Part 2: Identity, KYC, and AML

Chapters 6–11

"Know your customer. It sounds so simple. Three words that contain the entirety of the problem."

Why This Part Exists

The anti-money laundering compliance domain is, by most measures, the most resource-intensive area of financial crime compliance. AML programs at large financial institutions employ thousands of people globally and consume billions of dollars annually. Despite this investment, the United Nations estimates that only 1–2% of illicit financial flows are seized by law enforcement.

This gap between investment and effectiveness is the central problem that RegTech in the KYC and AML space is attempting to solve — not by spending more, but by spending differently.

Part 2 covers the full lifecycle of financial crime compliance as it is practiced in financial institutions today: from the moment a prospective customer first applies for an account (Chapter 6), through the ongoing monitoring of their transactions for signs of money laundering (Chapter 7), the screening of their names against lists of sanctioned parties (Chapter 8), the analysis of who ultimately controls the legal entities they operate through (Chapter 9), the assignment and maintenance of a risk rating that determines the intensity of ongoing monitoring (Chapter 10), and the processes for reporting suspicious activity to law enforcement when monitoring generates credible concerns (Chapter 11).

Each chapter contains significant Python code examples that illustrate the technology — transaction monitoring algorithms, sanctions screening fuzzy matching, SAR narrative generation — in enough detail to build genuine understanding.

Characters in Part 2

Maya Osei's Verdant Bank KYC program build-out enters its most intensive phase here. The KYC backlog is closed; now Maya must build a KYC process that scales sustainably. Her central challenge: designing an onboarding journey that is fast enough to satisfy digital-native customers and rigorous enough to satisfy the FCA.

Rafael Torres at Meridian Capital begins the AML monitoring overhaul that has been on the project plan since Chapter 1. The data migration issues from Chapter 5 are behind him; now the focus is on model calibration, analyst workflow integration, and the delicate process of transitioning from the old system to the new without creating compliance gaps.

Priya Nair is advising three clients through simultaneous AML implementations — a large UK asset manager, a Singapore-based bank, and a Caribbean money service business. The three implementations reveal how differently the same regulatory obligation plays out in different institutional contexts.

Cornerstone Financial Group provides the institutional-scale case studies — the beneficial ownership challenge in Chapter 9, a cross-border correspondent banking review in Chapter 8, and a SAR program quality assessment in Chapter 11.

Learning Objectives for Part 2

By the end of Part 2, you will be able to:

  1. Describe the legal basis and practical implementation of KYC obligations, including digital identity verification methods
  2. Compare rules-based and ML-based approaches to AML transaction monitoring and evaluate their relative strengths and weaknesses
  3. Explain the technical challenges of sanctions screening — fuzzy matching, false positive management, list calibration — and their compliance implications
  4. Describe the beneficial ownership disclosure requirements created by the Corporate Transparency Act and EU 6AMLD and the technology approaches to corporate transparency
  5. Apply the risk-based approach to customer risk rating, including PEP screening and adverse media monitoring
  6. Explain the SAR filing process, the legal protections and obligations it creates, and the role of case management systems in supporting it

Technical Preview

Part 2 contains some of the most practically applicable Python code in the textbook: - Chapter 6: Digital identity verification pipeline - Chapter 7: Transaction monitoring model training and evaluation - Chapter 8: Fuzzy matching for sanctions screening - Chapter 10: Customer risk scoring model

The code in this part requires pandas, numpy, scikit-learn, and (in Chapter 8) the fuzzywuzzy/rapidfuzz library. Ensure requirements.txt dependencies are installed before running.

Continue to Chapter 6: KYC Fundamentals →

Chapters in This Part