Danske Bank turns to Teradata for AI-powered fraud detection
The AI has “already reduced false positives by 50%”
Copenhagen-based Danske Bank has gone public on its deal with Think Big Analytics, the consulting arm of Teradata, to create and launch an artificial intelligence (AI) driven fraud detection platform.
The announcement is all timed sweetly to be part of the Teradata Partners Conference 2017 (22-26 October), attended by Banking Technology and held in Anaheim, California. Teradata also revealed its new data analytics platform at the show – and unveiled two other initiatives – Analytics-as-a-Service and a new software portfolio.
In terms of the engine, this uses machine leaning to analyse latent features, scoring online banking transactions in real-time to provide insight regarding true, and false, fraudulent activity.
Nadeem Gulzar, head of advanced analytics, Danske Bank, says by using AI it has “already reduced false positives by 50% and as such have been able to reallocate half the fraud detection unit to higher value responsibilities”.
According to Teradata, the bank’s original fraud detection system was “largely based on handcrafted rules that had been proactively applied by the business over time” and there were “record numbers of false positives – at times reaching 99.5% of all transactions”.
The Think Big Analytics team began working with Danske Bank in autumn 2016, to “augment their analytics team with specialist knowledge”.
The joint team began with building a framework within the bank’s existing infrastructure and then created machine learning models to “detect fraud within millions of transactions per year, and in peak times, many hundreds of thousands per minute”.
The engine includes an interpretation layer on top of the machine learning models, providing explanations and interpretation of blocking activity.
From a modelling point of view, Teradata says fraud cases are “still very rare”, with around one fraud case in every 100,000. The team has taken the false positives from the models and reduced them by 50%. At the same time, Teradata says they can catch more fraud – upping the detection rate by around 60%.
Danske Bank’s anti-fraud programme is the “first to put machine learning techniques into production while simultaneously developing deep learning models to test the techniques”.
