The procurement checklist, answered
VoxAide is built for regulated hiring and people conversations — consented, fair, auditable, and secure by design. Here is how each requirement is met.
Consent & recording
- ✓Every call opens with a recording disclosure and captures consent before any questions.
- ✓An append-only consent ledger records each capture (TCPA / recording / reference).
- ✓Reference & reactivation calls are opt-in first — no call before the person confirms.
Fair, auditable scoring
- ✓Scored on demonstrated answers against job-related criteria — never voice, accent, or identity.
- ✓Every scoring decision is logged in an immutable audit trail (model, prompt version, inputs).
- ✓Adverse-impact (four-fifths) bias auditing built in for AI-hiring oversight (NYC LL144 / EEOC).
Lawful by module
- ✓Frontline screening uses lawful, job-related questions only — no disability/medical inquiries; ADA accommodation path.
- ✓Negative outcomes at volume queue for human review — no auto-rejection.
- ✓Exit & listening are advisory and aggregated, never an individual employment decision, with anonymity floors.
Data security
- ✓Encryption in transit and at rest; row-level tenant isolation on every org-scoped table.
- ✓Call recordings served only through an authenticated, access-logged link — never a public URL.
- ✓Configurable retention with automatic purge + legal hold; optional field-level transcript encryption.
Working through a security review or DPA? Talk to a specialist — we support compliance and security questionnaires.
What VoxAide doesn't do
Trusting AI in hiring starts with knowing where it stops. VoxAide runs the first-round conversation and scores it — every judgement that matters stays with a person.
- Make the hiring decision — it scores and recommends; a person decides.
- Judge voice, accent, tone, or “culture fit” — only the substance of answers.
- Analyse faces or video — it's a phone conversation, scored on the words.
- Auto-reject — borderline candidates route to a human, never a silent no.
- Train AI models on your candidates' data.
The candidate experience
Before any call begins, every candidate is told they're speaking with an AI interviewer, that the call is recorded and transcribed, and that a real person reviews the result. Consent is asked for and recorded on every call.
They answer the same job-related questions everyone gets — on their own time, not just 9-to-5 — and are graded on what they say, never how they sound. If someone declines, the call ends and it's flagged for your team to follow up the usual way.