How Risk Scoring Works

Floodlight AML uses a transparent, explainable risk scoring algorithm that combines multiple signals.

Overview

The risk score is a numeric value from 0 to 100, where higher scores indicate greater risk. The score is calculated by evaluating matches across multiple risk vectors:

• Sanctions — Direct hits on sanctioned individuals/entities
• Entity — Entity-level risk attributes and profile data
• Transaction — Transaction pattern and behavior analysis
• Behavioral — Behavioral anomalies and unusual activity
• Provider — Third-party provider risk signals
• Country — Country-level risk factors and jurisdictional concerns

Overall Risk Levels

The highest match overall score across all results determines the risk level:

Match Score Structure

Each match contains a structured score object with sub-component breakdown:

{
  "score": {
    "overall": 94.0,
    "nameScore": 100.0,
    "idScore": null,
    "dobScore": 100.0,
    "nationalityScore": null,
    "algorithm": "BALANCED"
  }
}

• nameScore — Fuzzy + exact name match (0–100)
• idScore — Document/passport number match (null if no ID provided)
• dobScore — Date of birth match (null if no DOB provided)
• nationalityScore — Always null (nationality mismatch applies penalty but no positive score)
• overall — Weighted composite of available sub-scores (0–100)
• algorithm — LOOSE (high recall), BALANCED (default), or STRICT (high precision)

Null sub-scores do not penalize the overall — they simply mean the data dimension was unavailable for comparison.

Match Response Example

{
  "matchId": "match_001",
  "matchedName": "John Smith",
  "score": {
    "overall": 94.0,
    "nameScore": 100.0,
    "idScore": null,
    "dobScore": 100.0,
    "nationalityScore": null,
    "algorithm": "BALANCED"
  },
  "sources": [
    {
      "type": "SANCTION_LIST",
      "authority": "National Sanctions Authority",
      "program": "sanctions_list_main",
      "referenceId": "sanc-ref-001234",
      "publisher": "Government Treasury Department",
      "publishedDate": "2024-01-15"
    }
  ],
  "entity": {
    "type": "INDIVIDUAL",
    "nationality": "US",
    "dateOfBirth": "1985-03-15",
    "positions": ["Senior Executive"]
  }
}

Customizing Thresholds

Use minScoreThreshold to filter matches below a minimum similarity score, ormatchAlgorithm to choose the matching profile:

{
  "fullName": "John Smith",
  "matchAlgorithm": "STRICT",
  "minScoreThreshold": 80
}