A few weeks ago, I started running 2-3 Claude sessions to work on client deliverables simultaneously.
While I was eating lunch.
Watching anime.
And they're actually building stuff... They're all creating lesson plan drafts for a client's course project, generating structured output that I can review and 'quality check.'
You see, I've been building this system for a few months.
I started with a lesson plan builder using Claude Skills. Then, I built an AI client workspace so it's easier for me to juggle different clients and projects. Then I built a few more "skills" to help with other areas in my fulfillment workflow.
And yes, I understood what I was building. Conceptually.
But at that moment, I finally felt it.
And here's the honest part: the drafts weren't finished product. They still needed my eyes. Review, editing, rearranging until the lessons are cohesive and streamlined...that work didn't disappear.
But AI did most of the heavylifting.
My actual work now is: give Claude better context each session. Improve the feedback so the next draft is sharper. Then make the system better, so I get better quality outputs over time.
Having 2-3 AI agents working in parallel and actually getting things done? That's a whole different ballgame.
I'm not just talking about speed. It's about what the AI system can do while you're somewhere else entirely.
Here's the thing though, it only happened because of the unglamorous infrastructure work I did beforehand. I've built and rebuilt different skills. I've had multiple back and forth discussions with my AI to tighten the output.
And guess what, it's still a work in progress.
But I'm super excited by what I can still unlock with AI.
If you're at all curious by what I've built, check out my new video on AI delegation. I shared my ABC framework for building Claude Skills and how you can use Cowork to delegate tasks.
I know Claude Code can be scary for some people so in my new video, I focused on Cowork, which has a more user friendly interface.