At the end of October 2023, a pioneering data sharing agreement was implemented through a partnership with Pay.UK that gives Featurespace’s data scientists access to an expansive data set of Faster Payments from a large spectrum of banks and building societies in the UK. The aim of the data sharing project is to enable the use of innovative data science approaches to prove that value can be added on top of existing controls when the industry collaborates.
The partnership will allow for the discovery of cutting-edge ML techniques not yet in the market, leveraging the unique opportunity presented by having such a large volume of payments from a huge volume of customers spread across a sizable and diverse group of banks.
The launch of this innovative proof of concept comes at a time when pressure to combat fraud is rising for banks. The UK’s Financial Conduct Authority (FCA) is asking banks to “do more to detect, manage and reduce fraud and losses more effectively”. New insights into performance have also been published by the Payment Systems Regulator (PSR) who released the first set of APP performance figures in the UK. The figures show that while some Payment Service Providers (PSPs) have lower losses than others, no one has room to be complacent as nearly all will expect increased refunding rates and/or significant liability from receiving accounts. The other interesting view from the PSR data is that it suggests that some segments of PSPs have higher victim and/or mule propensity than others and this type of fraud is exactly what the machine learning model built for this proof of concept will aim to detect and prevent.
The Featurespace Innovation Lab is excited to get started on this proof of concept and there couldn’t be a better team to tackle the challenge based on previous inventions, such as Automated Deep Behavioral Networks and TallierLTM™ that use Artificial Intelligence and machine learning techniques to detect more fraud than other approaches.
Having developed a prize winning Privacy Enhancing Technology for data sharing, the Innovation Lab’s experience for developing future solutions is advantageous. With the introduction of the PSR APP reimbursement policy in October 2024, the reimbursement rate is expected to increase significantly and adds more incentive for financial institutions to detect and prevent APP fraud in real time. The splitting of liability between sending and receiving PSPs means that many are looking for real time inbound decisioning and the PSR figures suggest there are some great opportunities for improvement.
Featurespace’s response to support the industry to stay ahead of APP fraud was the recent launch of ARIC™ Scam Detect, which has shown uplifts in fraud detection performance for both outbound and inbound payments.
If you would like to know more about the Pay.UK POC and its results, or how any other Featurespace solutions can help you detect fraud and improve performance quickly, contact us.