DIREKT Fraud Prevention¶
Threats¶
- forged/edited documents;
- stolen identity;
- duplicate provider profiles;
- fake premises;
- location spoofing;
- reviewer/field-agent collusion;
- purchased/fake reviews;
- account takeover;
- payment/webhook manipulation;
- enforcement evasion;
- customer scam enquiries.
Controls by stage¶
Account¶
Contact verification, rate limits, session/device monitoring, recovery controls and privileged MFA.
Provider¶
Duplicate matching on normalized contacts/business identity/location/evidence hashes; manual review for ambiguous matches.
Evidence¶
File signature/metadata checks, immutable versions, issuing-body verification where authorized, visual/manual review, expiry and random rechecks.
Field¶
Assignment binding, check-in/out, structured capture, supervisor sampling and conflict disclosure.
Marketplace¶
Tracked review eligibility, anomaly detection, report/appeal and restrictions.
Payments¶
Signed webhooks, idempotency, ledger and reconciliation.
Fraud flags¶
Flags are internal indicators, not guilt. Each includes source, confidence, status, owner and resolution. High-impact action requires human review unless immediate protection is necessary.
Response¶
Preserve relevant evidence, restrict access, stop public harm, investigate proportionately, record decision, notify affected users where required and review systemic gaps.
Metrics¶
Duplicate rate, forged-evidence rate, reviewer disagreement, re-verification failure, review anomaly, account-takeover reports and loss/complaint impact.