# 184 How to make $5K per month per client selling managed AI agents

What if your AI agent was capable of building other agents building other agents that you could turn around and sell to business owners for $5,000 a month. Well, that's exactly what this week's guest is gonna show us. His personal AI agent named Dewey is out there Spending up entire managed agents for paying clients onboarding them into Slack even handling customer support all without him Lifting a finger now most people talk about AI agents right now and they're stuck on chat bots Well, Nick our guest this week is three layers deep. He's talking about agents building agents and getting paid five grand per month per client to do it He's the co-founder of orgo, which is the platform behind all of this and he just closed another client five grand a month this week And today he's gonna walk us through how Dewey his personal agent builds Deployes and manages these agents from scratch and if he stick around till the end He's gonna even break down his entire tool stack Which we're gonna give you guys access to including the exact pieces that you need to build your own AI employee and then turn around and sell them to clients So hope you enjoyed the episode All right, Nick is back on the pod last time we did a banger episode talking about managed agents how people can Deploy agents for clients make money with them. We've got to follow up to that episode today Nick What is the what is someone in the audience going to take away by the end of this episode? I'm here to tell you that the the service layer of really deploying these agents in the businesses like the hands-on work of coming in Building an agent out for a specific domain a specific use case and doing that That is the opportunity today. I'm gonna show you why I'm gonna show you how to do this in a leveraged way of like fulfillment and how to like actually deliver on this service But yeah, I'm just like really excited about right now you have the creators of the company her of the company Who built Hermes the Hermes agent? They're literally hiring right now for deployed engineers to deploy Hermes agents into enterprise So it's like the opportunity is insane and when you see when you see the creator of this product And that is their their source of attention on revenue is going into enterprise and deploying these agents It just tells you this is a venture backed company This is like this company aims to be a multi-billion dollar company. So to me that my ears perk up my eyes, you know start to widen and I'm like It just validates everything we've been saying up until now around managed agents. So Yeah, super excited to dive into all of this Well, I know we're gonna get in the weeds with it and I've seen you posting quite a bit about it on X I know you've got like your personal agent named Dewey that I mean it's my understanding that Dewey is the one Kind of like provisioning these agents standing them up getting them ready for your clients like you're not even really doing very much yourself when it like You close a client on a 5k month manage agent contract You're not the one then going behind the scenes and like Nick is the one Turning the knobs and pushing the buttons to get the agent set up. It's your agent Dewey is the one doing that work Is that correct and you're correct? That's that today exactly like it is magical obviously I posted on X I can I can share my screen a little bit to show like I posted this video of Dewey and He was He's onboarding an agent that he built out into Slack for a customer of ours. You can see here I have my my Twitter pulled up and I'll mute myself here and I'll make this bigger But you can see here essentially Dewey is using computer use and He's going through the Slack app. Slack.com to set up the bot token and all the settings that you need to add an agent into Slack And he's doing this for another agent not himself He's doing this for another agent onboarding that agent into the shared Slack with this customer of ours and he but he built the agent like I was actually recording the videos like he built the agent that they're going to be using He pulled in all the resources all the granola notes all the context from emails, etc And actually built the agent on or go and you can see that that second computer is the one he built So it's just it's insane this video is going viral obviously me and you We tweeted about it and and everyone is talking about it But I mean this is a 5k month customer and we have agents building their agents So the upfront work of just taking the time Building out your your assistant your executive assistant who has all the tools skill sets connectors everything it needs That upfront invested time is so valuable because everything downstream of that just becomes easier So yeah, it's literally agent section is what it seems like we've got agents spawning agents and who knows where this thing is headed So let's talk about let's get in the weeds and talk about like how how can someone do this? How can they make money with this? How are you making money with this? What are the opportunities right now around this kind of whole vertical? Yeah, so real quick like here's the company who makes the Harmeese agent news research I just want to show this because I think it's so cool like this company I mean there they're behind the most popular harness today more popular than open cloud at this point They're hiring for deployed engineers. What is the responsibility of these four deployed engineers deploying Harmeese agent into enterprise on the cloud? Okay, so integrate APIs essentially get it connected to tools You know work closely with customers to scope implement and iterate on solutions, you know improve deployment and improve improve observability Etc cetera, et cetera, so they're and they're clearly going after enterprise use case So they're going to go after the fortune five hundreds with these fds these four deployed engineers Imagine how big how much bigger the market is beneath that the mid market the SMBs the small medium businesses. I mean that is just up for grabs right now And it's up to people who are entrepreneurial and they're witty enough and they're clever enough to put together these tools And just like go take action and deploy the agents into businesses and so yeah, do we I mean you put up the upfront effort of Of doing this once and I have do we deploying agents for me and so I actually have been using this tool called Orca, so if you if you're not familiar, this is like a kind of like a coding development tool But they just have a nice like you I'll show you how to like you can spin up a project here and I'll spin up a project from like a blank slate And essentially it's just a terminal that you start with but what I do is I like I SSH I go into my or go computer and Now I'm inside of dewey's terminal computer and then I set up all these tabs here and I arrange it so that this is like my My workspace. I have dewey's computer up here. I have his terminal down here so I could talk to him and then I have a Grock over here on the right that's connected to the whole or go stack via the MCP. This is like my my work my work for my operation layer that I've been using lately and so What I'll do is I'll literally I'll come here and I'll tell dewey Let me switch the model I'll switch them over to oh, no, you know what? Let's keep let's keep Grock. I'll keep Grock. I'll say spin up a new Hermes agent in your brand agency workspace Just a demo showing the audience here So I'm literally telling dewey, you know spin up this agent for us and he's gonna go ahead and do that I'll go here into or go and I'll go in this like workspace view and pretty shortly We'll be able to see that he's gonna you're gonna see a computer start getting created here But while he's cooking on that it's like I would like to show you this is a can you see this this text message here would dewey? Yeah, oh, yeah, so I have dewey on iMessage obviously I Like texting him via iMessage. It's it's very like casual and cool And I gave him like the ability to make thumbnails via hex field and YouTube video like little shorts and stuff and so he just does everything dude like he's just doing everything at this point Right here you could see he's creating the Hermes demo computer and or go Literally just popped up out of thin air because he built it you're saying yeah exactly he's building it He's spinning it up from a template We make it so easy to spin up these agents. I mean we have templates here in or go You can select a Hermes agent open-clutch and clawed code agent whatever you want to use But he's making one right here. It's spinning up and and he made this entire one over here off to the right this Hermes customer setup for our brand design agency that we're working with and You know as a five-game month customer He built the agent Added it to the slack. He's gonna build out all the skills today all the integrations and so forth um, and so like you said it's agentception and to take it a step further I know I'm stimming out on this, but it's just like This is like talk about fulfillment leverage that used to be people you needed to deploy, you know now you have agents doing this There is let me show you let me pull up a Let me pull up this here so I Have do we Talking to customers as a i-message group chat. He is actually talking to customers helping fix customer issues on on group chats on i-message and He's like texting them in the group chat fixing things for them telling them all right You're all set you're good to go and so now he's not only building the agents out But he's doing the customer support for them as well and fixing them upgrading them managing the entire firm without me, you know, insane, insane. - So this is, it's crazy because all of this was not possible even like nine months ago. Like this is all so new. And so something I wanna clarify for the audience too, right? 'Cause we threw around a lot of names in different tools. So you said like Orca, you said Orgo. We were talking about like Dewey. So just to clarify, right? Like I want you to explain. So obviously you're the co-founder of Orgo, right? So take like 20 seconds, give us the elevator pitch there. And I can even speak like from our experience of me and my business partner Nick, when we build and deploy managed agents for both for ourselves and our clients. Like we've got seven managed agents for our own company that live in Slack that we built and deployed within Orgo. And so that's where we build them. It's super easy. I want you to tell the audience, like I'm just giving them the pitch, right? Like why build their versus like a VPS? What makes it different? Because again, I see the value. Some other people might not. I think it's just useful 'cause that question comes up all the time. So Orgo is a purpose built platform for deploying computers into the cloud that are secure like environment computers for these agents. And we give them not just a sandbox, not just some headless little VPS that is like, you can't see what's going on. We give them a full desktop computer. And that's really important because over here on Dewey's computer, you can see he was using computer use to sign up for Higgs field. So fun fact, I didn't sign up for Higgs field for Dewey. I told him to sign up for Higgs field. He has a card, he has an agent card, and he went and paid for his own account. He even made a weird username. He called it like, I don't know, Dewey, yeah, green, Baluga, 1607. I don't know. Like he just makes accounts for him. He's a funny, you know? He just makes accounts for himself. So he uses the computer. And what I'm seeing is with these new GPT 5.6 models, they're using computer use more spontaneously than they are other tools. So rather than wanting to use a MCP or a connector via Composio, my agent Dewey has been using more and more computer use because it's just, I mean, it's like, it's trained on human behavior, these AI's. They're like us. And we like to use a computer. They like to use computers too. And so that is how agents are going to be operating with software and computers is by actually having a computer. Now, Orgo makes it easy to manage a fleet of them. You can create workspace for a client, put their computer agents, their Hermes agents in those workspaces, and then manage the entire fleet just like you saw Dewey do just now with the MCP and CLIs for Orgo. So yeah, we just make it super easy to spin up a computer. You can just launch a computer. Once it's launched, you go inside of it, you can install a Hermes agent, an open-cli agent, whatever kind of agent you want, connect to it to your agent. And out of the box, it's like the simplest, easiest way to get going and start building these out for businesses. So. And you guys are your kind of positioning, Orgo, you're building it to be a platform for people who want to build and deploy these managed agents as a service. That's like kind of the niche that you guys are putting yourself into. It's like, hey, if you're selling agents as a service and Orgo is a place to do it, it's my understanding. So I think it's smart. And then we'll get into the weeds now on, you know, how do we set this up? How do we actually do this from scratch? How is Dewey facilitating this? But just in case people-- I'm sure people are hearing this and want to sign up for Orgo, we do have a pretty generous discount for my audience. I think I'm the only one that gets it. But it's-- I believe Nick, correct me if I'm wrong. First, three days of Orgo for free. And then the first three months after that, 20% off. And we'll put the link in the description for that. If you use the link and the code Corey, you get that discount, but you've got to use both. So let's dive in and let's check out-- let's see this in action, end to end. Yeah, so as far as-- Yeah, and that's correct. That's the deal is you get first three days for free if you use code Corey. And then you get 20% off the first three months. And so building these agents on Orgo is super-- both fun and easy. And like I showed earlier, you can spin up from a Hermes template right here to be able to launch a computer with Hermes pre-installed. Now, if you're on a scale plan for Orgo, you can actually use-- I can give you the link again to the GitHub-- but you can use my Nick stack here, which spins up-- I haven't updated it in a minute-- but it spins up a template that kind of installs everything that I use for my Hermes agents all pre-configured. And so you can see it comes with Obsidian pre-installed. It comes with Agent phone, Agent card, like all the tools in the stack that I use. It comes with all of that to boot. So I don't even have to configure any of it. And it just spins this computer up, snapshot it, ready to go with my whole stack. And that's really powerful. We'll put that link to his GitHub with-- so they can basically download that template from GitHub, imported into Orgo, and then one click deploy an agent with all those capabilities already installed. You're saying? Yeah. You could literally just give the link of that GitHub repo to your Cloud Code or your Codex or whatever, Grock build. And you just give it that link. And you say, all right, spin me up a template on Orgo for this repo. And you'll have this in your account. And it'll just work like that. Yeah. And then we'll put that in the description in the show notes so people can just go and download that for free and get after that. That's awesome. Yeah. It's super. We're trying to make it so that if you want to build a stack of what you want-- OK, if you want a marketing agent, you want a sales agent, you want a customer success agent. You could build the stack of what goes into that agent. You just come here to templates. You create a new template. And you could say, oh, I want a Hermes agent with G Stack or G-brain pre-configured. And you just tell the Orgo configurator to build this computer for you. And then it builds this spec out. You publish it. You build it. And then you could just use that repeatedly forever for a template. So-- And that's how you can productize these agents, right? Which everybody talks about selling a productized service, not reinventing the wheel every time and building custom every time, because that's how you get in a trap of not being able to scale if you're just constantly doing custom work. So this is your way of saying, OK, we can build this agent the way we want it one time. And then basically copy and paste it across clients. Exactly. You can, in one click, be able to spin up the ideal stack. And that's also why I think going after-- this is something that we don't actually do as much of, because we're kind of at a platform layer. But if I was spinning up an agency today, to just like that was the only thing I was doing, and I wasn't running a software company, I would just be making a very vertical, specific agent for insurance. And I would make a template on Oracle. I'd deploy that into businesses that a service layer first. And then I would productize it. And then eventually I would turn it into an insurance perplexity computer web app, you know, and have it super productized. And so you can see here, it builds this whole thing out. You can accept it. And then go on and build and publish the template. But yeah, like we're trying to make it super easy to just-- you build the ideal stack, you templatize it, and you could just spin it up to however many you want. I even have, you know, like the ability to clone a computer here. So I just made this Nick Stack computer. I can clone it and have another Nick Stack computer ready to go. So it's like, we just give you the whole fleet of ways to be able to build here. As far as getting set up with tools, connectors, everything like that, it really is as simple as-- I made this tweet describing the future of software. And it's happening today is understanding the idea that you don't need to know how to install the tools. And you don't need to know how to use the tools. I don't know anything about how to use Higgs field. I just know Higgs field is a creative generator AI tool. And I know that it's like, well, regarded. And so I just tell Dewey to go sign up for Higgs field and use that for all creative generation assets that we do. So what's happening is people are going to use more and more software than ever before. Some people say, oh, software is dead, not at all, not at all. Software is going to sell more than it's ever sold before. And you're just going to have to tell your agent what software you want it to use, aka Higgs field, or Composio for the connectors, or use agent phone for texting my agent. You just tell it what you want, and it'll configure it. And it's really that simple. Here I want to show you I have this telegram chat with Dewey. It looks like he-- did he send me some videos just now? I mean, he sent me some video. Let me download this. I just-- I live-- I live-- --means. So he made a-- OK, you can see here. He actually just made an explainer video. So I have a YouTube channel where I made a video on how to build an AI employee that does everything. And it kind of like walks you through how I built Dewey, for instance. And I pointed him to that YouTube channel. I said, like, go watch that video and create a little short that I composed on YouTube about this. And it looks like he made an explainer video. here using Higgs field and it probably has audio, I just have it muted. And he did this entirely himself and I can show you here. He has all these thumbnails he was making me. He's using vidIQ which is a YouTube tool to look up the highest most viral videos in my niche and he finds the thumbnails that work really well and he uses my face to create thumbnails for me in that niche that will do really well on YouTube. It's insane. So that's just like an example again of like, I don't know anything about how to use Higgs field, how to use vidIQ, anything like that. I just know that these tools exist. This is the use case for them. I point, do he at them? He uses them and I get the output. It's like insane. So we're really at a point where we're only limited by our imagination. It's more so knowing what to have the agent do because it can do just about anything. I'm going to just go back to that principle of focus, like basically telling the agent what to focus on. So that's like an insight that I got out of this is execution is cheap. It's just knowing what to focus on and prioritize. Now, so I want to switch gears a little bit because you tweeted, it was either today or yesterday, that you just landed another 5K a month managed agent client and it's funny because after the first video that we did, a lot of the feedback I got was, oh, well, like, you know, who's going to pay 5 grand a month for an agent, right? Like that's rip off. They could just pay 20 bucks a month for a cloud and get the same effect. So, you know, obviously, I have a strong opinion and rebuttal to that. But like, I get it is kind of a legitimate question to a degree. Like, what is making someone pay you 5 grand a month for one of these agents? Like, what is what's so special about the agent that you're creating that makes them pay that much? And then let's get into the weeds of like, what is it doing and how did you actually set it up for this specific client? Yeah. So, the key here is this, this is not an obvious insight, but it is one that I like kind of, it is discovered by being in the trenches of doing this so many times now for businesses. And it is that the customer, your customer, at the most principal level, they want to make money. So, if we're selling them an AI employee and it's not going to make them money, forget it, how do we get an AI employee to make a business money? One way that's been working really well for us is packaging up the AI employee that we built for them in their industry and their use cases and packaging that up after we build it for them for their end customers. For instance, this design agency that Dewey is building the whole agent for they want an agent that can do these kind of design templates and brand assets and so forth for their customers. So we're building it for them so they're going to build their agent for them so they can use it internally to build out these assets for their customers. But then their end goal is they're going to sell this agent that we just built them. They're going to sell that productized once again to their end customers as a revenue stream. And so this has happened across a few of our customers now. And so it's like B to B to B. And when your customers depend on you as a revenue stream for them, that is like these guys will never turn and they're making money from us and they're depending on us and our infrastructure and our deployments and our expertise of setting these agents up. So the key to reiterate is that we want something that is not just a one off deployment of an AI employee and that's it. We want it to be something that we can productize, package up and deliver to their existing customer base, A for reasons of scale and B for reasons of just stickiness. So that makes a ton of sense because when I think about, so like first of all, I agree with your point of like nobody is going to pay five grand a month for a managed agent if it isn't directly making the money. And so where my brain goes with that is, okay, well, that agent has to be bringing in revenue, right? And again, my just limited thought process was like, Oh, well, it's got to be like a sales agent of some sort. It either needs to be doing speed to lead or speed to quote or follow up or something that we can directly trace to revenue. But you've taken it even a step further and said, well, it doesn't necessarily need to directly make the money. It just needs to be something that they can turn around and package and sell to their clients. Like that's their way of monetizing it. So I, that's just like a whole different way to look at it. Like you said, it's not obvious, but like once you see it, you can't unsee it. Like that's genius to me. So that's, I think that's a really good insight. Now, so continue. I just, I just wanted to kind of touch on that. No, yeah, that's, that's exactly it. And, and it's like, you know, they, they depend on us and they'll depend on you as a, as a service as an agency to, to maintain the deployment and maintain the infrastructure and make sure that everything's running smoothly and onboard their customers for them, you know? And this used to be something that would take a lot of people to scale. But what I showed you earlier with, Dewey is like, you can have your agents help you with the customer support. You could have your agents help you with, you know, getting them set up with the, with their, with their, with their agents. So, yeah. And, and so there's that, there's that, you know, to recap, we're going to come into a business. We're going to build them an AI employee. They're going to use that employee. It's going to be catered towards towards their industry, towards their use cases. Now we're going to productize that. We're going to package that up into a template on or go. And we're going to go distribute that to their end customers by deploying those templates on or go and, and we're going to go onboard those, those end customers, either via ourselves or via our agent and get that into their hands. And that's still at a service there, but it's a productized service there. But then the extra step there after is to then say, well, can I make a one click, can I make a one click version of this? That's a truth web application, a true software. Can I make a perplexity computer for that use case as well? And have people just come, come to my web, come to this platform, sign up, pay 500 a month, and they get it even more productized out of the box. That's the spectrum. And it goes from, you know, one to few in terms of fulfillment to one to, one to, you know, some to one to many of how many people you can fulfill and scale with, you know, deploying these agents. So, yeah, I think the biggest thing is understanding that the time you take to kind of be, you know, the time you take to sit down, invest the time of setting up your own personal agent, getting it connected to all the right tools and anything that it might need to be able to do real work. And then just trusting that it can do that and letting it, letting it help you, you know, actually build out agents, reply to customer emails, help you scrape leads, use an appify connector to get leads and then enrich those leads via the MCP for clay. It's like you just have to know these tools, put them all together, invest that time and you have a co-founder that's going to help you run your business. Right. It's like knowing like where we as humans are valuable is knowing which tool to use for the job, not so much in knowing how to use the tool. Right. Like that's the main takeaway I'm getting from you. Like you said, it's like, I don't know how to use Higgs field. I don't know how to use clay. I don't know how to use appify. But I know that for this particular task, for this particular use case, it is the tool we need to use. So I'm just going to have the agent go figure out the execution. Right. And then once I've used the agent to execute on a problem that's valuable enough to this specific business, I can then go and sell that as a productized managed agent to that business. That business can, you know, together with that business, we can work out the kings. And then once we worked out the kings, we can have that business, aka our customer, turn around and resell it to their customers as our product that we own. And then, you know, from there, we can potentially offer it as like a SaaS solution. Like you said, once it's super dialed in. Now, could we go back to the graph that you actually just had up? I think that's important because people are going to want to know like, okay, like what, so you mentioned like knowing the right tool for the job, which tools the agent has access to to do its job. Is this like the full stack right here? I think that this is the big component. These are the big components. I can probably just. This is the 80 20. I can recite to like verbatim exactly what tools I use and I'm happy to do that. So go through these ones real quick. This is this is just what I use. There are alternatives to make that clear, but this is my stack so far. So you have a AI agent like OpenClar or Hermes. That's the harness. It needs to be powered by a model. So, you know, to supersede this, there's a, there's a harness here. Let me just say that harness like OpenClar or Hermes agent. A harness is just a framework of how your agent, your AI will use tools and how it'll gather memories and just kind of its personality and how it off like it's plumbing, right? Yeah, it's the, it's literally the plumbing. It's the architecture. It's the backbone of how it, how it operates, but the engine and the intelligence, the brain of it is the model. And so I have OpenClar or Hermes, but it's being powered by what? Grock or GPT 5.6 or whatever model you choose. And then okay, now I need a, I need a place for this this agent to live. It needs to have its own computer that's on 24/7. so you can click around and do things and actually have its own environment. That's secure. So that's where you use something like Oracle. Okay, I need to give my agent an email of its own. I'm not saying connect it to my email, yes, do that, but it also needs its own email so you can sign up for tools and platforms as well. Like you saw it sign up for Higgs field. It did that with its own email. So, okay, I use agent mail. Okay, it needs its own phone number. It happens to be a company called Agent Phone. It needs its own-- - Pro agent something. - Is the agent mafia? I'm good friends with these people to be clear. Like, we get together, Agent Card. So, you know, you can actually, and I think they even have, and we're gonna be a part of this too, but they have this thing called the agent, the agent bundle.com. I'm pretty sure, oh, did they get rid of it? I think it's called the agent bundle.com. There it is. And you can go here and you can get an agent card, agent mail, agent phone, and you're gonna have an org or computer as well. We're gonna integrate org or go as the runtime for all this. And you get it for like 20 bucks a month, you know? - That's so smart. - Yeah, so this is the agent bundle powered by the agent mafia. - As you're saying this, I'm sending a message to Nick. Literally, the message is just the agent bundle.com. - The agent bundle.com. And Nick is Cory's co-founder. Nick squared was two nicks in the room. And then I use Obsidian for like the knowledge base for my agent to be wired into. And then I also use this tool called hauncho.dev for the memory kind of like managed memory cloud platform for the agent to have persistent memory. So if I talked to my agent on telegram or I talked it on Slack, no matter where I'm talking to it, it has all this memory, all this context. I don't have to remind it about things. And then finally, the last 8020 tool here in the stack is Composio. So that connects Composio allows your agent to connect to tools like if you want it to connect to your Gmail, your calendar, be able to read and write access to all your platforms, all your tools that you use, you can use Composio and it's free to be able to do that. - No, okay, and this is awesome 'cause I think this is what people really enjoy is like, okay, what's a nitty-gritty? Like what are the tools? How are we setting this up? So just to clarify for my understanding, so like that client that you close yesterday, right? So like, you know, design agency, five grand a month, they're closed, they paid the first invoice. Now we've got to physically set up their new agent. So that looks like, so how do you go from like paid invoice to okay, their agent is ready and they're talking to it? - I, so I, sorry, I just heard the siren. I'm in San Francisco, I was always something on. So I always have a, first I have a discovery call. So the first call that I have is not a sales call. I don't know, I think in a lot of ways I vibe business. I don't have like a proper funnel. I don't have a proper like, you know, cold outreach like in a lot of ways I vibe business, but this is just what I do and this is what works. So first call is discovery call. We talk about, okay, what is the problem? Like what are we trying to solve for here? What is the use case? Is it valid? Is this customer even a, an ideal customer should we even work with them at all? Can they afford our thing? Or are they, are they gonna turn next month? And what you find is the ICP is usually a company that's doing between, you know, at the minimum, one to two million a year, at the minimum. And believe it or not, I actually like working with the SMBs and mid market companies more than with software companies and with AI. Like, like, like, they are, they will pay you more. They ask less questions on like, they're less needy. Like it is actually is the outcome achieved or no? Like, yeah, no, you know, exactly, exactly. So it's like, you know, over here in Silicon Valley, you're like trying to sell the startups and everyone's, you know, getting to a hundred million ARR selling the startups and it's like, it's a living hell because startups are broke, but you can go sell to, you can go sell to real businesses and their way more willing to pay money. So it's actually like a huge, huge opportunity right now. Once again, like I said at the beginning of the pod of, you know, selling to the SMBs and mid market. So that's the first call. Second call is usually I close the deal about, okay, like here's the proposal, here's what we're gonna do, here's the problems we're gonna solve. And then after that, I immediately spin up a trellaboard. I onboard them into that, I say, look, here's the backlog of items that we talked about. Here's what's being worked on right now. Here's what needs to be worked on next. And here's what we need to do together. Like I have like four columns. So it's like backlog working on now to do, and to do on a call together. Like that's what, those are like the four columns. And then after that, I, I mean, yeah, I always onboard them into a Slack channel and then I begin building out their agent or have do we build out their agent, add the agent into the Slack channel configured with the stack first and foremost. But for then next day, today is day two of that customer, right? So first day was yesterday, got the agent set up with the 80/20 stack here, added it to Slack. Today, we're gonna build out the skills, the automations, everything that that customer wants to be able to use that agent for, that's all happening today. And I think like the most powerful thing you can do is, you know, what I did yesterday, which is like the same day your customer wires you money, you turn around, you reward that, you give them something immediately. I'll never forget the feeling of that with myself, of like having a customer, I'm sorry, having a design company that we were working with, we were there a customer, we paid actually five K a month. And in the same day that we paid, we had something back already wired out, like being built on top of. So speed, speed of values, like super important, super super important. And then the process thereafter is just, usually it's like customer dependent, but it's usually like a weekly call, where you're like talking through the use cases that it's being used for, what upgrades to do next, what's working, what's not working. And everything in between is also just like being in the loop, add them to us a group chat with your agent to be able to fix things on the fly, things will break, and you need to be able to fix them quickly. Add observe, oh my God, I forgot to mention another great tool, it's called latitude.so, and they provide like observability into the agent, so you can see when the customer is upset talking to the agent, 'cause something's not working, you're notified, like the semantics around intent behavior. - Settivator analysis. - Yeah, exactly. Latitude is what I use for that. So, you could see all the things that go into this, this is not a $20 a month cloud subscription. This is like, we are creating a co-founder for a lot of these businesses. And so the value, like you said on Twitter, you're like, "Naked, you should be charging more for this." I think you're right, you know? Yeah, so. - So my next, and that's awesome, right? So obviously Dewey is the one kind of spinning up that stack, where I assume you're more involved is, and this is where I kind of got hung up initially, it's like, "Okay, well, you know, this stack is relatively productized, the part that's customer to customer, right?" Assuming you're not just working with like one specific niche every time, is the skills component, right? So it's like, "Okay, we just onboarded a new customer there, a design company, right?" Okay, the skills that we need to build for them might be completely different than the skills that we built for our like insurance customer, for example. So how are you guys approaching the skill building? I think at one point I heard you say where like, you know, on that discovery call, or maybe this is on like the follow up call with the client, you're just kind of having a conversation with them around the use cases, and then you're feeding that granola transcript to Dewey, and Dewey is then taking the transcript and then turning that into skills. Is that, is it, I mean, is it that simple, I guess is what I'm asking? That is exactly the, at the baseline of what's happening. Like, I literally have a granola note from the call with the customer, and they're talking through their workflow, they're talking through their use case, and what I have them do when I onboard them into Slack, they send, like if I open that Slack channel right now, you'll see they sent me, I literally tell them to just context dump, just context dump, all of your assets, past examples of materials that you've created for clients, give me everything, and I feed all of that, I don't even have to feed it, it's already just, you know, it's in Slack, and Dewey has access to Slack, and he can just download all those assets. And so then he sees all of that, and that is the reference material that we then build from to build out the skills and everything that they need. Yeah, so it's literally like, our job today is not to build, build is commoditized, that's it. Building is not the, that's not the skill set. It is actually insane that the job today is to ask the right questions, to have context over that, and to know which tools to piece together for solving these problems. A lot of that is like really similar to your audit. Like the audit is literally telling the customer, these are the tools you should use. They're not even building anything out for them, you're like, here's the tools you should use, and then just having those tools in mind of what to use to solve which problem, handing that off to your agent to assemble it together, to build an agent, that's the world we're living today, I'm living that. So. And it is like surreal. Like it is insane that I will go on a walk and I have customers texting in a group chat and I see their problems being solved on a walk. Like by my agent. By my agent in 2026. I actually thought that like you know, I thought AGI was a few years, right? I don't mean to be hypey, but I just genuinely mean it. Like to me this is AGI. Like this is what it promised, freedom. It promised you know abundance. So I agree. I mean, I'm not disagreeing with you at all. I care about your spot on. And that's why I'm so bullish on the audit or the assessment, whatever you want to call it, as kind of like a lead in offer. So we've done two paid assessments in the past three days and charging more for them than we ever have in the past. And it's funny 'cause all the good assessment is, is just asking the right questions. And it's amazing how much more the client trusts you after you just like basically grill them for an hour and just let them brain dump everything on you. 'Cause in their mind they're like, okay, like this guy just spent an hour asking me 40 different questions. He understands my business now. So who better to build these agents or build us a knowledge base or take whatever the next step is, than the guy that I just brain dumped on for 60 minutes straight. So and you can apply that to you know, you can do, there's so many different flavors of assessments and we've done a few at this point, but that's why I like your model 'cause it goes hand in hand with ours. And we're starting to do some what you're doing and I think vice versa. So I love what you're working on. I love Oregon. Like we said guys, we've got a discount code for Oregon in the description and in the show notes. You've got to use both that link and the code query and you'll get, I think it's what we say, first three days for free and then 20% off your first three months. And again, I can say that with confidence because we are a daily Oregon user. We've built seven agents on the platform so far and it's been amazing. So Nick, thank you so much for the time and this is awesome. Where can people follow you or find you or interact with you? Yeah, it's just my name Nick Vass Leskiew on Twitter, YouTube, what have you. I think the last two letters of my name, the CU are dropped off for YouTube and Twitter. But yeah, you can just type my name and you'll find me. I make content on how to build these agents out, how to deploy them into businesses and obviously co-founder of Oregon. So really just want to walk everyone through how to use the platform and get the most out of it. So yeah, thank you, Corey. It's always a pleasure, man. And we're excited to see what everyone does. Absolutely. Well, we're happy to have you back. I'm sure this will not be the last time. And guys go follow Nick, go check out Oregon and we'll be back soon. Thanks for watching. Cheers, guys.