Ant Bank Joins Hong Kong AI Sandbox to Advance 24/7 Liquidity Risk Management

Ant Bank and Bettr have been selected for Hong Kong’s GenAI Sandbox++ to explore how AI forecasting can strengthen FX and liquidity management in an increasingly real-time banking environment.

Ant International’s Hong Kong digital bank, Ant Bank, is turning to artificial intelligence to improve how it manages liquidity and foreign-exchange risk.

The bank and Bettr, Ant International’s inclusive embedded financial solutions provider, have been selected to participate in Hong Kong regulators’ GenAI Sandbox++, where they will focus on the liquidity risk management track.

Under the initiative, Ant Bank plans to use Ant International’s Falcon Time-Series Transformer (TST) 2.0 to improve its FX and liquidity forecasting capabilities.

The project reflects a broader shift in banking. AI is increasingly moving beyond customer-facing applications and into critical areas such as treasury, risk management and financial planning.

Ant Bank targets AI-powered liquidity management

Liquidity management is a core function for any bank. Institutions need to continuously assess expected cash flows, funding requirements and market risks to ensure they can meet their obligations.

For a digital bank operating in an increasingly connected financial environment, that challenge can become even more complex.

Ant Bank aims to use AI-generated forecasts to give its treasury team a more detailed view of expected liquidity conditions.

The proposed system will provide granular daily forecasts that can support more proactive liquidity planning.

Instead of relying only on historical information and traditional forecasting methods, AI can analyze large volumes of time-series data and identify patterns that may help financial teams anticipate changing conditions.

As a result, treasury professionals could gain additional information to support faster and more informed decisions.

Falcon TST 2.0 brings AI forecasting into banking

At the center of the initiative is Falcon TST 2.0, Ant International’s proprietary time-series transformer model.

The model is designed for forecasting applications involving financial and FX data. According to Ant International, Falcon TST 2.0 achieved state-of-the-art performance on the Mean Absolute Scaled Error metric in a leading public benchmark for time-series foundation models.

The company says the model achieved forecast accuracy exceeding 93% on the benchmark.

For Ant Bank, the technology will be explored specifically in the context of FX and liquidity management.

That makes the project notable because it connects advanced AI forecasting with a practical banking function.

The objective is not simply to generate financial predictions. Instead, the technology is intended to provide information that can support treasury planning within the bank.

Hong Kong regulators expand AI experimentation

The project is taking place through GenAI Sandbox++, a regulatory initiative designed to encourage responsible AI experimentation across Hong Kong’s financial sector.

The sandbox was launched by the Hong Kong Monetary Authority (HKMA) together with the Securities and Futures Commission, Insurance Authority, Mandatory Provident Fund Schemes Authority and Hong Kong Cyberport.

Its purpose is to provide financial institutions with an environment where they can explore AI applications while considering the risks and regulatory requirements associated with the technology.

For banks, this type of controlled testing environment can be particularly valuable.

AI applications used in financial institutions must deal with sensitive data, operational risks and potentially significant financial consequences. Testing new systems in a structured environment can therefore help institutions understand both their benefits and limitations before broader adoption.

Why liquidity management is an important AI use case

AI has already attracted significant attention across financial services, particularly in areas such as customer support, fraud detection and document processing.

Liquidity management represents a different type of opportunity.

Treasury teams deal with constantly changing information. Cash positions, customer activity, foreign-exchange movements and funding requirements can shift over time.

Traditional forecasting methods can provide valuable insights, but AI may help process larger datasets and identify more complex patterns.

For a digital bank, those capabilities could become increasingly important as customers expect services to remain available around the clock.

Traditional and AI-assisted liquidity management

The key point is that AI does not necessarily replace treasury professionals.

Instead, it can provide additional forecasting and analytical support while people remain responsible for interpreting results and making decisions.

Bettr supports the technology deployment

Bettr will support Ant Bank as its technology partner through its Platform Tech capabilities.

The partnership is intended to help deploy Falcon TST 2.0 for the bank’s treasury team and explore practical applications for AI-driven liquidity risk management.

This collaboration also highlights the growing importance of partnerships between financial institutions and technology providers.

Banks increasingly need specialized technology to adopt advanced AI without rebuilding every part of their infrastructure internally.

Consequently, technology partnerships can help financial institutions experiment more quickly while maintaining focus on their core banking operations.

Ant Bank’s digital banking model creates a real-time challenge

Ant Bank is positioning itself as a digital bank closely integrated with an e-wallet, with a focus on technology-driven and customer-centric financial services.

That operating model creates an environment where financial activity can happen throughout the day rather than following traditional branch-based banking patterns.

Customers can make payments, transfer money and manage their finances digitally at different times and from different locations.

For the bank, that creates a need for systems capable of monitoring financial activity and liquidity conditions with greater frequency.

AI-powered forecasting could therefore become an important component of the infrastructure supporting a 24/7 banking model.

AI moves deeper into banking infrastructure

The Ant Bank project reflects a broader trend across financial services.

Banks are increasingly moving beyond the first generation of AI applications, where technology was mainly used to automate customer interactions or assist employees.

The next stage is more deeply connected to the infrastructure of financial institutions.

Potential applications include:

• Liquidity forecasting
• Fraud detection
• Credit risk assessment
• Financial forecasting
• Treasury management
• Market risk analysis
• Regulatory monitoring

These use cases could have a greater impact on how financial institutions operate because they connect AI directly with core business processes.

However, the risks are also greater.

A forecasting error in a customer-service chatbot may create inconvenience. A serious error in a system supporting liquidity management could have much larger consequences.

That is why governance, testing and human oversight will remain important as banks deploy AI in increasingly critical areas.

Human oversight remains essential

AI can process large quantities of information quickly, but financial institutions still need experienced professionals to interpret forecasts and assess unusual situations.

Liquidity decisions often depend on factors that may not be fully captured in historical data.

Market shocks, unexpected customer behavior and changes in funding conditions can all affect a bank’s liquidity position.

Therefore, AI should be viewed as a decision-support capability rather than an automatic replacement for treasury expertise.

The most effective model may combine continuous AI-based analysis with human judgment and established risk controls.

Hong Kong strengthens its position as a financial AI hub

The GenAI Sandbox++ initiative also highlights Hong Kong’s efforts to encourage responsible AI adoption in financial services.

Financial institutions increasingly want access to advanced technologies, but regulators need to ensure that innovation does not undermine financial stability, customer protection or operational resilience.

A sandbox approach provides a middle ground.

Firms can experiment with new applications while regulators gain greater visibility into how AI is being used and what risks may emerge.

For Hong Kong, this could support the development of a financial ecosystem where AI innovation and regulatory oversight develop alongside each other.

What the project could mean for digital banking

The Ant Bank initiative could have implications beyond liquidity management.

If AI forecasting can improve the bank’s understanding of cash flows and FX exposure, similar technology could eventually support other areas of financial planning and risk management.

More accurate forecasting could help banks make better use of liquidity resources, respond more quickly to changing conditions and improve operational planning.

At the same time, successful implementation could provide a model for other digital banks exploring AI applications in treasury operations.

The project will therefore be worth watching as financial institutions increasingly search for practical ways to move AI from experimentation into everyday operations.

Future outlook for AI-driven liquidity management

AI is likely to become increasingly important in treasury and risk management as banks collect more data and develop stronger digital infrastructure.

However, adoption will depend on several factors.

Financial institutions will need to evaluate forecast accuracy, data quality, cybersecurity, model risk and regulatory requirements. They will also need clear processes for human oversight.

The Ant Bank and Bettr project provides an opportunity to test these considerations in a controlled environment.

If the approach demonstrates measurable benefits, AI-powered liquidity forecasting could become a more common feature of modern banking infrastructure.

Final Thoughts

Ant Bank’s participation in Hong Kong’s GenAI Sandbox++ shows how artificial intelligence is moving into one of the most important areas of banking: liquidity and treasury management.

By exploring Falcon TST 2.0 for FX and liquidity forecasting, Ant Bank aims to give its treasury team more granular information to support proactive planning in a 24/7 digital banking environment.

The bigger story, however, extends beyond one bank or one AI model.

Financial institutions are increasingly looking for ways to embed AI into the systems that support core financial operations.

As that transition continues, the winners may not simply be the banks with the most advanced AI models. They may be the institutions that can combine accurate forecasting, strong risk controls, regulatory oversight and human judgment into a single operating model.

AIAnt BankAnt InternationalArtificial Intelligencebanking technologyDigital BankingFalcon TST 2.0Financial AIfintechFX risk managementGenAI Sandbox++HKMAHong KongLiquidity Risk Managementtreasury management