1. What the tool consumes
The Vettika screening tool (the “AEDT”) takes a fixed set of inputs and produces a recruiter-facing report. Inputs:
- Recruiter-authored job description — the role title, responsibilities, and any required qualifications the recruiter pastes into the campaign wizard.
- Recruiter-authored scoring rubric — between 3 and 5 weighted criteria (1-5 weight each) and what each one should measure.
- Recruiter-authored screening questions — the 3-6 spoken prompts the interviewer asks, plus the private follow-up hints the AI uses to probe.
- Candidate audio — the candidate's spoken answers during the live call. Audio is transcribed in-flight by Deepgram and the resulting text is the source of truth for scoring.
- Candidate-volunteered metadata — name, email, phone, and time zone. Self-reported demographic data is not currently collected; an opt-in collection step, used only for the bias audit, is planned and this page will be updated when it ships.
The tool does not consume résumés, social media profiles, public web search results, prior employment data, or any third-party background-check data.
2. Qualifications + characteristics evaluated
The AEDT scores candidates against the rubric the recruiter wrote — nothing else. Common rubric categories we see:
- Role-relevant skills (technical depth, domain knowledge)
- Communication clarity and structure
- Ownership and follow-through
- Past achievements measured against the recruiter's success criteria
Each criterion gets a 1-5 score and a one-sentence justification. The overall recommendation (Strong Yes / Yes / No / Strong No) is a weighted aggregate. The model does not evaluate accent, voice timbre, perceived gender, perceived race/ethnicity, perceived age, or any protected characteristic — those are excluded from the prompt + the scoring rubric.
3. Retention windows
Default retention windows (a recruiter can shorten these in workspace settings):
- Audio recordings: not retained. Audio is transcribed in-flight and discarded — we keep the transcript, not the recording.
- Transcripts + AI-scored reports: 12 months from completion. (Planned: subsequent archival in a candidate-id-only form, PII removed, for the bias-audit cohort.)
- Recruiter-authored notes + decisions: retained for the lifetime of the campaign, then archived when the campaign is closed.
- Demographic data (audit-only): not currently collected (see section 1). When the planned opt-in collection ships, it will be retained in aggregated form only, with individual records dropped after the audit window closes.
Candidates can request earlier deletion via the candidate-data form — see Data Handling.
4. Where it lives
Audio is not stored (see section 3). Transcripts, structured reports, scores, and recruiter notes live in a Postgres database hosted on Neon in us-east-1 with at-rest encryption. Vercel runs the application and edge functions in iad1 by default.
Subprocessors and the exact data-flow diagram live in the Data Handling page.