Manual Review Process

The manual review interface enables reviewers to thoroughly evaluate verification cases that require human oversight. This guide explains how to effectively use the interface and make informed decisions.

Interface Overview

The interface is organized into several key sections for efficient review:

1. Navigation & Status Bar

• User identification and session details
• Current review status indicator
• Session navigation controls
• Quick action buttons (Approve/Decline)

2. Document Verification Panel

• High-resolution document images (front, back, additional pages)
• Extracted data for different document fields
• Inline editing capabilities for data correction
• Document authenticity indicators

3. Address Verification

• Extracted address from the document
• Advanced IP analysis with geolocation data

4. Facial Recognition

• Document photo alongside the user's selfie for easy comparison
• Facial similarity score
• Liveness score

5. AML Screening

• AML hits with their hit scores

6. Warning System

• Warning indicators
• Tooltip descriptions

7. Additional Information

• Session number
• Device & Browser
• Time & Location
• Vendor data
• Session info

Conducting an Effective Review

Step 1: Initial Assessment

1. Review any system-generated warnings at the top of the interface
2. Check the overall risk score and contributing factors
3. Note any previous verification attempts by the same user

Step 2: Document Analysis

1. Examine security features (holograms, patterns, watermarks)
2. Verify data consistency across all document fields
3. Compare extracted data with visible document information
4. Check document expiration and issuance dates

Step 3: Biometric Verification

1. Compare facial features systematically:
   • Eye shape and color
   • Nose structure
   • Facial proportions
   • Any distinctive marks
2. Consider age difference between document photo and selfie

Step 4: Risk Signals

1. Evaluate AML screening matches:
   • Match confidence scores
   • Date of birth proximity
   • Name variations
2. Analyze location data:
   • IP address location vs document address
   • VPN/proxy detection results
3. Review device intelligence:
   • Browser/device consistency
   • Session behavior patterns

Making a Decision

Approval Criteria

• All document security features present and valid
• High confidence in facial match
• No significant AML hits
• Consistent location and device data
• Clear, legible document images

Decline Criteria

• Suspected document tampering
• Low facial similarity score
• Significant AML matches
• Suspicious location or device patterns