I got 19,000 new subs, 11.1M views, and 60.4k watch hours in the past 17 days on a previously dead youtube channel and every single video was 100% AI generated (screenshot proof is attached to this post). And I’m going to walk you through step by step how I did it. No gatekeeping lol. Background: This is a client I'm working with who's in the franchise space. He had basically nothing on socials - 300 subs on YouTube with a couple random videos that had a few hundred views. Here's exactly what I did. Step 1: Content Research First thing I did was figure out what's already working super well in his niche. I went and searched terms like "franchise," "top franchises," "how to open a franchise," "franchise mistakes," "best franchises" on Instagram and TikTok. I went through and saved the top ~50 videos I saw with the most likes and views that were talking about franchise related topics. Step 2: Find the Patterns After I had those ~50 videos, I re-watched each one and wrote down the common patterns I saw. I figured out they fell into 4 main types of content that work really well in the franchise space: 1. Comparing franchises to other franchises (McDonald’s vs Burger King etc) 2. Case studies on a specific franchise (how much money a Popeyes makes) 3. listicles (top 3 franchises, worst 3 franchises etc) 4. Comparing franchises vs other opportunities (franchises vs real estate, etc) Each of these types of content also had 1-3 video formats that were popular. Talking directly to the camera, talking back and forth with a clone of yourself, etc Step 3: Plan the Content I took each of those 4 buckets and planned out 16 videos total - 4 per week. I paid really close attention to the specific elements of the videos from my research that did super well. I saw multiple videos pop off talking about "how much it costs to open a Chick-fil-A franchise." Then another one went viral about "how much it costs to open a Wingstop and how much it makes.” So I took the viral pieces from each video, and used those as my guide for coming up with the ideas.

No we are twins

In this video, I show you how I built a self-healing automation system using n8n and Claude Code. Whenever one of my n8n workflows throws an error, it automatically triggers an error workflow that calls Claude Code. Claude then uses its n8n MCP server to audit the broken workflow, understand what went wrong, and fix it, all without any manual intervention. I just get a notification that the error was caught and resolved. The next time the workflow runs, it works perfectly because Claude already patched it. It's like having an AI engineer on call 24/7 to maintain your automations.