Combating Money Laundering: Enterprise AML Technology

Money laundering is among the hardest activities to detect in the world of financial crime. Funds move in plain sight through standard financial instruments, transactions, intermediaries, legal entities and institutions – avoiding detection by banks and law enforcement. The costs in regulatory fines and damaged reputation for financial institutions are all too real. Neo4j provides an advanced, extensible foundation for fighting money laundering, reducing compliance costs and protecting brand value.

In this third blog in our six-part series on using graph technology to fight money laundering, we examine specific characteristics and components of the Neo4j graph platform that make it ideal for addressing the stringent requirements of anti-money laundering.

Neo4j: The Only Enterprise Graph Database

Traditional AML technologies store data in tables of rows and columns, or in documents.
Pattern matching requires the data to be retrieved, parsed and JOINed. To attempt all this,
they rely heavily on indexes. While fast performance might be possible for simple pattern
matching, it is impossible with complex patterns or large volumes of data.

In contrast to these other technologies, Neo4j stores data natively as a graph, an approach
known as index-free adjacency. Each node directly references its adjacent nodes, acting as a
micro-index for all nearby nodes. As a result, Neo4j traverses paths with blazing speed and
does not rely on indexes at all for the traversals, but rather, simply chases pointers in memory.

Because Neo4j natively processes data as a graph, nothing gets lost in translation while
interrogating the graph. Neo4j’s end-to-end processing and retrieval is blazingly fast and
accurate, even at extreme scale.

Neo4j is ideal for addressing the stringent requirements of anti-money laundering because it
delivers:

• Blazingly fast query performance at extreme scale for detecting suspicious patterns and
  behaviors in real time
• In-memory analytics pipelines to provide scored results in real time
• Highly optimized data science library highlights communities and central entities at a high
  level
• Causal Clustering to ensure an always-available database and disaster recovery
• Data ingestion and management to process billions of records a day
• Trusted integrity and dependability through its ACID graph transactions
• Enterprise-grade security that is schema-based for fine-grained access, traversals and reads
• Extreme agility, extensibility and customizability from its whiteboard data models
  and Cypher query language

Neo4j uses a native-graph language called Cypher. It is a descriptive yet concise programming
language that looks like plain English and requires only a few lines of code where other
languages require hundreds of lines. Industry experience shows that an application written
with 100,000 lines of code takes 100 times longer to develop and maintain than the same
application with 1,000 lines of code written in a more efficient language.

Neo4j: The Only Enterprise Graph Platform

Neo4j provides a flexible and robust end-to-end graph platform that extends from AML to
other global risk and compliance (GRC) areas that use a lot of the same connected data.
Neo4j further accelerates time to value by making it easy to integrate with its ecosystem.

Data Modeling Tool

Cypher Workbench
Model the graph and synchronize Neo4j with the model. Also generate Cypher code
and scripts.

Input APIs

Hop
Move source data the last mile into Neo4j via this robust, model-driven, open source ETL
and orchestration tool. Hop connects directly to data sources via connectors or can pull
files from a folder. Performs data validation, look-ups, normalization, scheduling, error
handling, parallelization, logging/lineage, auditing and other manipulations.

Native Neo4j drivers
Feed data from applications written in modern languages and scripts such as Python,
Java, JavaScript, .NET and Go. Community-driven drivers exist for other development
languages.

Output APIs

Neo4j BI Connector
Deliver direct access to Neo4j graph data from business intelligence (BI) tools such as
Tableau, Looker, TIBCO Spotfire Server and Microstrategy using the Neo4j BI Connector.

GRANDstack: GraphQL, React, Apollo, Neo4j Database
Build custom dashboards via the GRANDstack.

Bloom Self-Service Graph Data Visualization
Explore and discover with no coding required using Neo4j Bloom.

Neo4j Browser
Construct Cypher queries and visualize results in the Neo4j Browser.

Native Neo4j Drivers
Feed downstream applications using language drivers for Neo4j Database.

Analytics

Neo4j Graph Data Science Library (GDSL)
Deliver optimized graph algorithm performance at scale via GDSL.

Keymaker
Build queries for each step in an AML scoring process and pipeline the queries in memory.

Packaging

Package Neo4j frameworks for Windows, macOS and most common Linux distributions.

Neo4j Advanced Money Queries

The table below explains the basic types of money queries used in the Neo4j AML Framework
and how they can be used to build sophisticated AML detection algorithms.

Pattern Matching with Graph Algorithms

Neo4j is the only native graph database that has a robust graph data science library which:

• Frees analysts from moving large volumes of data to integrate scoring algorithms
• Leverages relationships that were previously untapped by AI/ML pipelines

Neo4j has a dedicated team of engineers developing and maintaining a comprehensive, high-performance Graph Data Science Library (GDSL). All of Neo4j’s graph algorithms – with the exception of similarity algorithms – directly leverage relationships in the graph. A majority are intended for monopartite graphs (those with a single-node label).

Similarity algorithms, such as Jaccard and Cosine, use relationships to preprocess data. These
similarity and graph algorithms traverse the graph, in some cases multiple times. For this
reason they all execute in an in-memory projected graph and their resulting computations
are either streamed or stored back into the Neo4j Database. This makes the Neo4j Graph
Platform optimal for executing, storing, retrieving and processing these algorithms.

Conclusion

Neo4j provides a flexible and robust end-to-end graph platform that extends from AML to
other global risk and compliance (GRC) areas that use much of the same connected data.

Next week, in blog four of our six-part series on fighting money laundering with graph technology, we will give an overview of a reference implementation for building a simple solution.