AI Sprint Planner
Tell it how long your sprint is and how many people are on the team. It reads your open backlog, picks a realistic set of work for that capacity, explains why it chose each item, and — once you're happy — creates the cycle and fills it.
Plan: Pro and above
What you can do
- Get an AI-recommended sprint drawn from a project's real open work.
- See a reason for every suggested item, so you can argue with it.
- Tick and untick items before committing to anything.
- Create the cycle in one click, with the selected items already in it.
Getting started
- In the workspace sidebar, open Command Center.
- Switch to the Planner tab.
- Under Sprint Configuration, choose:
- Project — the backlog to plan from.
- Sprint length (days) — defaults to 14.
- Team members — how many people are working the sprint. Defaults to 3.
- Click Generate Sprint Plan.
📷 Screenshot: the Sprint Configuration card with a project selected, sprint length and team size fields, and the Generate Sprint Plan button.
The direct route /<workspace>/sprint-planner still works if you've bookmarked it.
Reviewing the plan
The result has two parts:
- A summary — a short paragraph explaining the shape of the sprint the AI has proposed.
- Suggested Issues — the items it picked, each with a one-line reason (for example, urgent bug blocking sign-up; small estimate).
Every suggestion arrives ticked. Untick anything you don't want. The list only ever contains items that genuinely exist in the project — if the model tries to invent one, it's discarded before you see it.
📷 Screenshot: the AI Recommendation panel with the summary quote and a checklist of suggested issues, each with its reason underneath.
Creating the cycle
Click Create Sprint Cycle. This:
- Creates a cycle in the selected project called AI Sprint, starting today and ending after the sprint length you set.
- Adds every ticked item to it.
- Shows a success banner with a View Cycles link.
From there it's an ordinary cycle — rename it, adjust the dates, add or remove work, and run it however you normally do.
Tips
- The team size is a capacity dial, not a roster. Turning it up makes the AI select more work; it doesn't assign anything to anyone.
- Estimates help a lot. If your work items carry estimate points, the plan gets noticeably more realistic. Priority alone is a blunt instrument.
- Regenerate freely. Changing the project, sprint length or team size clears the previous plan — nothing is created until you click Create Sprint Cycle.
- Rename the cycle straight away. Every generated cycle is called "AI Sprint", so a second one will be confusing if you leave the name alone.
Limits & good to know
- It only considers work that's ready to start. Items in "unstarted" and "started" states are candidates; anything sitting in a backlog-group state is not offered.
- It doesn't assign people. The plan chooses what, not who.
- It doesn't set due dates on the items — only the cycle gets dates.
- It doesn't know your history. Past velocity, individual availability, holidays and part-time capacity are not taken into account. Priority, estimates, team size and sprint length are.
- Large backlogs are truncated. It works from up to a thousand open items per project.
- The cycle name is fixed at creation ("AI Sprint"). Rename it afterwards in the project's Cycles.
- It needs an AI provider configured for the instance (admin console → Artificial Intelligence). Without one, generation fails with a clear message.
- Treat it as a first draft. It's a starting point for sprint planning, not a substitute for the conversation.
Related
- Command Center — the hub this lives in
- Proactive AI — sprint health once the sprint is running
- AI Agent — ask it to tweak the cycle afterwards
- Pricing