The ONLY AI Agent I sell to make $25,833/Week

0:00I may just have $25,000 last week selling this one AI agent with all that money being upfront cash meaning I don't have to maintain the system. I don't have to do any monthly maintenance. All I have to do is build it, sell it, and I get all that cash upfront. So, in this ultimate beginners guide video, I'm going to show you exactly how to build the AI agent from scratch that you can genuinely sell the business owners for multiple five figures. Then, I'm going to show you exactly how to outreach and sell this AI agent step-by-step with my foolproof sales system. And by the end of this video, you're not only going to

0:33know exactly how to build the agent, exactly how to sell the agent, the exact niche you're going to sell that agent to, but it's going to transform the way you think about AI agents because right now within the space there is a ton of fluff. Everyone's telling you to sell AI receptionists. Everyone's telling you to sell these copy and paste AI systems, but that's not what business owners are looking for. They want custom agents directly implemented into their business based off their pain points, not some universal pain points that you probably

1:01think every business [music] faces. Truth facts based off the exact sales system that you can dissect within their business, understand what each company that you hop on with is genuinely struggling with, and then build this system for everyone. And here's a little bit of context about how truly crazy this opportunity is within the market and how you can position yourself. Right now, I'd say within the AI space, it's probably 99% dominated by technical people. These are people who are really good on the back end, really technical, really talented when it comes to developing, but they can't particularly sell and build relationships with business owners, which [music] is the

1:38most important thing within the entire AI market. So, my personal goal within this video is to not only show you the most profitable AI agent, how to build and sell it, but to position yourself and understand how easy it is to start an AI business without being the most technical person because you understand where the leverage is. You understand that non-technical people who can build relationships, who can sell really well, and actually portray these AI agents in a simple way to business owners so that they can understand it before they

2:06invest multiple five figures into it. That is where the real opportunity is. So, that is my overall goal with this video. It's not going to be an extremely long full course. [music] I'm going to timestamp it out for you, and I hope you understand everything that I personally know about selling AI agents so we can make as much money as possible. Just make sure you watch the video all the way through. I promise you it will be a game-changer, and this video is for everyone. Whether you're starting your AI agency from scratch, or you're looking to tack on an additional AI service that's super high ticket and you

2:34know will work within the market to what you've already been offering. Maybe you're selling lower ticket AI agents right now and trying to break into the enterprise level deals. All right, with all that being said, let's jump right into the video. But, before we walk through how you can actually build this AI agent from scratch as a non-technical person, or the exact sales system to sell this AI agent to high ticket businesses for a high ticket price, we need to go over what exactly the AI agent is and how it provides so much value to the business. Right now within the market, the problem with about 90%

3:06of service-based businesses, who are the ones who actually have the cash flow to invest into these AI systems, they don't run on centralized systems. They run on old outdated systems, and all their information is scattered all over the place. Their calendar is over here. Their CRM is over here. Their payment processor is over here. They're running on old outdated CRMs. They're running on old outdated payment processors, and nothing connects to each other, and

3:30nothing is all in one centralized place. And this is a huge pain point, especially for the business owner in particular, because they're the ones who have to jump into the payment processor to send an invoice. They're the ones who have to log all the information into the CRM. They don't have employees that can do all this for them. All right, so we have the problem, but how do we actually get this solved into a centralized

3:51system? [music] And we do that with a knowledge base. So, in this case, the AI agent is the knowledge [music] base, and we train it on all their current systems. We build the knowledge base, we build the agent, and we host it on a dashboard. So, the business owner can simply go into the dashboard, go back and forth with the agent, and get any sort of information that they're looking for. Now, on top of this dashboard, we include a WhatsApp agent. Now, what this WhatsApp agent can do is when they use a keyword in their WhatsApp group chat, it will trigger the agent, and the agent can directly type

4:20into the WhatsApp group chat and answer the question or do whatever it is that they ask. And on top of that, we can also sync daily updates to the WhatsApp channel in order to give a full daily recap of exactly what the business owner has to do that day, what the business inquiries are, what they're dealing with on a day-to-day basis, and just the overall updates of the business. So, for example, if we were building this for, say, a real estate business, it would be able to track which lead is the most intent to actually go forward with buying a house. It could track, okay, who do we actually need to follow up

4:49with? All this information in one place and constantly reminding us, and that's why these centralized systems are so valuable because when everything's spread apart, when everything's on different UIs, it's very hard to keep up with actually what's going on on a day-to-day basis within your business. It's very easy to forget about things, forget to follow up, forget to remind yourself of a client that you're about to close. But when everything's hosted on one knowledge base, it makes the entire process so much easier for businesses to run leaner and more efficiently. Another example, and these aren't just random businesses I'm talking about, both the real estate and

5:24a drywall company I've recently worked with. Let's say for a drywall company, they need to ask, "Hey, do we have this specific resource?" And the agent will go and actually pull from whatever it is that we train it on and tell them, "Yes, we do have this material. No, we don't. Yes, we're running low. No, we don't have We have a ton of supply." Whatever it is. Let's say, "Do we need more staff on this job?" Sometimes drywall businesses accidentally send two

5:47different crews to the same location. This never happens with the internal knowledge base agent because they can ask, "Hey, who's the closest [music] to this job?" And it can route and send people directly to the job. This agent is so valuable because it [music] saves the business owner so much time and so much money. Let's say they have a new hire. Typically, they would have to train that hire based off everything,

6:09right? Everything on their own accord. But with this knowledge base agent, they can literally just hand the UI over the dashboard over to the new employee, have them ask it questions, have them ask, "Okay, what do I do in this scenario?" And because we trained it on exactly how the business owner operates, it will give the exact answers he would give and save him all that time where he could spend actually doing a high-leverage task instead of training a low-level employee. It's also valuable because

6:35it's in one UI. It's in one dashboard. The business owner doesn't have to go in a bunch of different places to get a bunch of different pieces of information. They go in one centralized hub, and it makes the entire process so much faster and so much [music] easier. With an internal knowledge base agent, the owner stops becoming the bottleneck, and most of all, they get real-time data and real-time answers in a matter of seconds. They don't have to go, for quoting, for example, and write up a 15-minute quote after connecting to the parts supplier, looking at what the specific parts cost, looking at the labor rates, going into their

7:07calculator. They're on the side of the road. They're a service-based business owner. They're super tired. They're super annoyed. And they're sending out a free quote that doesn't [music] directly make them new revenue. This agent has all the data. It can whip up that quote in a matter of literally 15 seconds. Overall, I hope you can see why this agent and how this agent genuinely is so valuable for these businesses and how we're able to sell it for multiple five figures. Let's get straight into the build. Now, I'm going to build this demo for [music] an auto shop just because that's my go-to niche as well as that's

7:35the niche that I've done the most of these builds for. [music] So, let's pretend you're an auto shop owner. You have to not only do all the quoting for your business, handle all the call volume. Remember, people are calling [music] and saying, "Hey, do you have availability? Hey, I was wondering if you guys could fix a 2017 Honda Civic? Hey, do you guys replace rotors or just [music] brakes? What do you guys specifically do? Hey, what's your pricing?" You have to send out different quotes and inquiries. If you're a mobile auto shop, meaning you have vans that you send out to broken cars on the road,

8:04you have to worry [music] about location routing. You have to worry about sending different teams to different places. There is so much that goes into this one particular business, and that's why these knowledge bases and operating systems work so well. So, let's jump straight into the build. So, we're not going to use real data for this. We're just going to use some mock data just cuz it's a demo. But, remember, when you build this for a client, you're using all their information, all their data to

8:28make it 100% accurate. All right, guys. So, let's jump straight into the build. Now, I've streamlined this entire process with you guys. I built a UI and a dashboard with literally every single copy and paste prompt, step-by-step with the exact directions on how to build out this knowledge base. So, as I'm building this, as I'm using WhisperFlow, just keep in mind, click the first link in the bio, completely free. You can literally copy and paste every prompt and get this built in a matter of 10 minutes. All right. So, here's the link you guys can copy and paste, but I'm just going to use WhisperFlow for the

8:57build and then put all the prompts for you guys down below. So, "Hey Claude, I want to build a centralized knowledge base that looks exactly like Claude. For example, I should be able to create new chats, edit the name of these chats, delete the chats. Uh there should be memory in these chats for previous things that we've talked about. And I'm

9:17building this for an auto shop client. This auto shop needs to know, for example, everything about their business, how many quotes we send out every single day, how many people we need to follow up with, our profit margins, our parts suppliers, how much each specific part actually costs, everything that any auto shop owner

9:36would need to know about their shop. Make this a mock demo for this auto shop. Create a name, create a brand new brand, make up a bunch of information about how they actually run their business, but I need it to be thorough. Do some research, please, on real-world information about auto shops, so you can understand what they need. Also, for the tech stack, what they actually use within their business, their part supplies, etc., to actually build out their company. I want you to build this in Next.js, use local storage for now, use Open Router for the API to actually connect to the part supplier with the assistant,

10:16so we can pick through the models as we need to get. This needs to look exactly like Claude, like I said, from the sidebar to the animations after the time messages are being sent. Everything needs to look exactly like Claude. I need you to verify it, go into my browser, and screenshot the Claude interface if you need reference, and come back to me with the finished result. And we're just going to copy and paste this here. Once again, you guys do not need to use Whisper Flow. I have every single prompt for you guys. Just go into that

10:44dashboard. And once again, this is a pretty hefty prompt as the initial prompt just to give it all the instructions, so it's going to take a while, so we'll come back to it when it gives us the result. Okay, guys, so it came back, and believe it or not, this is not Claude. This is

10:59what it actually came back to us with. So, we can type into it, we can switch the different models. We could see the sidebar. It literally looks exactly like Claude, but this is actually specifically only for the client, and we're going to build out their knowledge base step by step. And since we have Open Router, which is essentially a tool that allows you to connect to any AI agent, and the client can basically use any model that they want. Okay, so no Open Router key setup. We're actually going to show you how to do that right now. And last but not least, instead of it looking like Claude, I want want to

11:30be actually branded to my agency just because if I show them the finished product and it looks exactly like Claude, they're going to be like, "What the heck? Did you just copy it?" Right? So, I want to brand it to my agency, Vocal Chat. Okay, so now that we're back in Claude, we're just going to say, "I like everything about this. The only thing is just remove the Claude branding and add my branding." And then we're going to copy and paste my website,

11:52vocalchat.ai. Try and download the logo from my site. If not, just go to my site and screenshot my logo and edit it to the branding. So, it looks exactly like my website and my brand. And as that's loading, the second thing we're going to do is actually go and get the Open Router keys. Okay, guys, so we headed over to Open Router. And remember, Open Router is essentially a tool that allows you to connect to any sort of AI model for one API key. So, you're just going to click sign in and obviously get set up with an account. All right, so once you get signed in, we're going to go

12:25ahead and actually create an API key. So, we're going to click new key and then we're just going to name it knowledge base video. And then for expiration, we're going to say, let's say an hour. And then for credit limit, we're actually just going to click custom amount and then just like $5. You're obviously going to use more than that, but this is just a quick demo that I'm going to show you one use case and then cancel. So, no point in making any more than that. So, we're just going to click

12:52create. And then perfect, gives us our key. So, we're going to copy this code here. Okay, so now that we have that link for the API key, what we're going to do is we're going to see this drop-down folder at the top left-hand of the screen saying env.local. And we're just going to simply copy and paste that Open Router link into that file. And make sure that you save it after you copy and paste it up here. And guys, while that prompt is loading, let's actually see what the update is for the branding for

13:20Vocal Chat on the UI and the dashboard. Because remember, it looked like Claude before. We changed it to look like my website so the client would know that it's from me. So, let's see what that looks like. And pulling it up here, it looks honestly perfect. It's pretty cool how it can literally screenshot my logo from my website and slap it on there like the advancements in AI have kind of gone insane, but let's just check the um left-hand sidebar here. Okay, so the um audit came back. Uh it fixed some things. So, now let's just give it a test in the little chat box here. And

13:54you know what? Just below the uh text box here, we don't have to um really type in anything. Let's just click on give me today's numbers. And boom. And look at that. Look how powerful this is for a business owner. Literally, we built this in a matter of like 15 minutes. And imagine if it was trained on the business owner's actual data and you could get a report this easily. It's it's ridiculous. And don't mind how the uh UI looks right now. It looks a little weird because we're in Cursor's browser, but when we full-screen it, it'll look much better. So, let's actually check this over. So, this is actually really

14:26awesome. So, quotes sent out today today through Sunday, September 20th. So, seven quotes were sent out for a total of around $9,000. Three were actually approved. Three are waiting on a reply from the business owner and one is actually declined. So, imagine you're a business owner for a second here. Think about how much time this would genuinely save you. Each quote for a service-based business owner takes 10 to 15 minutes because they have to check part supplier costs, check margins, check um how far the location is if they're a mobile business where they send out vans. That's all time that should be spent on higher-leverage tasks, but

15:03instead, now they can literally go in and type in here, "Hey, what is my profit margins on these quotes?" And let's see what it gives us. Okay, and look. So, today's uh seven quotes, the uh blended margin was um 57.1%. Um you can see the approved, you can see the waiting, you can see the declined, you can see everything, you can see how much money they actually made. This is a game changer and guys, I built this in a matter of 15 minutes. It's not trained on any data. Just kind of making up some BS right now, but like if you get this dialed in, that's why I'm saying like

15:38even a knowledge base can be sold for 20, 30, even $40,000 if you hit on the right pain points and if you provide that amount of dollar value back to the business owner by saving them either time or money, right? If this is directly replacing an employee who's sitting there quoting all day and they're spending 100, 200,000 dollars a year on this employee, that 40 grand doesn't sound so bad. If they're spending two, three hours a day as the business owner making six figures a month, that 40k to win their time back doesn't sound so bad. And guys, do remember this is a demo. You could

16:10literally build this in 15 minutes before your sales call and show this to a client and they will be like their minds will be blown and at the same time I've had clients say, "Hey, can I pay you multiple five figures for what you just showed me for something that I've built, someone who's not technical?" I have a team of developers who does all this stuff for me in a matter of 15 minutes. So, do remember that when you're actually building this out for a client. I hope you guys do have developers that you're actually delegating this work to because you want to run a business, you want to actually

16:40worry about cash collected, not the back end work. Do remember that you do need to connect to APIs. There's a lot of training that actually goes into giving 100% accurate quotes. You need to look at how they were previously quoting. You need to look at every step of the way that actually goes into a quote and you need to remember there's so much testing that actually goes into getting their quoting super accurate because the business once you actually develop this project from scratch, you don't hand it over to the business owner and implement it into their company right away. No, you hand it over to the business owner,

17:11to their employees and for two weeks straight they just need to train the agent, give you feedback and tell you what went wrong, what went right, why it's quoting too high, why it's quoting too low. And then as that gets dialed in within a couple of weeks, it'll quote super accurately and you can implement

17:27it straight into the business. And then heading up here, we can click on follow-ups and it will tell us exactly who we need to actually follow up with today for the business. So it came back to us and said today's follow-up list is 16 and it actually ranks it in a priority order. So the ones that we need to follow up with the highest intent leads are at the top and then it slowly goes down and down. And then for the follow-ups it gives us the phone number, the amount, the why now, like why exactly should they follow up in this exact moment, and then exactly

17:59what to say because it's trained on the previous conversations they had with this specific client. So this is honestly really cool. And then if we go down here and click on summarize our part suppliers, it'll actually do a summary of exactly who we're buying from in the last 30 days, the spend, the terms, anything below the minimum stock prices that needs to actually be reordered. So let's see what it comes back with. So it comes back with the specific part supplier and then what exactly it is that we're buying, the

18:27spend, the terms, and then the contact. This this is super cool. And remember, all this is achieved by getting the API connection, by actually connecting to their specific part supplier. So that's a necessity and there is times in which the API docs that the part supplier of the client's working with aren't accessible to the public which could be an issue, but we do have work arounds for that. So that's an example of what an API is. Let's say we're using Calendly for a meeting API, right? We can actually head over to their docs and actually grab that API right now. So I can just show you what this looks like

19:02to actually scrape an API off of the internet. It's not really scraping, it's just a lot of the times they have public docs that you can literally copy and paste which which you to connect to their specific platform. In tool? So, you literally just search up Calendly documentation, and then from there click on the first link for the developer documentation, and then you copy the link here, go back to Claude, paste it into the Claude chat, and bam, now you can directly connect to Calendly

19:30for booking meetings. It's ridiculous. But, going back to the build, the last thing that I want to do is actually connect WhatsApp to this agent as well, because a lot of these business owners run their company on WhatsApp. That's where their group chat is. That's how they communicate back and forth. So, adding this to the knowledge base would be super cool and super beneficial for them. So, to actually connect with WhatsApp, you need to go into what I like to call WATI. So, you need to make an account, and then once you make an account, you're going to click on add channel right here. So, from there

19:58you're going to get a trial account, and then um you're going to actually scan this QR code. It's super easy on your phone. So, again, this is what mine's going to look like. Again, yours isn't going to look like this until you scan that QR code and connect it to your phone number, but it's super, super simple, very easy. So, now we have WATI connected. Very simple, like I just said, you get a trial account, you scan the QR code, you connect your number, now you can access WhatsApp, which is super cool and super beneficial. And once you have that connected, we're actually going to copy and paste once

20:27again the um API docs from WATI like we did for Calendly, and go back to Claude. And we're going to paste these WATI docs back into Claude, and we're going to say alongside that, "Now I want to integrate my WhatsApp into the knowledge base, so any voice note, image, text, anything goes straight into the knowledge base. Images go into a separate folder, and voice notes are transcribed. Files and images will be together in its own tab for organization." And then paste that in here. And once again, what we're doing here is we're connecting WhatsApp to the knowledge base with WATI so that any voice note, message, image, document

21:09can be transcribed and added to the knowledge base automatically. And while that's loading, we're going to also type in here. Also, we want it so that when they use command, say Jarvis in the chat, it can call the knowledge base directly in WhatsApp. And this just means that now every single time they type in Jarvis with a question, it can automatically call into the knowledge base and get that

21:33information directly in WhatsApp. Actually, before we forget, we're going to go back into the ENV file and we're going to type in WAPI API key equals and then we're just going to copy and paste that API key into there. And then right after we do that, we're going to come down and type in continue and test with the WAPI API key

21:55in the ENV file. And see what that looks like here. And guys, you never want to leak the um API keys into the chat. That's what the ENV file is. It's for those credentials that you don't want to be leaked. Super, super valuable, only for you. And uh that's what the ENV file is for. So, I just want to make sure that you understand that and you're not copy and pasting the um API keys just straight into Claude chat. And again, remember, Claude is trained on data, so you don't want someone to hack your key because you put it in Claude chat and it's just a mess. Okay, so it is ready. So, I'm

22:27actually going to show you guys on my phone. So, I'm going to quickly switch it over to my phone and we'll jump straight into it. Okay, guys. So, we have the WhatsApp chat pulled up right here. Once again, it's nothing complicated. I just wanted to show you exactly what it does for them and how valuable it is. So, if we just copy and paste right here, Jarvis, how many quotes did we send out today and who

22:47should I call first? Paste it right here and bam. Once again, just pulls up straight to the knowledge base and tells us exactly what we need to do, the quotes at, who we need to call uh first, second, and third. So, pretty cool here. Okay, guys. So, that just about wraps up the build. Now this is the final product and like I said from the very beginning, this is a knowledge base. This is essentially a GPT for their business, a Claude

23:11specifically tailored to their business. So many business owners are looking for this right now. It's so valuable. It has follow-ups, quotes, part suppliers. Again, this is specifically for an auto shop. Guys, sell the simple systems. Sell what business owners are genuinely interested in implementing. You don't always have to go into the AI receptionist or the Google review agents or what people are saying online is popular and the best to sell. Sell simple systems that business owners can understand, use on a daily basis, and that'll allow you to upsell them for multiple five figures later on down the

23:45line.