How I Made $2,200 on Twitter With This One Strategy transcript

Last week, I wrote an article on Twitter and 13 million people read it by Friday. Yesterday, 4.4 million people read a single quote tweet of mine. Twitter has paid me $1,400 so far last month and I built an AI agent that tells me exactly what my audience wants me to post and when. This is how it works. Before I built this, I had a very specific problem. The days I had the most meetings, the most back-to-back calls, the most fires that I have to put out, those were the days my Twitter engagement was absolute worst.

Because I'd finish the day, sit down, [music] I'd open Twitter to post something and I just stare at a blank screen because I honestly had nothing. I mean, I had no idea what was trending that day. What were people even talking about? What could I react to? I'd been stuck in the day-to-day of running a business and let's be honest, Twitter runs and moves very fast. I mean, really fast. By the time I caught up, the moment was already gone.

So, I wasn't posting or I was posting lazy stuff and it showed. I'd get one piece of content that actually went semi-viral maybe once a month, maybe on a good month. And that was the problem I needed to solve. I want to be very clear about something before I explain this. Sage, my AI agent, does not post on my behalf, ever. That is something I will never let it do. The most important reason people follow you is because of you, your perspective, your voice, your takes.

The second you let AI post for you, you lose that. So, whatever you do, don't do it. Don't let AI post for you. What I designed Sage to do is for it to solve my specific problem as a content creator, which for me is content ideation. And this is what I want you to do on your side. Ask yourself, what is your biggest bottleneck as a content creator? For me, it was always the the blank screen. I mean, I had plenty to say.

I just never knew what the right thing to say right now was and that's what I created and made my AI agent automate for me. So, without further ado, here's the actual workflow Sage runs every single day. Step one is the brain update, which Sage runs every day at 6:45 a.m. Before I wake up, Sage has already analyzed my last several months of tweets. He breaks them into the top 20% performers and the bottom 20%. He then studies what topics I cover, what copy structures I use when I'm at my best with the top 20%.

He's basically reverse engineering my own voice and figuring out what version of me hits the hardest. Then, he does the same for my competitors. He looks at what they posted in the last 24 hours. He scans Twitter for trends that are in my niche and he looks for anything trending that I would have a natural right to talk about. Then, step two would be the morning brief that Sage runs every day at 7:20 a.m. >> [snorts] >> I get a Telegram ping with reactive opportunities, stuff that just dropped that I can respond to.

And this is huge because reactive content is some of the most powerful content on Twitter. >> [snorts] >> And you can only do it if you're really fast. And Sage is always fast. So, this is the brain update, for example, for today. You can see him saying step zero, tweet sync, step one, competitor scrape, uh which he skipped today because I'm out of credits with Grok. I need to uh refill that. And then he goes through here.

You can see step two, rejection analysis. Then step three is voice examples. Step four is edit history analysis and then step five is performance analysis. And that is his morning brief. Then with the reactive opportunities, let me also show you that real quick. Here you can see some of his uh pings, for example, from today. Him pinging me a reactive draft. Zara tweeted 1 hour ago and along with an angle, a tweet that I can create, an idea that I can uh post.

And what's great about this is literally, I could be working all day. I just open my chat with my AI agent, Sage, and he has the reactive drafts ready, the things that he thinks are worthwhile for me to react and talk about. Then, step three is the drafts and ideas that Sage gives me. Three times a day, Sage sends me five content ideas. So, that's 15 in total. These are ideas and angles, topics with suggested framing.

He gives me raw material and ideas that I then can take, if I like them, and write it in my own voice or spin up my own angle. Here, let me show you real quick. There you go. That's an idea he just gave me, a draft he just gave me. Someone stole my face for their thumbnail and I went viral calling it out, blah blah blah, and and I can pick out whatever I like, whatever I dislike, and move on like that. This is the part most people miss.

The system only gets good because of the feedback loop. Every draft that Sage gives me has two buttons that you can see right here, in green and in red, approve or reject. If I approve, he logs it. He scores that topic, that tone, that angle as something that I liked. He'll give me more content like that in that direction in the future. If I press reject, he asks me why, to state a reason. Not just that I rejected it, he's asking why.

Maybe the tweet topic is fine, but maybe I don't like the framing. Maybe I don't like the topic at all because I think no one will care. Maybe I think the topic sounds too salesy. Whatever the reason, I tell him and he learns. And over time, he gets sharper. He stops pitching me the same dead angles every single time, repeatedly. And the first few weeks, maybe 30% of his ideas I actually found useful. Now, it's closer to 70 to 80% because he's had 8 weeks of learning my taste, 8 weeks of that feedback loop, me rejecting, approving, teaching him as he goes.

So, let me show you what this actually produced. First up, first case study. We have the reCAPTCHA tweet, which is sitting at 13 million impressions. Says, "You've been training Google's AI for 15 years. You had no idea." That's it. One article. 13 million people read it. Sage simply spotted a moment. The reCAPTCHA angle was trending in some AI circles and he knew I had an audience that cares about AI. He flagged it.

I wrote my version of the take and it went everywhere. The next day, I followed up with uh DNA test angle. The same format, same idea. This one had 2 million impressions. He caught the moment from the first one and said, "Oh, this works. Let's do it again, but with a different story this time." Oh, this is one of my favorite examples of how this system compounds. I made a post referencing Karpathy's auto researcher, this post right here.

Basically, bunch of AI researcher stuff. And I mentioned that I was using it to train my trading bot. People started commenting on it and asking if I can make a full video on this. Sage spotted those comments. He pinged me on Telegram saying something like, "People are asking for a video on this." Then, and this is the part that I love, he sent that signal to Nova, my YouTube agent. Nova picked it up, came back to me with a video concept.

We filmed it, posted it, and that video did well, very well, both on Twitter and on YouTube. Literally, one of my top videos to date right now. And that's the part that you just have to love. Two AI agents talking to each other. One that caught the man signal and one that built the content. I just had to show up and film. All right. I want to say this one more time because I mean it. Do not let AI post on your behalf.

I know some of you are watching this and thinking, "I'll just let it do everything for me." Please, please, please, don't. The reason your audience follow you is because of you, your point of view, your voice. Sage gives me the ideas. I still write every single tweet I post. I still decide what goes up and what doesn't. Think about your own specific personal bottleneck. Maybe your problem isn't ideation like it is for me.

Maybe you have too many ideas and you need help prioritizing them. Maybe you need help with the research part of content creation. Maybe you need someone to watch your analytics and tell you what's working and what isn't. Whatever your personal weakness is as a creator, build around it. Automate around that. Don't just copy my setup blindly. Know your weakness and build for it. Also, another reason not to let AI post for you is it will 1,000% sound like AI.

Trust me, it shows. It's all I read on Twitter. The reason this works, it's not the AI, it's the system. Sage doesn't make me a better writer. He makes sure that I show up. He makes sure that even on my worst, most chaotic days, the days where I didn't sleep, right, and I had 12 back-to-back calls and my brain is fried, there's still something on my desk that says, "Hey, here's what to post today and why." Before this, I'd get one viral tweet a month, maybe.

Now, it's every week, sometimes every day. $1,400 in Twitter monetization last month so far. And 13 million impressions on a single tweet. In fact, 38 million impressions in total in the last 2 weeks alone. A YouTube video that came from my agent noticing a common thread. All that because I stopped trying to do everything in my head and I let AI build the system for me. And just before uploading this video, let me show you what just happened.

The last 2 weeks' creator payout just came out and it came down to $2,231. That's legit all thanks to Sage. So, I don't know what more proof you need, but there you go. If you want to build something similar to what I have going on, I'm going to upload my repo on GitHub. Literally, you can open that link, send it to your open claw, your AI agent, whatever AI agent you have, and ask it to install the same exact system for you.

Completely free of charge. If this was useful, subscribe. I post every single week about building AI systems that actually do something for you. Oh, and if you are a founder or building a company and are looking for content systems that scale, that generates you leads, that make you money, that run automatically, please reach out to us at founderfunnel.com. I got you covered. That being said, see you guys next week.