Ant International Wins Global Bank Adoption for FalconTST 2.0

Cross-border payment providers must forecast cash flows across currencies to decide when and where to fund accounts and how much foreign-exchange exposure to hedge. Errors can leave capital idle, create liquidity gaps or produce excessive hedging. Ant International first used FalconTST internally for hourly, daily and weekly cash-flow and currency-risk forecasts. Its rollout to global banks shows how purpose-built time-series AI, rather than general language models, is moving into treasury and risk-management workflows.

Ant International announced FalconTST 2.0 on Aug. 20, saying Barclays, Citi, Deutsche Bank and Standard Chartered had integrated the model for cash-flow forecasting and FX liquidity management. The company said the system posted a Mean Absolute Scaled Error, or MASE, of 0.666 on a public time-series benchmark and consistently achieved forecast accuracy above 93%. Barclays linked it to BARX NetFX, Citi paired it with Fixed FX Rates, and Standard Chartered uses it alongside SCALE FX in the Monetary Authority of Singapore’s PathFin.ai program.