About Fraud Prevention

TrustGate provides comprehensive fraud prevention tools to detect and block fraudulent applications, protecting your business from identity fraud, synthetic identities, and account takeover.

Fraud Landscape

Common Fraud Types

Fraud Type	Description	Detection Method
Identity Fraud	Using stolen identity documents	Document verification, biometrics
Synthetic Identity	Fabricated identities using mixed data	Data correlation, device intelligence
Document Fraud	Altered or fake documents	AI document analysis
Account Takeover	Hijacking existing accounts	Device fingerprinting, behavioral analysis
Application Fraud	Multiple fraudulent applications	Velocity checks, deduplication

Fraud Prevention Stack

┌─────────────────────────────────────────────────────────┐
│                  FRAUD PREVENTION                        │
├──────────────┬───────────────┬───────────────┬──────────┤
│   Document   │   Biometric   │    Device     │ Velocity │
│   Analysis   │  Verification │ Intelligence  │  Checks  │
├──────────────┼───────────────┼───────────────┼──────────┤
│ • Tampering  │ • Face match  │ • Fingerprint │ • Rate   │
│ • Forgery    │ • Liveness    │ • Risk score  │ • Volume │
│ • Expiry     │ • Deepfake    │ • Geo/IP      │ • Pattern│
└──────────────┴───────────────┴───────────────┴──────────┘
                            │
                            v
┌─────────────────────────────────────────────────────────┐
│              RISK ENGINE + ML MODELS                     │
│      Combines signals for fraud probability score        │
└─────────────────────────────────────────────────────────┘
                            │
                            v
┌─────────────────────────────────────────────────────────┐
│              AUTOMATED DECISIONS                         │
│      Accept / Review / Reject based on thresholds        │
└─────────────────────────────────────────────────────────┘

Key Features

Document Fraud Detection

AI-powered analysis detects:

• Physical tampering: Cut-and-paste attacks, laminate manipulation
• Digital manipulation: Photoshop edits, font inconsistencies
• Forgery indicators: Template mismatches, security feature absence
• Data consistency: MRZ validation, date logic checks

Biometric Verification

Verify the person matches their documents:

• Face matching: Compare selfie to document photo
• Liveness detection: Confirm live person, not photo/video
• Deepfake detection: Identify AI-generated faces
• Passive liveness: Detect spoofing without user action

Device Intelligence

Assess risk from device signals:

• Device fingerprinting: Identify unique device characteristics
• Fraud scoring: Risk assessment based on device history
• Geolocation: IP location vs claimed location
• Proxy/VPN detection: Identify connection masking

Velocity Controls

Detect rapid fraudulent activity:

• Application velocity: Multiple apps from same device/IP
• Document reuse: Same document across applications
• Data clustering: Shared data points across applicants

Getting Started

Enable Fraud Checks

curl -X POST https://api.bytrustgate.com/v1/verifications \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "applicant_id": "550e8400-e29b-41d4-a716-446655440000",
    "checks": [
      "document_verification",
      "biometric_verification",
      "device_intelligence"
    ],
    "fraud_detection": {
      "enabled": true,
      "sensitivity": "high"
    }
  }'

Fraud Score Response

{
  "verification_id": "ver_123456",
  "fraud_assessment": {
    "fraud_score": 85,
    "fraud_risk": "high",
    "recommendation": "reject",
    "signals": [
      {
        "signal": "document_tampering",
        "confidence": 0.92,
        "details": "Photo area shows signs of digital manipulation"
      },
      {
        "signal": "device_previously_fraudulent",
        "confidence": 0.88,
        "details": "Device linked to 3 previous rejected applications"
      }
    ]
  }
}

Fraud Score Ranges

Score	Risk Level	Recommendation
0-20	Very Low	Auto-approve
21-40	Low	Standard process
41-60	Medium	Additional review
61-80	High	Manual investigation
81-100	Very High	Likely reject

Fraud Detection Settings

Configure Sensitivity

curl -X PUT https://api.bytrustgate.com/v1/settings/fraud-detection \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "sensitivity": "high",
    "auto_reject_threshold": 85,
    "manual_review_threshold": 60,
    "checks_enabled": {
      "document_tampering": true,
      "biometric_spoofing": true,
      "device_intelligence": true,
      "velocity_checks": true,
      "data_correlation": true
    }
  }'

Sensitivity Levels

Level	False Positive Rate	Detection Rate
Low	~1%	~80%
Medium	~3%	~90%
High	~5%	~95%
Very High	~10%	~99%

Fraud Prevention Metrics

View Fraud Statistics

curl -X GET "https://api.bytrustgate.com/v1/analytics/fraud?period=30d" \
  -H "Authorization: Bearer YOUR_API_KEY"

{
  "period": "30d",
  "total_applications": 10000,
  "fraud_detected": 450,
  "fraud_rate": 0.045,
  "breakdown": {
    "document_fraud": 180,
    "biometric_fraud": 120,
    "synthetic_identity": 80,
    "application_fraud": 70
  },
  "false_positive_rate": 0.028,
  "financial_impact": {
    "estimated_fraud_prevented": "$2,250,000",
    "avg_fraud_value": "$5,000"
  }
}

Best Practices

1. Layer defenses: Use multiple fraud signals together
2. Tune thresholds: Balance fraud prevention vs customer friction
3. Review edge cases: Manually review borderline scores
4. Monitor trends: Watch for new fraud patterns
5. Update rules: Adapt to evolving fraud tactics

---

Next Steps

• Device Intelligence - Device fingerprinting and risk scoring
• Email Verification - Email risk assessment
• Phone Verification - Phone number validation
• Risk Signals - All fraud signals explained