The complyAdvantage Knowledge Graph

Last updated: September 3, 2026


The ComplyAdvantage Knowledge Graph is the proprietary risk database that sits underneath everything ComplyAdvantage offers. It is the structured, always-updating picture of the people, companies, vessels and aircraft the financial system needs to be careful about, and crucially, the relationships between them.


Good to know

A normal sanctions list is a tabular list of names. Our Knowledge Graph is a connected map. It tells you not just that a person exists and is risky, but who they are connected to, what companies they direct, who their close associates are, where the risk originated, when it changed, and which underlying article or government list it came from.

Three Problems a Graph Solves

Financial crime risk doesn't sit in single rows of single lists. It sits in the interactions portrayed as relationships between people, companies, addresses, dates and articles. Here are the three practical problems the Knowledge Graph solves that a traditional list-based approach simply cannot.

Problem 1:

The Static-List Problem

Traditional list-based screening checks a customer against a snapshot. If a director of a previously-clean company is added to a sanctions list tomorrow, a snapshot-based system has no way to know that the customer's relationship to that director has just become a risk.

The UK Financial Conduct Authority has flagged exactly this — updates to one entity frequently do not cascade to related entities in a portfolio, leaving compliance teams reactive and exposed. A connected graph propagates that change automatically.

Problem 2:

The Shell-Company Problem

Money launderers obscure beneficial ownership behind chains of companies. The UN estimates around $1.6 trillion is laundered each year — about 2.7% of global GDP — and around $7 trillion in private wealth is hidden in haven countries.

Resolving who actually owns or controls an entity requires linking registry records, directorships, addresses and aliases across borders. That is graph work, not list work.

Problem 3:

The False-Positive Problem

Industry estimates put the share of alerts that ultimately turn out to be false positives at over 90%, with roughly $100 billion of compliance effort wasted on them annually.

Graph context — "is this John Smith the one who lives in Manchester and directs a UK company, or the sanctioned John Smith in Moscow?" — is one of the most direct ways to compress that number, because it adds disambiguating signal to an otherwise ambiguous name.

By the numbers

At a high level, the Knowledge Graph today contains:

  • 25M+ entities — people and companies connected by various relationships between them

  • ~2.5M sanction & PEP records — from sanctions lists and thousands of political exposure sources

  • ~5M watchlist records — from warning and fitness-probity lists and most-wanted registers

  • ~37M adverse media mentions — classified into 34 distinct financial crime categories

  • 8M documents per day — processed by the AI-powered ingestion engine