AI response
Recommended AI: ChatGPT, Claude
Input: "Next week, discuss the new onboarding flow · 5 attendees (PM, design, 2 devs, product) · 60 min"
Definition of success (what's in hand when it ends)
Name the single highest-drop-off step among the current 5 onboarding steps and decide whether to fix it next sprint. Not a vague "share directions" — one decision in hand.
Decisions that must be made
- Which of the 5 steps do we touch this sprint? — Decider: PM
- Is the fix a 'UI change' or a 'step removal'? — Decider: PM + design agreement
- Into this sprint, or push to next? — Decider: dev lead (capacity call)
Timebox (total 60 min)
| Item | Min | Purpose | Owner |
|---|---|---|---|
| Skim drop-off data | 10 | Share -> consider replacing with a doc | PM |
| Narrow high-drop candidates | 15 | Discuss | All |
| Decide fix type (edit vs remove) | 15 | Decide | PM/design |
| Sprint inclusion | 10 | Decide | Dev lead |
| Actions/owners/dates | 10 | Decide | PM |
The '10 min skim' becomes 0 if you circulate it as a doc before the meeting. See pre-work.
Pre-work
- PM: share the per-step drop-off table the day before (seeing numbers for the first time in the room burns 15 min).
- All: read it and come with 'one candidate you'd pick'.
- WARNING: without pre-work this meeting slips 100% into brainstorming — no decision comes out.
Fallback for absence/overrun
- Dev lead absent -> defer decision 3 (sprint inclusion) to async and close after settling decisions 1 and 2.
- Overrun -> drop 'fix-type detail' first. If you only name the step (decision 1), this meeting succeeded.
Meeting waste gauge
35% — the first 10 min (data share) is doc-able and can be cut, but the 3 decisions need real-time agreement, so a meeting fits. Circulate the pre-work and it shrinks to a 45-minute meeting.





