liveBook · Manning

Know your customer

As explained in Fight Fraud with Machine Learning

Know your customer is a customer-identification and verification process used by financial institutions and other regulated entities.

Know Your Customer (KYC) verifies customer identities and helps regulated businesses prevent financial crime before and after onboarding. It combines customer identification, customer due diligence, and ongoing monitoring. section 7.0 section 7.4

KYC divides operational checks into three connected areas:

  • Customer identification checks whether a customer is who they claim to be, typically using an identity document and selfie. section 7.0
  • Customer due diligence assesses trust and risk from metadata such as device type, location, credit score, and salary. section 7.0
  • Ongoing monitoring repeatedly checks transactions and customer-data changes for fraud. section 7.0

Automated selfie–ID verification

A digital banking signup can collect a selfie and passport photo. Face detection locates each face, face recognition produces embeddings, and cosine similarity determines whether the faces match. MTCNN and InceptionResnetV1 support this workflow through facenet_pytorch. section 7.4

Figure 7.1 The 3 different components of KYC along with the various types of KYC tasks listed per component.

Use the component map to navigate customer identification, due diligence, ongoing monitoring, and their associated KYC tasks. section 7.0

How Fight Fraud with Machine Learning covers it

How deep it goes
Decomposes KYC into three components and specifies automated selfie, document, metadata, and face-matching checks.
Examples
Uses digital-app signup, selfies, identity documents, metadata, transaction monitoring, face embeddings, and cosine similarity.
What you can do after
Build it from scratch by combining document, selfie, metadata, and monitoring checks with suitable machine-learning models.
How it connects
Connects customer identification, customer due diligence, ongoing monitoring, computer vision, tabular machine learning, and face recognition.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Fight Fraud with Machine Learning. The section links point to the parts of the book it draws on.