An approved pattern for safe email drafting with AI
A blessed way to draft email with a model that keeps a human on send. AI writes, you review and send — nothing leaves the outbox on its own.
This is where good ideas earn their place. Teammates propose AI features and workflows; we vote to prioritize. What rises to the top is what we build next.
Anyone can add an idea or change their vote.
How it works
Upvote the ideas you want. Click again to take your vote back. The board sorts by votes, so the most-wanted work sits at the top.
Have something that isn’t here? Add it. A new request joins the board with your vote already on it.
Propose an AI feature or workflow. The team votes; the top ideas shape what we build next.
A blessed way to draft email with a model that keeps a human on send. AI writes, you review and send — nothing leaves the outbox on its own.
One place for prompts that already work — reviewed, named, and reusable. Stop everyone re-deriving the same instructions from scratch.
A shared research workspace on a plan with a clear data posture, so research questions don’t get pasted into random consumer accounts.
A short guided check that tells you whether an integration is fine as-is or worth a closer look — before you grant any standing access.
A quick, consistent way to strip names, numbers, and secrets out of text before it reaches a third-party model. Same habit, every time.
A safe environment seeded with realistic-but-synthetic records, so we can try autonomous workflows without ever touching real inboxes or deals.
A baseline editor and assistant configuration we agree on — what’s on, what’s off, and what never gets indexed — so AI-in-the-IDE is consistent across the studio.
A standing slot to look at what’s new, decide what to adopt, and keep the handbook current — instead of tools creeping in one DM at a time.
The categories