HR & RECRUITING

Transcribe and anonymize each Zoom exit interview

When a Zoom offboarding call finishes recording, transcribe it, strip names and other identifiers.

CategoryHR & Recruiting
Enginesim
Difficultyintermediate
Triggerevent
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerZoom recording completed for offboarding callZoomZoom
  • ActionTranscribe audio with Hugging Face speech-to-textHugging FaceHugging Face
  • ActionRedact names and identifiers via NER modelHugging FaceHugging Face
  • LogicHold transcript if any identifier remains, else proceed
  • OutputWrite anonymized transcript to Coda tableCodaCoda

What it does

Turns every completed Zoom exit-interview recording into a redacted, analysis-ready transcript. Speaker names, manager names, team identifiers, and contact details are removed before the text is ever stored, so the raw, attributable recording never leaves Zoom.

When to use it

Run this as the foundation of any exit-interview program. People Ops wants honest signal about why employees leave, but employees only speak freely when anonymity is guaranteed. This enforces that guarantee mechanically instead of trusting a human to scrub each file.

How it works

  1. 1A Zoom webhook fires the moment an offboarding-call recording completes.
  2. 2The audio is sent to a Hugging Face speech-to-text model to produce a verbatim transcript.
  3. 3A redaction step uses a Hugging Face NER model to detect person names, emails, and team references, replacing each with a stable role token (Departing Employee, Manager A).
  4. 4A logic check confirms no high-confidence identifiers remain; anything ambiguous routes to a hold list instead of the clean store.
  5. 5The anonymized transcript is written to a Coda table keyed by a random interview ID, with the raw recording reference deliberately excluded.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect ZoomMeetings, recordings, transcripts.
  2. 2
    Connect Hugging FaceModels, datasets, spaces — the open-source hub.
  3. 3
    Connect CodaDocs, packs, automations.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.