TLDR: Ant International asserts its Time-Series Transformer (TST) AI FX Model, a numerical AI for foreign exchange, will revolutionize financial services much like OpenAI’s GPT transformed the broader business landscape. This model achieves over 90% accuracy in forecasting cash flow and FX exposure, significantly reducing hedging and risk-premium costs. The company is also expanding its AI initiatives with platforms like Alipay+ GenAI Cockpit, focusing on automated workflows, fraud detection, and democratizing AI for emerging markets, while emphasizing robust AI security measures.
Ant International is making bold claims about its Time-Series Transformer (TST) AI FX Model, a numerical artificial intelligence solution designed for foreign exchange. The company believes this model will have as profound an impact on financial services as OpenAI’s large language model (LLM) GPT has had on the broader business world. Kelvin Li, general manager of Ant International’s platform tech business unit, describes the TST model as ‘another track to AI,’ distinct from the content-generating LLMs that many financial institutions currently employ primarily for risk minimization, customer interactions, and regulatory compliance. Li emphasizes that while LLMs have largely remained at the periphery of core financial operations, the TST model is designed to tackle fundamental financial tasks such as trading, pricing, and transaction processing.
The TST AI FX model demonstrates impressive capabilities, forecasting Ant International’s cash flow and FX exposure on an hourly, daily, and weekly basis with over 90% accuracy. This high level of precision enables more accurate predictions of trading volumes and leads to substantial reductions in hedging and risk-premium costs for banks. The model is already serving a diverse range of clients, including banks, airlines, online travel agents, and e-commerce platforms. Notably, British bank Barclays has integrated TST into its offerings to help businesses mitigate FX-related costs and risks amidst global market volatility. Ant International has also indicated an upcoming partnership with a low-cost airline in the region.
Beyond the TST model, Ant International is actively pursuing a comprehensive AI strategy, which includes the launch of platforms like Alipay+ GenAI Cockpit. Introduced in June 2025, this AI-as-a-Service platform is designed to empower fintech companies and super apps to develop their own AI-powered financial services. The platform streamlines automated workflows and task orchestration across various fintech operations, encompassing payment orchestration, customer onboarding, compliance checks, fraud detection, and dispute resolution. Jiangming Yang, Chief Innovation Officer of Ant International, envisions a future where ‘agentic AI’ will autonomously carry out tasks within complex financial and compliance contexts, while also continuously learning and evolving with increasing precision.
Recognizing the escalating threat of AI-driven fraud, which has seen a tenfold annual increase, particularly with deepfake technology impacting the financial sector (with 22% of businesses reporting AI-generated payment fraud), Ant International has developed an ‘AI SHIELD framework.’ This framework is designed to manage risks across system architecture, data processing, model training, and inference operations. It provides real-time dynamic risk assessment capabilities, including the detection of adversarial prompts and sensitive data leakage, utilizing over 100 recognition models and 600,000 risk lexicons.
Also Read:
- Grasshopper Bank Pioneers AI Integration, Securing Early Lead in Fintech Race
- Tencent Expands Generative AI Initiatives Globally, Focusing on Southeast Asia
Ant International is also bolstering its investments in human capital and technology, particularly at its Digital Business Centre in Kuala Lumpur. This center, established in 2024, now houses nearly 800 engineers and professionals dedicated to AI R&D, technology, security, and product development. The company’s overarching mission includes democratizing AI for emerging markets and small businesses, with a commitment to developing a ‘future-ready Trusted AI architecture’ in collaboration with partners and regulators.
