Draft your product roadmap from a pile of Notion notes

The problem. Your actual roadmap lives across a dozen Notion pages, half-finished docs, and things you said out loud in a meeting three weeks ago. Nobody, including you, could recite it cleanly right now.

The play. Point Claude at the real folder of notes and make it do the compiling and prioritizing you keep meaning to sit down and do — with your GOALS.md as the filter for what actually matters this quarter.

Setup (5 min). Get everything into one folder or Cowork Project — export the Notion pages, drop in any docs, don’t bother cleaning them up first. Have GOALS.md and USERS.md in the same place.

The prompt:

Read every planning note, doc, and fragment in this folder. Also read
GOALS.md and USERS.md for what we're actually prioritizing and who we serve.

Do this:
1. Pull out every distinct product idea, feature, or initiative mentioned
   anywhere, however rough
2. Merge duplicates and near-duplicates
3. Group them into Now / Next / Later based on how well they map to GOALS.md
   and how directly they address something in USERS.md
4. For anything you're placing in "Now," state which goal or user problem it
   serves
5. Flag anything that shows up often but doesn't map to GOALS.md at all —
   that's either a signal our goals are wrong, or scope creep

Cite which document each item came from. Don't invent initiatives that
aren't actually in the notes.

What you get. A real Now/Next/Later roadmap, grounded in what you actually wrote down rather than what’s freshest in your head. The “doesn’t map to GOALS.md” flag is often the most useful line in the output — it’s where you find out your planning docs and your stated priorities have quietly drifted apart. For anything landing in “Now,” the PRD playbook is the natural next step.

What to watch for.

• This reflects what’s in your notes, not the truth — an idea nobody wrote down doesn’t exist to this workflow. Good notes in, good roadmap out.
• The Now/Next/Later grouping is a first draft for your judgment, not a final answer. Read the “why” behind each placement before you commit to it.

Make it repeatable. Run it monthly, right after your weekly review, and watch which items keep sliding from “Next” to “Later” without ever landing — that’s usually a sign to either commit or kill them.