NatWest Group’s purpose is to champion potential by helping people, families, and businesses to thrive. A key way NWG realises that purpose is by protecting its customers from financial fraud, which cost people in the UK a total of £1.2bn in 2022.
This is a significant challenge, as fraudsters continuously devise innovative ways to circumvent existing fraud detection systems, creating new types of fraud that NWG’s fraud detection systems might not recognise. This leads to a “whack-a-mole” situation, where the machine learning models that underpin NWG’s fraud detection systems must be regularly updated to adapt to the changing tactics of the fraudsters.
The data used to train these models comes from a wide variety of sources, such as transactional, personal, and device data. However, the vast amount of data available to inform machine learning-based fraud detection algorithms is not fully utilized due to the difficulty in connecting and creating features from non-structured data sources, such as suspicious activity reports and Companies House records. This project will focus on developing cutting-edge artificial intelligence algorithms that provide a step-change in the capabilities of systems for detecting fraudulent transactions.