AI moves into the payment engine, reshaping how global commerce gets done

Payment platforms are deploying AI to route transactions, detect fraud and automate disputes, while emerging agentic commerce pushes the technology beyond checkout.

Payments are becoming easier for shoppers, but the machinery behind each tap, scan or click is growing more complex. A cross-border purchase can move through merchants, acquirers, card networks and issuing banks in seconds, yet the money itself may not arrive for days. In its report, TMTPost described how that process is now being reshaped by artificial intelligence, with payment platforms increasingly using models to choose routes, assess risk and even support transactions before and after checkout.

The article places Antom, the merchant-facing arm of Ant International, at the centre of that shift. According to the report, the company is using AI across the full payment journey, from merchant onboarding and market research to fraud detection, dispute handling, reconciliation and system operations. Antom says its tools can compress tasks that once took weeks or months into minutes, while also helping merchants navigate a fragmented global payments market in which consumer preferences, regulation and settlement norms vary sharply by country.

That fragmentation remains one of the biggest obstacles to international commerce. As Antom’s guide to global payment processing explains, businesses selling across borders must deal with local payment habits, different security standards, processing fees and settlement arrangements that can affect both conversion and cash flow. Its separate guide on failed payments says merchants often lose revenue to authorisation declines, chargebacks and delayed settlement, and recommends measures such as smart routing to local channels, 3D Secure and account-to-account transfers to reduce losses.

The TMTPost report argues that AI is now helping payment firms tackle those frictions in real time. In Antom’s demonstration, an AI agent called Autopilot was shown analysing new markets, recommending payment mixes, testing checkout flows and then launching changes automatically. The company claims that what once required extensive research can now be reduced to minutes, allowing merchants to make faster decisions about which markets to enter and how to localise checkout.

Security is another area where the stakes are rising. The report says Visa has tightened monitoring rules for acquirers and that fraudsters continue to evolve their tactics, including automated card-testing attacks that can flood merchants with thousands of attempts in a short period. Antom said its AI-driven systems now watch transactions continuously, generate rules when anomalies appear, simulate the likely effect of those rules before deployment and then review their performance on an ongoing basis. The company said this approach has sharply reduced fraud and chargebacks for some merchants.

The same logic is being applied to disputes, reconciliation and operations. Antom’s tools are designed to help merchants prepare evidence for chargeback claims, automate the paperwork involved in disputes and shorten reconciliation processes that can otherwise take days. The TMTPost report also notes that payment platforms are increasingly expected to monitor channels around the clock, reroute traffic when outages occur and intervene before service failures hit transaction volumes.

The bigger change, however, may be that AI is moving payments earlier in the commercial journey. The report says the industry is now talking about “agentic commerce”, in which an AI assistant searches, compares and even completes purchases on a user’s behalf. That creates new demands around pricing consistency, risk controls and settlement. Antom has responded with products that combine payment processing, checkout, marketing and funding tools, reflecting a broader push by global payment groups to become infrastructure providers for commerce rather than mere transaction handlers.