The Most Overrated AI Agents (+ What To Sell Instead)

Hey, so a lot of the automations and AI agents that I'm seeing people build online right now demo really well, but they don't actually solve real business problems. They don't actually deliver real value. And as I'm sure you can imagine, that's a problem when you sell it to a customer and they're expecting an ROI. So, what I want to do in this video is I want to show you which things not to sell, which things don't really provide the value that they might be marketed as having, and which systems you can actually go out and sell for anywhere between $1,500 to $10,000 a pop. My name is Nick. If you don't know who I am, I scaled an automation agency to 72K a month. This is basically all I do and I've seen more or less everything in this industry at this point. So, I'm just going to be breaking down the most overrated AI automations and agents and then what to sell instead. So, to get

started, this is your average AI agent seller right now, a little bit burnt out and probably not having a good time because the vast majority of their clients are turning. And then this is hopefully going to be you. Super cool. Giga chat. I was told that I should try using memes in these videos. So, engagement rate wise, uh, that's what I'm doing. Okay. So, I got a lot I want to cover. I mean, like, if we just zoom out a bit, I have a bunch of systems on the right hand side that I'm going to show you that you guys can sell. And then I also have a bunch of systems on the left hand side that I'm going to highly recommend that you don't sell. But let's just start with the very simplest permutation of this, which is

what makes an AI automation or an AI agent or a system useless, outdated, or overrated. Okay? And there really a few things the way that I see it. The

biggest is false autonomy. So, you know, you'll get somebody interested on the other end of the line for some automation, keeping in mind that the whole idea is that it's automated and then you'll sell them a system and then that system won't actually be automated at all. It will essentially require some sort of constant human monitoring in order to do the thing that uh you sold the person as it being able to do. So, I'm sure you can imagine if you guys sell something that you market as autonomous and then later have to do a bunch of maintaining um you know, the person that you sold the thing to is probably not going to like it. That's going to lead to low retention. It's you not getting a follow-up project with the client, and it's going to lead to just like general dissatisfaction, which I think we should just avoid, right? We want to deliver good things to clients. The next big thing is

unreliable execution. So, what I'm seeing a lot of these AI agent flows do in particular is, you know, they they're really flexible and they're great, but that flexibility with AI is always at odds with reliability, right? Um, you can see flexibility is both a pro and a con. A lot of the time for an established business, they have some business process that already generates them money and they just want you to automate it and scale it up. So, if you think about it in that way, flexibility a lot of the time is not a benefit. It's actually like a drawback, right? You don't want the system to be flexible. If you have a pipeline that's working freaking amazingly, okay, and I don't know, I'm just going to draw a little pipeline example from left to right. I swear that is not what you guys all probably think it is. Okay. Oh my god, that's getting even worse. Okay. Anyway, um let's say you have a pipeline. Okay. I'm never beating the allegations. And uh the way this pipeline works is you know you have a new leads or whatever on one end and then on the other end you have a bunch of money. If you think about it this is more or less like a business. Okay businesses you conceptualize any business is just a pipeline where you get some lead in at the front end and then you perform some process in the middle and then that process results in money. This is a very particular sort of process. Why would a business that already has one of these things that's making, I don't know, let's say 100K a month or something want an AI agent that starts, you know, putting stuff in other pipelines that may not deliver the same sort of result. Why would we want something that is inflexible? If this is even a 1 to 5% error rate on both sides of this, okay, that's like, you know, you're immediately shaving off like two to maybe 10% of your profit and you're diverting a lot of value from this really well-established pipeline into these errors and rabbit holes and other sorts of processes. Businesses don't really want this in practice. I know that wasn't the best um explanation, but if you guys want to be automation engineers, automation specialists, don't focus on like building new things from scratch for customers to start. focus on taking an established process that's already generating some sort of revenue for them and then just making that a little bit better because a 5% improvement to something that makes you know 100 grand a month is going to make the customer at least 105 grand a month and then you can justify charging some fraction of the five grand a month in value you're providing the third major

problem that I see is no tangible output okay and this is a really big problem at the end of the day the automations that we're building are pretty ephemeral they're pretty abstract right they're basically just a bunch of cloud services that just connect between each other and send and shuttle data to and from If you think about it from the business perspective, most business owners do not inherently understand what these systems are doing. They find it really hard to ascribe value to the systems because there's just a bunch of cloud server HTTP calls and stuff, right? So, the number one way to make your system a lot more valuable and charge a lot more money for it is when you provide some sort of tangible output to the customer. What I mean by tangible is focus on things that they can like see, touch, or feel. So, stuff like a document that your thing outputs, stuff like an update, some notifications, stuff like some metrics, a dashboard, a visualization. You know, if you generate a lead for a client, don't just generate a lead in a CRM or whatever. and then give them an update saying, "Hey, we just generated you a new lead. " And what I see is a lot of these agent systems, they I don't know, they're just not really well thought out for that purpose. I think a lot of the things that people are talking about now are probably they're cool and they're focused on new developments in the space, but they're not actually focused on things that like make money, you know? They're like they're shiny objects essentially, which is actually the next thing that I wanted to talk about. So, a lot of these

systems are shiny. They're focused on novelty and like, "Wow, this new gamechanging platform just came out. You should really use it in your business. " But it's like, you're kind of putting the cart before the horse. You know, you're losing the forest for the trees. Just because this new technology came out doesn't mean you should apply it to your business. What you should do is you should see if that new technology allows you to solve a pre-existing business need better. And what are some pre-existing business needs? Like revenue, like savings, payroll, headcount, right? These are things that actual good and valuable automations

actually focus on. Another one is high maintenance. This sort of feeds into that first one, but essentially a lot of these flows that I'm seeing need constant fixing, debugging, or handling. So you might build a cool chatbot or something for a customer. And then at the end of the day, you know, after the first day, like a client's going to be like, "Hey, can you make this adjustment to the chat bot? I don't really like how it um how it talks, but like you know, put yourself in the shoes of the" and you know, a lot of you guys are probably like actually experiencing this right now cuz a lot of you guys got stoked on selling chat bots and are doing it. How much control do you actually have over the output? Did you train the AI model? No, you did not. Did you put all the infrastructure and all the GPUs in place to do it? Did you give it the data set? No, you didn't do any of that stuff, right? We're engineers that are taking a thing that's been developed and then we're applying it, but we actually have limited control over the thing itself. Sure, we could prompt better and sure we could tell it, hey, can you please not say that thing that the client doesn't want me to say? But you can't actually fully control it. So then you need to constantly maintain it. It's super annoying. You constantly debugging. Another major issue is that it's easily commoditized. So what I mean by this is there's just

some lowhanging fruit here where you know things are going to be replaced by features from OpenAI, Microsoft or Google pretty soon. So a good example of that is like the image generators, right? Like if you build your entire business and AI automation isn't exactly this, but let's say you started like a SAS company and you build your entire business around AI image generation and then a week ago GPT4's AI image generation drops and it's just like way better than everybody's and you could do it through like a chat interface and now they're dropping like a visual drag and drop thing like you know how sustainable is that business over the long term? Probably not. Another one is uh that it

solves non-urgent problems. A lot of the systems that I'm seeing don't tie directly to revenue, cost or time leverage. These are the three big things that I always cite in my YouTube videos and my communities. Maker School, make moneywithmake. com. Essentially, these are just things that business owners actually care about. Like, let's be real. We're all in business here to make money. Like, it'd be nice if we could do so with like our Optimus robots serving us food and, you know, us listening to AI generated music and all this stuff, but all these are really distractions at the end of the day. Like, the technology itself is not the end goal here. I don't think for most of us. The end goal is us building, you know, a lifestyle that enables us to not have to be wedded to work. And then the fact that we can do it all from our laptops is dragging and dropping some modules across the screen and having some sort of minimal contact with clients. Like that's the end goal, right? Uh security, safety, feed our families, experience things that 99% of people will never be able to experience. That's the goal. I don't think the goal is actually just like hunching over some new piece of technology and trying to understand like the transformer architecture, right? As dope as that is and as enjoyable it is to be curious about that stuff. Most people here aren't here for that. So, just want to make sure that we're solving the things that matter and the

things that actually drive us towards generating revenue. Another one is lacks defensibility. So, anyone can copy or undercut it with a cheap subscription. Um, I just mean like there's some SAS products out there that just do some of the things that we are, you know, that I see people selling and they just do it way better. So, it's like why don't I just spend $5 a month for a SAS subscription with a 99. 9% uptime guarantee. Then finally, there's a lot

that have bad risk to rewards ratios. This is one of the most important parts. I put this at the end here. Um, but uh this is one of the most important parts. It's like I see a lot of people trying to automate things that will save them a tiny bit of money. A good example of this is like an AI outbound caller or something. Saves them a tiny bit of money and the cost is $100,000 a month in opportunity. Okay, I'll cover more about exactly which systems have all of these problems. Um but you know, just from a bird's eye perspective, um yeah, I just wanted to give you guys sort of a run through um of what that looks like. So AI agents don't make money. Change my mind. I mean, I know this guy's blowing up all over the place, or he was a few years ago, but yeah, this is basically me here. Agents don't really make people money right now. So, that's the reality situation. So, let's start it off with some things that you should not be charging money for or things that I don't personally recommend that you build as of the time of this recording. Um, I should also just add a disclaimer like technology is going to get better and so some of the issues that I'm describing with these software platforms and tools, they are going to improve. So, you know, if you're watching this in like 2032, like odds are a lot of these problems have been solved, and maybe you can actually go about selling them and delivering an ROI. But I care about like what we're doing today, and that's really the value here. Um, the first is basic chat bots and FAQ systems. So, the idea is that you have some sort of Q& A system on your website. It's usually some sort of GPT powered bot, you know, through like an NAN agent that just regurgitates fat content. And the issue with this is it's usually like rag. Rags stands for retrieval augmented generation which uh basically means like your chatbot just performs a mini Google search over your own data to try and find the snippet of text that is relevant to answer the question. Once it finds that then you inject it into the prompt then the chatbot uses that alongside its prompt to answer. So I mean you know it sounds really cool and stuff but you know I've sold a few of these at this point and um the promise which is that they're 24/7 customer support replacements. This is the promise that most people are making not the promise that I make. the promise that it's a 247 customer support replacement. Uh, I mean, it's it's not really there. It's not really at the point where it's a 247 customer support replacement. Sure, you can answer really simple questions, but I got to be honest, if somebody's at the point where they're like talking to a chatbot to try and answer a question, if they're that motivated to answer a specific question and like there's a section on your website that just answers that question, then they're just going to answer it in on the section on their on your website. You know what I mean? Like if you have a section of your website on your lander called like the FAQ, like they're just going to go to the FAQ, scroll down, they're going to see their question, they're going to click it, they're going to read it. Like you don't actually need to have this back and forth cool, you know, chat experience. This to me is like the definition of losing the cart for the horse. You know, you have some cool shiny object in the bottom right hand corner that just takes away from like all the information that you are laying out in your website anyway. Um why have like a crappy answer when you could just have the best answer ever? So, in practice, uh, you know, people use these and some companies are using these like semi-successfully, but definitely not for 24/7 customer support replacement. The main cons are they have pretty low engagement rates. In practice, most visitors ignore them because they're just like, "Oh, this is an AI bot, so it's just bound to be low quality. " And most people, to be honest, just still don't really trust AI. They don't really think AI is as good as a human being. And I think most of them are right for it. It also can hallucinate answers and outside their knowledge base. Has a very minimal impact on conversion. like I don't think this increases it at all, but I did say that, you know, like could increase it a little bit. Um, but again, I don't think this is actually really improving things. I also find that in order to like make it good and consistent, it does require extensive prompt engineering which just ends up taking a ton of time and then there's always that like back and forth with the client on on, you know, this isn't saying that sort of thing that I want to do. Um, a lot of people like jailbreak it and then end up just using it for their own stuff. So, I don't know, like generate me hot pics of Scarlett Johansson. Most businesses also already have FAQ pages that do the same job or FAQ sections in their website. Or if you're smart, to be honest, you will just answer all the questions that a person could ask the bot just through the website itself. Like they read it top to bottom. They know 90% of the stuff and then they're motivated to buy. This thing isn't really doing a lot that um good website design or good copywriting will solve. Obviously, there are some limited pros, which is that it takes like 5 minutes or 13 minutes or however many minutes. in my previous video where I showed you guys how to do this setup can handle some basic repetitive inquiries and it can reduce support tickets for common questions. So, um this is actually pretty good for like a back-end system. So, like if this is like a customer support agent and they're already a customer and they're asking like, "Hey, can you do this super templated process? Can you initiate a refund? " Like Amazon has a great example of this. They have like a refund bot basically where if you go down the refund route, you're like, "Hey, I'd like a refund. " The bot literally just calculates. It's the funniest thing. The bot literally takes how much money have you spent with Amazon in the last year or something and then how much money have you requested in refunds and if that ratio is below a certain thing, they will just give you your money back immediately. Like that's a great example of a bot that's probably going to work 99. 9999% of the time because it's like a procedural equation. Amazon's just like, "All right, like let's just keep our refund uh total down to like 3% or less. If it's 2. 5%, we'll just give them the refund and then they don't need to worry about it and hassle our customer agents like our actual humans, right? Like that that's a good example of one that works. But usually on the front end, you're not doing that. You're not giving people money and you're not actually going through that process. So yeah, the promise here is that it's 24/7 customer support reduces support costs, increases conversions. In the reality, um it does basically none of those things. Okay. Um there are a couple of alternatives to that. Obviously, human support for complex queries, like if there's a human being that's online that happens to be able to answer that question, people are going to want to communicate with that more and you're going to get more perceived like brand value if there's a person. Um, you could also just get like a well structured FAQ page or FAQ section, I should say. And ideally, you just answer those questions before, right? Okay. Another really big issue is appointment setting agents or AI. So, um, you know, the idea behind this is an AI system like a go highle system or something that handles booking, rescheduling and the managing of appointments. Um, these things are often voice or chat based interfaces like a Vappy agent or something. And the whole idea is that the promise is that the they're essentially a virtual receptionist or a scheduling assistant like, hey, you know, just hire virtual Peter. Peter will do everything for you at a millionth of the cost of an actual person. So, we'll just install that guy in your business and then you, you know, we'll charge you $3,000 a month for maintenance. It's like, that sounds really nice. And the person gets to let go of one of their staff members and maybe start making a little bit of money on the arbitrage. Um, this is exactly what I was talking about when I said losing the forest for the trees. Here's how the math works out. Okay, let's say we have two options, okay? Like this is human and before. Okay, let's say every time a new um a lead comes in previously the new lead was converting into a meeting at something like a 30% conversion rate. Okay. Um this is reasonable in practice, you know, maybe 30 to 50% depending on how quickly you can get back to them. and then you know this meeting um should it go through should 30% of the meetings go through and you land a client I don't know maybe you just make $1,000 very simple math okay let's say you get a hundred of these a month so basically your business if you think about it you get 100 people they go through the 30% thing and then you know uh you basically get uh $30,000 a month that's how much money this business is making in addition this $30,000 let's say the actual customer support agent cost $5,000 so you know what you're really making is you're making $25,000 after a cost of marketing and stuff. And you're thinking, "Oh, that's great, but we could probably save that $5,000, right? Maybe I'm just going to build an AI system to do it all for me. " Well, here's how that math actually usually works out in practice. Okay? You have a, you know, this is the AI version. You have a new lead that comes in and then instead of giving it to a human, you give it to an AI. Okay? The thing is when you give it to an AI, AIs make mistakes and people really don't like being sold to by AI. So, what is inevitably going to happen? You may have eliminated this customer support person completely and saved yourself $5,000, but you will also substantially decrease your conversion rate. So maybe now you only convert 20% to meetings. Okay? So 20% times 100 leads is only $20,000. So $20,000 minus I don't know, you spend like uh $2,000 or something instead of the $5,000 on an AI person to handle it uh you know, like one of us basically to handle it for you. Your net is now $18,000. So, what actually happened? Sure, you saved $3,000 by replacing this person with an agent, but then at the end of the day, you lost $7,000 because you lost um your inbound. So, this brings me to a general rule of thumb, which is like at the first point of contact with a customer, you want that first point of contact to seem as human as freaking possible. And if AI and humans are still not exactly at the same point just in terms of like their quality, their ability to respond, um how the person on the other end of the line feels like you want to make sure that people feel respected when they reach out to your business. If they just feel like another cog in the wheel, the likelihood that they're going to goes down a lot. So the main issue is if you replace that with an AI agent, then basically what you're doing is you are you're saving a tiny bit of money, but then you're losing a massive amount of opportunity cost, aka money that you would have made had you actually had a human being in the process. So, don't

lose the forest for the trees, I guess, is what I'm trying to say. And these appointment setting agents have a number of problems which interfere with their ability to do things. Now, I'm not going to go over all of these because, as you can see, um, we're getting kind of wallet texty, but they have a big failure rate. Also, you guys are going to have this whole document down below in the description, so feel free to take a look at it. U, they have a big failure rate, especially with complex scheduling requests. Um, limited context understanding, which leads to more failures, okay? And it's difficult for them to do things like prioritize urgent versus regular appointments. If you try and build that sort of logic in, like there are some appointments that are significantly more urgent that you need to deal with than other appointments. And so in reality, if somebody's going to pay you $10,000, you want to make sure that they get dealt with first, right? This sort of second and third order logic is very difficult to implement right now with these schedulers. So you end up with these like very brittle systems that don't actually allow you to flexibly um deal with the front end of your business, which to be honest, I think is the most important part of your business. So don't sell appointment setting agents either. The marketing promise is that you get a virtual receptionist that eliminates scheduling, has 24/7 booking capability, but in reality you have very high failure rate, very poor rescheduling, you can have minor issues like time zone confusion and then also you kind of lose context in the clients which ultimately results in you know human intervention having to come in and you just like defeating the whole purpose of the savings. On the same vein, AI sales agents, so instead of inbound um schedulers, but like actual sales people suffer from all the same problems that I showed you guys earlier, except 10x that many problems. So they don't just suffer from, you know, like a 30% to 20% conversion rate. But considering that this is the entire sales arm of your business, this is even further ahead if you think about it than that. This is like the most important part of your entire business. If conversion is over here, okay, then this is the part of the process that we are essentially trying to outsource and the closer something is to conversion generally like the less you want to outsource this, you know, less outsource. That doesn't make sense, but you guys catch my drift, right? So, there are some, you know, vague supposed benefits of this, like it can handle lead qualification. Um, if you're smart, you can use an AI sales agent to like immediately send somebody a text message or SMS the second that they like get to your business or whatever. These are minor little benefits, I would say. And there are, of course, some nuances, but yeah, you definitely don't want like an all-in-one agent that qualifies leads, handles objections, and closes deals because the gap between them and um a real uh human being that is like a closer is so much bigger than the gap between a real human being that's a setter. Uh, so if we were losing $7,000 in that example before, we're probably going to be losing like $15,000 here. Just doesn't make sense for most businesses. Another big issue

is outbound voice callers. So I see this a lot now because I made a Vappy video where I covered, you know, how to have like an inbound agent essentially. Listen, outbound voice callers can work for particular purposes. One big cool purpose they can work with is end of year reactivation for like some sort of health clinic where you have some sort of benefits or insurance or something. You know, it's like December 9th, let's say. you have a secretary that's like overwhelmed and there's 500 people on a list that have insurance benefits expiring that you just need to find a way to call. Right? In that situation, it makes sense to get yourself one of these outbound AI callers and just have them blast the same message to all 500 people on that list. Save your secretar's time and then, you know, have some percentage of those people want to rebook some appointment to take advantage of their insurance benefits. That makes sense. I've sold that myself. I've made money off that. And a lot of people here have sold similar systems. Voicemail drops and stuff depending on legality and what not. that can make sense. Onetime offers for I don't know like um free teeth whitening or like electric toothbrush or I don't know a free laser uh skin appointment or something that stuff can make sense. But actually having like an outbound caller for I don't know like cold calling or whatever. Dude, that makes no sense right now. That makes no sense. It's not very good. And the vast majority of human beings the second that they smell that there's even a 1% probability that this thing is an AI, they're going to take their hands off the phone and they're going to click cancel. Okay. It just doesn't make sense for the vast majority of situations. How many AI phone calls do you guys currently get a day? And how many of them do you actually listen to? I probably get at least 50. Probably more than the average person because I expose my goddamn phone number all the time inadvertently. But how many of those do I actually listen to? Like one. It's when, you know, Safeway calls to tell me that they have an order for me or something. That that's it. So, you know, tons of cons here. Um the promise which is that there are scalable alternative call centers or SDRs. This is just not true right now. Maybe it'll be true in the next like two years. Who knows? But right now, we're not at that point. And if anybody's trying to tell you that they are, just be very careful. You're probably seeing like a manicured top. 01% demo. You're not seeing like actual boots on the ground um grassroots uh systems that people would actually build in practice. Okay. And yeah, these are the limited pros. This is what I was talking about with the initial outreach attempts and then some sort of regulated industries, usually health, and then simple information collection and controlled environments. That might work as well. Anything else though, not really. So, here's that marketing promise again. Unlimited outbound calls, nope. Perfect compliance, consistency, nope. Cost effective call center? Nope. The reality is that everybody hangs up. More and more systems, calling systems are blocking them. Usually have some pretty crappy voice qualities that are still in the uncanny valley. And then you have um I don't know, poor conversation handling, I want to say. Okay, let me get to a pretty controversial point here. um because I made a system that was an u an SEO content producer and that was one of the systems that I think made me grow pretty big on YouTube. So these SEO AI content producers, they just don't really make sense for most businesses unless the content that they're producing is like hyper relevant and actually being read. If you're trying to game a system right now, aka the system of Google with SEO AI content, I don't consider that to be a very valuable thing that you can actually sell to customers at this moment in time. So what is the promise here? The promise is basically, hey, instead of you doing all this stuff manually, let me do um AI content generation, okay? And I'll make a content factory for you that'll just drive massive amounts of organic traffic. In order to do that, we're going to use GPT. some similar models. It'll generate articles, blogs, and landing pages. And, you know, we'll essentially just be able to take in a bunch of keywords and a bunch of topics, and then we'll just be able to do thousands of blog posts a day or something like that. Um, in reality, Google is now actively penalizing AI generated content. But more important to that is that AI generated content just doesn't really like people are looking for alternatives to this stuff anyway at this point. AI generated content for most purposes is like it just tends not to be very deep. If a person's at the point where they're like clicking on like the f, you know, bottom of the first page of Google to look for something, they probably want like a very nuanced deep answer. and AI is just not really yet at the point where it can do that at scale consistently with most of the approaches that I'm seeing. The reality is if I were to just like pump in the same search query to chat GPT or GPT 4. 5, I'd probably get like a better answer than I would if I went on that. And I'm not trying to say that like just because I could do that invalidates this as a business process. But the way that I see it is there was basically a short period of time where it made sense to do this and if you did this you could actually go and you know you could make money and I think personally maybe some people are going to call BS and that's fine. Um, this was around like 2020, maybe like 2021, maybe like 2022. Okay. And I think that there was like kind of a big opportunity here, but then right over here, chatbt came out and a tons of search engines and everything started penalizing AI content. And I don't really think like if you consider this as the value of the system over time. I don't really think, you know, here us being right now, I don't really think that there's much value in this at all. So, my reasons for this are that you typically get poor engagement metrics on this content. And whether that's because Google or other search engines are actively penalizing um AI generated content or just because most of it sucks, I don't know. Um I do find that most implementations of AI content generators do blow. And so, because they blow, you just don't really get very good content. And if you don't get very good content and Google sees that not a lot of people are spending time on your website, that just tanks everything else. I think that um the good stuff requires human editing. And you know, I wrote defeating the purpose here probably a little bit preemptively. It doesn't defeat the purpose. You could use like AI content in concert with human beings and you could have that be pretty fine, but um you know, you are going to have some sort of editing essentially to make this viable. Now, there are some limited pros here. U my content writing company, 1 second copy, does this and we've been doing this for the better part of many years now. um we will have AI generate or pre-generate content for you and then we'll have human editors and human writers go in and then refine the content but usually if you think about it like we'll do like 50% of the work with AI and then the other 50% of the work will be human and that's a far cry from what most people think which is most people want like 90% AI and then like 10% human so the bar is increasing right um I think this could work for extremely longtail or low competition keywords we're talking like really procedural stuff like um you could have AI generate like uh 5,000 small little regions or something like that in Arkansas or something and then you could like actually rank a very specific zip code based or region based search that might apply to like a very particular subset of people. I think that might work. Um and I do use it to generate content outlines or research summaries. But yeah, just in terms of like mass SEO content websites and stuff, uh that opportunity was here. And I think a lot of people are still trying to sell the opportunity over here as if it was here. And it's like, no, no, man. There's a big gap basically between what people are saying that the systems are going to produce and what they actually produce. And guys, if we just get in the habit of selling systems that don't work to customers, you also think about it just strategically, our entire industry gets devalued because there are a lot of empty promises and fake things being thrown around. If there is any industry to be devalued, to be honest, it is our industry because we're working with like the shiniest technology on planet Earth. But just wanted to show you guys my thoughts there. Okay, the next one are email agents and GPT responders. Um, I mean, like, you know, I've sold a bunch of these and I use a couple of these in my business. So, I'll show you guys what not to do and then what to do. Um, and then again I have a bunch of alternative systems that you guys can actually sell later on in the video. But uh, essentially these are AI systems that will draft, send, and then respond to emails completely autonomously. The idea is to process inbox coms with minimal human input. And the whole marketing idea is never miss an email again. In reality, you can use them to filter emails and stuff. That's fine. But actually going through the entire process of sending and responding to emails and like dealing with your whole email inbox, I think is just very silly because if you're at the point where somebody is sending a customer an email themselves, okay, that person is similar to a lead. It's like the highest quality sort of communication you can realistically get. It's the closest thing to money that you could ever get is like a customer sending you an email in your inbox or one of your, I don't know, in my case, community members sending me a message or something. Why would I automate that? It doesn't make any sense to automate. uh you can hand that off to another human being. You can delegate that, but the probability that an AI will actually be able to reasonably fulfill that request right now is still substantially lower than what a human could do with similar amounts of information. And so considering that this is one of the closest things to money, this is our conversion event again and it's dollar signs and this is like over here like it's really close. You don't want to automate that. Okay. In addition, a lot of the time like you'll see it come out and then you'll have to edit it. It misses a lot of the relationship nuance. You'll get an email from somebody that you've known for like two and a half years or whatever and you'll respond to them like you've never talked to them before. And there's just like really weird tone mismatches all the time where I don't know your model will start using emojis or something even if you train it not to some percentage of the time because it screws that up. You can actually destroy client relationships like who's actually emailing you guys? Who's actually emailing you? Most of the time, if you're a business owner, it's customers, current customers or it's prospective customers or prospects or it's, I don't know, um, people that have a problem with what you're doing for whatever reason. You know, they have a problem with your product or service or whatever. These people are money. These people are also money. And these people, if not tended to correctly, are money. So, why would you have AI with a lower percent probability of being able to do it? Let's say even 20% versus I don't know I think like a human being might answer it 50% of the time fine that 30% gap means that you are losing onethird of all of your money. I mean that's crazy to me right? But people don't really think about it this way. They think about in terms of savings. Oh I get to save 10 minutes in the morning. you know 45 minutes going through my email at the end of the day. It's like well that 45 minutes makes a third of all of the money that you make. So why would you try and farm that out? Sure, you'll save 45 minutes, but you're going to lose 33% of all of the dollars that are uh new dollars that are coming into your business. That is not worth it to me. Anyway, there obviously some um nuance here. You could use a system like this to handle some basic acknowledgement emails. If you receive the exact same templated email from some service or something saying you get a new lead, obviously you can have this come in, you can process that, which is procedural, basically deterministic, and then have AI write some short little snippet at the beginning of an otherwise templated message and send that back. You can do that. I do this. A lot of people do this. This is like a very high ROI way of leveraging this technology. But don't just have it do everything for you. Okay? If you're going to have AI do stuff, have it generate some draft and then read through the draft, make edits, and actually make sure there's a pair of human eyes on everything that goes out. All right, so I mean, you know, the promise here is that you get to inbox zero automatically, which is just not true. We got to get back to red, which is scarier. The reality is that, um, you know, a lot of the time it's generic. A lot of the time you get weird tone and missing contacts, and then you hurt relationships with the people that matter the most to you, which is money. All right. Another big issue that I'm seeing is the idea of these all-in-one AI agents with false autonomy. So these are uh I think we've all seen these. These are like the really big crazy chunky N8N graphs with like you know a quadrillion nodes and then it looks like a giant piece of spaghetti. These are marketed as fully autonomous business assistants or I've seen these marketed this way. This thing will run your entire life. Sorry, I dropped my pen. You'll never have to do anything again. I don't know do your dishes cuz my AI agent's going to do it for you. It's like, do you really think that that's accurate? I don't. These models are not yet at the point where they could respond to your emails correctly. How are they going to run your whole life? And why would you want something that is like 50% as capable as you across every metric to run your whole life right now? Come on, man. You're not going to sell that to a business owner. And you're certainly not going to use that yourself for more than like super isolated things. So, yeah. Like the reality is these models um we tend to anthropomorphize them as much as humanly possible because that's just how humans work. anthropomorphiz just means we try and make them out to be like people, right? But they think very differently to people and they handle things we're just not yet at the point where it makes sense. I realize that this entire time I've been like a doomer, you know, being like these technologies aren't there. But like I want you guys to know I'm one of like the biggest proponents of AI. I have been invested in this industry for the better part of the last six maybe seven years. I went to school because I wanted to work with artificial intelligence. I dreamed of artificial intelligence like Openheimer dreaming about nukes in that uh um you know the recent movie. Okay, this is all I think about. There's a reason why I'm going so harsh on all of these systems that I'm seeing people sell, and it's because they just don't work right now. This technologies will get to the point where you can have an all-in-one AI agent assistant probably at some point in the next like few years, but we're just not at that point yet. It doesn't make sense to sell people false promises. And the reason why is cuz you're not going to make any money off that. You're going to sell one crappy project. You're going to get a terrible experience with the client and then you're just going to have wasted like a good 3 weeks of your life. focus on real business problems like what I'm going to be talking about in a moment. Okay. So, what are some issues here? 1 to 5% failure rate per query. Um, if you are losing 1 to 5% of every query, you're not your money. Like I had somebody point out a couple of videos ago in my comments. Uh, you know, if you screw up once for a client, okay, you could be screwing up that entire relationship. Okay? If out of a hundred inquiries, you screw up one of them and you screwed up really massively, that client may just like take that entire business away from you. So 1% failure rate per query does not equate to like a 1% loss in revenue. No, it might equate to like a I don't know like a 30% loss in revenue or something. Is that really something that you can confidently sell to a client? I don't think so. Also, a lot of the time, you know, there's some unexpected scenarios that'll come in. Again, these are business scenarios. There's always like an ROI attached to you making a good move or making a bad move. Um, and then there's also some security risks which um I haven't really touched on a lot, but when you give an AI agent like full control over everything and you try and have it like do your invoicing and have it send and receive transactions and when you just like have it run your entire life basically, there are of course massive security issues with that. You're giving one company like every aspect of your entire life. Is that really something you want to do? Probably not. Um, moreover, like if there are any bad actors that take some of this information, you're screwed. And is that something you want to subject a customer to? Probably not. Yeah, limited pros-wise, like it's good for demos. It's good for proof of concepts. We're at the proof of concept stage, okay? We're not at like the let's use it in like a $100,000 a month sort of business yet stage. So, the claims are that it'll run your business. It'll um work without supervision. It will, I don't know, please your wife, like I said in that other video, which people find hilarious. It'll just do everything for you. But that's not true. In reality, they need constant monitoring. They have regular failures. They have the illusion of intelligence, okay? not actual intelligence. Okay. And then why don't we cover the last point here which is NANA agents. So um you know these are like I' I've sort of talked about NANA agents throughout this whole video. Nad's blown up and obviously the agentic feature is a big reason why but NANA agents just the way that they're constructed tend to just fail um a lot and they might be one of the most mature agent implementations currently available today. I think there are like a few better ones but most people are using NA AI agents. In reality just NA agents they just ain't it yet dog. Like I mentioned in one of my previous videos, they have very impressive demos. Um they do combine a bunch of interfaces into just one and they have like awesome debugging and stuff and I think there's some personal productivity use cases in which they can work but like you don't want to do that in a real business. Okay, so that 1 to 5% failure rate, same issue as before. You know, expensive development for minimal business advantage. All right, so now that I just spent half an hour talking about what not to do, let's actually get to the point where we talk about what to do. The reason that I tickled all of your hypothetical and proverbial balls with that was I just wanted everybody to be abundantly clear that you know if you sell these sorts of systems you're probably not going to make that much money and I just want you to make a ton of money. That's all I care about. So that is what not to do. What do we actually do now? Well, here's what a high ROI automation actually looks like. And let's switch to green cuz we're happy now. Okay. The first thing a high automation does is it solves a dire business need. There are three major business needs. The first is revenue generation. The second is cost savings and the third is time leverage. Revenue generation means stuff like increasing the leads flow. Revenue generation means increasing the average customer lifetime value. You'll see that acronym a lot. Um but that just stands for how much money you make in in general um throughout a customer life cycle. Revenue generation means increasing the conversion rate. That is another three-letter acronym you'll probably see often when discussing sales. Okay. Um, and there are variety of other ways to increase revenue as well. And I think I'm going to be pushed off the side here. So, let's just do an arrow down here. Um, increasing retention. Okay. Which feeds into CLV, but a little bit different. So, these are all ways that you can actually add real value to businesses by providing mechanisms that increase the number of leads they're making, increase the retention of clients because you do a better job, increase the total lifetime value of clients, which is interrelated with this, and then increase the conversion rate on the front end. We're going to talk about all these. The next way you can add value is you could save money. So, how do you save money? You decrease, okay, headcount. When I say headcount, what do I mean? I mean people. Let's just, you know, we're under no illusions here. A big chunk of the work that I do has led to job loss. People leave like or they are let go because they can't justify the the expense that they're providing the employer. If the employer can just hire a system or build their processes in such a way that we do uh five times the amount of work with half the staff members, what do you think is realistically going to happen to that organization? It's going to grow substantially more efficient and leaner and better. It's going to be much more profitable, but at the end of the day, we are going to lower the headcount, which is one of the biggest and worst expenses most businesses have to pay. Uh what are some other ways you could decrease let's say software expenditures? You could decrease like operational expenses but basically just decreasing opex. Okay. And then time leverage you are basically doing things that scale and you're just doing them a lot faster. Okay. And a lot of the time you're also specifically related to the founder time because the founder a lot of the time of a business is a bottleneck. So you are increasing the free available founder time. So these are all of the dire business needs basically that automations that make money solve. Okay, we're going to cover them in a moment. Now what else does a higher automation do? It produces clear deliverables. So it produces a tangible output. What's a tangible output? It might be like an email. Okay, might be like a notif of some kind on their phone. It might be like a document. It might be a PDF. It might be anything like that, but something that the customer it might be like a faxed out sheet. It might be like a like a book or something, right? I've helped people automate whole books. uh this is something that people can actually touch and feel and you know as a result it also needs to be visible to the client in that way. If it's not visible to the client and produces a deliverable doesn't matter. So we got to make it a deliverable and then we got to actually make it visible to the client. Ideally it's also templatable because if you build a templatable system then the next time you get a client you know instead of you going from zero to 100%. Okay you actually cut that down. You just go from 80% there to 100%. What is this in layman's terms? That's what we call 5x leverage. So instead of you having to build a system out from scratch, every single time you build a like a system that's almost done and just make some minor tweaks to it. And then quick deployment feeds into that as well. That's valuable because the more templated it is and the quicker you can deploy it, the more perceived value you can offer a customer to reduce their buyer's remorse. So what that means is when somebody sends you money, there's sort of like a gap right now. There's like a hm I've given Nick a ton of money, but I haven't really gotten anything in return for that money. Huh. You don't want people to feel that way. and you want to minimize how long they feel that way because the longer they feel that way, the less likely they are to give you more money in the future. So, what templated systems do that you can put together really quickly are they allow you to minimize that time between when somebody gives you money and then you deliver the result. Okay? Cuz that's just how services work. It's not like a product where like you go to a store, you're like, I'd like that shirt to please and then you give them money and then you receive the shirt in one go. You know, you give somebody money and then it takes a little bit of time for that service implementation. So, the shorter we can make that turnaround time, the better. And then finally, ideally, good systems are platform agnostic. Uh that's just a fancy way to say you could build the same thing out on multiple tools so that if one tool stops being accessible you could do it on another tool or maybe instead of you doing it through like a make. com module you do it through an API call because the API call doesn't change maybe the make. com module does and then not vendor locked is important as well. So you know basically you have a template for I don't know some system that does it on instantly and then you also do it on smart lead as well. All great ways that you could do this. And then I'm not going to cover the low ROI traps just because I think you guys are probably seeing the parallels here. Okay. All right. Let's now talk about some criteria for high ROI automations or agents and what you can do to make tons of money and be like Leonardo DiCaprio. This is pre25 is the oldest age I will ever date, I think. So, he's actually still he's still cool here. We still like Leo here. Hasn't been cancelled yet. Okay. So, uh I'm going to get cancelled for that. uh here are things that actually matter. We're going to solve a dire business need. We're going to produce clear deliverables. We're going to template out most of it. They're going to be platform agnostic. We're going to focus on reliability to a pre-existing business process over flexibility. We're going to focus on measurable impact. We're going to focus on things that require human oversight but then apply it at particular points allow us to massively improve leverage. Then ultimately we're going to do some scalable implementations that can be deployed quickly with minimal customization. the very first probably one of the highest ROI systems that I put together now and it really is taking off is this idea of a deep personalization cold outreach system. So we're moving away from all those fancy shiny objects towards systems that actually help with something that a lot of businesses are using today to make money. So what is one thing that a lot of businesses are using today to make money? Cold outreach. And what does the system do? It takes a pipeline of cold outreach which is working okay I guess. Okay. you know, maybe it's generating 10K a month. And then what it does is it massively increases the diameter of that pipeline so that we can make I don't know 50K a month or something instead. This is what all successful systems basically automation systems do. They just scale up. So the way that this works is you have a system that works which is usually some sort of custom outreach where you will customize some message that you're sending or customize some video or whatever and then with depersonalization you use AI to do automated research on the prospect something like perplexity or something like OpenAI's now web search API you get all this information and data on them maybe use Apify for this maybe use some other phantom buster scraper to grab their LinkedIn fields and then you have it take a template that is working right now an email that you are sending that is delivering results and then just modify it a little bit so that instead of it being like hello Peter um I do X Y and Z and I would be very happy to assist you with this you say hey Peter I love that you do X Y and Z I've been following you for a really long time just listened to your podcast with blank talking about this I wanted to put something in front of you to hear your thoughts on it okay you take a general template and you convert it into something super customized so this combines scraping AI generated personalization and then even automated email sending. And this significantly boosts reply rates. I said 20 to 40% improvement. I've seen reply rates of about 20% off of a cold email, which is, you know, for most people unheard of. Saves a massive amount of time compared to manual personalization, which a lot of people are doing before. Provides a measurable ROI through increased booked meetings and closed deals. And it reduces dependence on large sales teams for that initial outreach piece. It's easily templated. You just build out a system once in NAND. I've done so. You guys could look at one of my previous videos. I think it's titled deep personalization system. Integrates with tons of CRM. does email marketing automation tools and so on and so forth. So, how does the system actually work in practice? Uh, it's pretty simple. Usually feed in some sort of lead list and usually the lead list has like a name. Sometimes it'll have an email. Sometimes it'll have other fields like I don't know their job titles or whatever. You can get this list in a variety of ways. You could use a service like Ampify which scrapes Apollo. You could do something like LinkedIn. You could scrape off of like Instagram. I mean there, you know, just check my last video which talked all about how to scrape every social media platform if you want an example of that. What you do is you take all that and then you use it to scrape public data using something like Perplexity or um OpenAI web search and then you pass it through an AI where basically you feed it a template with a bunch of these curly brace variables and you say, "Hey, can you fill these variables with hyper relevant information from all the data that I just gave you? " You then generate a custom message. Okay. And then what you do is you feed that into a platform like instantly or smart lead or whatever. When somebody gets back to you, then you update a CRM automatically. Okay, really cool system. There's tons of value that you could have um that you could provide for people that are running businesses based off cold outreach. And I want to say that like there are tons of businesses that are crushing it with cold outreach. If you were to take a deep personalization system and put it in front of them, they'd immediately save a ton of time because they're no longer personalizing all this outreach themselves. But then they'd also make a ton more just in terms of the opportunity cost because you know if they're sending templated emails and they're achieving reply to 3%. You could send that exact same volume of templated emails just using a depersonalization system and then you could have like 15%. You're literally 5xing the size of their business like this example right over here. Okay. All right. The second major system and these are all systems that people are selling right now is some sort of AI powered invoice and payment automation. Okay. The reason why this works is mostly because you are able to recoup a large portion of the invoice outstanding payments that people are not aren't giving you money um for early enough. Like what I find a lot of businesses do, especially like digital agencies, is they do what's called like net 30 or net 60 terms. Okay, for those

of you that aren't familiar, that means that um you will provide a service, you'll sign a contract, you provide a service, and then 30 days later the person will pay you for it. This is in um contrast to what I was just talking about with the buyer's remorse. I always do my payments upfront, at least 50% of the payment upfront depending on the size of the project. And if it's a monthly, I just always get the payment up front. But this is like the old school way of doing things. And a lot of agencies businesses that you are going to be working with, especially like um uh business models that have like time and materials, so like um general contractors, you know, like a lot of the time these like construction adjacent businesses, they're going to do some sort of net 30 or net 60 terms. The issue with this is the longer that you have to wait as a business for cash flow, okay, the less money your business makes. you have to wait 30 minutes to get a certain amount of uh 30 days cash flow. That means you have 30 days that you can't invest back in the growth of the business. So in practice, what these net 30 and net 60 and whatever payment terms usually do is, you know, if this is the future where everybody gets paid on time and this is the revenue of your business, okay, this is like kind of what actually happens, okay, you end up with these plateaus. Why do you get these plateaus? because you just don't have the money yet. And so it just takes your business way longer to get to an equivalent growth level. I mean, this is almost 2x the time, right? So, what do these systems do? Oh, another thing is a lot of people just forget. And there's also a lot of issues where like if you just don't get paid for 60 or 90 days, like it's no longer a concern to the customer. Like the likelihood of somebody just skipping out on payments goes way the hell up. I've uh read so many stories of like construction businesses in particular that have failed because of these net 30, 60, 90-day terms where they're just waiting for money and because they don't have any money, they can't take on new projects, they're promising their team members stuff. Anyway, to make a long story short, what the system does is it just standardizes invoices. So, it automates the creation of invoices, the sending, the follow-ups, and even the tracking of these payments. It'll use AI to detect overdue payments, and then trigger personalized follow-up messages, very similarly to what I was talking about before. Okay, I cannot draw a straight line for the Okay, I'm just going to use the straight line thing. Um, a lot of the time, uh, you can also build it in so that it integrates seamlessly with common accounting software platforms like QuickBooks, Zar, etc. And the real value is it significantly reduces late payments, also just reduces unpaid invoices, right? Uh, it improves cash flow because now you're getting proactive follow-ups. The likelihood of you receiving something when you should have received it is way higher. Um, you save time spent on manual invoice tracking, collection activities, and then you provide clear tangible deliverables that are visible to the client. What are those tangible deliverables? Invoices, payment emails, payment statuses, their CRM, their QuickBooks, the money that they collected. This is money. That's what customers care about. So, this is super templatable. You just make one for basically every invoice platform, QuickBooks, Zero, whatever. Um, you can easily adapt it for various industries with minor changes in the wording. Platform agnostic, works across tons of different CRM systems, and it directly impacts cash flow, which is the hottest button business need. This is a more or less a breakdown of how it works, but I've actually created a video showing you guys how to build more or less this exact same system that is agnostic of platforms and you can use for anything. So, you deliver a product, you automatically generate the invoice, you'll send the invoice to the client, and then if the payment has not been received after, let's say, 3 weeks, you'll have AI powered follow-up start. And then you basically just go this way over and over again until you receive your money. Okay. Then finally, they do you update the accounting software with the payment and then you make that moola. This is super valuable to sell to clients. If you're selling to a business that's $100,000 a month and they're in one of these industries that I was talking about and they are losing like5 or $10,000 a month to this problem that I'm talking about. You solve that $5 or $10,000 a month problem. You can charge like5 to $10,000 a month to to fix it. Right? You're solving today's problem, but you're also solving future problems of scale. you're not only increasing um immediate cash collection, but you're also decreasing time that founders and other administrative staff have to spend constantly following up. So, you kind of deliver a ton of value

with this. All right, another simple automation that I've put together recently and a lot of people in my community, Maker School, are putting together right now are just some general CRM automations. So, basically, you know, you get some new lead coming into that beginning of your pipeline. Usually have some sort of criteria based off that. So, industry, budget, engagement, uh project type, number of bedrooms in a house, whatever the business that you are working with does, home rena, or they do some sort of PPC, they're usually going to have some sort of like inbound funnel where they ask various questions, right? And the way that people answer these questions sort of slots them into like different um I want to say like different routes. Okay? And if you're smart about it, what you do is you typically as a business, I don't know, imagine this is a marble falling down one of these poles or whatever, right? One of those games that I think a Plinko or whatever, I don't know. Um, usually based off the route that they take, you'll do differential comps with them, right? So the value with this is once you're done with getting somebody in with a CRM automation, what you can do is you can actually enrich the lead data. You can use public profiles or third party APIs. You get a bunch more information about that person, which could substantially improve your ability to close them. And you are just using a CRM now. And so a lot of teams, they have these super big, chunky, amazing CRM, they just never use them and they never actually update them. So what this does is it just eliminates that issue where like when new leads come in, at least they're in the CRM. And usually once they're in the CRM, then clients can um client sales teams can actually start to work with them. So this works with basically any CRM, HubSpot, Piper, Salesforce. I built this on clickupmonday. com. Um same energy across the board. You guys can watch some of my earlier make. com videos for this. Uh, essentially the way that it works is you'll get some new lead maybe through like a Facebook lead form. That's my usual example. Somebody clicks on an ad for I don't know some plumbing or HVAC system or whatever. Um, comes in and then you enrich the lead data. You can actually use AI based qualification or you could actually send them a message using AI immediately after which we'll talk about. And then basically if it's qualified then you add to CRM with some tag or if they're not qualified you discard or tag them as some low priority or something. I don't actually recommend using a priority uh like the priority columns that are in most CRM just because I find if everything is a priority, nothing's a priority. But you can find different ways to do this. Bump it up to the top of the queue. Notify a salesperson. Do whatever you have to do. Okay. A fourth major system um that I've sold many, many times. I mean, I've probably made over $200,000 selling this exact system is a proposal and document generation automation. So the idea here is you are generating a proposal or some sort of PDF estimate or some sort of quote or some sort of white paper and the whole idea is you do it based off of form inputs or CRM data. Okay, this includes client specific data, can include pricing, uh you can create beautiful deliverables and then you just have a PDF at the end of it which you can send via email, you can upload to a shared drive, you can add to something. Hell, you as I mentioned can print it out and then mail it to somebody for Christ's sake. So, there's so many different examples of this, but basically, let's say somebody's on a sales call. A good example of this, and actually where I started this way back in the day when I was doing videography sales, and I was actually inspired by my roommate at the time, who showed me what he was doing with this form system. But essentially, you fill out a form when you're on a sales call. The form is, hey, so like when is this wedding happening? When is this happening? When is when do you want the video shoot done? And blah blah. You add all of that information into a form. Okay, this will insert all of that into a template. Your template is like a PDF. It's like, I don't know, um I don't know, it's a Google slide or it's a Panda doc or something. And then you generate a branded beautiful proposal using a PDF or a doc form that you can then use make. com, n any code tool to send to a client or upload to a drive immediately. Then you can update some CRM status. You could just start printing money. Um the value there is, think about it, sales team is getting all that information anyway, right? And if you get all that information, you could now just add that to a CRM, couldn't you? So, you can actually combine this with basically every other system that I've showed you here. And actually, what I would do personally is I'd round up every system that I've just showed you guys, and I would combine that into like one mega automation package. Then I'd try and see if you could sell that to somebody. My inclination is if you rolled all of this stuff up into some very, very mega big package, you could get in the habit of making these 30, 40, and 50k deals that I'm seeing people in make moneywithmake. com do. But I don't know, just a recommendation. Okay. Another one is a onboarding and client intake automation. This automates the client onboarding process after a deal is closed. You guys are probably seeing some similarities here. Every one of these systems is actually about a process that already exists in a business. It's not about new processes. It's about invigorating and improving old processes that people are using to make money. So, this is you've already signed a client. How do you improve the probability that client is going to have a great experience and pay more money with you later? What this does is it collects client info via forms. Usually when you sign a new client, you would then send them an onboarding form, collect a bunch of key info, then use it to auto update a CRM or PM tool, and then you can do things like send automated welcome emails, schedule kickoff meetings by sending them a calendar link, create a bunch of PDFs and customized docs. Hell, you might even be make an API call. This is something that I saw that once that was really cool, make an API call to a service like Swag Box or something and actually generate custom branded assets. A little hoodie with your freaking company logo on it. A little mug that says, "Thanks so much for being a client. " Okay? Little place business card holder for them to put on their desks. These are all so high value and basically nobody's doing them, which just breaks my freaking heart. So, when you do something like this, you create a professional onboarding experience, which people like. You eliminate all the back and forth that most people have for basic info collection. Usually, you'll start a business like, "Hello, thank you so much for uh working with us. Here's a list of questions. Could you please answer these questions? " You got to answer the questions. Then the secretary or front desk person has to read through the questions, add them to a system. Why don't you just eliminate that step and have them add to the system themselves, right? You reduce missteps and onboarding workflows. And yeah, it's just like super practical. You make a ton of money doing it. So you can build this out in Typform, um, Google forms, uh, Air Table, Notion, ClickUp, literally any CRM or any system. Um, as long as you have a form that like feeds into it, you're good. You deliver visible, tangible results, and people really like that. Again, I've built this system out on video numerous times. Also, I have like an end to-end walkthrough of exactly what this looks like in Maker School. So, if you guys like this sort of thing, definitely check out my community for more information on that. Okay, the sixth one

is some sort of customer support ticketing automation. Remember how earlier I was telling you guys um how Amazon does their stuff? This is essentially what I was referring to. So, you know, the idea is you have some sort of customer support ticket. You're running some SAS company or something, okay? And these tickets start taking a vast massive amount of time. They're usually like three or four categories, but they're just a little bit different. So, it's hard for you or your company to like deal with these things reliably. And this is a major cost. You now have to hire a bunch of customer support people. Any major SAS company has massive customer ticketing teams. When you hear about like a company uh that's a software business that has a thousand employees, usually the way it works out to is they have like a founding team of like three people, then they have like 10 engineers or 20 engineers, and then the other 970 are customer support staff. So, a good system like this can actually deliver massive savings and eliminate a lot of payroll BS. So, what's the idea here? Well, basically you have some sort of support u route. Okay? And then you feed that into AI which categorizes where that support route goes. Is it a problem with the software itself? payment? Are they asking for a refund? Are they doing one of various things? Okay. After that, then you either assign priority or department and create a ticket in a help desk tool and then notify a specific agent and then finally track the status and response time or what you do is you just say, "Oh, is this a refund request? " Okay, if so, then why don't I calculate how much money they've asked for refunds for divided by how much money they spent, and if it's below a certain threshold, I'll just give them their money back. If you do stuff like this, you could save a good 20 30% of all of these customer support inquiries and requests because refunds are in practice a very big chunk of why people want uh you know want to contact you. Okay. Uh we got a couple more here. The next is recruiting and candidate screening automations. So the idea here is this is another process that most businesses suffer from or have to go through. So they recruit things, they screen candidates and so on and so forth. Um, what this does is it automates the job application intake, the candidate pre-screening, and the scheduling of interviews almost entirely autonomously. In the past on my channel, I've referred to this as my hiring system. So, feel free to check that out if you guys want. Essentially, what happens is when a candidate submits an application, okay? Usually, they'll fill out some form or something. Okay? So, then we'll extract fields from that and then screen and score them. Now, I don't use AI to do all of the screening and scoring, but I'll just use AI to look at a couple of fields, right? Like, do they have a portfolio, for instance? how much do they want? You can either use AI to do that. You can also just do this entirely procedurally. The point is you get some threshold. If they're above the threshold, then you go to your applicant tracking system, which in my case is usually like a ClickUp hiring pipeline or something or a monday. com hiring pipeline. Then if not, you can send them an automated rejection email and archive. And you could even tell them why. This is really valuable because the vast majority of recruiting companies and stuff, they don't actually tell you why your application wasn't accepted or whatever. You can earn a ton of goodwill in the market just by having an automated email that says, "Hey, thank you very much for applying. You know, we looked over your application. Unfortunately, at this time, we're not able to like have you proceed with the process, but I just wanted to let you know right off the top why we didn't so that you could take this and then maybe use it to improve in your other job. " Instant way of getting a ton of goodwill, instant way of impressing a lot of people that maybe later on down the line you will end up having as candidates for some other service or something. So, super quick and easy, saves a lot of time. And then these applicant tracking systems are usually just nightmares for most recruiting companies to deal with. If you have a simple process like this that you can build entirely custom yourself in like two hours with my video or my series of videos, then uh you can deliver a lot of value to recruiting companies super quickly and super easily. Okay, next up, sales pipeline and follow-up automations. So this is something again that is a pre-existing process that most businesses will do. Most businesses will follow up with customers and the money is in the follow-up. it's like the third, fourth, fifth, nth time that they talk to somebody, the person will actually decide to pay them. So the whole idea behind these systems of these uh classes of automations is they automate the movement of deals through a sales pipeline based on the activity of the person or just a certain amount of time. You know, once a week they'll check to see how long a person has been in a stage. The person hasn't moved in a stage, then they will just send them a follow-up. You do this automatically in any major CRM platform today. Super valuable and super easy to get up and running with. From there, maybe you will send a specific sort of email for this stage to that stage, another email from depending on whatever. And then what you can do is you can take the same approach that I took earlier with depersonalization, then you can actually customize it based off of a log of all of the customer activity. Okay? So instead of saying, "Hello, I've noticed that you have not proceeded with our blank," you say, "Hey Peter, loved our chat last week about X, Y, and Z. Just wanted to like double check and see if there was anything I might be able to help you with cuz I noticed you hadn't filled out the next step in our program. " These are all super valuable, as I'm sure you can imagine. And you're treating a customer like a real human being, which they tend to respect, while also knocking that follow-up task off of the chart. It's no longer something that human beings necessarily have to do. And you get to double or triple dip because now your CRM is updated and it reflects deal progress and it's really clean. Okay. The way that this works is the simplest variation is you literally just have your CRM and then you have some sort of like you know number of days in field um stage which you can just use as formulas in most CRM. So just see you know what date they moved to the stage and then um how many days it's been since that stage. Uh once you know that then just every day you run an automation that just goes through your whole pipeline and says okay how many days have all these people been in these stages. Then if you are um I don't know x days, let's say like three days, if you haven't moved or done anything in three days and there's no emails in the CRM, then you send a follow-up email using AI to customize. Then you update the CRM stage. If no, then you wait for response or activity before going back here.

here. Okay. All right. Finally, let's talk about the last system, which is just a reporting and analytics automation. Essentially what you're doing is instead of you having to have some business development person go through and then compile a report, you just do it completely autonomously and and automatically using AI and then some sort of data extraction. So the idea is you are automating the generation of weekly or monthly reports using data from multiple sources. For instance, you might take data from their CRM, Stripe, you might take data from their Google Sheets, you might take data from various project tools, whatever. But you have all this stuff dumped into some central database. I'll maybe use Google BigQuery. Um, sorry, I'll use BigQuery or I'll use something like Google Sheets. And then it'll format and deliver that as a dashboard in a platform like Looker Studio or Plecto or one of the very many alternatives. Um, so that the person can just look at it. When I say generation of weekly or monthly reports, by the way, I don't mean that you do this once a week or whatever. I just mean like you have a dashboard that you can pull from once a week. But that same dashboard is actually accessible anytime. It's not like, you know, you have to wait a whole week to get a PDF or whatever. It's like, no, this is a real live dashboard on Looker Studio or Plecto and just once a week you get an email being like, "Hey, here is a link to the dashboard with the filter setup for progress over the last week. " This saves so much time pulling manual data across all platforms. Saves so much time because you're standardizing the exact same thing every single week, which most small to mid-size businesses making less than maybe 200k a month just really struggle with. And it makes reporting feel proactive instead of reactive. You're not generating reports that you're looking at. What you're doing is you're looking at a real time updated dashboard and then just getting notified. Okay. The way this works again, schedule trigger, pull data from all these tools, calculate the KPIs and metrics, and then generate a report. I recommend that you personally use a dashboard wherever possible, just way cleaner, then just, you know, email it over to stakeholders at decision makers. Okay, as you guys could see, there is a lot more BS on the left side of this graph. Okay, My earnest recommendation for you guys is when you are building out systems, focus on real pre-existing customer problems. Focus on situations where the customer is already making money or doing a good job in this sphere. So, how can we make them do an even better job by sprinkling a little bit of automation on it? Don't put the forest before the trees. Don't get so caught up in how amazing and enjoyable and exciting automation is that you guys forget what you are put on this earth to do with this business model which is to take a working process and just scale it up. Okay? Don't get caught up with a lot of the crazy things you guys are seeing on YouTube that show amazing demos of technologies and stuff like that, but instead focus on actual practical ways of implementing triedand-true methods like I've shown you here to make more money. Because if you can make your customer more money, happy. And if you can not sell your customer then they're going to make you very happy financially and also just in terms of your relationship in return. Thank

you very much for watching. If you guys enjoy this sort of thing, I highly recommend you check out Maker School. As of the time of this recording, we are just under 1500 members, and I'm increasing the price the second that we hit to 1500. So, Maker School is my all-in-one 0 to1 roadmap with daily accountability to show you guys how to sell the exact same systems that I just ran through here. I give you guys practical, actionable dayby-day essentially check-ins over Loom where I guide you through step one, step two, step three, and by the end of it, you guys will have a fully functioning and not just functioning, but thriving AI and automation business. Really appreciate all the time. If you guys have any questions, just feel free to drop them down below. You guys have any recommendations for future content, please do the same. And I will catch youall in the next video. Thanks so much.