I Used to Add Tools. Now I Add Agents. Here's the Difference.

Hinzugefügt: 2026-05-21

[00:00:01]Okay, we're live. It's James speaking.

[00:00:03]Uh, let's have some fun. Um, I haven't been on YouTube in a while. Uh, and in fact, I think I did one little kind of live stream last week or the week before, but really I haven't been posting any content on YouTube. And if you've been a subscriber to the channel, maybe you've seen before, you've wondered, "What the heck's going on, James? What have you been doing?" Well, I want to share with you something that happened over the last couple weeks and how I think it's well, I know how it's going to change my business and I think that it's going to be changing yours as well. So, uh, I want to go over what happened during the last three weeks and I just want to explain to you why I stopped using tools over those last three weeks and I think you should stop looking at all of the tools that are available to you and you get bombarded with each and every day. Uh, and I'm talking about today, for example, because what's going to happen, it's uh, it's the Google IO conference is today, May 19th. We're going to see a thousand YouTube videos about all the great tools from Google. Every day, there's some new LLM coming out. There's some new tools that are working on it. And all of these are supposed to be able to do all sorts of work for us. And over the last couple weeks, um, I did a whole bunch of work, but I was in Portugal. uh I wasn't here at the office and in fact I didn't use any tools. What I did use is uh my AI operating system which is I'm calling campus AI and the reason it's a campus is I am still uh in the market of helping 45 plus experts who have this expertise who just want to be able to share it create their own personal brand and their platform for being able to whether it be teach it or coach it or sell it but just be able to spend time doing what they want to do and let agents do the rest. And in fact, that's exactly what I did over the last couple weeks. So, what I'm going to do is I'm just going to show you uh basically what happened. We'll go through it. Uh and then I'll give you a couple options at the end uh as to how you might want to consider implementing this. And I want to say up front, I'm not 100% anti-, you know, tools or anything like that. I think there's still, you know, some tools that uh should probably be used or can be used. and I still use the odd tool. But over the last two and a half, 3 weeks, um I've done a whole bunch of stuff for at least my agent operating system has without me touching a tool.

[00:02:35]So, just to give you an idea, I've been away uh on a holiday. I was in a completely different continent and I actually have since then during that holiday, I've got 14 salailable products that are now for sale that weren't before I left. And that includes six actual agent plugins that fit into the campus AI operating system. And I'm talking about being able to have finished products created for me. Those are available. They're ready for sale right now. I did a whole bunch of uh additional features and updates to existing products and our services that we're doing. I had a whole bundle of skills added to a library. And in fact, in total, I think there was 74 unique skills that got created while I wasn't here. It's something that uh Claude was able to create. And during that time, I had nine recurring tasks running in the background being able to do all of it.

[00:03:34]So, here's a quick kind of short view of it. So, SIP, six new agent plugins with 74 new AI employees added to the library. Uh, I launched the campus AIOS as on uh, pardon me on GitHub. So it's open source. You can go and download the uh the introductory level. Uh so I'll put the link to those of course as well.

[00:03:56]Uh I ran a live sprint. So we did a three-hour session where the people in the session actually created their own AI operating system, got it set up, configured, and now they can do exactly the same uh thing that I'm doing. uh I have a whole complete system I set up for every plug-in that I create in the future shares the onboarding and the agent operating system infrastructure is the same on every plugin. Um and then the other thing is is this allows me to duplicate the process or at least have Claude duplicate the process for any idea that I come up came up with. Um, the other thing that happened, number five, is I I diagnosed a huge problem on one of the goals that I had, one of the objectives that I had while I was gone.

[00:04:42]Uh, I thought it was something else, but when I ran the analysis or had Claude run the analysis, it identified where the breakdown was. Again, that was done without me. uh sales page playbook and also the sales pages that I had previously for all of the products, all of the services. They've all been updated using a brand new playbook or a brand new framework to make them more effective. And here's the one that I think is the most exciting in the AI operating system. The one thing that I did add to it now is it has a self authoring skills loop. And what this means is it automatically goes and creates skills and saves it to their own skills, its own skills library if there's one that can be applied to a task that is being worked. And here's just some of the different uh products that I built. So, uh I have a live session session agent. And if you haven't seen uh the way that this particular uh setup works, let me just pull this open.

[00:05:43]U let me just open one here. We'll do uh I'll open this one up here. So, the Campus AI operating system, what it is is I actually have the blueprint for how it works. There's an org chart for it.

[00:05:56]And it's really important that any of these agents or the agent operating system has a set level of memory almost like a filing cabinet that everything works on it. And if you have that solid foundation, all of the agents, skills, any work that you do, Claude, is always based on that bedrock, uh, that layer of memory that is consistent across every task that it does. So, you know, these are some of the pieces that are in place here. And one of the ones that I got really excited about is making it proactive. And that was the last piece there because now when it's working on that bedrock, any of these agents that I use are using that foundation to make sure that the work that it does follows my ideal customer profile, is working towards my goals, knows and understands all of my settings, that things are set up. So I have one for a live session agent. This is a department or digital employees that work in my AI operating system across departments. And this is designed specifically any time that I'm doing any kind of live presentation like this one as an example that all of the pages that actually get created for me automatically. Uh and in fact, this page that we're looking at, it was created for me based by that particular agent.

[00:07:16]So that one's available. It's live. I've got an outbound sales agent. This is a one that saves me hours of time because now I can just say I want 15, five, 50 qualified leads a day. These are automatically researched, scored uh and all of the outbound marketing or reachout marketing is created for whether those be direct mail or even uh uh messages, direct messages. I have a content scout now that that agent and all of those employees goes and find me all the content that I need for all of the products that I'm creating, but also all of the marketing that I'm doing automatically. I've got one that is class prep, so specifically if we're doing coaching or live sprints or teaching is done. I've also got the AI admin tool that is basically done and this fine-tunes all of the work that we're doing. That one's free. You can actually do it. And I have my campus ambassador which is now available as a product. So if you have customers and you want an AI agent to take care of customers, clients, members, and be proactively involved on a onetoone personalized lab uh level, these are all up and running and there's 74 individual AI employees running in these six products and these are all standalone products that run by themselves. Uh and just to give you a couple examples of different things uh that were done, this is the one that I'm most proud of. We've got the notebook staging pipeline. The one thing that I did do when I say I use tools is I still use Notebook from Google and I actually have Claude hooked up to work with notebook. That was one of the things uh that we did. Uh two of the big ones here. This is the first one, the skills authoring loop one. This one actually has built into the operating system. It keeps track of all of the work that's being done. And what it does, it says, "Hey, these are some ideas of tasks that you're doing. And if we added a skill that does this, it would remove you having to be involved.

[00:09:11]Uh, so this is up and running now in the agent OS that's actually running. And if it's done on a repeated basis, the default is three times, it'll proactively say, I'm going to create a skill for it. What do you think? Here's what I have in mind. I also have proactive outcome tracking. What that means is if I have goals on how many sales I have, how many sales I want. It actually tracks to find out am I drifting or am I away from my goals.

[00:09:36]This one is now set and it knows when I send emails out, who's clicked on what, who's purchased what, and it knows what's going on in the cart and what's going in in the CRM. And it is able to go, hey, you're not getting the results.

[00:09:50]Let's go and do something about it. I've also got the repo prospector, which is a really cool one. Because even if you're not technical, one of the things that I'm really learning about is that there is a wealth of information with agents and skills in something called GitHub, which is the one the programmers use.

[00:10:08]But I'm not technical and the majority of these are all like programming or software engineering uh plugins or services that are available on GitHub.

[00:10:17]Uh I just wanted the ones that are marketing related. So what I did is I actually created a plugin that goes, this one's internal right now, but what it does is it searches GitHub for specific softwares and/or agents that I can bundle and unbundle to improve my own or actually go and create them. So this is something that is done for me now and in fact I have the ability to actually create the plugins and all the marketing material. So this is a standalone done for me product pipeline that is available using other content that is researched and scored and put into combinations where the licensing is involved. Uh as well I also added a couple really neat things that I picked up from some other people. Uh one of the ones here is I've got actually some admin tools about making sure that I can now fact check. Remember I was talking about that memory level. It'll actually fact check everything that gets done.

[00:11:15]I've also got ones where it will actually look and pick um the best situation or it's able to make sure that what I'm doing is on compass with my ideal uh customer profile. So that happens automatically and I've got also one so sharpen which is a prompt enhancer.

[00:11:33]Sometimes it can get a little bit long giving the prompts or asking it what you want to do. Uh and in my case when I say asking it what I'm want to do I'm talking about in the AI operating system I only talk to one person who is Dean who runs my business. He's a chief orchestration officer. I just want to make sure I'm able to ask him to get finished outcomes done or call one of these other uh departments that have built together for example these plugins. Uh so I have this one that actually clarifies what it is that I'm saying so that he will understand it.

[00:12:08]And there's a whole bunch of additional ones uh that are happening there. And of course I was awake. I wasn't here. What was running in the background? Morning intelligence report. Every morning at 7 I think it was at 7. Yeah 7 a.m. for example. That particular one was sending me an iMessage with stuff that was being done. daily briefing stuff that's going on in my community, my revenue, any of the AI news, YouTube outliers, competitor moves, that was done without me. Campus ambassador twice a day. I have my campus ambassador, which is also available for sale. Now, I have it in my community. If you have a membership site, if you have anything with customers and you have a a group of customers in a CRM, for example, this one will actually go and be proactively involved on a one-to-one basis based on what is the person doing, what are they not doing, do they have questions, being able to reach out, send message to it, making sure that they're following the instructions or steps that may be laid out in a course. All done. Uh, I've also got the weekly partnership scan. So, this is actually going out and finding opportunities for me to do podcasts or YouTube videos or joint venture with people all running in the background.

[00:13:22]Exactly the same as having uh an outbound email team. And of course, I've also got all of this being ingested directly in to the wiki, which is Andre Kaparthy's memory level from all of the or archive of all of the content that I've been doing and outputting. And it's actually allowing all of that content to be ingested so that anything that happened, even if I wasn't here, is automatically put into this huge library of my outputs. And it allows Claude to pull or extract information from that entire uh wiki or that index and file.

[00:14:02]Also, the sales and copy updates, all of this pieces here. Now the thing that's interesting is I you know I I didn't use any tools for this. Uh I didn't go and find a particular tool for doing infographics.

[00:14:17]I didn't do anything to get a tool for creating web pages. I didn't do anything for a tool for writing sales copy. I didn't even use uh any of the programs like Notebook LM. I didn't log into any of those. Not a single one of these. But I'm really excited because I was able to get all of this done while I was on holidays. And I'm not saying just on holidays to be uh you know to kind of I don't want to say word to brag, but I did this as a project for myself because I want to be able to focus on making YouTube videos and being able to help people and talk to people. But this is stuff you still have to get done in your business and you have to make a choice.

[00:14:56]Do I want to be the one that is doing the work and using the tools, which is what I was doing previously to this, or do I want to have my own AI operating system with all of these agents and teams and ability to work together from a shared memory and a shared foundation?

[00:15:16]Do I want that part of my agentic business in its own operating system to go and do the work for me? And this has worked over the last 3 months. Is it perfect? No, it's not perfect. Is it iterative? Yeah, absolutely. You get better at each time. Some stuff works the way you want. Some stuff doesn't work the the way that or where or you have to make changes so it works the way you want. But once you have this, and I was talking about memory levels, there's kind of three memory levels that I get discussed when we do these agent sprints. Um, you know, there's the foundational stuff about your ideal customer. That one's in the filing cabinet. We have a living memory, which is a stuff that Claude understands from a memory.md file. And then we've got the wiki, which is that complete index and archive of all of the outputs. So that memory level and that recording of all of the work that you're doing allows Claude to get better each and every time. So what I did over these three weeks, even though maybe not all of it worked, a lot of almost all of it worked, but not all of it. The thing is that as I continue to use it, it's going to get better and better and involve me less. And that means that I'm not the bottleneck from having to go and use a tool. The only job and the one that I'm struggling with right now is I have to figure out and just make sure that I'm actually telling Claude in this case what I want done, not asking how to do it. I can plan. I can even go and say I want this done and it's figuring out based on that AI operating system and all the memory that it now has available to it. It's figuring out how to do things. And the cool thing is I just started it out uh in the last day or so.

[00:17:01]It has the ability to create the skills and the agents that it needs to actually figure out how to do things that I haven't asked it or given it instructions on how to do. So this is a learning system now and it's available for you. Now if you're just getting started and you're thinking hey I want to play around with it and you are comfortable with using something like GitHub uh I'll put the link below so that you have it. There's just a simple command and it will actually install the AI operating system for you based on the campus and having a privately branded campus and an educationbased business.

[00:17:38]This isn't for building software. This is about people who are marketing their expertise and want to be able to create content and market their content and sell their content and be able to spend time having fun as opposed to doing all of the other work in your business. So, it will set that up for you. Uh, and it's available in GitHub for free.

[00:17:59]Again, I'll put the actual uh details for you. If you are serious about it, what you can do is you can go and basically, let me just pick it out here so we're in the same place. I'll put the sales page in as well. I do have a full version of it that is completely set up and we'll pull that one up so that we have it.

[00:18:20]And I just want to make sure I have the link here for you.

[00:18:30]There we go. So, this one is available to you. It's $97 and this one includes 20 different AI employees that are ready to go in four different departments.

[00:18:41]It's just an install and it's completely done for you, ready to use. Uh, and again, I'll put the links to these ones so that you have them available. Just think of it, no more tools, no more downloading or trying to figure out or pay for a subscription on something. No more you blocking or being the bottleneck in the work that you do.

[00:19:01]Start building out your own AI operating system that is unique to you, your outputs, your expertise, your own brand, and start having fun and taking advantage of it. So, hope you enjoy this little session. It's jamesraining.io.

[00:19:15]Like and subscribe to the channel, of course. And if you haven't joined the community, it's free. It's right there.

[00:19:19]There's the link that you can get started with. uh and get in there and ask some questions about this next 2026 blast because today I'm really excited to see what's going to happen because an AIOS doesn't matter if you're using chatpt codeex or claude code or co-work or hopefully today the Gemini desktop app.

[00:19:47]this will run on all of them and you're going to be in a unique position that you're now not going to have to worry about tools or which is the best particular LLM. You've now got your business running on an operating system.

[00:20:00]So, hope you enjoyed it. Take care.

[00:20:01]We'll have some fun. I'll be back tomorrow with another video because I have fun making these and I'm not on holidays anymore. So, take care and expect the best.