New Agent OS Just Changed AI Forever!

I'm about to show you how to give your AI agent superpowers, agentic operating systems. Today I'm answering the biggest questions about them. And by the end of this video, you'll know how to run your own AI team that works for you day and night. You'll see the amazing systems people are building with this right now.

You'll see how to make every AI agent share one brain so nothing gets lost and you'll learn how to run it all completely for free. And there's one setup that makes your AI agents run every day without you touching a thing. I'll reveal it later. So stick with me to the end and your AI will be a hundred times more powerful, no matter who or where you are in your AI journey.

Today we're going to be answering some of the latest questions about agentic operating systems. Now, if you're not familiar with what this is, this is basically a system where you can have all of your agentic harnesses, all of your agents, all of your systems plugged into one place where everything is just one click away. So for example, if we have a look at this system here, this is Hermes Astros. If we scroll down, we can see that it comes up with customized content ideas based on our competitors and our keywords monitoring everything.

And then what it also does is it makes sure that it gives us unique content angles and titles based on that. And we can then just plug this into our SEO system, our video agent system, our notebook for example. And we've also got a powerful memory system. Everything learns from so that it has one system where it can plug all its memories, all of its learnings inside one places, get automatically updated.

And then every time something new comes out, for example, like Jayco just dropped super powerful system where it can code way, way faster. We've got a terminal, we've got a workspace, you've got the history all plugged into one place where we can get the most out of this stuff, right? And look at this stuff, right? You can create amazing websites, landing pages, whatever you want using this system.

And it's all saved in one place ready to come back to whenever you need to. So what I do each day is I answer the latest questions inside the AI Profit Boardroom. So if you have questions for me, feel free to post inside the community link in the comments and description, or just go to the a8born.com. I'm going to get straight into some of the latest questions, what people are building with this and what's working right now.

So let's get straight into this. So you can see here, Mohammed has actually built out an amazing SEO machine based on our system. So what we actually do is we give people inside the AI Profit Boardroom, our agentic operating system. It's like a zip file you can get and install and then people create and customize that and build out their systems from there.

And also they get the latest updates that we build in. So if you have a look at this system that Mohammed has shared, like he's created an awesome system, um, where he's got Vera HQ and then it's got a full SEO machine. So for example, as AI workspace, a command universe operations, automation builder. So you can actually build new automations based on this system over here.

Pretty amazing stuff. And he shared what he's learned. So for example, he was saying what's working right now is he has a mission control that gives him one live view of his SEO and AI workspace that shows all the departments of specialist agents, a command universe that maps out every department agent service automation, life system in one place. So shout out to Mohammed.

Like this is absolutely amazing. Well done and taking action and actually implementing it. I also really liked the fact that he's basically taken the system, made its own, uh, improve the UI or change UI how he wants to. And he's also got it in responsive mode too, which is absolutely amazing.

That's one of the best things about the AI Profit Boardroom is like people share their journey. They share what they're creating. I actually think the command universe is something we could build in. So you can see here, it's got the different operating sectors here and the different agents and tools that he's using inside this system.

Let's have a look at the next question. So Lucas just posted this, by the way, look how fast we reply. Lucas literally just posted this and we're getting back to him now. So he says, I have a technical question.

I have many, many chats like 20 running simultaneously on Claude. The problem I face and would like to solve is that they would know and interact what happens in one or the other, call it live single, um, at least a daily between the chats in a single project, as it's a pain for every new or old chat to explain what I did in other chats. How do you solve this issue? So one of the first things that you can do is just have them all working from the same project.

So if you don't already, you can have them all working from the same place. So if we create a new conversation inside Claude, we can just make sure that working inside the same project folder, if you're not already doing that, definitely recommend it. Next up, what you could actually do is just have them put inside one place, a daily log of everything they've worked on. So for example, I use Obsidian for this inside our agentic operating system and our agents, each time we interact with them, they log what they've done that day with the date.

So you can see, for example, the Friday, the 7th of August, here's what we're doing, right? And so when we're doing that, the great thing about that is that our agents understand what they've been working on previously across multiple conversations. And that doesn't matter which agent you use or which API you use. You can just have them all sync up together.

Finally, you could actually just fork the conversation. Each time you start a new conversation and that way, when you fork the conversation is picking up where the other one left off and just simplify the amount of conversations you have. So for me, for example, I usually have like maximum five different conversations per day with Claude or different conversation threads. And then they're all just forks of each other, which means that they still have context of what they've been working on recently.

So that's three different solutions right there. Make sure they're working inside the same project folder, make sure they're logging a daily note into the Obsidian folder and make sure that you fork the conversation. Each time you set up a new conversation and minimize this and fly the amount of conversations you have with Claude as well. Great question there, Luca.

We've got a question from Brian about NA10 workflow help. So for example, like this one, anything you see us post, we just have it inside the NA10 section. So every NA10 workflow is inside the NA10 section. So you can find it inside the classroom for the NA10 folder.

We've got another question from Leon, which is any tips on getting Hermes agent to make better images? And what Leon says is I am struggling with Hermes agent. We've opened our codecs off connected to create decent images. I have the image generation tool enabled, but half the time I get these basic diagrams or skills and that sort of thing.

So there's two things that I use to get better images from agents, whether that's Claude or whether that's Hermes or whatever you're using. So number one is just a templated prompt that I can use where I give it the image and I ask it to iterate the image based on what I know works. And GPT image V2 is the best generation tool. So if you're definitely using that, I don't think you have any issues there.

So a thumbnail prompt that we often use is this one. Whatever agent we use to create the image based on that prompt above is very, very repeatable. So we will get the same outputs no matter which agent we use, as long as we're using the same API for generating images, which is normally chat GPT image V2. The other option that you have is that you create a custom skill with Hermes and each time you iterate and give it feedback based on what you want and what you prefer.

So for example, if you say, okay, next time make the colors more contrasting, make sure that it updates and saves the skill. So step number one is making sure that you get Hermes to create a skill for generating images. And step number two is to make sure that when you're generating images, it updates the skill every time you give it feedback so that it gets better and better over time. The third option that we actually use is our studio inside the agent OS.

So we have an image generation studio inside the agent OS and we do the same with videos and images for Hermes and then it can generate beautiful images and we just type in the prompt that we want and it looks pretty nice each time. So that's another option. We typically generate the images of either minimax or grok plugged in to the system. And those seem to be the best that work with Hermes when it comes to like an image generation studio, but GPT image two could work as well.

So three options there. Number one is the, you give it a templated prompt. Option number two is that you create a custom skill and update it every time. And option number three is that you use the image studio or the thumbnail studio, which we've got over here inside the agentic OS.

And that can create beautiful images like you can see right here. We've got another question from Warren, which is, do I need open core if I want to use Hermes? Absolutely not. You definitely don't need open core.

In fact, I haven't touched it for months simply because Hermes is so good and Hermes basically replaces the fundamental need for open core. So if you have Hermes running, you're happy with it, then you definitely do not need to install open core. In fact, open core can cost you a lot of time in terms of repairing it, fixing it, making sure it works. And from my own tests over the last few months, particularly about a month or two after it launched, it just stopped working and it broke a lot.

And you actually don't see many people talking about it. So I think that's why Hermes is, is goated right now. We've got one from Timo who says a question regarding the proactivity and autonomy of Hermes agents and says, I've been using a Hermes system for some time to handle various tasks for a new project. The focus is marketing, social media SEO optimization, blah, blah, blah.

But he says, nevertheless, compared to Claude, I somehow expected more autonomy and independence from Hermes. Find it very difficult for Hermes to handle tasks in a truly autonomous manner. For example, if I want 50 contacts research and qualitatively assess each day, it doesn't really work well or reliably. There is a specific skill for this task.

And as far as I see, it runs as a cron job. I'd be interested to know which part of that process is not working on a daily basis. So for example, if we set up a scheduled task with Hermes Oracle inside the agent OS, it tends to work pretty nicely and smoothly. It just runs like clockwork every single day.

So you get updated on a daily basis every 24 hours and we get new outputs from it. It's not that autonomous in terms of being proactive, but it is autonomous in terms of just making sure that it runs that system each day on a schedule without us. So if you are struggling to get Hermes to run a schedule, I'll be interested to know why. If you just want it to be more autonomous and move forward without you, I would actually give it a scheduled task and bear in mind, this is only if you want to.

So only do this if you're comfortable. But what you could do is you could give it a task each day, whereas I on a schedule every 24 hours, improve all of my automations to schedule tasks with you so that you run more autonomously and proactively about me. You could ask it to do that. Now, I don't recommend doing that because then it could go off the rails without you.

But if you wanted it to be more autonomous, if you wanted it to be more proactive, that's something that you could do. Got a question here from Robert, which is how do you update your agentic operating system? So we're always adding new things, new, interesting features. Number one, don't feel obliged to update.

Like if you, if you feel like you're happy of where it's currently at, you don't need to update, but if you do want to update, you can just get the latest version with the zip file inside the classroom. You don't need to reinstall it or set it all up again. Literally there's just an update.md file in the new zip file. So each time you want to update it, you can use that.

And we got a question from the legend himself, Christian, who's a member with us for months inside the community. So shout out to Christian. And so here's the question. Those of you running isolated products, but wanting centralized orchestration, do you go to MCP server per client or fully separate control plane?

Not 100% sure what you mean, but basically, if you're wondering, okay, do we have like separate MCPs for each client? Or do we use like the same MCP for each client? Typically I would just use a separate MCP depending on the tool. So for example, if you need to operate a client's NA10, that's going to be totally different from the MCP that you set up for another client's NA10.

And so like, there's going to be different MCPs for each client. So that's the way that I would look at it is you will have different MCP servers or the same MCP server, but a different connection depending on the client. So like to access one client's MCP server on NA10 is going to be totally different from another client's MCP server because they're like different end points and different access points. We've got one more question, which is, you know, once you've logged into the AI profit boardroom, is there like a implementation guide on the agent OS?

So yeah, so if you go to the classroom here, and then you go to the agent OS section, we have a full guide here for implementation, but also we have a full video tutorial here as well. So if you want to go really in depth, there's a one hour tutorial. If you just want the quick guide, here it is, right? And if you have any questions you could post inside the community as well.

So, you know, however you want the help, we've, we've basically set up. So it's as easy as possible to understand this stuff and get the most out of it. And then this is all available inside our AI profit boardroom community. So that's all of the questions from today.

If you want me to answer your questions with a video tutorial and help you, you can post a question inside the community here. If you want to get our agent OS system, it's inside the classroom, go to the new daily updates and you'll find over here. And then we add new daily guides and useful updates based on what just drops inside the classroom as well. You can get access to all of our best trainings, all of our best systems and everything else.

And inside the calendar, you can jump on weekly coaching calls, ask questions, get help and support in real time inside the map. You can meet people in your local area who are building out with AI agents like you. And this is all available inside the AI profit boardroom link in the comments, description, or go to the AIprofitboardroom.com. Some people as well, like they say, you know, it doesn't run in an agent operating system, use up a lot of tokens.

We actually have free models and free AI coders inside here. So for example, we have the free AI coder here and we also have the local coding engine over here. Both of them, both of them are free. And that means you can run this without using up tokens or without costs.

Also some people say, well, isn't this like super technical and difficult to set up, but actually it's just a quick zip file that your agent sets up for you. So if you're wondering, okay, can anyone else get the most out of this? Well, you've already seen examples today, but also we've got over 207 pages of wins of testimonials of people crushing it with this stuff. So if you want to join us, feel free to check it out inside the AI profit boardroom.

And thank you for watching. See you inside there.