Safe hiring
Risk assessment
Automated checks across collection, scoring and results that flag candidates trying to game the system.
- <20 minto first interview
- 4.5★candidate rating
- 11fraud signals detected
Risk analysis across every signal
Admin dashboard
Risk signals for every candidate at a glance. Sort, filter and drill into evidence.
- Risk score per candidate
- Filter and sort by risk type
- Drill into evidence and recordings
Video analysis
Video analysis spots reading from screen, multiple faces and identity drift.
- Eye-movement reading detection
- Multiple-face detection
- Biometric face match against KYC
Audio analysis
Audio analysis catches second voices, AI-generated speech and voice mismatch.
- Multiple-voice detection
- Biometric voice fingerprinting
- AI voice synthesis detection
Sorry, can you repeat the question?
Stall tactic
Conversational analysis
Detect tactics in the conversation flow that suggest external help or duplicate identity.
- Stall-tactics detection
- Asking for help
- Multiple candidate profiles
LocationFrankfurt, DE
VPNDetected
TorActive
Session analysis
Connection and device signals reveal location obfuscation and unusual sessions.
- VPN detection
- Tor and anonymized networks
- Suspicious session patterns
Candidates increasingly use AI to game interviews. Enable risk assessment and suspicious conversation, audio and video signals are flagged on the result screen, with thresholds per role.
All detected risk signals
What the AI flags, with each signal linked to evidence in the session.
| Signal | Category | What it detects |
|---|---|---|
| Reading from screen | Video | Eye-movement analysis flags reading prepared text instead of speaking naturally. |
| Multiple faces | Video | More than one person visible in the candidate's camera frame. |
| Face mismatch | Video / KYC | The interview face does not match the candidate verified during KYC. |
| Multiple voices | Audio | More than one voice detected in the candidate's microphone stream. |
| Voice fingerprint mismatch | Audio | Candidate voice does not match the enrolled biometric fingerprint. |
| AI voice synthesis | Audio | Audio shows characteristics of synthetic or AI-generated speech. |
| Stall tactics | Conversation | Candidate stalls or asks unrelated questions to buy time during answers. |
| Asking for help | Conversation | Candidate appears to direct questions or signals to a third party in the room. |
| Multiple profiles | Conversation | Same person applies multiple times under different identities. |
| VPN connection | Session | Connection routed through a VPN that may hide the candidate's true location. |
| Tor / anonymized network | Session | Connection routed through Tor or similar anonymizing proxies. |
Ready to interview smarter?
Run AI-powered interviews on your next role in minutes. Test for free, subscribe when ready.