PRD
Help me spell out what I'm building for the AI: the problem it solves and what counts as done.
“After six meetings a week, just organizing the notes and forwarding them one by one takes 40 minutes.”
State who has what problem in which situation, and label what is fact versus assumption.
A PRD aligns decisions from background through acceptance. A technical proposal can link into it, but cannot make those decisions for it.
Start with the user problem and goal: Name the user, situation, and obstacle, then state the result the work should change.
Set the scope and non-goals: State required cases, constraints, and what this release will not address.
Agree on evidence: Use observable acceptance criteria or success signals, and keep assumptions and open questions with the request.
You can say this to an AI Agent
Turn “one-click meeting-notes export” into a one-page PRD covering the target users, current problem, goal, scope, non-goals, and success criteria. Mark unsupported details as open questions instead of deciding the technical solution or drawing wireframes for the team; then check that every requirement can be verified.
Select page elements for precise AI edits
Click a button, text, or card, note the change, then copy its exact location to AI.
Open Selector Pro website↗A team is preparing a one-click meeting-notes export. Which set of content best forms the core of its PRD?
ANS Original Methods