Preparation

Were the vision, backlog and capacity ready before day one?

Half the features had no acceptance criteria, so the first breakout went on questions for Product Management.
We got the vision deck the night before. Nobody had time to read it.
Capacity was worked out properly this time, with holidays included, and it showed in the plan.
The two days

How did the event itself run, in the room or online?

The business context briefing ran 40 minutes over and ate into the first breakout.
Remote team members couldn't hear the room during the draft plan review.
Product Managers walked between breakout rooms, so questions got answered in minutes.
Dependencies and risks

Which dependencies and risks did the board miss?

The API dependency on the platform team was on the board, but nobody from that team agreed the date.
We found the shared database migration on day two, after most teams had planned around it.
ROAM was rushed. We marked the vendor risk as owned, and the owner left the company in week three.
Objectives and confidence

Did our PI objectives and confidence vote hold up?

We voted four fingers, then missed two committed objectives. The vote didn't show what people really thought.
Our objectives read like feature lists, so Business Owners found them hard to score.
Keeping one objective uncommitted gave us room when the support load spiked.
Next PI Planning

What will we change before the next PI Planning?

Product Management shares the vision and top ten features two weeks before the event.
Every cross-team dependency needs a named person from each team before it goes on the board.
Add a short ROAM session at the end of day one as well as day two.

What is a PI Planning retrospective?

PI Planning is the biggest event in the Scaled Agile Framework (SAFe). For two days, every team on an Agile Release Train (ART) plans the next Program Increment (PI) together. The agenda ends with a short retrospective led by the Release Train Engineer (RTE), and that slot is easy to rush when everyone is tired. A PI Planning retrospective gives the planning process proper time, so the next event runs better than this one. The board has five columns that follow the event in order. Preparation asks whether the vision, backlog and capacity were ready on day one. The two days covers the event itself: the agenda, the breakouts, the draft plan review and the tools, in the room or online. Dependencies and risks looks at the ART planning board and the ROAM session. Objectives and confidence checks whether the PI objectives and the confidence vote held up once the work started. The last column turns all of it into changes for the next PI Planning. In TeamRetro, everyone adds ideas privately first, then the group sorts them, votes and assigns owners. Run it at the end of day two while the event is fresh. Then run it again near the end of the PI, before Inspect and Adapt, when you can see which objectives were met and which dependencies slipped. The second run answers the question the first can't: was the plan any good? Each run gives the RTE and the teams a short list of fixes to make before the next PI Planning.

PI Planning retrospective format

Preparation

Were the vision, backlog and capacity ready before day one?

Many PI Planning problems start in the weeks before the event. Ask whether features were ready (sized, with acceptance criteria and an owner), whether teams knew their capacity, and whether the vision and top priorities arrived early enough to read. Push for the specific gap, such as which feature wasn't ready and who needed what, so the fix can go into the preparation checklist for the next PI.

The two days

How did the event itself run, in the room or online?

Look at the agenda, the timeboxes and the logistics: the business context briefing, team breakouts, the draft plan review and the management review. For remote or hybrid events, ask about time zones, tools and whether remote people could follow the room. Separate what was merely annoying from what changed the plan, and spend the time on the second.

Dependencies and risks

Which dependencies and risks did the board miss?

The ART planning board shows feature dates, milestones and the dependencies between teams. Ask which dependencies were found late, which were on the board but never agreed by the other team, and which risks were marked resolved during ROAM but came back. When you run this near the end of the PI, look at the dependencies that actually slipped and trace each one back to how it was planned.

Objectives and confidence

Did our PI objectives and confidence vote hold up?

Right after PI Planning, ask whether the PI objectives were clear, measurable and agreed with the Business Owners, and whether the confidence vote was honest. Near the end of the PI, compare planned with actual business value and ask what the vote missed. A high vote followed by missed objectives usually means people didn't feel safe voting low, or the plan left no room for unplanned work.

Next PI Planning

What will we change before the next PI Planning?

Turn the strongest themes from the other columns into a few specific changes, each with an owner and a date before the next event. Many belong to the RTE or Product Management, such as the agenda or feature readiness, and teams own some too. Add them as actions in TeamRetro and review them at the start of the next PI Planning retrospective.

When to use this retrospective

  • At the end of day two of PI Planning, after the confidence vote, while the details of the event are still fresh.
  • Near the end of the PI, before Inspect and Adapt, to check whether the plan, its objectives and its dependencies held up.
  • After a low confidence vote or a plan that needed rework, to find out what went wrong in planning.
  • When your ART moves PI Planning from in person to remote or hybrid and wants to learn from the first run.
  • When a new Release Train Engineer takes over and wants the teams' view of how planning works today.

Suggested icebreaker questions

  • On a scale of one to five fingers, how much energy do you have left after PI Planning?
  • What's one thing from the last two days you never want to see on a planning board again?

Want a warm-up your team plays, not just answers out loud? Browse free icebreakers →

Ideas and tips for your retrospective meeting

  • Hold the retrospective inside the PI Planning agenda, before people leave. A follow-up a week later gets fewer people and vaguer answers.
  • Invite the whole train: team members, Product Owners, Product Management, Business Owners and the System Architect. Each sees a different part of the event, and remote attendees often see problems the room misses.
  • Keep it about how the train planned, not about what is in the plan. Scope and feature debates belong in backlog refinement.
  • Bring the evidence: the ART planning board, the ROAM risk list, the PI objectives with their business value and the confidence vote results.
  • Use anonymous input. People who voted two fingers or fewer in the confidence vote should be able to say why without being singled out.
  • Vote on the top two or three changes and give each an owner, usually the RTE or a team, with a date before the next PI Planning.

Frequently asked questions

What questions should a PI Planning retrospective ask?
Were we ready when the event started? How well did the two days run? Which dependencies and risks did we miss on the ART planning board? Did our PI objectives and confidence vote hold up during the PI? What will we change for next time? This template gives each question its own column, in that order.
What are some PI Planning retrospective ideas?
Run it twice, once at the end of day two and once near the end of the PI, and compare the answers. You can also open with the actions from the last PI Planning retrospective, start from the confidence vote results, or follow one late dependency from the ART planning board through the PI to see where the plan broke down.
Who runs the PI Planning retrospective?
In SAFe, the Release Train Engineer (RTE) leads it as the last item of PI Planning. On a large train, each Scrum Master or team coach can run it with their own team first, and the RTE combines the top themes.
What are PI objectives?
PI objectives summarize what each team intends to achieve in the coming PI. Teams split them into committed and uncommitted objectives, and Business Owners give each one a business value score from 1 to 10. Comparing planned and actual business value at the end of the PI shows how realistic the plan was.
What is a confidence vote in PI Planning?
At the end of PI Planning, everyone shows one to five fingers for how confident they are in the plan. SAFe suggests accepting the plan when the average is three or more and reworking it when it is lower, and anyone who votes two or fewer is asked to explain why.
How is a PI Planning retrospective different from Inspect and Adapt?
Inspect and Adapt happens at the end of each PI and looks at the whole system, with a PI system demo, a review of the measures and a problem-solving workshop. A PI Planning retrospective looks only at how the train planned. Run it at the end of PI Planning, and again before Inspect and Adapt so its findings feed in.
Ready to run this retrospective?Try this retrospective live