Ant Bank and Bettr have joined the Hong Kong regulators' GenA.I. Sandbox++ to trial AI-driven liquidity risk management.
The initiative will see the two companies apply generative AI forecasting tools to treasury operations, with the stated aim of building a 24/7 liquidity risk management ecosystem.
Ant Bank plans to apply Ant International's Falcon Time-Series Transformer (TST) AI Model 2.0 to its foreign exchange and liquidity management functions. The model is intended to generate granular, statistically grounded daily forecasts to support real-time treasury planning. According to the companies, the deployment is meant to reflect customers' continuous digital financial activity, moving liquidity planning away from static, periodic assessments and toward ongoing, data-driven monitoring.
Ant Bank describes itself as Hong Kong's only digital bank integrated with an e-wallet. Bettr, as Ant Bank's technology partner, will support the deployment of Falcon TST 2.0 through its Platform Tech infrastructure, working with Ant Bank's treasury team. The two companies said they intend to develop shared practices for applying AI to liquidity risk management within the sandbox framework.
Falcon is Ant International's proprietary time-series transformer model, developed to support foreign exchange risk management in cross-border payments. In addition, the company stated that Falcon TST 2.0 achieved state-of-the-art results on the Mean Absolute Scaled Error metric, a benchmark used to evaluate time-series foundation models, with forecast accuracy above 93%.
Regulatory background
The GenA.I. Sandbox++ was launched by the Hong Kong Monetary Authority (HKMA) in coordination with the Securities and Futures Commission, the Insurance Authority, the Mandatory Provident Fund Schemes Authority, and Hong Kong Cyberport Management Company Limited. The programme is designed to give financial institutions a controlled environment in which to test generative AI applications across banking, securities, insurance, and pension-related use cases, under regulatory oversight.
Ant International said its participation reflects continued engagement with the HKMA and ongoing involvement in the development of Hong Kong's financial technology infrastructure. The liquidity risk management track specifically targets AI applications capable of supporting continuous, real-time treasury functions, an area regulators have identified as relevant given the shift toward always-on digital banking and payment services.
For the wider industry, the sandbox participation illustrates how Hong Kong authorities are structuring AI adoption in financial services through supervised testing rather than open deployment, allowing forecasting models such as Falcon TST 2.0 to be assessed against regulatory expectations before broader rollout. The outcome of the trial may inform how other digital banks in the region approach AI-based liquidity and treasury management.
