Scoring and debrief

Finish your scorecard while the reasoning behind your notes is still fresh. When you explain a rating, point to something specific: in a Dispatch Service exercise at Lanefold, a fictional freight company, that could be the candidate's retry design or a moment in the drawing.

Scoring requires interviews:run and an interviewer or shadow seat on the panel. Reading the debrief requires interviews:read and either panel membership or interviews:results.

Submit your scorecard

After the session ends, select Score in Interviews, or Open debrief on its setup page.
Under Your scorecard, choose a rating using each competency's anchors. Write the supporting Evidence.
Choose an Overall recommendation and add a Summary.
Wait for Draft saved, then select Submit scorecard and confirm Submit and lock.

Submission requires at least one rated competency and a human recommendation. You don't have to rate every competency. Ratings and the recommendation belong to the named person submitting the scorecard; AI never supplies them.

A submitted scorecard is locked. Reopen scorecard is available while the interview is awaiting feedback. Once all interviewers submit, the interview completes and reopening closes. A shadow's submission doesn't delay completion.

Shadows should submit while the interview is awaiting feedback, before all interviewers have submitted. The Submit scorecard control is no longer available after completion.

With Blind independent scoring enabled, panelists without interviews:results wait until every interviewer has submitted before seeing the panel debrief. Submitting your own scorecard alone doesn't end that wait.

Someone with interviews:results who is an interviewer on the panel must submit their own scorecard first, then can see the other scorecards before everyone has submitted. Someone with that permission who isn't an interviewer on the panel can read results without waiting. Turning blind scoring off removes these submission checks.

Draft evidence with AI

Select Draft evidence with AI to fill the evidence and summary from your notes, the timeline and a canvas summary. You need Team or above and ai:use, and Debrief drafting must be on in organization settings. Each use consumes AI credits.

The draft goes straight into your scorecard. Read and edit it before submitting. It doesn't choose ratings or the recommendation, and the scorecard records that an AI draft was used.

Compare the panel

Once you can read the debrief, Panel matrix shows the named panelists' ratings by competency. The means and recommendation distribution count submitted interviewers; shadow ratings are shown for calibration but excluded from those totals.

Divergence marks a competency where counted ratings differ by at least two levels. The weighted mean uses the competency weights in the kit. Both are calculated from the ratings people entered.

Replay the drawing

In Session replay, select a note or a timeline marker to jump to that moment. The replay shows saved canvas snapshots, so it depends on Record canvas snapshots being enabled during the interview. It isn't a video recording.

Where stage snapshots are available, select a stage under Stage continuity. State at end uses the stage's completion snapshot, then falls back to the final session snapshot, or the latest saved snapshot if no final snapshot exists. Added during stage dims content from other stages. The replay canvas needs a viewport at least 768px wide.

Export the debrief

Under Export and webhook, select Markdown, JSON or PDF. Copy scorecard copies the matrix's ratings and evidence as text. Exports follow the same result access and blind scoring checks as the debrief.

Practice debriefs show a Practice badge. Their Markdown and PDF titles begin with "Practice interview:". You can export them, but they never send completion webhooks.