Introduction to Moov Watchman

Core Functionality

Watchman is a robust compliance screening tool that provides:

  1. Data Management:
    • Automatic downloading of sanctions lists (US OFAC, US CSL, UK, EU, etc.)
    • Regular refreshing of data to maintain compliance
    • Custom data file support for specialized screening needs
  2. Search Capabilities:
    • High-performance indexing and search functionality
    • Advanced fuzzy matching using Jaro-Winkler algorithms
    • Multi-field search with entity type filtering
  3. Integration Options:
    • HTTP API for web and service integration
    • Native Go library for direct implementation
    • Webhook notifications for automated workflows

Available Libraries and Components

Component Purpose
OFAC Library Parse and process US OFAC sanctions data
US CSL Process the US Consolidated Screening List
UK/EU CSL Handle UK and European sanctions data formats

Search Methodology

Jaro-Winkler Similarity Algorithm

Watchman uses the Jaro-Winkler distance algorithm to score the similarity between search queries and list entries. This approach:

  • Matches the methodology used by US Treasury’s OFAC Search
  • Is specifically optimized for person names and other proper nouns
  • Produces scores from 0.0 (completely different) to 1.0 (exact match)
  • Has been validated by academic research as effective for compliance screening

Search Customization

Watchman offers environment variables to adjust search behavior:

  • EXACT_MATCH_FAVORITISM: Controls weight given to exact matches (recommended values: 0.1, 0.25, or 0.5)

Common Questions

How are entities prepared for the search index?

Entities undergo a multi-step preparation process before being indexed:

  1. SDN Name Reordering 
    "MADURO MOROS, Nicolas" → "Nicolas MADURO MOROS"
  2. Company Name Cleanup 
    "ACME CORPORATION, INC." → "ACME CORPORATION"

    Legal suffixes like “CO.”, “INC.”, “L.L.C.” are removed

  3. Stopword Removal 
    "TREES AND EQUIPMENT LTD" → "TREES EQUIPMENT LTD"

    Common words like “and”, “the”, “of” are removed

  4. UTF-8 Normalization 
    "Raúl Castro" → "raul castro"

    Punctuation is removed, text is lowercased, and diacritical marks are normalized

The resulting normalized names enable more accurate matching across different formats and variations of the same entity.

What’s the difference between Watchman’s search and standard text search?

Standard text search typically relies on exact matches or simple wildcards, which can:

  • Miss alternative spellings
  • Fail to handle name inversions
  • Be overly sensitive to typos
  • Require multiple manual searches

Watchman’s fuzzy matching approach allows for:

  • Identification of similar names despite variations
  • Tolerance for typographical errors
  • Handling of word order differences
  • Normalization of international character sets
  • Confidence scoring to prioritize results

This produces more comprehensive screening with fewer false negatives while still providing the tools to manage false positives effectively.

Next Steps