Fraud queue

Review careers applications flagged by fraud scoring and confirm or dismiss.

Feature flag: dei_compliance — Application fraud scoring and the flagged queue ship with the DEI & Compliance package.

When candidates apply through the careers page, InsightHire scores applications for spam and abuse signals (device fingerprint velocity, suspicious patterns, etc.). High scores land in the fraud queue for human review before recruiters invest review time.

Open the queue

Path: /dashboard/candidates/flagged

If dei_compliance is not granted, the page shows "Fraud review is not enabled."

Each row displays:

Candidate display name (or email)
Position title
Fraud score (numeric — higher = more suspicious)

Actions

Action	Effect
Dismiss	False positive — application proceeds through normal pipeline
Confirm spam	Marks application as fraudulent / rejects spam

Mutations call compliance.dismissFraudFlag and compliance.confirmFraudFlag; the list refreshes after each action.

Where scoring originates

Fraud scoring runs on careers apply submissions (/careers/[orgSlug]/[positionId]). Invited candidates and LinkedIn Apply Connect flows use separate integrity checks; they may not appear in this queue unless the same scoring path applies.

Threshold constant: applications at or above the platform FRAUD_REVIEW_THRESHOLD surface here (see application-fraud.service in the API).

Operational cadence

Check /dashboard/candidates/flagged daily when running paid job board campaigns.
Pair with Analytics → Sources (/dashboard/analytics/sources) — sudden careers spikes often precede fraud bursts.
Document dismiss vs confirm ratios for tuning campaigns.

Privacy and fairness

Fraud scores are risk signals, not automatic rejections. Always human-review before confirming spam. Do not use fraud scores as proxies for protected-class filtering — they measure abuse patterns only.

Applying — careers apply flow
Compliance overview