Singapore’s Ant International has deployed a new time-series transformer (TST) AI model, FalconTST 2.0, at Barclays, Citi, Deutsche Bank and Standard Chartered to improve forecasting accuracy and currency fluctuation management capabilities.
FalconTST 2.0 is designed to help businesses manage complex cash flows and financial exchange (FX) exposure. The platform reduces misleading forecasting patterns by distinguishing missing data, like absent bank transactions over the weekend, from actual zero values. It also has the capacity to learn common time-series patterns across new business scenarios and support multiple time frequencies.
In a statement, Ant International asserts that FalconTST 2.0 delivers “more accurate forecasting in real-world FX risk management of cross-border payments”, with the vendor claiming it consistently improves forecasting accuracy by at least 93%.
The platform is being utilised in different ways at the partner banks. At Barclays and Citi, FalconTST 2.0 has been integrated with the banks’ respective FX hedging platform BARX NetFX and Fixed FX Rates solution to facilitate FX risk management on e-commerce platforms or airlines.
