What the PM must have ready before sprint planning
- Finalized PRD: All 10 sections complete, reviewed by engineering lead and design, version-locked.
- Acceptance criteria for every user story: In Given/When/Then format. Every criterion independently testable.
- Edge cases documented: At least 3 to 5 per feature, with expected system behavior for each.
- Dependencies resolved: Every external dependency has a confirmed owner and status.
- Open questions closed: No unresolved decisions that would block implementation.
The sprint planning agenda that actually works
| Step | Time | Owner | Purpose |
| Backlog review | 15 min | PM | Walk through what's in this sprint and why (RICE rationale) |
| Story refinement | 30 min | Engineering | Engineers ask clarifying questions, PM answers from PRD |
| Story pointing | 20 min | Engineering | Estimate complexity with full requirement context |
| Sprint goal definition | 10 min | PM + Engineering | Single-sentence sprint goal that connects to product strategy |
| Commitment | 5 min | All | Team commits to sprint scope, changes require PM sign-off |
Sprint planning anti-patterns
Starting planning without a reviewed PRD. This turns sprint planning into a requirements meeting, which takes 3× longer and produces worse output.
PM-led story pointing. Story pointing is an engineering estimation exercise. PMs who push back on point estimates in real-time undermine the trust that makes sprint planning work.
Adding stories mid-sprint without scope change process. Every mid-sprint addition displaces something else. The PM needs a documented process for evaluating scope changes, not an ad-hoc approval flow.
No sprint goal. A list of stories without a sprint goal is a task list, not a sprint. The goal connects individual stories to the "why" that keeps the team oriented when edge cases require judgment calls.