Manager Reviews

Draft, edit, and complete comprehensive manager performance reviews in minutes with Windmill’s AI-assisted writing, peer feedback context, and review templates.

Manager performance reviews bring together all feedback from self, peer, and upward reviews with continuous context from the review period. Instead of starting from a blank page, managers have everything they need to write thoughtful, evidence-based assessments in minutes.

The review packet

When it’s time to draft reviews, managers access a packet for each direct report. The packet is comprised of all the review types and the supporting context Windy uses to help managers write:

• Summary report — An overview of the employee’s review period, pulling from self, peer, and upward reviews, continuous feedback, recaps, pulse responses, 1:1 notes, goals, and activity from connected tools (Slack, Jira, Salesforce, GitHub, Google Workspace, and others)
• Self review — How the employee describes their own growth and accomplishments
• Peer reviews — Feedback from colleagues on collaboration, strengths, and opportunities
• Upward feedback — If the employee manages others, feedback from their direct reports
• Historical context for reference — Past 1:1 notes, past reviews, past continuous feedback, and past recaps from prior periods, so managers can ground the current review in trends over time

All of this context is automatically gathered and organized, so managers can focus on writing rather than hunting for information.

To access a packet, go to Performance Reviews, select the active cycle, and click into any direct report.

How inputs surface in a review

Most review inputs are writing aids — context surfaced privately to the person drafting, never published into a shared review artifact. The only inputs that become part of the shared review are the answers people explicitly write: self review answers, peer review answers, upward review answers, and manager review answers.

Input	Where it surfaces	Who sees it
Self review answers (written by the employee)	Self section of that employee’s packet; [Self] observations as a writing aid in their manager’s draft	The employee’s manager, skip-levels, and cycle admins. The employee sees their own.
Peer review answers you wrote about a colleague	Peer section of that colleague’s packet; [Peer] observations as a writing aid in their manager’s draft	The colleague’s manager, skip-levels, and cycle admins. Never the subject.
Upward review answers you wrote about your manager	Upward section of your manager’s packet; [Upward] observations as a writing aid in their manager’s draft of them	Your skip-level and cycle admins. Never your manager.
Manager review answers	The reviewing manager’s draft of the packet	The reviewing manager, skip-levels, and cycle admins. The subject sees them only after the manager shares the review, and only for questions marked Visible to employee.
Calibration notes and rating adjustments	Calibration view of the packet	Calibration facilitators and committee members. Never the subject.
Private note you wrote tagged to another employee	Writing-aid context in your own draft of any review you write about that person — your manager review (if they report to you), your peer review, or your upward review of them	Only you. Never published into any shared review artifact.
Private note you wrote about yourself	Writing-aid context in your own self-review draft	Only you.
1:1 notes	Writing-aid context for the reviewing manager: [1:1] observations in their manager review draft, and the summary report	The reviewing manager only, within the context of the review cycle. Not visible to the subject, peers, skip-levels, or cycle admins as a review artifact — though 1:1 notes have their own native visibility outside of reviews.
Continuous feedback, recaps, pulses, connected-tool activity	Writing-aid context for the reviewing manager: [System] observations in their manager review draft, and the summary report	The reviewing manager only, within the context of the review cycle. Not visible to the subject, peers, skip-levels, or cycle admins as a review artifact — though each underlying source has its own native visibility outside of reviews.

Writing aids never automatically appear in the published review. The reviewer has to select or copy them into an answer for them to become part of the shared artifact. For the full visibility model — including the rule that you never see feedback written about you — see Who sees what.

Drafting a review

Manager reviews are structured around specific questions configured when the cycle was created. These might include:

• Overall performance rating
• Key accomplishments and impact
• Areas for growth or development
• Goal progress and alignment with company values
• Promotion readiness or compensation recommendations

For each question, Windmill provides AI-generated observations to help managers get started.

AI-suggested observations

Observations are short, objective snippets Windy suggests on a per-question basis. They draw on the same context as the summary report — self, peer, and upward reviews, continuous feedback, recaps, pulse responses, 1:1 notes, goals, activity from connected tools, and historical context like past reviews, past 1:1s, past feedback, and past recaps — and additionally use the framing of the specific review question to generate guidance tailored to what that question is asking about.

In other words, the summary report gives managers a holistic picture of the review period, while observations narrow that same context down to what’s relevant for the question being answered.

Each suggestion is tagged by source so managers know exactly where the insight came from:

• [Self] — From the employee’s self review
• [Peer] — From peer feedback
• [Upward] — From the employee’s direct reports, if applicable
• [1:1] — From 1:1 notes
• [System] — From integration data (activity in connected tools), private notes, recaps, continuous feedback, and pulse responses

For example:

[Peer] “Consistently delivers high-quality work on tight deadlines, as noted by three peers who worked with them on the Q3 product launch.”

[Self] “Took initiative to refactor the billing system, reducing processing time by 40%.”

[System] “Received positive feedback in Slack from the design team for collaboration on the new onboarding flow.”

Managers can:

• Select suggestions to include in their answer
• Edit suggestions to add context or adjust wording
• Write their own observations from scratch
• Add custom observations alongside AI suggestions

Generating draft answers

Once managers have selected observations for a question, they can generate a full draft answer. Windy synthesizes the selected observations into coherent, professional prose that:

• Weaves together multiple observations
• Balances strengths and growth areas (based on the rating)
• Maintains a professional tone
• Preserves source attribution

The draft serves as a starting point. Managers have full editorial control and can revise any part before submitting.

Adjusting tone

The same feedback can land very differently depending on how it’s framed. Windmill’s tone selector allows managers to fine-tune how draft answers are written—ensuring feedback is both honest and effective.

Tone options

When generating a draft answer, managers can choose from three tone options:

Balanced (default) — Neutral tone that matches the rating and observations

Glowing — Celebrates outstanding performance with enthusiastic language

Corrective — Emphasizes areas needing improvement with factual language

How to adjust tone

1. Select observations and rating

   Choose the AI suggestions or write custom observations for the question. If it’s a rating question, select the appropriate rating.

2. Choose tone

   Before generating the draft, select the desired tone from the dropdown.

3. Generate draft

   Click Generate to create a draft answer using the selected tone.

4. Review and edit

   Review the generated draft and make any manual edits as needed. You can regenerate with a different tone if the first draft doesn’t feel right.

What stays the same

No matter which tone is chosen:

• Review content remains grounded in evidence — Windy doesn’t invent accomplishments or exaggerate impact
• Data sources are preserved — Each observation still traces back to self, peer, system, or upward sources
• Managers retain full editorial control — The draft is a starting point; managers can edit freely

Tone changes how feedback is framed and emphasized, not what feedback is included.

Submitting reviews

The subject can’t see any part of their manager review until it’s shared with them. See Who sees what for the full visibility model.

When a manager is satisfied with all answers, they click Submit. Once submitted:

• The review is marked as “Submitted” in the manager’s dashboard
• If calibration is part of the cycle, the review moves into calibration
• Managers can still edit the review until it’s locked

Reviews are not automatically locked when submitted or calibrated. Cycle admins can lock reviews at any point to prevent managers from making further edits. Calibrations can still be done on locked reviews.

Review status

Manager reviews move through several states:

• Not Started — The packet is available but the manager hasn’t opened it yet
• In Progress — The manager has started drafting but hasn’t submitted
• Submitted — The manager has finalized and submitted the review
• Locked - The review cannot be edited by the manager, only via calibration.
• Calibrated — If applicable, the review has been through calibration
• Shared — The review packet has been shared with the employee

Managers can see the status of all their reviews in the My Team view. Cycle admins can monitor org-wide status in the All Employees view.

Learn more about sharing reviews with employees in Sharing Reviews.

Writing effective reviews

Even with AI assistance, the best reviews are:

• Specific — Reference actual projects, outcomes, and behaviors
• Balanced — Acknowledge both strengths and growth opportunities
• Forward-looking — Include development goals and next steps
• Fair — Based on evidence, not recency bias or gut feeling

Windmill makes it easier to write high-quality reviews by:

• Surfacing specific examples — AI suggestions include concrete work products and outcomes
• Balancing strengths and growth areas — The draft structure and tone options ensure reviews aren’t one-sided
• Grounding feedback in data — Every observation traces back to a source
• Saving time — What used to take hours now takes minutes

FAQs

How does Windmill save time for managers?

When it’s time to draft reviews, Windmill consolidates data that’s been collected throughout the review period:

• Employee self reviews completed in Slack
• Peer feedback from colleagues
• Upward feedback from direct reports
• Context from connected tools (Jira, Salesforce, GitHub, Google Workspace, etc.)
• Recaps, pulse surveys, and 1:1 notes

By surfacing this context in one packet, Windmill eliminates the need for managers to rely on scattered notes or memories.

Instead of spending 1-2 hours per review, managers can complete a full, data-backed assessment in as little as six minutes.

Can I skip AI generation and write my review entirely myself?

Yes. In the Observations step of the manager review, leave any or all questions blank and click Continue to skip AI generation. Any questions left blank on this step will not be AI generated. You can then write your answers from scratch in the finalize step, using the side panel for relevant context (self review, peer feedback, system signals, etc.) as you go.

This is useful when you don’t want Windy rewriting your wording.

Can managers edit reviews after submitting?

Yes. Reviews are not locked when submitted. Managers can still edit reviews after submission until a cycle admin locks them.

This allows for:

• Final edits before sharing
• Corrections or additional context

Cycle admins can lock reviews at any point to prevent managers from making further edits.

How are AI suggestions generated?

Observations use the same context as the summary report — self, peer, and upward reviews, continuous feedback, recaps, pulse responses, 1:1 notes, goals, activity from connected tools, and historical context like past reviews, past 1:1s, past feedback, and past recaps — and additionally use the framing of the specific review question to generate question-specific guidance.

Each suggestion is tagged by source ([Self], [Peer], [Upward], [1:1], or [System]) so managers know exactly where the insight came from. [System] covers integration data, private notes, recaps, continuous feedback, and pulse responses. Managers retain full control and can select, edit, or write their own observations.

Does Windy use our career leveling framework when drafting manager reviews?

Yes. If you’ve added your leveling framework (or other rubrics, job descriptions, or competency matrices) to Company Knowledge, Windy uses it as context when generating AI-suggested observations and draft answers in the packet.

Suggestions and drafts are framed against the expectations for the employee’s level, so feedback lands in the right context. The framework isn’t injected verbatim into the review as explicit criteria—to make level expectations show up explicitly in the published review, also reference them in the manager review questions you configure on the cycle.

Can I customize the questions managers answer?

Yes. When creating a cycle, you configure specific questions managers will answer about each direct report. Questions can be:

• Text questions — Open-ended written responses
• Rating questions — Single-select ratings (e.g., 1-5 scale, Exceeds/Meets/Needs Improvement)
• Yes/No questions — Boolean responses

These questions are fully customizable to match your company’s review process.

Learn more in Cycles.

Related reading

• Performance Reviews — Overview of the complete review process
• Self Reviews — Employee self-reflection and performance assessment
• Peer Reviews — Collect structured peer feedback
• Upward Reviews — Feedback from direct reports about their manager
• Calibration — Ensuring fairness and consistency across ratings
• Sharing reviews - Share finalized packets with managers and employees
• Progress tracking — Monitor review completion across your team