Product

Fraud shows up at every stage of the funnel. So does Palladium.

A single checkpoint — an ID scan at application — misses most of it. Fraud rings specifically route around one-time checks. Palladium watches the whole interview process instead of one moment in it.

Where fraud happens

Five stages, five different risks

StageRisk
ApplicationAI-written resume, stolen ID
ScreenDeepfake video, voice mismatch
Technical roundProxy solves the assessment via remote access
Final roundConfidence/skill jumps vs. earlier rounds
Day onePerson who shows up isn't the person interviewed
The four pillars

What Palladium actually watches

Identity-document verification is already commoditized by ATS-native tools. Palladium covers the layer they don't.

1

Live deepfake detection

Real-time analysis during the actual video call — lip-sync artifacts, lighting/shadow consistency, blink rate, and compression artifacts characteristic of face-swap pipelines. Not a one-time selfie check.

2

Proxy & remote-access detection

Flags remote-desktop signatures, unusual display configurations, and screen-share anomalies during live coding or technical rounds — the setup proxy rings actually use.

3

Cross-round consistency

Compares voice and behavioral signatures across every interview round. A confident round-one screener and a different person in the final round is exactly the pattern proxy fraud produces.

4

Risk score in your ATS

Every signal feeds one composite risk score, surfaced on the candidate's card in Greenhouse or Ashby, with a full evidence trail — not a black-box pass/fail.

How it works

Connect, capture, detect, surface

StepWhat happens
1. ConnectInstalls as a video-call add-on and an ATS marketplace app. No change to the recruiter's workflow.
2. CaptureDuring each interview round, audio and video signals are analyzed in real time — nothing recorded beyond what the ATS already stores.
3. DetectDeepfake, remote-access, and voice-consistency models score the session; anomalies are timestamped and flagged.
4. SurfaceA single risk score and evidence trail appear on the candidate's card, next to the scorecard recruiters already fill out.
Where we're starting

MVP: narrow, fast to trust, easy to integrate

V1 — in scope

Real-time deepfake detection on live video, remote-access & proxy detection during technical rounds, cross-round voice consistency, risk-score dashboard embedded in Greenhouse & Ashby, basic resume/ID red-flag screening.

Phase 2

Full government-ID/biometric document verification (partner, not rebuild), keystroke-dynamics assessment integrity, day-one post-hire re-verification, additional ATS platforms, enterprise SSO.

Early access

Be first in line as a design partner

You're on the list — we'll be in touch.