Much of this year’s agentic payments discussion has focused on protocols. At its merchant focused Voyage event on September 9 in Shanghai, Ant International took a broader view. Its senior executives argued that the most immediate AI opportunities lie in automating merchants’ existing payment and finance workflows. At the same time, the company is building the identity, billing, and authorization infrastructure needed to support the future of agentic commerce.

Taken together, the company announced close to 100 key product updates in a single release cycle — what Ant International calls its “full stack of AI-native solutions across payment, account, FX, treasury, and growth operations.” The AI strategy is worth attention because it targets the entire money flow and prior and post payment financial workflow that agents will transact on, not just the payment execution itself. Four themes around AI stood out to me.

Payment And Finance Automation Is The Immediate AI Opportunity

Antom, Ant International’s merchant payment service provider arm, has applied its AI agent Autopilot across the entire merchant payment lifecycle, from onboarding, integration, orchestration, risk management, to troubleshooting and analysis, revenue recovery, reconciliation, and market expansion. It reports that over the past 12 months, 89.5% of its merchant clients adopted its Autopilot which executed 81.4% of merchant payment and operational tasks.

Merchant onboarding with Antom showed the clearest gains. In a Southeast Asia case, merchants previously entered 20 to 40 fields by hand, waited three to seven days for underwriting, and spent five to 10 days on integration. Antom says the Autopilot can cut lead time to first transaction from days to minutes. The number of merchants using AI integration quadrupled in last 12 months. By the end of 2026, 95% of the new merchants are expected to complete onboarding with Antom within 1 hour.

These figures come from Antom’s own data. Results will also vary by market, merchant category, and risk profile. Merchants can get the clearest picture from a proof of concept for three to six months. The test should track approval rates alongside fraud and chargeback rates, especially where authentication flows change.

Payment Data Feeds AI Models And Powers A Higher-Margin Economic Model

To Ant International, payment processing can provide tailored insights back to a specific merchant. Every transaction helps its FX and liquidity forecasting model- Falcon TST, a time-series transformer model, sharpen its understanding of the merchant’s own sales patterns, currencies, cash flows, cash positions, and future financial needs.

FalconTST shows how this can change the economics and monetization model of Ant International. More payment data improves WorldFirst’s (Ant International’s Global Account service provider arm) ability to forecast a merchant’s future currency inflows. A more accurate forecast helps the merchant hedge less, while reducing WorldFirst’s cost of covering any shortfall. The resulting FX economics can outweigh the margin from payment processing. Payment processing becomes the customer-acquisition cost, while AI-powered FX, treasury, and risk services become the profit pools.

This model also rewards scale twice. More payment volume generates more fees, but it also supplies richer data to train the models. Better models improve forecasts, lower the cost of financial services, and support stronger pricing. The real advantage lies in connecting payment data with higher-margin financial decisions across the customer relationship.

Payment companies should therefore stop judging each product in isolation. A low-margin payment relationship can create substantial value when it presents opportunity for monetization from FX, treasury, and risk services.

Serving AI Platforms Beyond Payment Processing

AI platforms such as DeepSeek and Kimi from China can reach international scale within months. They need more than a payment checkout. They also need card and local payment method coverage, usage metering and invoicing with tax, controls against fraud and token abuse, and cross-border settlement.

Usage-based pricing is a tailored pricing model for AI platforms and is the most complex piece. Customers may pay per token, API call, agent action, or completed outcome. Each event has to be captured, priced, invoiced, and collected. This helps explain why Stripe acquired Metronome and Adyen acquired Orb earlier this year.

Antom is building its Meter capability in-house. Management sees accurate, high-speed processing of usage events as the core challenge and plans to draw on Ant’s experience handling massive payment throughput during Double 11 shopping festival transaction peaks to its Meter product. As Antom supports AI platforms for large token usage events, its billing product will also need prove its capabilities to handle negotiated contracts, prepaid credits, mid-cycle plan changes, and revenue recognition beyond just payments and settlement handling.

On model routing, Ant International takes a different view from Stripe’s OpenRouter acquisition. Rather than reselling model access, Ant International wants to link model choice to specific business workflows and recommend models based on cost and success rate. This product hadn’t launched at the time of the event, and the market will watch how it develops. Both approaches point to the same trend: payment providers want a role in how AI usage is measured and priced, not just in how it’s paid for.

Building Trust Across the Agentic Payment Chain

Verifying an agent tells a provider which agent acted. A payment provider also needs to know who authorized it, what task and budget it received, and whether the payment matches the original intent.

Ant International’s agentic payment solutions address this with agent-callable interfaces, task-based authorization, and controls that run across the whole workflow as shown in its industry first Account For Agent product. The company uses blockchain to link the agent’s identity, mandate, actions, and payment in one verifiable record. Running controls throughout the workflow, rather than only at the payment instruction, gives providers a fuller view of agent activity. As the approach scales, implementation will need to account for transaction latency and data localization requirements across markets.

The know-your-agent interoperability framework with Visa and Mastercard carries significant weight. It aims to align Visa’s Trusted Agent Protocol, Mastercard’s Verifiable Intent, and Ant’s Agentic Mobile Protocol (AMP) around operator traceability, shared certification, and continuous monitoring. A common framework could spare merchants and PSPs from repeating agent verification for each network. It could also bring digital wallets, the main payment credential for large consumer populations in Asia and other emerging markets, into emerging agentic standards alongside cards. The next milestone will be shared technical specifications that the industry can implement for broader use cases.

Conclusion

Ant International has deeply embedded AI across payment and finance automation, higher margin services like FX, value-added services for AI-platform, account built for AI agents, agent trust protocol, wallet interoperability, merchant fund guarantee against agent specific risks and others. Few payment companies match Ant International’s breadth across global wallets network (Alipay+), merchant payments (Antom), B2B payments and business accounts (WorldFirst), and SME lending (Bettr). Its full-stack AI approach brings these capabilities together into one AI-driven payments and financial infrastructure.

Its progress will depend on execution: merchant results that hold up over time, billing and metering capabilities that are proved by more clients, and transparent and interoperable industry standards that more partners can adopt.

What To Read Next

Forrester has dedicated research and blog posts on payment innovations, such as agentic payments, including:

The State Of Agentic Payments For B2B

Consumers Aren’t Ready To Delegate Payments To AI Agents

Predictions 2026: Payments

75 Days After Launch, Open USD’s Payment Strategy Takes Shape

Stripe Sessions 2026: Stripe Is Rearchitecting Payments For An Agentic AI Economy

Ant International’s Playbook On AI, Blockchain, And Wallet Network

Forrester clients can set up an inquiry or guidance session to discuss these topics with us.

Share