Case study
HIVE: One Group Chat with Your AI Team
The problem
If you use more than one AI assistant for serious work, you know the drill. You lose track of which one discussed what. You copy threads back and forth for a second opinion. You ask a simple question and get three long answers when one good one would do.
My role
Product owner and designer. I wrote the product requirements document (v1.0, September 2026), with several AI assistants as contributors.
The process
- It started as an experiment. I was already running my own team of specialized AI agents with defined roles, handoffs and approval gates. HIVE grew out of what worked, and what was annoying.
- One-sentence definition: HIVE is one continuing group chat with the AI assistants you choose. It brings in the right assistant, leaves the others quiet, and remembers what happened inside HIVE.
- Rules on paper: what the first release must do, what waits for later, and what HIVE won't do (like scraping other sites or taking outside actions on your behalf).
- Power users first: people who already use two or more AI services for personal, creative or technical work.
The result
The PRD is done, and HIVE is in development.
What I took from it
The best AI setup isn't the smartest model. It's a team with clear jobs, a shared memory and a human who makes the final call.