Peer feedback

Reviews are stronger when they include input from people the employee actually works with. PerfCopilot's peer-feedback flow is built so that:

Nomination is a one-click decision for the manager
Responding is a one-click decision for the peer (no PerfCopilot account required)
Anonymity is the default — managers see the feedback, but not who said what

Nominating peers

From an employee's dashboard, click Invite peers in the peer-feedback panel. Add up to 10 peers by email; for each one you can optionally set a display name (otherwise we use the email's local-part).

Click Send invites. Each peer gets an email with a unique single-use link to leave their feedback.

Some details worth knowing:

Nominations are scoped per (employee, cycle). Inviting the same peer twice for the same cycle is silently deduped.
You don't have to wait for an AI review draft before inviting peers — nominations live independently. The review reads matching peer feedback when it's generated.
The employee being reviewed is not notified of the nominations. They only see peer feedback if you choose to surface it during your 1:1.

The peer experience

The peer clicks their email link and lands on a PerfCopilot page that shows:

Who they're reviewing (the subject's name and role)
Who invited them (you, the manager)
What cycle the feedback is for
A short form: a 1–5 rating, strengths, areas for improvement, and an optional overall comment
An Anonymous toggle — on by default

Submitting takes 2–3 minutes. The link is single-use; reusing it returns "Feedback already submitted." If they don't want to participate, they can decline (the manager sees a "declined" status without details).

Anonymity

The peer's anonymity toggle is the source of truth:

Anonymous (default) — the manager sees the rating + comments but not the peer's name. Internally the link is still tied to the nomination so you can track who responded vs. who didn't, but the peer name is suppressed in every UI.
Named — the manager sees the peer's name alongside the comments.

Anonymity applies to AI prompts too — the AI sees "peer_name": null for anonymous responses, so it can't accidentally reference a name in the generated review.

Where peer feedback shows up

Peer feedback flows into three places:

The peer-feedback panel on the dashboard — a roll-up of submitted feedback for the selected cycle. Average rating, total responses, plus each individual response (strengths / improvements / overall).
The signals strip — each submitted peer response shows as a peer_feedback signal, alongside GitHub commits, Slack messages, etc. Click to see details.
The AI review prompt — included as a [PEER FEEDBACK] block. The AI is instructed to weave peer observations into strengths/improvements where relevant, with their rating influencing (but not dictating) the AI's rating.

Fan-out across cycles

If you run multiple cadences concurrently — weekly + monthly + quarterly — a single peer response surfaces in every cycle whose window contains the response date. The peer fills out the form once; the manager sees it in the weekly view, the May monthly view, and the Q2 quarterly view (assuming it falls in May).

This means peer feedback collected for one cycle's review feeds the others naturally — no double-collection, no asking the same peer the same question twice in two months.

Cleaning up

Two manager actions on the peer-feedback panel:

Resend — re-emails a peer who hasn't responded yet. Same single-use link, no duplicate created.
Cancel — deletes a pending nomination (one that hasn't been submitted yet). Submitted nominations are immutable; if you need to discard a response, you'll need to delete the underlying signal through support.

Re-running the peer-feedback flow in the next cycle is a fresh round — old nominations don't carry forward unless you explicitly include them.