Performance Review Calibrations
Calibrations that are 3x faster and fairer
Pre-reads that highlight discrepancies and flag biases, so calibration meetings are grounded in truth, not politics.
































































AI-powered calibration pre-reads

Whoa the calibration reports! What an unexpected moment of delight 😅
Whoa the calibration reports! What an unexpected moment of delight 😅
Setting ratings is a breeze

Four views
One session, four ways to see it
People
Review one packet at a time, with manager, self, peer, and upward responses side by side.
Grid
A 9-box or 25-box built from any two axes. Facilitators drag reviewees along editable axes and toggle between initial and adjusted placements.
Table
A sortable roster for scanning the whole session, with question, stat, and supplemental data columns.
Insights
Rating distributions before and after proposed changes, so the committee sees the impact of every decision.
Built for real committees
Runs live, async, or both
1. Discuss live or async
Committee members comment on packets on their own schedule in the Discussion panel, or walk the grid together in a live session.
2. Propose changes
Facilitators propose updates to ratings and other structured answers. Every comment and change stays attached to the reviewee.
3. Submit with control
Apply changes to manager reviews directly, or send recommendations that managers approve one by one.
How it works


throughout the year, then automatically generates calibration materials after reviews are submitted. 1. Continuous context gathering
Windmill collects performance data year-round from your tools, 1:1s, and team feedback.
2. AI detects discrepancies
Windmill instantly identifies similar performers with different ratings, manager rating patterns, and potential biases.
3. Pre-reads generated
Committee members receive comprehensive briefs with rating distributions, flagged outliers, and discussion areas.
4. Focused calibration meetings
Leaders spend time making data-driven decisions instead of debating, so calibration sessions run 3x faster.
Why traditional calibrations fail
Frequently asked questions