Ant International launched a full suite of nearly 100 AI-native products on September 18, spanning payments, accounts, foreign exchange, treasury and merchant growth tools for global businesses. The suite runs on two in-house models: Antom 3-in-1 Transformer, a payment model with more than 10 billion parameters that processes 90 trillion data points a year, and FalconTST, an 8.5-billion-parameter forecasting model that Ant says cuts corporate FX hedging costs by 30 to 60 percent. “The future fintech leader delivers two things: a future-ready trust infrastructure and foundational FinAI models that combine complex types of deep data and rich domain expertise, and a full stack of connected AI-native solutions across payment, account, FX, treasury and growth operations that prepare them for today and the fully agentised future,” said Peng Yang, Ant International’s chief executive, in the launch announcement.

The headline product is Account for Agent, which the company describes as the industry’s first account type built for an AI agent rather than a human user or a registered business. That is a bigger structural claim than a product launch usually carries: it means Ant is proposing agents as a distinct class of accountholder inside its financial infrastructure, with their own controls and risk treatment, rather than treating an agent as a permission layered onto an existing human account, the same distinction this publication raised when it asked whether consumers would let AI agents pay, not just shop.

The original insight is in the adoption numbers Ant disclosed, not the launch itself. The company said 89.5% of clients on its core merchant payment service had already deployed its AI agents in the past 12 months, and 81.4% of payment tasks on that service are now completed with AI help. Those figures describe a rollout well underway before this week’s announcement, which makes the launch less a bet on where agentic commerce is going and more a formal naming of infrastructure already built and adopted, a pattern this publication has traced across agentic commerce’s reasoning and payment layers splitting apart this year. Merchants evaluating agent tooling now have a real usage benchmark to hold vendor claims against.

Source: PR Newswire