Ant International has launched a full-stack suite of AI-native financial solutions spanning payments, accounts, foreign exchange, treasury and business growth, marking a broader move from AI-assisted fintech workflows toward agentic financial operations.
The next phase of financial automation may be less about individual AI tools and more about connecting them across the financial workflow. Ant International is taking that approach with a new stack that combines payment intelligence, FX forecasting, treasury automation and agentic commerce infrastructure.
Unveiled at the company’s 2026 VOYAGE merchant event in Shanghai, the suite includes close to 100 products across Ant International’s Alipay+, Antom, WorldFirst and Bettr businesses. The architecture is built around two proprietary models: the Antom 3-in-1 Transformer for payment operations and FalconTST, a time-series forecasting model for FX exposure and liquidity.
Ant International says the Antom model processes sequential, tabular and graph data within a single architecture. The company says it has more than 10 billion parameters and processes about 90 trillion data points annually, supporting applications including fraud detection, abuse prevention and payment-success optimisation. These are company-reported technical specifications.
FalconTST takes a different approach, using time-series modelling to forecast cash flow, FX exposure and liquidity. Ant International says the model has more than 8.5 billion parameters and has achieved over 93% accuracy in its FX and cash-flow forecasting tests, while reducing corporate FX hedging and allocation costs by 30% to 60% in reported real-world cases.
The stack also introduces infrastructure designed specifically for AI agents. Alipay+ Agentic Mobile Protocol (AMP) provides a framework for agents to initiate payments through wallets and other mobile interfaces, while Ant International’s AgentSafePay and know-your-agent architecture are intended to address risks associated with autonomous transactions.
AMP is particularly notable because it treats the AI agent as a new participant in the payment flow rather than simply another user interface. Ant International says the protocol can reduce the steps needed to connect an agent with a wallet by 50% and supports transaction permissions designed to let users authorise specific tasks without handing over unrestricted account access.
The company is also extending AI into treasury operations. WorldFirst for Enterprise Treasury Autopilot combines cash visibility, forecasting, FX management and payout routing. Products including World Map, Falcon Forecast, Falcon FX and Whale Pooling are designed to help treasury teams understand liquidity positions and automate decisions within predefined controls.
At the settlement layer, WhaleRTP provides blockchain-based wholesale connectivity. Ant International says the platform interoperates across blockchain assets from 23 financial institutions and processed 45% of its cross-border volume in 2025.
The broader significance is the attempt to connect these layers. Instead of treating payment, treasury, FX and account management as separate systems, Ant International is positioning AI agents as an orchestration layer across financial operations.
The approach also reflects a growing industry requirement: autonomous financial systems need identity, permissions, monitoring and intervention mechanisms alongside AI capabilities. The Bank for International Settlements has similarly examined how AI agents could automate elements of payment and liquidity management while stressing the importance of safeguards and human oversight.
Market Landscape
AI is moving deeper into financial infrastructure, from fraud detection and compliance to liquidity management and payment execution. The BIS has examined AI agents for cash management in wholesale payment systems and found potential for automating routine liquidity decisions, while highlighting the need for regulatory safeguards and oversight.
The emergence of agentic payments adds another layer. An AI system that can select products, initiate transactions and manage financial workflows requires more than conventional authentication. Identity, permission boundaries, transaction monitoring and accountability become part of the infrastructure itself.
Ant International’s approach therefore combines two trends: AI-native financial operations and payment systems designed for autonomous agents. Its use of blockchain-based settlement adds a further interoperability layer as financial institutions experiment with tokenised assets and programmable transactions.
The BIS has noted that digital innovation can improve payment efficiency and enable programmable transactions, while also raising questions around governance, financial integrity and resilience.
Top Insights
- Ant International launched an AI-native financial stack spanning payments, accounts, FX, treasury and business-growth operations.
- The architecture combines the Antom 3-in-1 Transformer with FalconTST for payment intelligence, forecasting and liquidity management.
- AMP introduces a payment protocol designed specifically for AI agents operating through wallets, devices and other digital interfaces.
- WorldFirst’s treasury tools extend AI automation into cash forecasting, FX risk management, liquidity operations and payout routing.
- The platform pairs agentic capabilities with KYA, AgentSafePay and governance controls aimed at managing autonomous transaction risks.
Get in touch with our fintech expert
