AI assistance
Hidden answer generation
Surface patterns consistent with off-screen copilots, teleprompters, and externally generated answers while the interview is running.
Evidence attached for reviewLIVE INTERVIEW FRAUD DETECTION
Know who joined.
Know whose answers
you are hearing.
Sherlock helps Talent, Compliance, Risk, and Security teams detect fraud in video interviews, from hidden answer assistance to identity fraud, with evidence your team can review, trust, and act on.
Identity fraud Misrepresentation fraud Reviewer-ready evidence Live & post-interview workflows
Listening↪Candidate entered the Interview11:28
⊘Unauthorized AI usage detected11:32
↔Candidate left the interview window11:32
↗Candidate switched to another tab11:29
↪Candidate entered the Interview11:28
Watching candidate activity in real time
Google MeetZoomMicrosoft TeamsHuman review stays in control
LIVE ANALYSIS
See suspicious signals while the interview is still happening.
Sherlock evaluates the session in context and turns scattered activity into a small number of reviewable moments. No single signal decides the outcome.
Michael L · CandidateCandidate activityReal-time integrity signals
Unauthorized AI usage detected12:21:27 · High-confidence signal
Candidate left interview window12:21:16 · Context captured
Tab switch or secondary display detected12:20:45 · Context captured
Identity remains consistentContinuous participant verification
Candidate entered the interview12:20:12 · Session initialized
Watching candidate activity in real time
WHAT SHERLOCK CHECKS
Four live signals. One evidence trail.
Sherlock combines multiple indicators and preserves the moments that matter, so reviewers can inspect the evidence rather than trust a score.
Response integrity
Scripted reading
Review repeated gaze, timing, and delivery patterns in context instead of treating one glance or pause as proof.
Evidence attached for reviewIdentity
Proxy candidates
Track identity consistency across the session and flag evidence that the participant may not be the expected candidate.
Evidence attached for reviewMedia authenticity
Deepfakes
Detect visual and audio inconsistencies that can indicate synthetic, reenacted, or manipulated interview media.
Evidence attached for reviewDEEPFAKE DETECTION
Detect synthetic identity signals in real time.
Sherlock examines the live video and audio stream for inconsistencies that can indicate facial reenactment, manipulated media, or cloned voice.
- Facial boundary and texture anomalies
- Audio-video timing inconsistencies
- Participant identity continuity
Human-reviewed evidence A flag is a reason to inspect the session—not an automatic accusation.
Realtime active
Real-time Alert Log
FaceHIGH13:42:33Face Anomalies
FaceHIGH13:42:30Face Anomalies
FaceHIGH13:42:27Face Anomalies
FaceHIGH13:42:24Face Anomalies
FaceHIGH13:42:21Face Anomalies
Feedback Analyzing
COVERT AI ASSISTANCE
Uncover the tools interviewers cannot see.
Detect signals consistent with AI copilots, teleprompters, audio relays, and hidden desktop assistance without asking interviewers to play detective.
- Hidden AI and background-process signals
- Off-screen answer reference patterns
- Covert audio and display assistance
Context over shortcuts Sherlock looks across the response instead of treating one glance as proof.
INSIDE THE SIGNAL LAYER
Evidence a reviewer can understand.
Every flag should answer three questions: what happened, when it happened, and why it requires attention.
Timestamped moments
Jump directly to the relevant part of the session.
Multimodal context
Review video, audio, identity, and behavior together.
Clear explanation
Understand the signal without decoding a black-box score.
Human decision
Keep hiring policy and final judgment with your team.
LIVE INTERVIEW REVIEW3 moments to inspect
Review statusNeeds human review
74/10012:14 Off-screen answer reference24:08 AI-assistance signal31:42 Identity continuity review
FROM INTERVIEW TO DECISION
A simple workflow for a complicated problem.
Sherlock adds an evidence layer to the live interview process without replacing the people responsible for the decision.
01
OBSERVEInterview
Sherlock joins the live interview workflow and observes the session in context.
02
ANALYZESignals
Identity, behavior, video, and audio signals are evaluated together as the interview unfolds.
03
EXPLAINEvidence
Relevant moments and signal context are attached to the session for authorized reviewers.
04
DECIDEHuman review
Your team clears, re-verifies, holds, or escalates according to its own hiring policy.
SECURITY & PRIVACY
Candidate data deserves enterprise-grade care.
Sherlock is designed for sensitive hiring workflows, with controlled access, inspectable evidence, and security practices built for enterprise review.
Protected data transport Authorized reviewer access Customer-aligned controls
Visit the Trust Center 
GDPR
ISO 27001
SOC 2
CPRAFAQ
Questions serious teams ask before rollout.
Clear answers for Talent, Compliance, Risk, and Security teams evaluating live interview fraud detection.
No. Sherlock is a decision-support system. It surfaces signals and the relevant evidence for an authorized human reviewer; your team remains responsible for the hiring decision.
Sherlock evaluates patterns across time and across modalities. A single glance, pause, or behavior is not treated as proof. When the evidence is incomplete, the result is routed for review rather than forced into a verdict.
Uncertainty is made explicit. Reviewers can see the moments that influenced the result and decide whether to clear the session, request re-verification, or investigate further.
Sherlock analyzes participant consistency and media-authenticity signals throughout the live session. It can flag evidence consistent with proxy participation, synthetic facial reenactment, manipulated video, or cloned audio for human review.
The workflow is designed to stay in the background. Interviewers continue their conversation while Sherlock records integrity signals and prepares reviewer-ready evidence.
Start with a controlled pilot using sessions where your team already knows the outcome. Compare known-good and known-bad cases, inspect the evidence, and agree on review and escalation policy before broader rollout.
LIVE ANALYSIS
Stop guessing whether the interview was real.
Pilot Sherlock on live interviews and inspect the evidence with your own hiring, compliance, and security teams.
Google Meet · Zoom · Microsoft Teams · Human-reviewed evidence