Back to use cases
Recruiting & Onboarding

Per-interviewer stance across the loop

How each interviewer landed, and where the panel disagrees: before the debrief.

Best for
Pre-debrief
Primitives
ActorsStanceDecisions
Token savings
48%

The prompt

You are preparing the debrief for {{candidate_name}}'s loop. Using Sentra: 1. Pull each interviewer's written feedback Interaction (scorecards, Slack DMs to the recruiter, notes docs). 2. For each interviewer, extract: overall vote (strong yes / yes / no / strong no), the two dimensions they emphasized, the one concern they noted, and a direct quote that captures their bottom line. 3. Identify where the panel disagrees and on which dimension. Output: - Per-interviewer card with the fields above. - A "disagreement map" — which dimensions split the panel and how. - Three questions the debrief facilitator should put to the panel to resolve the disagreement.

See it work

Run it from

ClaudeCursorChatGPTAny agent · MCP

How it runs

  1. 01Copy the prompt into any agent connected to Sentra over MCP.
  2. 02The agent reads the company memory, scoped to what you are allowed to see.
  3. 03One grounded answer comes back, with every line traceable to its source.

Or automate it with Actions

The same prompt runs as a Sentra Action: on a schedule, or on a semantic trigger. Output lands in Slack or email, and every send waits for your approval.

Trigger idea: pre-debrief

Sentraover MCPPer-interviewer stance before the debrief

Platform lead loop · where the panel stands

2 strong hire

debriefs

Both cite systems depth; one calls the incident walkthrough "the best I have heard this year."

1 neutral

debrief

No technical concerns; unsure what the candidate uniquely adds to the current team shape.

1 no hire

debrief · quote

Pairing round: "talked past the interviewer twice." The only communication datapoint in the loop, and nobody else tested it.

Shape of the disagreement

One axis: communication under collaboration. A focused follow-up conversation resolves it; another systems round does not.

The prompt above becomes this. Fictional data, real mechanics: every line cites the meeting, thread, or record it came from.

Without Sentra

  1. 01Pull every interviewer's feedback artifact — scorecard, Slack DM to recruiter, notes doc.
  2. 02Read each to extract the overall vote (strong yes / yes / no / strong no).
  3. 03Identify the two dimensions each interviewer emphasized.
  4. 04Pull a representative quote that captures their bottom line.
  5. 05Cross-compare to identify panel disagreement by dimension.
  6. 06Compose the debrief prep doc.

~50,000 tokens

With Sentra

  1. 01Pull each interviewer's feedback `Interaction`.
  2. 02Each has typed `vote`, emphasized `dimensions`, `concerns`, and a representative quote.
  3. 03Identify cross-interviewer disagreement on dimensions automatically.
  4. 04Compose the debrief.

~26,000 tokens

Book a DemoStart Free