I'm an operator, not a reporter. I run several small businesses and I build the automations they actually run on — client intake, reporting, the follow-up that used to eat a Tuesday.
That vantage point is the whole point. The demo always works. What decides whether a tool sticks is the thing that breaks in week three: the seat limit, the export format, how the admin console handles someone leaving. Those show up later, and they are what you actually live with.
So I publish what I build. My MCP servers and n8n connectors are public, my SquadScore ratings show their weights and version numbers, and I've mapped 4,067 of the real questions people search about AI and automation — including when each one first appeared, which is the part nobody else can tell you.
What I'm doing now
- Writing The AI Five — five runnable prompts every weekday, then what actually happened in AI, each item paired with something you can use.
- Building Automation Squad: tool ratings for people who have to decide, and implementation for people who have to build.
- Publishing MCP servers, n8n connectors and agent skills at AutomationSquadOS.