# The AI Systems That Run My Entire Business

Channel: Jordan Rafealov
Video: https://www.youtube.com/watch?v=pY7wBIzm130
Duration: 49 min
Language: English
Words: 8879
Transcript page: https://viewrankai.com/tools/youtube-transcript/pY7wBIzm130

---

[0:00] For the first time ever, I'm going to take you behind the scenes and show you 35 of the AI systems that runs my multi6figure AI agency. I'm going to show you each of the AI systems one by one and how you might be able to build them yourself. So, if you want to make more money without hiring a ton of people while working less, then this is the video for you. Let's dive in. So, this is going to allow you to juggle dozens and dozens of clients with one

[0:27] VA. And so I have mapped it into seven different categories. First we have the offer and copy systems, the lead and infrastructure systems, the outreach and reply systems, and then the sales systems, the delivery systems, and the content and launch systems, as well as the operating systems, and the guard rails that all tie it all together.

[0:48] So first up, we have the offer writer. Now, this is for cold email outreach. It's very important that we have a compelling offer and we don't trigger any spam words otherwise we're just going to end up in spam and not in primary inbox. So it's very very important that all of this is taken into equation. So you can see down here that I use Claude. I call Alex Hormozy's value equation. I tear the offer. Then Claude writes the fourline email as short as possible. Unlike other emails where long form copy works really well on brand new cold traffic, you haven't earned their their attention yet. It

[1:28] needs to be short. And then it runs it through a spam checker and then it spits out clean copy for you to load into your campaigns. And so this is an example of a cold email that might go out if you run a cleaning business. And the second part of the offer system is the pitch deck. Right? So if you are putting together an offer, you need a beautiful way to present it. And so the way that I have built this is that it pulls in the call recording with my prospect. It runs all of their numbers and then it builds live JavaScript slides and then it builds a a password

[2:09] behind it and it pushes it to GitHub behind a locked a locked screen with a specific password that I can then share with the prospect and say here is the link click on this link and it's mobile optimized so they can open it immediately within their email and the password is John 2026 for example And so if you were to build this yourself, this is the sort of flow that you would go through. Now, if you were to hire a designer or a VA to build this, you might be looking at something crazy like 45 bucks an hour. However, this is all built into the pitch deck

[2:48] system. And so it even has dragable uh numbers that ask the prospect to go, okay, what is one client worth to you? How many calls do you close? And then if we give you this amount of calls, this is the amount of money that you should be making. And so these are all numbers that I picked up in the discovery call,

[3:06] right? So these are incredibly powerful. If I just show them, these are all the numbers you said, John, this is how much money you're actually going to make. What's stopping you? Now, within that, I have the lessons file as well. And so this is a system that pick that picks up on all of the things that I might have done wrong in the past. I think there are like 500 lessons plugged into this system that sort of orchestrate the entire business and make sure that I don't make the same mistakes over and over again. And so if I had a consultant, for example, it might cost

[3:37] me 50 bucks an hour. However, here it all runs for me automatically in the back end. And so these are uh things like it stops emails going out to prospects that I've already contacted. it stops a campaign if a if the wrong inbox is loaded in or if it hasn't gone through the 14-day warm-up period first. All of these lessons are now hard rules and it's impossible for any of them to be broken within the Claude code terminal and all the orchestration and

[4:08] operations of all of these AI systems. And so if you are running any sort of outbound agency, any sort of cold outreach at all, right, LinkedIn, email, even phone calls, even going and knocking on doors, what you will find is that the constraint is actually having enough human beings to contact, right? Once you once you finish your warm list, going out into cold is fantastic, but if you run out of people, if you run out of leads, who are you going to contact? Your business stops, right? And so, this was an absolute gamecher for me. I can't tell you how many iterations I went through before I built out this

[4:50] pipeline. And so, this is the pipeline that you're going to want to run. So, first you go to Clay and Clay will actually scrape tens of thousands of leads for you initially top ofunnel and it'll do it for free as long as you have installed this GitHub repo here, which is the agent plugins. And so, I will leave this link down below along with my affiliate codes for all the clay stuff

[5:14] so you can get a bunch of free credits. But this is how I run everything. And so then after that it goes through AI arc to enrich the email. And then for the ones that it misses it goes through Prospio. And then for verification it runs it through million verifier. And then this was a game changer for me that

[5:32] I didn't realize is actually possible. But when you run it through million verifier, there's a whole bunch of leads that are catchalls they call it. And those catch alls are considered unsafe. And because of the way milling verify is set up, it cannot verify those emails. So it discards them. What you can do is you can take those emails, which might be like 25% of the actual total emails, and put them into bounceban. And bounce has some wizardry where they're actually able to go and verify those catches. And so you might get an extra 10 to 15% of your leads. And so if you start with

[6:09] 20,000 people that you can contact, it's actually a significant uplift in the amount of people you can reach out to if you're able to capture another 10 15 20% of those leads and you've got emails that you can actually reach out to. So that was a game changer for me. Here's just a quick quick tip. And then everything is done here automatically uh end to end for you. And then it even loads them into instantly. And so I use these tools here because they have APIs for me to plug my Claude code to plug into Claude code and they are the best in the market. And like I say, if you

[6:44] want to use any of these guys, I have affiliate links down below. I highly recommend you check them out. And this honestly saves me probably 15 hours a week and hundreds and hundreds of dollars in VA cost that I would usually outsource, right? Or like pulling my hair out trying to figure out where all these leads are going to come from. I just now type slash lead engine and everything is done for me. And so when I type slashleen engine, this is the process that happens inside of claude code. I point it at a client. It pulls in the context about who that client is, the ICP, etc. Where to find these types

[7:19] of people and then it uses the lead engine to run it from end to end starting from clay, ending with bounceban and uploading it into instantly. All done for me. And so in that process there is the ICP gate. So if you don't know what ICP is, it means ideal client profile. And so we don't want to be emailing people that are a complete wrong fit. And so it used to be very very expensive to figure out if somebody was a fit or not. Or even worse, you'd email them and then they'd reply with something horrible like, "Ah, why are you emailing me? Take me off

[7:54] your list. I'm not even associated with that industry." Right? And that happens all the time. inside of cold outreach because none of these databases are perfect. But the best solution I have found and it's actually free is to use something called HTML to text. So if you just type this into Google, you should

[8:11] be able to find something like this. Okay, so I believe it's this one, but do your own research first, but I believe it's this one. It's completely free. You download it and it is a skill that you can plug into uh Claude. And what it'll do, it'll go to the website and it'll turn all of the HTML into text so that you can ingest it into your LLM, in this case Claude, and it'll do it completely for free. Now, if you were doing this with a provider like Firecrawl or something, it would cost you uh hundreds of dollars if you were to do it with tens of thousands of websites like we do

[8:46] sometimes. So, we always start with HTML to text because it reads the website data and pulls in all of the context for us. And then we get a pretty good idea if they're an ICP fit or not. And then we we rate them from one to five. And obviously we kill off the ones that are a one or a two. And we look at the three, fours, and fives. And if that fails from the entire lead list that we load, we use something called

[9:11] spider.cloud. And so it allows you to point an agent or an AI at a website and it's just a much more comprehensive version of HTML to text, but it costs credits. Now, it's not crazy expensive, but it's not free. And then after that, we use Firecrawl, which maybe some of you have heard. It's a fantastic product, but it's very

[9:34] expensive. And then after that, we use Exa.ai, which is also a fantastic product, but very expensive. And then everything else that cannot be found through HTML to text, spider.cloud, firecrawl or exa, we we use some version of claude to send an agent to that website to see whether or not they are a fit based on everything that's on their website. And usually you get a really good idea about that just from their website. like uh if you're targeting B2B SAS founders, right, and we go to their website and all they talk about is providing a service or they give away or they they do events or

[10:12] something like that, then all of the copy on the page will be geared towards delivering that service or selling that product and your Claude should be able to easily identify whether they're a fit or not and then you can exclude them, right? And so within the claude code terminal window, this is something that you might see. Good fit, not a fit, not sure, three out of five, five out of five, something like that. And then we have something called signals. And so this is a massive game changer if you're doing any kind of cold outreach because when somebody is at a different stage of awareness, they're much more likely to

[10:48] respond positively to your cold outreach. And they don't put their hand up and say, "Hey, I'm willing to buy your service." but they give off different signals that might suggest that they would be more open to it. So, let me explain. First, I use Ampify and then I even am able to use things like the SEC uh public company records. And for this particular pipeline, I use something called import Yeti because this is for a client that uh they sell a solution to importers and brokers who bring things into the United States. And so I need all of those import records and a lot of them are publicly available. And so I

[11:28] use Ampify and I use uh import Yeti and then I use Google Maps as well to find different local businesses and then I can infer different signals based on those things. And then I can change the way I reach out to people based on all of the signals. And [snorts] so if somebody is uh paying a ton of import tax, I can hit them with my offer. If they're hiring a sales rep, I can insert my offer. If I find a local business and my client is in that particular place and they provide a service for them or or a tool or something, I can reach out

[12:00] to them. Now, I have put together a list of I think not 27, but something like 50 57 different intent signals. And so, I call it social hijacking. Um, and so if you want this, leave me a comment below and I'll try and put this for free inside of the dopamine digital school group. So the way that I have built this is if if uh somebody liked a pain keyword post on LinkedIn, then they are much higher to much more likely to respond to outreach around the solution to that problem. If they commented on a competitor's posts, they're much more likely to respond positively. If I say, "Hey, I saw that

[12:46] you commented on XYZ company's post about not having enough leads or how Gentic AI is uh terrible because of this reason. We've actually solved that and here is the unique mechanism. Would you be against getting on a quick call?" And I'll show you how it works. Right? Can you see how like you're putting an offer in front of a completely different awareness stage prospect if they have reposted or shared something? And so I do all of this through Ampify and Exa. And again, those are linked below. If you use my code Jordan inside of Ampify, you're actually going to get, I think, 20% off, which is

[13:24] crazy because I've been using this and spending thousands of dollars on Ampify forever. And uh they were kind enough to give me a code for me and anybody that wants to use it. And there's no catch. Um, if they are a new seuite or VP hire, then they're likely shaking some things up inside of the company. If they are a former customer or alumni, if they're from the same company promotion or it's a job change or it's their first 90-day window, they are likely to take bold action. And so, you can reach out to

[13:58] them with a different message. If you look at they have just raised funds and so this is something that a lot of people do and um it's part of it but it's not like a fantastic signal in my opinion if they are hiring a bunch of SDRs if they're launching a new product if they've just won an award all of these different types of things give off different signals that that then infer that they might be in the market or at a different stage of awareness for a

[14:26] particular solution. And so there are 57 of these if you want them. Listen, I will I will uh package all of this and I'll put it in the free dopamine digital school group uh which you can access here. There are 8,200 other business owners just like you. And I try to post as much juicy stuff in here as possible. All the resources from the videos, like for example, the last couple ones that I made. Here is a like a tracking sheet and a PDF. And here is a Claude skill that I put together that shows you how to go from zero to $16, $35 a month in a 100 days. And you can

[15:04] literally just type /undday plan and Claude will tell you, okay, day one do this, day two do this, day three do this. It's pretty crazy. And now this is maybe the biggest unlock that I have found in a cold email outbound, which is what I run primarily. So what you will find is that is a juggling act between two things. First is having enough leads to email, enough human beings to actually put your offer in front of and secondly managing the infrastructure so that you can send those emails to those human beings. And so the infrastructure

[15:39] is a gigantic part of cold outreach. Most people say to me that they're just going to purchase five to 10 inboxes and they're just going to blast out hundreds of emails and they're going to hope for the best. And that's why cold outreach doesn't work. So this is how you do it correctly. You set up the infrastructure end to end all done correctly. And this was the biggest pain in my butt before because I'd have to go and manually do this across multiple platforms. So I'd have to go and purchase the domains, set up the infrastructure, do all the DNS records, load everything into instantly, start the warm-up. Now I just type slash

[16:18] purchase infra and then I tell it which client it needs to purchase for and then it purchases everything for me. It runs it through spaceship. It runs it through scaled mail. It it sets up the DNS records. It loads them into instantly and then it sends me a message on Telegram telling me when it's done. It's like a dream. Again, if you want to use any of these services, I've left

[16:42] affiliate links below. And so, you literally just type slash infra purchase infra. And then Claude will do everything for you. It'll pull all the APIs, it'll purchase the domains, it'll purchase the inboxes, it'll set up everything, and then it'll push everything to uh instantly. And it's all done and dusted. It is like magic. And now you have all of these inboxes set up. So sometimes I have 150 300 inboxes and they all decay just like everything. And so I need to understand what is landing in spam, what is not. Uh has my domain been put on a blacklist somewhere because I'm paying for all of this infrastructure and I don't want to

[17:25] be paying for things that are not delivering. Doesn't make sense, right? So how do you do that? Well, don't use Instantly's uh spam checker because it's absolute garbage. Deliverability monitoring or whatever they tell you it is, it's all rubbish. You need a third party. And so I have built an end-to-end system that every few days it tests all of my inboxes to see if they are landing in spam or not. If they're on any domain blacklists and then it removes them and it tells me, "Hey Jordan, stop paying for these 12 inboxes. This one is

[18:01] landing in spam. This is on a blacklist. You can't do anything about it. Remove them from your rotation." And so this is the endto-end way to do this. Now I use something called email guard and everything is done for me in this pipeline. It even stops me from using an inbox that isn't warmed up correctly. And so I put together this like email monitoring uh dashboard as well uh which is part of the internal wiki system that I'll show you in just a second. But you can see that, you know, I can I can basically see all of the clients here, all the accounts, all the domains, and which

[18:39] ones are burned and which ones are not, and which ones are landing in spam. And so, when you start signing a bunch of clients that are enterprise, you'll realize that their biggest fear is that you're going to reach out to an existing client or a previous client or they're going to damage their reputation somehow. And so what I do before I load any leads in anywhere is that I cross reference them against the do not contact list and I have built a shield that does this every day. So not only when I load the leads in, but it'll check if they have just signed a client

[19:13] and they've told me I add them to an air table and then it'll constantly check. And also, if you run multiple lanes of outbound, so if you're running email as well as LinkedIn, you don't want to email them and then send them a LinkedIn request the same day with the same message. So, how do you solve that? The

[19:32] DNC shield. And so, this will take people out of sequences if they're on the do not contact list, if something slipped between the cracks. And so, this actually updates every hour. And this one is super exciting. This is the AI sales rep. And so every time that we get a yes from a prospect or a we need more information. And so what's the most important thing about getting somebody booked on a calendar? It's getting speed to lead. So, you might have heard this term before, but it just basically means as quickly as possible, can we contact that prospect who has said yes to more information or they want to be sent

[20:13] something or they want to book in a call within 60 seconds. Hormosi says ideally within 5 minutes and that's his benchmark. Uh, but we want to get speed to lead and so how do we do that? We use AI instantly finds unanswered replies. Hey Reach also pulls in uh LinkedIn replies immediately. Claude reads it and rates the fit and then it calls Calendarly to see where are available times that we can suggest to book in for that prospect and then Claude writes the reply and then it's sent. And so this is what it looks like in practice. So the prospect Sam Carter at Northwind said yes, happy to look at more. So he is a

[20:58] good fit. Therefore, it moves to the next step which is to uh send a call Calendarly and see if either I'm booking the call for me or a prospect. So if the or a client so if the client has that those two times on his calendar, it pulls it from Calendarly and it double checks it and makes sure and then it suggests it in Sam's time zone and it does all of this in the back end like that and it and it sends him a reply

[21:24] automatically within 60 seconds. So, this is a gamecher if you're trying to book calls and you can't figure out why. It's because you're replying seven hours too late. And so, this is a follow on from the one that I just showed you. This is the reply desk. So, every reply from both channels lands as one card with an owner and a timer. And so, I have noticed that sometimes we want to

[21:47] be replying manually to things. Sometimes we want to let AI handle it. So, the system understands all of that and you're able to come in and claim a reply or it escalates it to the AISDR. Okay. And the next system is the cold calling system. And so, I've gone full screen because I want to stress the importance of this because nobody is picking up the phone and calling anybody anymore. Everybody is too afraid to do that. They're all afraid of getting rejected. And as soon as I added a layer of call calling on top of my email outreach, my bookings went through the roof. And the amount of

[22:31] clients that I signed that said to me, "Oh my god, you're going to call the prospects. Yeah, let's go." Like, we need that. And so I built a system end to end that allows me to use my desktop computer to call from anywhere in the world to anywhere in the world. and it just costs a few cents. So, this is what it looks like. It looks at ClickUp to see all of the people who need a phone call and then Claude puts together a bunch of notes before each call. I use Twilio to set up the actual dialer. Uh, I have Calendarly so that I can have the

[23:08] calendar appointments next to the calls and then N8 logs every calls in the ClickUp so it can see when I called, how long I called. that even puts the transcript and a uh the audio recording of the call into ClickUp for me to look at so that everything is recorded within one card in my CRM. And so this is what it looks like. But let me actually show you in real life what this is. And so

[23:36] this is the system that I have built. You can see I've loaded this up with $32. And all down here, these are all the people who are ready to call. And so I've even got a way to get their number and it calls different APIs. I believe it calls three different APIs to try and find their phone number based on their name and email and LinkedIn. And so you can see that this one here, I can give Philip Green Witch a call by clicking this button. And it'll just call him

[24:03] straight away. Um, and then over here the calendar is hooked up. And then over here it loads all the context of this lead. And so if Philillip and I have gone backwards and forwards a couple of times, um, then all of that data will be here. You can see it's qualified him. His LinkedIn is here. I can open his ClickUp card. Uh,

[24:25] here is like a TLDDR of who this guy is. And here is a script based on Philillip's uh, Philip's uh, business and the offer that I'm selling to Philillip. And here are objections that I can even handle. And you know, I can do cool stuff like go from dark mode to light mode. Um, and then over here, I can change which client I'm calling from. So, it's pretty crazy powerful what you can do. And this has been one of the biggest game changers. I can even change the name. I can change like the settings here. Uh, I can do lots of crazy stuff, right? And so, this is so

[25:02] powerful. You show clients this and they lose their mind. Now, if you are just getting started, you might be using Upwork. And I I I'm a big big fan of Upwork personally. That's how I first started. It's how I first scaled to $10,000 a month. I had a lot of clients come in through Upwork before I started

[25:20] spending money on cold outreach. And so what I have built is an endto-end system that looks every day on Upwork for specific jobs that might be a good fit for my talents. And then Claude goes through and scores everything. And then I get a full printout inside of Slack and it gets added to a task inside of ClickUp. And then Claude Code drafts me a reply or a proposal that I can send to that person. And then I can just manually like uh like take it and paste

[25:54] it in. And so this is what it looks like. Uh somebody posted a job that says build our full sales funnel before a January January course launch. So, it scored a five out of five because I'm fantastic at building sales funnels. And then I can click on is it a good fit or not? Is it not a fit? I can even open the job from here. If I click on it's a good fit, it'll ask if I want to generate a proposal and then it'll do it all for me and I just copy paste and send. And so, this is pretty crazy powerful. If you are brand new getting started and you

[26:30] don't have any case studies, you don't have any social proof, you don't want to spend thousands of dollars on cold email infrastructure, all of your clients are on Upwork looking for you. You just need a systemized way of reaching out to them. And look, this is the next one called the inbound desk. And I get a lot of inquiries from lots of different places. And some of them are important, some of them are not important. And so I've hooked up Type Form, Claude, Stripe, Telegram, Slack, so that all of

[26:57] it runs through this this like AI brain. And then it understands which ones are important and it escalates those notifications to me which are. So for example, I get notifications that's like, hey, somebody just filled out a form on your website. They're a really good fit. You should reach out to them. or it should be money just arrived in the account or there's a payment that's that was failed or this client hasn't paid or they are overdue or this card was declined and so I don't need to

[27:25] check my emails or anything like that. Everything is all delivered to me inside of Telegram. You can deliver it inside of Slack or WhatsApp, whatever, but Telegram is the easiest for me. And so when me or the team books in a call, this is the pipeline that kicks in. Every new booking is automatically logged. It adds it to the ClickUp card and then it creates uh it starts an N8 workflow that asks Firecrawl to read their website and then it builds out a qualification and a uh a justification for booking the call and it creates this dossier like a full paged researched dossier. And so I get a

[28:09] ping inside of Slack saying hey new booked call and you can also tag the client. So, hey John, we've just booked you a new call. Here's all the information you need. Their website, their LinkedIn, and this is the dossier. So, that's the first part of this process. This is the second part. It runs through firecrawl exa runs through the harvest API and then Claude goes and scour the internet and tries and finds every single thing that they've ever done and puts it together into a comprehensive re research document so that the client can come in with a very clear understanding of is this person qualified or not. And

[28:46] so like here is a real life example of [clears throat] a strategy dossier. So, the client can come here and they can click this and it downloads a claw.md file of this prospect. Uh, you can see that there's a summary here. Here's all of their details. Here's their website. It runs it through Homozy's core 4. Uh, it's found a bunch of u marketing gaps based on things it could find. Here's their profile and company intel. This is their business model. This is the TAM math. So, I need to make be making sure that I'm only speaking to people who have hundreds of thousands of people in

[29:27] their TAM. And so, this is important for me. Are they qualified or not? This is a key part of my qualification process. And then here are a bunch of pain signals and gaps that I could potentially squeeze to close a sale. And then this is their sales motion. This is contradictions that we found uh that don't make sense that we might need

[29:48] to ask them like what's going on here. And then these are the discovery questions that I can ask them. Like can you imagine how mindblowingly easy discovery calls are when you have this? Forget about an AI sales coach over your shoulder which I have built. But these are all of the discovery questions that you need. Imagine being able to give

[30:09] this to your client. Then on every single call that I get on, every single one, it creates a summary and a task list because sometimes I'm on a call with a prospect or a client or my team and they say, "Oh, we need this, we need that." And it's like a five minute task and I write it down and I forget it

[30:29] or I don't write it down, even worse. And so what it does is that it runs through the entire uh the entire transcript and it pulls out action items similar to how Fathom does, but it hooks into my ClickUp and it creates a task. And so I can click here and it'll add this task to my ClickUp and it'll remind

[30:51] me periodically to do the thing. Gamecher for productivity. Now this is the call coach system. So this is part of the sales system where every sales call gets marked out of 120. So I can ex I can see exactly where deals are leaking. And if you're not watching back your sales calls, you're making a massive mistake. But if you're not watching it and you're not taking notes on what you could be saying and doing better and you don't have somebody else reviewing it, it's very difficult to improve. So it runs through Fathom Claude, Slack, N8N, Google Docs where it writes the report and then it gives me a

[31:28] score out of 120. And so uh it automatically posts it in the Slack for me. And then I can get a full report. And this is what like a this is what the tool looks like. And so I can this is all done automatically in the back end for me. And it just it spits it out inside of Slack. But if you want to do

[31:45] it manually, this is what it looks like. And so I can paste a Fathom link in here or I can paste a transcript and then it'll give me a score out of 120 and tell me what I did right and what I did wrong. Um, so it's it's like it's pretty invaluable. Next up, we have our first delivery system. So this is the client portals that I have built for each one of my clients to house all the documents, all the reports and all of the pipeline that I generate for them in a beautiful, easy to understand, accessible place from their phone or their desktop that they can give access

[32:25] to uh anybody on their team that they can all look at. So it runs it through well, it runs it through all of these steps here. I won't bore you with all of that, but it saves me hours and hours a week by putting together these reports in a Google doc that no one's ever going to read or see. And so it it allows me to create a client hub like this where they can log

[32:47] in with a specific username and login. And this is all of their documents all housed here for them. and they can see the reports um and like monthly pro monthly performance reports and all of the the things that are going on inside of the business for their outreach. And then the weekly reports is another part of that delivery system, right? So every client gets a weekly report built from real numbers. And so I don't have to go to instantly and be like, "All right, how many emails did we send? What was the reply rate? How many calls did we book?" I don't need to go to the air

[33:22] table or anything like that. Everything is automatically sucked in for me and it looks like this. And so here is an example of a report where they can see how many messages were sent, how many replies, how many booked calls and here is their pipeline. And so uh we booked a call with Sam, we booked a call with Priya. This is somebody that we've got in the pipeline right now and they're only a three out of five star and next week this is the plan. And so when you're able to come to a client meeting with stuff like this, it's so much more powerful. You look so professional. And

[33:57] something else we have been doing to massive effect is reaching out with different types of lead magnets. And so um the actual building of the lead magnet is done through ClickUp, Claude, GitHub, Cloudflare, and Slack. And then once they once the prospect opens the lead magnet, we get a ping inside of Slack. and we're able to pounce on it and and and say, "Hey, I saw you just opened the report. Would now be a good time to book in a call and I can walk you through it." And so the report tracker will ping us and say, "Uh, you know, Sam just opened the report that we

[34:31] made for him or one of the lead magnets." And so this is the draft message. And I can click send it now and the AI SDR will automatically send it or I'll do it myself or I can even give him a call and find his number and then load it into the dialer. And if you're a Shopify store or if you sell in e-commerce, we have built the journey scanner. So it looks at their shop. It uses Playright and Cloudflare to go to their website and actually go through joining their email newsletter. And then it pulls the email newsletter and it finds all the mistakes in the email

[35:08] newsletter and then it creates a comprehensive report on what could be done better. Right? So, this is what it could potentially look like where it actually creates a report saying, "We got this email from you. Here are all of the problems that we identified. Here's how you could do it better. Do you want to get on a call and I can show you how we can implement this for you?" Or if you are doing any kind of conversion rate optimization, you can actually send AI to go to their website and then crawl everything and show you where the things are being done correctly, where where are they missing certain things like are

[35:45] things not presenting above the fold in the way that they should. Is the call to action unclear? All of these sorts of things. And then put together a comprehensive report based on that. And so it looks hyper hyper personalized. And now we have some really juicy stuff around content. And so this was one of the biggest game changers for me because I want to make content, but it's really difficult to find the time and to put together all of these slides and to edit everything. And I used to have to use two other people to do this. And now I just do it from one terminal inside of

[36:18] Claude. And so this is what this looks like. I use uh Whisper from Chat GBT. I run it through Claude and then I use something called FFmpeg and then I use Remotion and all of this is is free. You just pay for credits and this is the output. So some of you may have seen this already but all of this was done by talking to this camera and then I just plugged it into this skill

[36:44] and it edited this all for me from zero. Like if you were a video editor, you should be shaking in your boots because this took me two minutes. This would have taken me two or three days and cost me hundreds of dollars in in fees if I was using an editor. And this is one of my favorites. And I actually used this thumbnail studio to create the thumbnail for this video that you are watching right now. So, if that's exciting for you, like this video or leave me a comment or something because I want to see if you made it this far in the video and you saw the thumbnail and you

[37:24] clicked the thumbnail and you made it this far in the video. Obviously, the thumbnail has worked, right? And so what it does is it pulls from multiple different channels in my niche and then it creates a list of uh inspiration for me to potentially use and then it allows me to put my face into the thumbnail studio and it uses Nano Banana and GPT and it creates 4K thumbnails for 30

[37:49] cents. And so one of the most recent videos I did, I used uh Dan Martell as an example here and I created this thumbnail and it managed to get a one out of 10 for me, right? And so here is the studio. And so you can see here that I have Hozi, Imanzi, Cole Gordon, a bunch of other people here loaded in. And I can come

[38:13] into this one and I can load in my face. And I can ask AI to change this book to something different. change the background to something different. And then here are the generations that I have made. And so you have likely seen this thumbnail where I talk about revealing all 35 of my AI systems. And now you are watching the video from this thumbnail. So I like to create the title and the thumbnail first before I create the content. Otherwise, I won't create the piece of content. And so this is the thumbnail that got you to click. And I'm and this is the system that I use to

[38:49] build the thing. Pretty crazy, right? And then so system 26 is the launch funnel. And so AI builds the sales pages, the upsells, and the ads, and then it makes sure that it's optimized for mobile. And I use Appify, Claude, Nanobanana, Chrome, obviously, and GitHub pages. And so these are some examples of the ads that I have built which you are running right now which are crushing it for us. And then here is an example of a funnel that I built from scratch using Claude using all of these skills. You can see I built this entire website from scratch all from within Claude. I'm not a designer

[39:33] and I I think it looks pretty darn good. And you can see it's optimized for mobile. And so everything needs a command center. Everything needs a dashboard. Otherwise, you've got all these disjointed systems that don't talk to each other. And so, everything is housed inside of the operating system, which is system number 27, the command center. So, every lead reply lands in Slack with real buttons. I can answer from my phone. All the notifications from other channels or other types of

[40:01] outreach are all within Slack for me. And so, this is what it looks like. A reply comes in. I can click send now. I'll edit it. hold or something like that. And also other notifications from LinkedIn, uh, emails that even come in instead of me going to Gmail. Now, everything is just all housed within Slack for me to basically have an all-in-one command center. On top of that, I need to know what's going on in my business, what things are falling between the cracks, uh, if this team member is doing a good job or not, uh, things like that. And so I have a daily brief every day which checks my

[40:42] meetings. It checks my emails that are might be waiting for me. It checks uh late overdue tasks inside of ClickUp. It looks at my instantly sends and sees if everything is working there. And then Claude picks my top three actions for the day. And then I get a notification inside of Slack being like, Jordan, these are the things you need to focus

[41:03] on. So good morning Jordan. this is your day and these are the three things that you need to do. And then it also does something similar with my CRM, right? So, it's looking at all of the people who are prospects who are warm and they are almost ready to close and then it's looking across my entire pipeline for opportunities for me

[41:28] to make money and close deals. And so, it's telling me that two people said no. uh one said no or this is the closest to being a a yes call this person and then here are the three best moves to do today and then so this is some of the unsexy stuff but it's really important I need to make sure that all my data matches across all of these systems and so it's checked twice a day and backed up nightly where it checks across clickup inn it files it to air table it checks for gaps inside of Slack twice a day and it backs everything up to GitHub so

[42:10] everything is in sync and there's only one master data file everywhere and so it looks at lead lists, it plugs different gaps and it looks at uh all of the different air tables and it makes sure everything is congruent. Now some of you may not believe this but AI is not infallible. Sometimes it makes mistakes. And the amount of times that I have lost hours, sometimes days worth of of projects because AI just decided to just delete a whole bunch of stuff because it didn't understand my command or it didn't prompt it correctly is outrageous. And so I've had to install guard rails to make sure that those

[42:51] things do not happen. And so this is what that looks like. Claude code cannot send, publish or delete anything massive unless I give it the go-ahehead first. And so it runs through this pipeline and it blocks it from actually happening. So for example, if I publish a API key to GitHub publicly, it won't let me do that. Or I've created guard rails around some spam triggers that make my emails instantly go to uh the spam folder. It won't let me do that. And then again, these aren't very sexy, but they keep everything running. And man, when I didn't have these things in place,

[43:32] everything would break every second day. And so, I run all of my agents off a Mac Mini. And sometimes it just goes down. Sometimes GitHub pages stop working. Sometimes the AI just decides to not wake up and not work. And so, every five minutes, it runs a check to see that everything is working end to end. And so it might go down at 3:00 a.m. and then this system helps me bring it back

[43:59] online automatically. If you have staff, the worst thing in the world is spending money on them so that they can sit on Tik Tok and do nothing. And if you have staff that don't have actionable tasks to do every day, they will just sit on Tik Tok and do nothing. And so I have built a way to keep them accountable. And so it looks at emails that are really being sent. It looks at automations that have really run. It looks at tasks on their list that have sit there and have grown stale and it looks at hub staff which I use to record their screen. And so it tells me

[44:33] if there are any anomalies and if they haven't been working or doing the things that they say they've been doing and then it sends me a report every day saying, "Hey, here is an anomaly. Uh John is not doing the task or Sally uh reported that she's doing 20 hours a week but actually it looks like she's only doing seven. there's this massive gap

[44:55] and so this is maybe what it looks like. So it's really important for me that all clients are at full capacity and we are sending thousands and thousands of emails every day. And if that's not done, that's my team's job to fix. And so if there are replies waiting or something or we have not used AI to reply to these people, that's my team's job to fix. And if they're not doing it, I shouldn't be paying them. And so this allows me to have transparency through

[45:23] that. And you have seen me go a couple times to this internal wiki which looks like a notion uh platform that I have built from scratch cuz I hated notion. And so I call it the internal wiki. Every SOP, every system, every program, every document, every uh skill file is all housed inside the internal wiki. And so

[45:48] let me show you the internal wiki here. You can see I can view all the clients here. I can view all the SOPs. So let me open this up. So you can see all the SOPs here, right? You can see that this is how we run everything. These are the messages that you send. These are the

[46:06] copy pasteable things that you can run. This is how we book podcasts. This is how we create content. This is how we run sales. This is how we deliver for clients. All of this stuff was generated by Claude Code in a click of a few buttons. That took me It took me months to build this thing. Like, as ridiculous as it sounds, even down to the the way that this purple gradient flows across this card. It took me it took me hours to perfect that. Um, and all the AI skills that I have that I have built out, right? So like how do you set up

[46:46] cold email infrastructure? Well, here is the MD file to teach your claude code how to do that. And so having all of this housed in one place has been so valuable that I can just share with my team and then we don't have multiple Google Docs and multiple Google drives. This is just the

[47:04] one single source of truth. And then the last system that makes sure that nothing gets lost is the core install where Claude Code saves where I am before it runs out of memory and it's allowing me to pick up where I left off or pick up where I uh pick up tomorrow. And so sometimes I'm in a window and I close it down and now I've lost all that context. Well, this backs everything up periodically. It allows me to save where I am. Think of it like a a save point in

[47:37] a video game. This is how I'd recommend you build all of these systems if you want to build all of them. So, offer and copy, leads and infrastructure, outreach and replies, sales delivery, content and launch, operating systems and guardrails. Now, if obviously [snorts] if you don't want to do all of this because literally this is how much all of these systems are worth, I have packaged all of that together for you and you can get access to everything inside of the AI system society. This is a crazy special deal. It's usually 800

[48:09] bucks a year. You can get it for 247. So, you're going to save 70%. Lock that in for life. I'm going to leave that link first link in the description. But if you want all these systems to run your own business, you can literally just copy me, plug and play all the systems, plug them into your business. Claude will walk you through customizing them for your specific business, tailoring them however you want. And if you don't want to spend all of this money, like it probably cost me5 to $10,000 in just claude code credits to set up all these systems. Forget about all the money it would have cost me if I

[48:47] was to get a developer and you're getting that for for 247. It's outrageous. So, hopefully you enjoyed this video, guys. Make sure to leave a like, subscribe, and join the AI system society in the first link down below. And I'll see you in there. case.

---

About this transcript
Read from YouTube's own caption track and laid out by ViewRank AI (https://viewrankai.com).
ViewRank AI finds the videos already beating a creator's own average on Instagram, TikTok and YouTube Shorts, transcribes them from the audio itself in more than 60 languages, and turns what worked into new ideas and scripts.
Free transcript tools, no account needed: https://viewrankai.com/tools
How to read any video this way: https://viewrankai.com/llms.txt