Human oversight and supervisors
Automation handles repetitive document checks and data extraction, but humans review every unclear, risky, or high-value case. Supervisors provide final sign-off and escalate sensitive outcomes to compliance teams.
Oversight layers
Automated risk flags
Systems flag incomplete checklists, sanctions hits, expired documents, mismatched data, or low-confidence biometric scores. Flagged cases pause and route to specialist agents.
Agent escalation protocols
Agents review flagged cases, request additional documents, re-run checks, or escalate to supervisors. Every action is logged with reasoning and evidence links.
Supervisor final approval
Supervisors review the full case history, check evidence quality, and approve or reject outcomes. High-risk approvals require two supervisors (four-eyes principle).
Quality assurance mechanisms
Daily sampling and calibration
Quality teams sample 5% of completed verifications daily to check for missed red flags, over-reliance on automation, or inconsistent decision-making. Findings trigger agent retraining.
Weekly calibration sessions
Supervisors and quality leads review edge cases together to align decision standards. Session notes update internal playbooks so future cases are handled consistently.
Audit trail visibility
Every verification logs who viewed it, what checks were run, when evidence was uploaded, and who approved the outcome. Regulators and clients can request audit trails for specific cases.
Escalation to compliance and legal teams
Cases involving sanctions hits, adverse media matches, serious criminal records, or disputed identity are escalated to Cleared\'s compliance and legal teams. These teams decide whether to release results, request additional evidence, or decline the verification. Escalation decisions are logged and available for regulatory audits.