# AI Agents vs AI Automations in 2025 ## Метаданные - **Канал:** Nick Saraev - **YouTube:** https://www.youtube.com/watch?v=r3tAJ8Oj4J4 - **Дата:** 04.04.2025 - **Длительность:** 27:36 - **Просмотры:** 25,848 - **Источник:** https://ekstraktznaniy.ru/video/12416 ## Описание Join Maker School & get automation customer #1 + all my templates ⤵️ https://www.skool.com/makerschool/about?ref=e525fc95e7c346999dcec8e0e870e55d Want to work with my team, automate your business, & scale? ⤵️ https://cal.com/team/leftclick/discovery?source=youtube Watch me build my $300K/mo business live with daily videos + strategy ⤵️ https://www.youtube.com/@nicksaraevdaily All Make.com & N8N templates mentioned in the clip ⤵️ https://leftclicker.gumroad.com/l/xovbb Summary ⤵️ What are the key differences between AI Agents, Traditional Automations and AI Automations? I explain in this video. My software, tools, & deals (some give me kickbacks—thank you!) 🚀 Instantly: https://link.nicksaraev.com/instantly-short 📧 Anymailfinder: https://link.nicksaraev.com/amf-short 🤖 Apify: https://console.apify.com/sign-up (30% off with code NICK30) 🧑🏽💻 n8n: https://n8n.partnerlinks.io/h372ujv8cw80 📈 Rize: https://link.nicksaraev.com/rize-short (25% off with promo code NICK) Follow me on other ## Транскрипт ### Introduction [] AI agents and AI automations are fundamentally different things. They are tools in your kit. And just like with any tool, I see a lot of people using the wrong one for the job. So, in this video, I'm going to run you through the differences between AI agents and AI automation, which different problems these two tools solve, and how to choose the right approach that delivers the best results for your clients or just your specific business needs. This is especially relevant with new technologies like MCP or model context protocol, gaining a really wide market share very quickly. So, if you guys want to stay ahead of the impending AI agent boom, it's important to understand how all this stuff works and how to realistically choose which one to use. By the way, my name is Nick and I've sold several hundred,000 worth of AI and automations and agents over the last few years and I coach over a thousand other people to help them do the same. So, as I mentioned before, the current iteration of AI agents, specifically NADN AI agents, isn't yet at a point where you can realistically and reliably produce an ROI for a business that you couldn't do through easier, better, and more conventional automation means. I'm also not saying this because I hate AI agents or I think AI is bad technology. I run my entire life using artificial intelligence at this point. Makes me all my money. It's all I talk about. Um I've discussed this many times, but basically what I really don't like are hype bubbles and common misconceptions. So I just want to clear up both right now. So ### Understanding Traditional Automation [1:10] why do people get confused? Well, the biggest confusion I see is people not understanding the fundamental differences between three types of systems. We have traditional automation, AI automation, and AI agents. These three categories are a lot simpler than most people make them out to be. So, let me break this down for you. First, let's talk about what traditional automation actually is. Now, traditional automation is your conventional procedural automation. The old left to right flow where you execute some trigger, then a bunch of actions occur immediately after, and none of those actions include AI. We've been able to build these things for decades now. And in terms of no code tools, you can think Zap year, make. com, nonen where essentially you build these linear workflows that connect a bunch of different tools together via API. And they're ultimately reliable, predictable. You know, they've been around for years. We've been using a lot of them. For example, let me show you guys a simple automation that's traditional that I built in make. com for invoice follow-ups. Essentially, what happens is every day, okay, at 7:30 in the morning, we will grab the current date. We will then check a Google sheet, which is this Google sheet right over here. We'll get how long it's been since we've sent an invoice essentially. Then if it's been, I don't know, 7 days since we've sent the invoice, we'll send an email that'll look like this. If it's been 14 days, we'll send another email that might look like this. It's been 21 days, we'll send another email. It'll look like this. And so on and so forth. Okay? Now, we've had the ability to use these sorts of flows. In this case, this isn't actually going to send the email because none of the dates match our time period, but we've had the ability to do this sort of thing for a very long time. There's nothing special here. There's nothing magic. This sort of automation is actually now built into a large number of the tools online that people uh use, you know, like go high level, various inbound and outbound email tools, various CRM and so on and so forth. And you can think of these sorts of module flows as like a pipe, okay? Where essentially things will start at one end. So you have some sort of input coming in. In this case, it's just the current date. And then data will flow through this pipe left to right before determining which route it wants to go down. So it's not making any complex decisions. Uh it's incredibly reliable. And for most businesses, this type of automation can easily generate an additional 10 or $20,000 a year just by ensuring that invoices don't slip through the cracks. I consider it super simple but super effective, which is ultimately what I'm all about. Next up is AI automation. And this is where we ### Exploring AI Automation [3:29] take the same procedural flow, but we add an AI component somewhere in the middle. Maybe instead of just sending an email, we have a large language model that personalizes the content based on something about the client or the prospect. Now, something like this, which I built out an NADN, would be considered an AI automation flow. What we have again is the same left to write sort of system where in our case, we click a test workflow button that's down here. Then we proceed to do a bunch of things. We get some data set items. We have some eliminter. We do some loop logic. we do an if then finally we hit an API and then here's the important thing we use AI as one of the nodes in our step before finally adding a row and uh sort of working through like that. So this works essentially the exact same as you guys saw previously. We have some sort of left to right flow. Really the only difference is we also have an AI node that does something inside of it. It does some sort of flexible in our case personalization where uh the output is we end up with a bunch of highly customized fields that we can use in an email. Okay. Hey, I mean, I sell automation systems like this from anywhere between $1,500 to $10,000. These are very, very simple to use. As you can see, exact same logic as in traditional automation. We just sprinkle in a little bit of AI magic. Now, we also have a subvariant of that, and ### Dynamic AI Automation Explained [4:41] that's the dynamic AI automation. Instead of just using AI as a component within a fixed workflow, a dynamic AI automation actually allows the AI to make decisions about which path to take. You can see what makes dynamic AI automation different in this example. What we have here is we have a flow that I put together in NADN which auto labels and drafts incoming emails. Essentially, when a new email comes in, what we do is we pass all of those emails through a text classifier. The text classifier's job is to take an email, then put it into one of these categories. Transactional, invoices and receipts, cold air outreach, sponsorship/ affiliate requests, and QA advice, and community inquiries. Essentially what's happening in this case is AI itself is deciding which route to go down based off of the logic that we put in. It might choose transactional emails depending on whether it thinks that the input to this node is a transactional email. Might choose QA advice. It might choose unknown. Right? And based off of that AI logic we're doing the routing. Essentially what I'm trying to say is AI is making decisions about the flow of data rather than just processing data within a predetermined sequence. And as I'm sure you can imagine, this adds a ton of flexibility in your automations. But I want you guys to focus on this for a second because here's where a lot of people get confused. are calling this thing right over here an AI agent. Okay? But it's not an AI agent because at the end of the day, the AI is still operating and constrained within a set of these pathways that we put together. We've basically taken the time to bake in all of this logic. say you can go down this pathway, that pathway, and that pathway, and you get to choose which one. Um, AI in this case is really powerful, but it's ultimately just a router. It's a decision-making machine. It doesn't have the freedom to define new paths on its own, and it can only go down the sequence from left to right. Now, that's where AI agents come in. When we talk ### The Rise of AI Agents [6:28] about AI agents, we're really referring to something fundamentally different. An AI agent is a system that uses a large language model to decide what's called the control flow of an application. In plain English, it just means the AI gets to make decisions about what actions to take and importantly in what order. Okay? With an AI agent, instead of you designing the whole workflow like I showed you guys a moment ago where we have left to right nodes, you're giving the AI access to various tools, then letting it figure out the most efficient way to complete your request using those tools. And really, in a nutshell, if I just strip away all the complexity, all that means is at the end of the day, the AI agent decides what your flow looks like. You just give it the ability to call a tool when it needs to. Let me show you what I mean with a practical example. I recently built a website AI agent NADN that essentially functions as a chat button. Unlike traditional automations, this agent will receive a chat message right over here. When that chat message comes in, as you can see, we have an AI agent node which is right next to it. Okay. But there's no linear flow. I mean, you know, Naden allows you to create a linear flow if you want, but the linear flow isn't really the core part of this, nor is it entirely required. What we have instead is we have a bunch of tools that this a agent knows about and is connected to and can call depending on your query. Essentially, it's deciding the tools to use, the tools to select based off of the logic that the uh large language model has. Not the logic that we've baked into, but the logic that it is deciding for itself. So, for instance, if I were to open this up here and then chat, say something like, "Hey, could you book me a meeting for uh let's check what day it is today? " Let's say I got to make sure that it's a Monday because I'm my calendar blocks out um Sundays. Let's say Monday, March 24th, 2025. Now, when I call this, essentially what's occurring is this model has a choice. And that choice is in our case which tool should I call? And because it has access to these tools, it's asking us to fill in information that tool needs in order for us to execute it. So now we have a back and forth with this model. What it's asking us to do is could you please provide the start and end time for the meeting as well as the meeting summary or title. If you try and think about how we would have baked that in ourselves using some sort of traditional or AI automation, um essentially we would have had to have you know like a new chat message received then we would have to have some other question where the AI asks another sequence of specific things based off that and then finally you'd be able to proceed. But in our case we were able to do that just with one AI agent node. Okay. So, if I go back here and then I say, "Yeah, sure. I'm happy to. Let's do 3 p. m. to 4 p. m. and call it Nyx N8N agent AI agent seminar. " Now, what's happening is it's taking that information and then it's auto autonomously choosing and deciding to call the Google calendar create event node. it goes receives output from the Google calendar event node, processes it, chooses to do something else with it, which in our case is respond to us with that and then you know now we have u I guess a confirmation message essentially where at the end it asks if there's anything else I would like to do. So there in lies the difference. Okay, it's autonomy. Traditional ### Comparing Reliability of Systems [9:45] automations and AI automations are always going to follow predefined paths that you build. But AI agents have decision-m capability. That's what ultimately makes them an agent in the first place. And this autonomy is exactly what makes them both very promising for future applications, but also quite problematic right now, which a lot of people don't really seem to understand. So why are traditional and AI automations, in my opinion, better right now than AI agents? Well, now that we understand the key differences, um I want to dig just a little bit deeper into traditional and AI automations. First of all, despite all of the AI agent hype, these traditional and AI automation systems are actually still the backbone of the vast majority of the systems that businesses use. These conventional left to right flows where you choose the way the automation executes. These are basically what I'd consider the old faithful of the automation world. You will trigger an event and then a predictable series of actions will occur in a specific order and that series of actions occurs the exact same every single time. Meaning that you can build these systems and they'll typically run for years without a hitch. What you're doing is you're building a pipeline. What makes traditional and AI automation so valuable? The key thing is reliability. These systems typically have a very low failure rate. They approach zero. It's 0. 0001% depending on the uptime of the various services involved. And when you build a flow that checks your CRM for new leads and sends them the same email every time, it's going to do exactly that consistently and reliably without a question. So the only time these sorts of systems are going to break is just when the APIs change or there's some external system modification or whatever. And that does happen in practice, but it is super rare. Beauty of traditional and AI automations is in their simplicity. There is a clarity to the logical flow, right? Like when X happens, I want you to do Y, then Z. And that makes these systems incredibly easy to debug when something goes wrong. You can see exactly where in the chain the failure occurred and fix it. I like to think of it as tracing a wire to find out where that wire is broken. Traditional and AI automations are also super easy for developers set up or extend or maintain. It's basically like working with Lego blocks if you use one of our no code tools. Um, each piece just snaps together in a predictable way. That's how APIs work. You can swap with the components whenever they're needed. And this modularity makes these systems super versatile and adaptable over time. Now, here's the problem with traditional AI automations. And here's really the dream of AI agents and where they're supposed to come in, and that's flexibility. Okay, reliability and flexibility are essentially at two odds of a spectrum. If you wanted to handle different scenarios or edge cases or different input schemas like different types of parameters that you feed into an input without necessarily defining all of the logic manually yourself, AI automations and traditional automations are unable to adapt on their own to learn from these situations or apply even very elementary logic to them because you're the one that set the route. built the flow. The thing is most people overblow how valuable this is. In reality, usually if you have a business that's making a lot of money, which is ultimately all we really care about, right? We're just here to build systems that enable businesses to make more money, whether it's somebody else's business or our own business or whatever. If you have like a system that a business is already using and it's printing money, and even if it's a manual or whatever, and you want to sprinkle some automation in there, if you think about it, that system is a defined series of actions that occur very similarly every time, right? So that system is sort of like a manual automation. All you are doing by automating it and putting it on the cloud is you are building a predictable, reliable, repeatable flow over and over and over again. The last thing that a business wants if it has a pipeline that continuously and consistently generates $100,000 a month is something that mostly works or something that, you know, works 95% of the time or something that's super flexible and has all the capability to do all this cool stuff, but we only want to do that one thing and it just doesn't really very well. The way that I see it, and this might date me and make me a little bit older than I actually am, but it's the different it's the difference between like a general purpose CPU and kind of like an application specific integrated circuit. General purpose CPUs can handle any problem that you throw at them, more or less, they're typically just going to be a lot slower at doing that. And you know, because of their uh substantial reduction in speed, very few people use them for these hyperspecific applications. If you consider like an application specific integrated circuit, essentially what that means is we've actually built the architecture of the logic manually. Like we we've hardwired it into a circuit board and it just does that one thing every time. It's just a series of um electrical impulses essentially that follow from one transistor to the next. Okay. Anyway, nerdy example over. Essentially, what I'm trying to say is traditional and AI automations are the backbone of most successful business operations today. Because if you guys have an established process that makes a company 100,000 bucks a month, 200,000 bucks a month, and then you start introducing even the slightest bit of error like AI agents tend to do, maybe 5%. It's not that you're just losing 5% of your revenue. You're losing uh I mean, honestly, up to like 100% of that revenue, you're potentially costing that company their entire topline if you build a system that undermines their ability to deliver consistent results. Now, I've sold several hundreds of thousands of dollars worth of traditional NA automations over ### The Current State of AI Agents [14:29] the years because they deliver measurable consistent results without the unpredictability that comes with more flexible systems. And for most business use cases, this predictability is worth its weight and gold. This is what businesses want at the end of the day. So, where exactly do these AI agents come in then? Um, AI agents, you know, everybody says that they're these revolutionary pieces of technology. They're going to change automation forever. I think they will, don't get me wrong. And I don't want to be like that lidite that stands against this torrent of technological progress and tries to stop it, but they're not really at that point right now. And because of that lack of reliability, because of the fact that these agents are essentially just glorified chat bots at this point, they're very unlikely to really be able to be installed in a company and then drive immediately an actionable revenue. They're very unlikely for you to be able to define a process that prints money on a cyclic basis or whatever um and then hand it off to the age and then not be concerned about its ability to do things. Okay, so let's look at this example again. I know I just showed this to you before, but essentially walking you guys through this. What we have is we have a chat interface that I can open by clicking this chat button. We're storing memory, aka we are adding messages to some buffer using this window buffer memory option. We're using OpenAI to be our chat model um in our case GPT4. And then we have a Google calendar tool, a ### Evaluating AI Agent Performance [15:46] Gmail tool, HTTP request tool, a calculator tool, a Wikipedia tool, then we also have a vector store tool. Okay. Now, when you send a message to this agent, what it is doing is it's deciding which tools to use and also in what order based off what you're asking for. I'm sure we can all agree this stuff looks super cool from a bird's eye perspective. If I ask it to book me a meeting tomorrow at 3 p. m. like it did yesterday, it'll interpret that request. It'll check my calendar availability. It'll create the event. It's a super undeniable highv value wow factor, and it makes for great demos. But the issue is the reliability cost. If I were to run this exact agent just like I did before a 100 times, it would fail somewhere between one to maybe five times. Okay? And that's as of the time of this recording. There's a infrastructure framework that's gaining a lot of popularity recently called MCP. The whole idea is it makes things substantially more reliable. We're not entirely there yet. I'll make some videos on MCP pretty soon. Um, but I just wanted to point that out. Now, one to five errors out of a hundred requests might not sound like much, but that is astronomically higher than a traditional automation to the tune of several thousand times. If we go back to something like this, okay, if I were to run this 100 times, it would fail zero times. If I were to run this a thousand 10,000 or 100 thousand times, might fail once. Okay, this is an order of magnitude better and more reliable ultimately for businesses that ultimately care about uptime and stuff like that. Um, obviously businesses are going to they're going to prefer simple and repeatable systems. Now, anyway, the point that I'm trying to make is since you're communicating with this model, it's not always going to interpret your request correctly. Sometimes you're going to say something like, "Hey, could you book me a calendar event? " But then instead of booking you a calendar event, what it's going to do is it's going to go down this HTTP request route instead. Now, in our case, this is actually doing an API call where it gets the current weather. Okay? But maybe just because uh you know large language models are flexible and they have the opportunity to make a mistake. What it'll do is it'll make an HTTP request to that get current weather API and then it'll add like the city and make it a calendar instead. It'll basically hallucinate and invent stuff. Doesn't happen super often, but it happens enough that uh a lot of businesses are I want to say skeptical of using these sorts of things for applications that are high cost um and not necessarily high ROI. Now, I've also previously touched on a cost that a lot of people don't care about, and that's the prompt engineering of these models. Uh, like if I were to, I don't know, ask this to send me an email. Hey, can you send an email to Nicholas Sarifgmail. com, congratulate him on his recent career move? He started a new company called Leftclick, right? If I tell it to do this stuff, which is one of the most common um I want to say examples of AI agents where you have it send some update or whatever, you know, it'll do the thing most of the time. In my case, it'll send an email to Nick congratulating him on the recent career move. But if I actually go back to my email inbox, I just open this up for you guys and we actually look at like the quality. Essentially, what we have is, "Congratulations on your new venture with these capitalizations. Dear Nick, I hope this message finds you well. I wanted to extend my heartfelt congratulations on your new career moving the second launcher company leftclick was a fantastic achievement I know you make a significant impact in your new venture wishing you all the best as you embark on this new journey best regards your name so I guess what I'm trying to say is imagine if you know you sold this to a business owner or something and the business owner just didn't fully understand the capabilities of a chat model like this because they're not you know AI automation specialists they're not developers they don't really get how all of this stuff works deep down that's why our value is our value and imagine if they sent an email like this to some past client or something. Okay. How do you think that client is going to take something like that? Pretty poorly. I guess the point that I'm trying to make is if you try and use NADN AI agents in some sort of customerf facing application, there's a high likelihood that this is going to negatively impact how you're perceived, which can have massive downstream effects if you're not careful. What if instead of all that fancy overengineered LLM stuff with the AI agent and the tools, right? I mean, that's the definition of overengineering. you are building the wheel of all wheels and all you need is like a little round pebble or something like that. Um, what if instead of that we had a simple set multiple variables node in make. com that just says call with Nick March 25th, 2025, 3 p. m. to 3:30 p. m. We had two steps where the second is we create a calendar event directly in Gcal. And what if we just simplify it and cut down all of the complexity. Okay, this is going to work basically every time. It's substantially easier to set up. It's infinitely more scalable and it's also much more maintainable. Sure, it doesn't have a cool chat interface, but what do you think is more valuable to most businesses at the end of the day? Is it a nice chat window or is it actually getting things done reliably? I mean, how many situations are there where people want to book a meeting with you? Don't focus on the user interface and making the user interface as complex and cool and sexy as possible. Well, just make sure the thing that you want them to be able to do is doable. So, in our case, instead of offering them a chat window, just give them a calendar, right? It does the exact same thing. And people are much more used to using calendars than they are chat windows anyway. So, that's the fundamental difference. You know, with AI agents, you're giving up maybe 5% of your business reliability for what ultimately amounts to a cool little chat box. And that 5% might represent $50,000 or maybe even $500,000 a year. Any serious business owner is not going to be making that trade-off. Does that mean that AI agents are useless? Absolutely ### When to Use AI Agents [21:01] not. They're incredible for personal assistance, experimentation, non-critical business processes, and I also think they have that they have their place in sales. Um, they're definitely going to become a lot more valuable because of stuff like model context protocol. And I mean, I use AI constantly in my own work. But for core business processes and operations where reliability is very important, traditional automation and AI automation is going to reign supreme for quite a while still. So, now that we understand the key differences between traditional automation and AI agents, how do you actually decide which one is right for your specific situation? Let me walk you through my own thought process when I'm evaluating this sort of thing for clients. To make a long story short, given the benefits and the drawbacks of traditional and AI automations versus AI agents, here is how I choose. I use traditional automation when one, the ### Choosing the Right Automation Type [21:44] process is well defined and follows a predictable pattern. Two, reliability is non-negotiable like financial transactions, legal documents, or flows that are producing a high ROI already. Or three, the decisions required are clear-cut. like if it's A then B or C then D, I'm going to be using traditional or A automations. Or maybe four, uh if I need to maintain precise control over outputs, that might be particularly important in um I don't know like uh like enterprise applications. And also five, if the process rarely changes. Okay, if all of those or hell, if a couple of those are true, I'm going to be using traditional or AI automations to do that client request just because it's easier. It's Lego blocks. there physical little pieces that you are building that you can then interchange swap out modules when necessary and then just like you know build repeatable templates off of super easy and super scalable. Now I'm going to be using AI agents when I'm building something that's customerf facing something where the wow factor of a chat window is very valuable and important. Uh something where the cost of an occasional error is minimal aka it's not something integral to the lifeblood of your business like a booking flow like I was showing you guys earlier. You just use a calendar for that. Uh, but it's something that's just low stakes. It's, you know, you trying to get information, it's you browsing through a website and using some sort of chatbot to tell you more about the company, that sort of stuff. Uh, and also when the workflow needs to be able to adapt to a wide variety of unpredictable inputs. So, like a chatbot, like a customer booking or customer support agent, somewhere in that process, little cracks are going to start to show up and you're see where businesses solve their problems organically rather than systematically. You want to focus on situations where the customer tried solving their problem organically because it means that they didn't really put a lot of thought into it. It's basically like a fire that they had to put out immediately in front of them. Um and you know when it's a fire you're less likely to think through things like a systems or a process engineer you're less likely to think through things in the sense of huh you know how can I set up my system in such a way that doesn't just solve this problem once it actually solves all iterations and instances of this problem forever which is more you know processoriented. So uh yeah, after that you just go through my little checklist essentially and if your solution is more similar to what a traditional or AI automation would be able to do, then great news because you can build extraordinarily high ROI flows super quickly. Um whereas if you know you have multiple uh inputs, if a chat window is really important and so on and so forth, then I'd probably build an AI agent for them. Here's my prediction. Over the next year or two, we're going to start seeing AI agents get significantly more reliable, and they're eventually going to reach a point where they become viable for uh more and more business processes, including those higher cost, higher ROI ones that I was mentioning before that I think that traditional NAI automations are better for. The way that I see it, okay, is like it's like a freshly paved road. You see the new ash fault glimmering in the sunlight. How tempting is it to drive over that thing? Extraordinarily tempting. But if you're a wise driver, you're probably going to wait until the asphalt sets to avoid getting stuck or avoid screwing up your wheels. Very similar situation that I see with AI agents. I think the ashvault is setting. It's in the process of becoming very robust. MCP and a bunch of other protocols are going to help with that. But as of the time of this recording, we're not really there yet. So, the future will likely involve hybrid systems. Traditional automation handling the mission critical backbone operations with AI agents providing some sort of flexible interface and then handling the more unpredictable parts of the process. um MCP um a couple of the other infrastructural paradigms uh those are going to help with that but they're only going to be a part of it not necessarily the whole thing. Finally, here is the most important thing that AI agent promoters won't tell you. You guys can always just add traditional automations into a future AI agent once the technology matures. So you can always just add a flow, okay? like spend all that time creating a flow and then when the AI agent technology is at the point where it can actually like choose reliably which tools to use, you could literally just add that as a tool later on. So it's not necessarily that like if you spend your time building solid reliable automations now, it is a waste of time. It's precluding you from being able to incorporate more advanced A capabilities later. The, you know, secret sauce really is that it's going to give you a stronger foundation to build upon and more cool things that your Agent can do when they eventually get to the point where they're reliable enough that you want to build them for businesses. Now that you understand the real differences between a agents and traditional automation, I like to think ### Equipping for Future Automation Challenges [25:52] that you guys are equipped to make better decisions about which one to implement in different scenarios. Um, ideally this knowledge is going to help you build systems that deliver measurable results for businesses, not just things that look cool and make for great YouTube demos. Ultimately, these successful automation specialists, in my humble opinion, aren't just technically skilled. They know how to identify highv value problems and implement solutions that deliver a genuine return on investment. They understand exactly when to use traditional automation for reliability, then when to leverage new and upcoming technologies like AI agents for flexibility. If you guys like this, I got over 160 videos on my channel showing exactly how to build these automations and use them for business step by step. Everything from invoice collection to lead generation to client onboarding. All of it is there for you. But if you guys are serious about turning your automation knowledge into an actual business, I highly recommend ### Building Your Automation Business [26:35] you consider joining Maker School. It's where I help AI automation agency owners land their first client in 90 days, guaranteed. Maker School gives you guys everything you need to succeed. Complete business roadmap with templates, proposals, and scripts that I use to scale my agency to 72K a month. You'll get ready to deploy automation systems for lead genen sales and client management. You also get direct coaching through live calls, hot seats, and daily personalized Q& A. And ultimately, you're going to get a focused community of over a thousand serious AI automation entrepreneurs committed to growth. Whether or not you guys join, and I'd like you to, but you don't have to, my recommendation is you start applying these automation principles in your work today. Don't just watch this video and let life and this super fast-paced industry pass you by. Build solutions that solve real business problems using the right approach for every specific challenge today. Take what I've shown you guys about these linear uh traditional and AI automations and actually put a couple of pieces together. That's how to get somewhere. You essentially just have to get started. So, if you've been on the sidelines up until now, this is me giving you permission to get going. Thank you very much for watching. Peace out.