Run a hiring debrief
Collect every interviewer's take independently, then synthesize against the rubric.
An interview debrief is where good evidence goes to die: the loudest voice speaks first, everyone anchors, and the quiet interviewer's red flag never surfaces. This workflow collects every take independently before anyone sees anyone else's, then synthesizes the evidence against your rubric.
What does this workflow do?
After a candidate's loop, the agent sends every interviewer an individual information request and collects their written feedback privately. Each take stays private until all are in; nobody anchors on the first opinion because there is no first opinion. Then the agent synthesizes: evidence mapped to rubric competencies, observations separated from conclusions, disagreements laid side by side with the question that would resolve each.
It runs the bias checks a good facilitator runs: is one strong session bleeding into unrelated competencies, which competency did nobody actually test, and is the most confident voice the one with the most evidence. The output is the debrief meeting's agenda, with the decision left where it belongs: with the hiring manager.
How does it work?
- The loop ends and collection starts. The agent sends each interviewer an individual information request; they submit written feedback privately, tagged to the competencies their session covered.
- Takes stay sealed until all are in. The agent holds each interviewer's feedback private, and nudges stragglers, until everyone has submitted. Independence first, discussion second.
- The agent maps evidence to the rubric. Per competency: what was observed (quotes, examples) and what was concluded, kept separate. "Seemed junior" is a conclusion; "couldn't walk through their own architecture decision" is evidence.
- It surfaces disagreements and gaps. Divergent reads side by side, with whether they saw different evidence or read the same evidence differently. Untested competencies get named; a gap is a gap even when everyone feels good.
- It delivers the debrief agenda. Delivered into a shared debrief conversation the whole panel joins: the evidence-weighted read per competency, the case for and against (both steelmanned), and what the evidence cannot decide. The hiring manager makes the call in the meeting, which is now short.
Why collect feedback independently?
Because the first opinion voiced becomes the anchor for every one that follows, and interview panels are exactly the group setting where that costs the most. Structured hiring research has argued this for decades; the hard part was logistics. Information requests make independence the default: the agent asks each interviewer individually, holds the takes privately until the set is complete, and the shared debrief discussion starts from organized evidence instead of a first impression. ATS tools like Greenhouse and Ashby collect scorecards and are stronger for pipeline management at scale; the synthesis, the bias checks, and the sealed-until-complete mechanic are what this workflow adds. And because the workflow remembers past debriefs, calibration compounds: what "strong technical" meant last quarter is what it means now.
Works with
Slack, Notion, Google Docs. Run per candidate after the interview loop; the debrief document archives to Notion or Google Docs.
Frequently asked questions
How is interviewer independence enforced?
The agent sends each interviewer an individual information request, so feedback is submitted privately and each take stays sealed until all are in. The synthesis only starts when the set is complete.
What if we don't have a rubric?
The agent builds a provisional one from the role description and says so. A debrief without criteria is a vibes contest; a provisional rubric is the minimum fix, and it becomes the starting point for the real one.
Does the workflow recommend hire or no-hire?
No. It delivers the evidence organized, the disagreements framed, and the cases for and against steelmanned. The decision stays with the hiring manager, on purpose.
Can it work with our ATS scorecards?
Yes, paste them in or connect the docs where they live. Greenhouse and Ashby are stronger at pipeline management; this workflow takes the raw feedback wherever it's written and does the synthesis.
How does calibration improve across hires?
The workflow remembers past debriefs: how competencies were graded, where the panel disagreed, what the eventual outcomes were. Rubric language sharpens hire by hire instead of resetting each time.