold on, what are those? Like, you just talked about two things.
[00:10:39] Speaker B: Okay, okay, so, so here. Okay, so if you find yourself saying the same things over and over again. So I work with restaurants and when we use ChatGPT to help us come up with content or captions, I say the same shit again and again. And these are things like, I want no fluff, nothing cringy, be concise and simple, do not add unnecessary bullshit, et cetera, et cetera. Add all these things, right? Because I want an output that is human and. Or it just, it sounds like something that a busy restaurant owner. And this is another thing that this part of the context, I say you are a busy restaurant owner who really cares. You really want to grow your community and provide an amazing product and experience, blah, blah. So these are the things that we say again and again. And so inside of ChatGPT you can create what are called custom GPTs. And these are little things where you can give 8,000 characters of instructions. You can also upload, I believe, up to 50 documents of PDFs and things like this or yeah, documents. And then every time you go to use that custom GBT to create, say content or something, it already has all that built in background so you don't have to repeat yourself again and again. Does that make sense?
[00:11:57] Speaker A: It does. So like a custom GBT is just that it's customizing the GPT to your brand, your tone, your voice, right? And then ideally you're creating an AI brain that had, that is knowledgeable about everything that you do that your brand does. So it's going to, it's going to spit out content that is, is in the same tone, the same. It's just going to feel like you're writing it, right? Because it's learned off of you, it's learned off of these documents. Like you could feed it emails or you can feed it voice, like just like transcripts of things just to give it an idea. Or you could send it somebody else's and say, I really love the way this guy writes. That's how I want our writing to be. And then give it a little bit of your tone in there as well and kind of mix the two and then have it to create this new voice almost where you want to tell it. The tone is professional or casual or funny or whatever. Right. I will, I kind of do informative professional. And informative is I want to be casual, but I, I still want to be professional. It's not like I'm talking to a buddy. And so that's, that's how the content that I want for clients to come out. But for each client, we have an AI brain that we've created for that client with all of their materials and stuff. So when we produce some social content or page content or blog content, then the AI already knows the flow, the sound, the tone, the feeling, you know how to lay out that document and how to make it sound. So it just makes your job a lot easier when you know, like you're saying you're doing all this repetitive stuff now you're building in that repetitive stuff inside of ChatGPT and that's something simple. You can just go to chatgpt.com there's a GPT, literally a GPT link in the left hand side, click that, click create and walk through it. It's super simple to do. And I mean it's nothing more than uploading some documents. And then it'll start asking you questions like the GPT will start saying, hey, what else do you need? This is what I've done, I've worked on this. And then you go test it and it allows you to test it right there. Yeah, the GPTs. Custom GPTs are great. And then Claude is another large language model and it's from a company called Anthropic. And Claude is actually my favorite writer and actually coding for coding. So I mean, I'm not a coder in any sense of the word, but these guys can write code for you and chat or in and clone happens to be very good at the coding aspect of it and then better than chat GPT in my opinion for the writing that it produces just to make it more Human and just more natural sounding. It just sounds so much more closer to something that I would actually write for a client or for myself.
[00:14:42] Speaker B: Totally, I agree with that.
[00:14:44] Speaker A: Right. Where did I go here? Okay, what is. So that's a large language model, right? And we can consider a large language model to be an AI brain. So people talk about an AI brain, but you're just, you're just powering the brain, right? So a brain is just anything that you'd feed information. So it doesn't have to be a custom GPT like in Clode they have projects. And so this project is the same thing. It's. It's a folder by itself where you can upload a bunch of other information, PDFs, URLs, other documents and then train it. You're just training that, that agent or that AI brain on whatever it is that you're uploading. Right. So then you can now query that brain for anything, any of the information that's uploaded. Right. So very helpful if you're, you know, you upload a bunch of, let's say, export your data from Google Ads, right? And then you have all of the data, all the exports of all the different reports that you can put out from Google Ads. And now you can talk to this AI brain and say, hey, what were the missed opportunities? What kind of keywords are really the cream of the crop? Which ones should we be putting out as negative keywords? And it'll actually give you a list, here's a list of all the negative keywords. You can upload this CSV file so you can ask about sales and seasonality and which ones are performing the best. And you can get it to learn from itself by continually uploading more current stats and documents and stuff.
[00:16:05] Speaker B: I actually think that is where I would start as a local business owner is I would start by taking as much data as you possibly can and feeding it into an LLM. If you're paranoid about the LLM training off of your data, which is just like.
So that's how these, that's how a lot of LLMs learn is we feed it data and it learns from it. And so some business owners are paranoid. Oh, I don't want to give it, I don't want anyone finding out about my data. It's never going to surface like your actual data. But if you are paranoid about that, Claude actually doesn't, I think you can, with a paid version, you can make it so that it doesn't train its models off of your. Anything that you put in there. But at any rate, if you just feed, read the LLM, your data, it could be your, it could be your accounting, your profit, your P L statement, balance sheet, it could be transactions going back years and years. It could be your guest or sorry, your customer information, it could be your ad data. And just putting that all into say one conversation, you can then start to draw insights and connect the dots that you never would have been able to otherwise. At least most normal human beings can't. Like I mentioned, we're not generally good at processing a ton of data and that is where so much of the power is just in being able to talk to your data, to talk to your business. And, and so that, that's where I would start is just start giving it raw numbers and have it help you find the insights, the golden nuggets, the growth opportunities that are for most just hiding in plain sight.
[00:17:49] Speaker A: Yeah, I like that too. So I think I would start with a cloud project. That would be a good starting point. So get a, what is it, $20 a month US I think something like that for cloud. So get a, a membership with there and then and just start creating pro projects and like a project could be your PPC data, your like Google Ads data or your Facebook ads data or it could be your sales data or it could be your traffic stats for from your Google Analytics export. Like anything that you can gain an edge, right? What do you think about what you normally do with your data during meetings or day to day or quarterly or.
[00:18:25] Speaker B: What you don't do with your data or ignoring your data. This is like a golden. If you are that person and I have been that person where it's just like, oh my God, you don't have the bandwidth. Right. It's oh, what am I going to do? I mean you're just going to sift through this data and put it into spreadsheets. I suck at spreadsheets.
[00:18:40] Speaker A: Now you have, yeah.
[00:18:42] Speaker B: You have this amazing partner that you can just. And this actually blows my mind a lot. Okay, caveat. You have to have a paid membership with like ChatGPT. I think the plus again, anyone who doesn't. Okay, anyone who doesn't spend the $20 for ChatGPT, I'm just like, that blows my mind. I feel like it is the highest leverage, highest impact $20 you could possibly spend just for this nugget right here, which is you can just take screenshots and put it into ChatGPT and say hey, what am I looking at? You could take screenshots of your QuickBooks dashboard of your, your spreadsheets and just say, hey, help me make sense of this data. What else can I give you so that we can make this really actionable and high impact? And it'll tell you what about just.
[00:19:31] Speaker A: Like uploading your data and saying what? What can I do with this? I'm a local, whatever business.
Here is my traffic stats, here are my ad stats, here, whatever. What can I do with this? What can I learn? What can you teach me from this? What can I do? How can this benefit my business? And it'll tell you like it, it's like the smartest little assistant that you never had. It's, it's this brilliant little mind that you just need to, you just need to access. Right?
[00:19:56] Speaker B: Exactly. Exactly. That is, I think a lot about where work is going. Like how are human beings going to continue to be valuable in this sort of new world? And obviously we're thinking about local businesses. And I think that if you are someone who can help a local business tap into intelligence and AI, because what that'll enable you to do is go further, go deeper, deeper, go better because you can just draw more insights from the data and then if you can take that and you can use automation to have repeat tasks just get done so that you can do more things where you build once and have it continuously reap rewards.
That I believe is how human beings can continue to be valuable. And I'm not, obviously there are all kinds of ways to be valuable. I don't want to, but like a lot of people are like, well, hey, AI is coming for my job, it's coming for my business. Well, instead of having that mindset, why not instead be like, I'm going to wield these tools to go further, faster, better. And then I'm also going to combine that with automation wherever possible so that repeat tasks get automated away. And I just, I think if you move forward with that sort of one, two, punch it, you're more likely to be set up for success.
[00:21:21] Speaker A: If you're, if you're on the picket line with your like, no AI here, no AI here, you are so close to getting.
[00:21:29] Speaker B: Yeah. Or if you behave, if you operate like a tool, like I am just a graphic designer or I am just a website designer, you are like, you're already done. You have to, I think you, you need to become more of an orchestrator or you need to become more outcome oriented. It's okay, you're a graphic designer to what end? And then whatever that end is, have AI and automation help you, help whoever your Customer is get to that end more efficiently and better. I don't think it's about if you're an artist, you're a painter or whatever. I'm not saying that you can't still do that. And I. This also. We're also not here just for artists, right? We're here for really. We're here for local businesses. So that's kind of the lens through which we look at this.
[00:22:14] Speaker A: I had a conversation with my web developer kind of today, actually. We were talking about something else, but it kind of moved on to his. Is would be one of those businesses that you would think you're a web designer. Well, how long do you think you have left in the world before AI takes out your business? Right. And so we were kind of discussing how is it turning out so far? Because we're starting to see a lot more traffic from LLMs, people searching different ways and stuff. And then now, I mean, even yourself, you're making. Creating websites right out of thin air, almost. Right.
Cursor or something of the light and then uploading it. Right. So that, that there, there is no web developer there. So you became web developer. So I said, is it. How is this changing for you? Are you concerned? And he says, you know what? You know what? We're seeing a ton of people using that, not knowing what they're doing, putting it up on Vercel or something on some sort of cloud hosting platform. Something goes wrong and then, oh, I have no idea where to go from here. I can't fix it. The client's complaining or whatever the situation is that he's getting tons of calls to fix things now and he's actually busier and he's expanding his team because of like, very unexpected results, I guess, coming from this. Right.
[00:23:22] Speaker B: So, yeah, I mean, I think that's with where AI is in many cases. I think it does an amazing job of like right away impressing you to the point where you're like, I could just do this myself. And then as you go further, like you get to a certain point, like using just AI to get things to the finish line right now, I think is still like really hard. And I think a lot of people are probably wasting a buttload of time just trying to like, like nitpick their way to get something to where they want it to go. Because it's certainly, it's far from perfect right now. But I think folks say this is the worst it's going to be.
And, and ignoring it or not harnessing it I think is just like this time, two to three years from now, you will be far, far behind.
[00:24:17] Speaker A: If you're sitting, if you're sitting down waiting for this to get perfected, to make sure, oh well, I don't want to see any mistakes, any hallucinations because that happens, right? Like you come up with the wrong number or just makes make some stuff up and that's legit. But if you're not using it for what it is and, and you're just saying, oh, I'll wait until it's perfect, I'll use it then. Do you know how far behind the rest of the world, every other competitor you're going to be? Do you think that it's supposed to be three to five years until it comes like full circle and everybody's using AI and kind of in their day to day lives and stuff. If you wait three to five years, do you think you're the leader of the pack? Are you kidding? Like this technology can do stuff right now that is just mind blowing. Can it create like the most awesome picture that you've ever seen and it's the best thing and no one ever could create something better than that? Yes or no?
Yeah, I can do both. Yeah, maybe it does and, or maybe it screws it up. But there's so many things that it can do and so just ignoring it because it makes mistakes is just, it's the wrong way to look at it. But my tried this with my sister. She had heard that you can create images with chat GPT and stuff and then so she tried this probably, I don't know, you know, maybe six, eight months ago or something like a while back where fingers were still a problem. I mean there's still a, there's still a problem, right? AI imagery has problems with fingers and extremities and stuff and they kind of screw them up, put five fingers or three or whatever. So she tried it two different images and it screwed up some hands and she's. That again. Yeah, it's just not ready yet and then that's it. So she, that's it. She's not going to look into it any further.
No use for it. That, that mentality is crazy, right? That's, that's not the winning mentality that you're looking for to stay on top of the competition, right? Like why not use every tool available to you for US$20 a month to, to leverage it and to get ahead of the competition. And we're talking about being able to grow your business so that there's an exit there, right? Having systems and processes in place so that there's an exit. Like you're building the business so that you can step out and somebody else can step in and replace you. Well, you need systems and processes and you need to understand your data if that's going to happen. It can't just be. You can't be the business, right? Then you can never get out of the business. Use the tools to create the business that you can actually step aside from, whether that's you coming, letting your son, your daughter run it, or bringing somebody in to see as a CEO or selling it off, whatever it is, it doesn't matter, but it's still built the same way. Like, like you're going to sell the business. So now start using AI to create the systems and processes that you can put in place that will save you time, will increase money, the traffic and conversions and stuff that you're able to get to a position where you can double or triple the income that you're making from your business. And you're not putting in twice or three times the amount of investment to get there. Like, literally the investments, $20 US. We're not taking a cut off of that. Like, just go try it.
[00:27:15] Speaker B: Yeah. Okay, let. Let me take this episode to the finish line and paint a picture for you. Just think about what it takes to create a standard operating procedure, right? Like, standard operating procedures are the classic way to build systems and processes. So it's every single time you do something in your business, ideally, you capture that in a standard operating procedure so that it isn't always dependent on. On you to get the thing done. Now, back in the day, to do that for a guy like me who's just like, very forward, let's go. For me to take a step back and be like, okay, Kyle, here's so, okay, so this is what I did. This is the playbook. This is the goal of this standard operating procedure. Like, when you do this, here's what success looks like. Here's how much time it should take. Here's step one, step two, step three, step four, step five, blah, blah, blah, blah. Here's a little video of me showing how it's done, right? All this stuff takes a ton of time. And it is the reason most local business owners have, if any standard operating procedures they have, like, they're kind of bad. And so now what you can do is you can.
[00:28:23] Speaker A: You.
[00:28:23] Speaker B: You can just be going through something with a team member and you can record that call, right? We all know how to record meetings. Then what you can do is you can upload that call into. You can just have it make a transcript for you, right? And then what you can do is you can copy and paste that Transcript into say ChatGPT and you can literally say to it, please turn this into a no nonsense, fluff free, very well structured, standard operating procedure. So that my team, whether they're, they've been with me for several years or they're brand new, know nothing about what our business is, they can take it and they can see it through the finish line in a way that is successful and that, and that's it there, boom, you have an sop and then what you could do is you can actually put that into say a cloud project or maybe it's even, I don't know, a custom GPT or something so that eventually maybe all it is like your new team members, they just talk to a custom GPT to get the information that they need to make the business go. And so I think that that kind of thing is, it's crazy how cool that is.
[00:29:30] Speaker A: So that, that brings back to that little tip that you had about the screenshots, right? So if you run into any situations, any errors, whatever, screenshot something, maybe you can circle it and be like, hey, fix this. This is where. This is what happened. How do I get past this kind of stuff? But yeah, it's so either recording yourself, going through it, right? Like you, if you're going to do it and you, and this is something that you do all the time, or something that another employee does all the time, then just have them record their screen as they're going through it. And then, yeah, just as you said, go with the transcript and then build up. If you have a transcript of something, which is easy to get, a bunch of any whole slew of tools or whatever, even YouTube, I think you can download the transcript there. If you have a transcript from a video, I mean you can create any type of content from that, from training content to social media content or whatever, right? So that would probably be, yeah, a very good beginning, I think, as you said, to upload your documents, start talking through your documents and then creating, using AI to create SOPs and maybe even start with hey, chat GPT.
What kind of SOPs would be required for a plumber, plumbing business, dental business or whatever and have it start and then look at them and then. Okay, does that make sense? Is that the type? Do I need that? Do I use that in my business? Yes, it is. Okay. And then walk through each of those things, right? And then create that, that sop, those are really some of the most valuable stuff that you can have for your business. Right? Like so on every single thing that you do. And yeah, I struggle to put these things together. They're massive documents. But not anymore. So, so use AI. That way we can just end on an AI agent, right? So we, we talked a bit about the LLMs, right? Large language models. So ChatGPT and Cloud and then an AI brain kind of using projects inside of cloud or using custom GPTs for chat GPT.
And then an AI agent is kind of like a next level LLM or next level AI. So an LLM is singular in the sense that it can only do really one thing. So you tell it to do a thing and then it'll output this, whatever you're asking for, but it's not going to, sorry, it's not going to go out and do things for you. Right? Like, so an agent is like a digital employee that will, that'll go out and, and do things. An AI brain is just like smart and helpful, but the AI agent thinks about what you're asking, it decides on what it needs to do and then it'll go out and do the thing that it needs to do to make it happen. A real world analogy like is if you, you own a bakery, a chat GPT style AI brain is like a smart intern and that person's helping you write posts about muffins and baked goods and stuff. An AI agent is more like the virtual assistant and says, I noticed we're low on muffins. Muffin orders today. I've already scheduled a face a Facebook post. Email your customers and updated the menu online.
So it looks at like the entirety of the situation. We're running out of Facebook posts, for example, maybe. And okay, now I know that we only have two more scheduled. I need to create these posts. And then it'll actually go and create the post and, and then you say you want to store those posts in a Google sheet and then it can automate even the scheduling of those posts and then update your menu online. It can do all of those things together.
And so it just kind of runs in the background making these things happen. And some of the times you need to make connections and stuff and there's some things to build out, but you can always ask, I don't know how to do this. What is the next step? I don't know what you just said to me. I'm not a coder.
Tell me what you mean and just get it to tell you you can always ask for clarification and stuff. So it makes it really easy to use them. And then one, the one that I'm using right now is, is Manus.
M A N U S dot. What is Manus dot? I am, I think, hold on a second. Manus I. So Manus is a. An AI agent and it's. It performs a slew of tasks like whatever you can think of, just start playing around with it.
So how to create little resources or how to create little tools that you can even use or do a site audit for your website or optimize some pages or some content that you currently have on your website. Hey, analyze this URL. And we're trying to rank for this particular keyword. Tell me what do I need to do to optimize that page for this keyword to be able to rank in the top of Google. And it'll tell you. Right? So AI agents are super advanced LLMs essentially that can do a bunch of other stuff. And the crazy thing is Manus is again like 20 bucks.
[00:34:11] Speaker B: $20 US yeah, but it eats up credits. Like it's insane. I find.
[00:34:18] Speaker A: I think I get a lot of credits. Like for some reason I use it a lot and because it's brand new, they just keep on feeding me a bunch of credits. So I've never ran out yet. And I've never.
[00:34:29] Speaker B: I have not found that to be the case. I feel like every time I use it, I get cock blocked by it asking me for more credits. It actually I often. I pay for it, but like I'll use it maybe once a month. I find for the. I find genspark. Genspark AI kind of a similar agent tool. I personally find it is better value.
[00:34:49] Speaker A: Okay, interesting. I mean, so I think it's because I had a super beta invite right to madness before.
Yeah, you know what? To get that. So you either had to know somebody who had a code that they could give you, share with you, and they were making it like. Like you're getting into a really exclusive event or something, right. And so you either had to find somebody that had that code or you had to basically write an application, give your resume to them and say, hey, pick me to be special and give me access. So I had to write. I was pretty in depth about it and I talked about different tools that I was creating, custom DPPs that I had created that didn't really work out for one reason or another. And that's the reason kind of I was hoping that I could use this to solve those issues and they gave me that. They're like, oh, you're in. And then with that, there was a whole bunch of credits that went along with it. And so I've never.
[00:35:39] Speaker B: Lucky duck.
[00:35:40] Speaker A: I've never actually gotten that message.
[00:35:42] Speaker B: Oh, dude.
Every single time it's in it, actually, because. And it's very.
[00:35:47] Speaker A: It's.
[00:35:47] Speaker B: I find it expensive just because of like the rate at which it swallows up credits. So I don't know. I'm certainly not a menace expert I have. But it. It is powerful. Like, you can for sure.
I think you. One time you said something that it put out in a few minutes was basically the equivalent of a $5,000 consulting contract or something. Just in terms of the quality and depth.
[00:36:09] Speaker A: Yeah, it was like an. It was SEO audit for a national company. And it was so in depth. And. And then not only did it have. Here are the 10 different CSV files that you need, like of all the URLs that you need to fix or broken URLs or dead pages or whatever else. All of the CSV files to go and upload or whatever. But the. A PDF, like overview of the summary of the actual site audit. Walking through everything and then a website like it create. It created a website to. Here you can go show this to the client and the client can go interact with the website, download all the documents. Like, it's crazy. And that's all from just basically one prompt and then kind of walking through and hey, is this what you're looking for? And oh, I wanted it more focused on this or that. And I think it took, I don't know, let's call it two hours. Right. Like, you have to wait. It starts working in the background, right? And it's so like a. An agent isn't as quick as an LLM because an LLM is only doing commercial of one thing. An agent might take 30 minutes, 40 minutes, 20 minutes, whatever, to. To go through and work through the task and then come out with an output. So an AI agent. Yeah. Is just like a next level AI brain. So the differences between the two is like an AI brain is. Can write stuff and you can ask it questions and it can tell you a lot of stuff. But an AI agent can take the action that it needs to go and do those things. So it crawls the website, it looks at all the URLs, it goes through an act. Like it'll go through Reddit posts, right? And search all the Reddit posts and stuff to get all the information that exists and then pull it together. And then start creating the reports and it'll ask you questions along the way.
So it can use tools, it can decide by itself. Okay, I think this is probably the best way to go and then make that decision and then just kind of do it. So on a local, local business level, we'll leave it with this and I'll let you finish on what AI agents can do for a local business. Just as an example. Kyle.
[00:38:07] Speaker B: Well, I mean I, I would start with, I would start with anything you're thinking of hiring entry level labor for, whether it's admin, like responding to emails or at least drafting emails. Whether it is responding to your leads in a timely manner, whether it is drafting up content or posts or articles. All kinds of.
[00:38:33] Speaker A: Yeah, further with it, like with appointments, right? You can integrate it with your calendar and have it book appointments for you. There's, I mean even now voice agents, right, where they can answer your phone and then book in your calendar and calendly or whatever, you can follow up with customers through text message. You can have automated follow up email message to say, hey, your appointment's coming up soon. Hey, it's in 30 minutes. How was your appointment after the fact? Post content on your socials for you. It can look at and analyze your Google reviews so it can like it tell you one. Of course. Here's a notification. You got a Google review. Analyze all of the different reviews for sentiment and say, hey, this is what people are saying. And maybe if you're like a restaurant, you get a ton of reviews, right? Restaurants get crazy reviews compared to other businesses and there's going to be some negative reviews in there. So maybe the AI agent is reviewing all of the content of the reviews. Picking out. Hey, did you know that three times some people have commented that the bathrooms were disgusting? Oh, okay. Well, we need to have a better cleaning schedule for the bathrooms, right? So it can tell you all of those things and then you can take action on it.
So it can do pretty much anything that you could think like a low level employee could do.
[00:39:44] Speaker B: 100%. On that note, Reese, shall we close?
[00:39:48] Speaker A: Yeah, I think that's it for this one. It was a good episode. We'll see.
[00:39:52] Speaker B: Great episode.
[00:39:53] Speaker A: See everyone next time.
[00:39:54] Speaker B: All right, cheers.
[00:39:56] Speaker A: Ciao.
[00:39:57] Speaker B: That's it for this episode. If you're ready to grow smarter, not busier, head to localaishow.com sign up for the newsletter and get the first three moves guide. It's the fastest way to start using AI and automation to save time and grow revenue. This week. Remember, you're on Main street, but you don't gotta play small. See you next time on the local AI Show.
