Having more ideas is a wonderful problem until every idea becomes another thing you haven’t finished.
That was the problem I wanted my AI council to help with. I didn’t need another enthusiastic response to whatever had caught my attention that afternoon. I needed a repeatable way to ask: does this deserve a place in my life right now?
A board built around my actual weaknesses
My personal setup uses different voices to challenge different parts of a decision. I also made a council about pacing, chaired by an imagined version of me at fifty. That changes the conversation. An exciting project looks different when the person asking questions cares about the life I’m building over several years.
The useful part is the job assigned to each seat. Who challenges the scope? Who asks what gets displaced? Who notices when preparing has become a comfortable substitute for doing?
These are simulated perspectives inside an AI conversation. They aren’t the real people they might be named after, and five voices aren’t five independent expert opinions. The value is in the questions they help me remember to ask.
What happened when I used it
The council didn’t invent this channel. It helped me see how a channel could give my curiosity somewhere useful to go: try something, explain what happened, share the part someone else could borrow.
Then it caught me doing the predictable thing. I was planning the channel instead of recording it. More preparation wasn’t the missing ingredient. Making the first episode was.
That’s the test I care about. Did the conversation lead to a decision and an action, or just a nicer-looking plan?
Build a small version of your own
The starter prompt begins with an interview. It asks about your goal, recurring distractions, unfinished ideas, energy, and limits. Then it proposes four seats based on your answers and helps you choose the voices.
Four seats are enough for this starter version; you don’t need to reproduce my whole setup. Bring it one real decision. Ask what to do now, what to park, and what to drop. The prompt also limits your active priorities to three: a useful constraint, not a rule that knows your life better than you do.
I demonstrate a browser-chat version in the episode. My more elaborate personal system uses Claude Code. Start with the small prompt in a chat tool you already use before deciding whether you need anything more.
If you finish with one fewer active project and one clearer next step, it has done something useful.
Once you’ve tried that, the six follow-up prompts cover the weekly run, new ideas, and a check on the pace. The editable file template is there if you want to keep the setup outside a chat.