Most people and companies are still getting their value from AI just by getting good at ChatGPT, Claude, Copilot or Gemini. (That’s great, and that’s exactly where most people should be.)
But out at the edges, I’m watching the emergence of something else starting to show up. Instead of just using AI as a co-intelligence tool, people are now asking it to build bespoke tools and software that fit how their teams actually work.
Things that used to require big software development projects and huge budgets can now be created in hours, by one person, for cents or a few dollars.
Not demos or experiments, but powerful internal tools people actually use.
The interesting part is you don’t need everyone doing this. You probably only need one or two people who know the business well enough to turn workflows and SOPs into usable tools and interfaces.
I’m calling this the rise of the “AI Operator”.
This week’s video essay walks through what that looks like in practice, and why it matters so much more than just better prompting into AI chatbots.
