Antom, a merchant payment and digitization services provider under Ant International, has launched an agentic payment solution that supports secure checkout through alternative payment methods (APMs) and cards.

The new system allows AI agents to complete payments using multiple methods, including digital wallets and cards, and is open sourced on GitHub.

Antom said the solution uses an AI-ready payment mandate model and enhanced payment asset management to improve user intent recognition and transaction security.

The company is among the first to work with Mastercard and Visa to explore and pilot tokenized card-based payments for AI agents in Asia-Pacific.

The platform supports both confirmed and pre-authorized transactions, such as purchases within spending limits or scheduled sales, and aims to reduce integration costs for merchants and developers.

🔗 Source: Antom

🧠 Food for thought

1️⃣ AI payments evolution reaches autonomous transaction capabilities

The launch of agentic payment solutions represents a significant leap from AI’s traditional role in payments, which historically focused on fraud detection and risk management.

Earlier AI implementations in payments primarily addressed fraud detection challenges, where approximately 35% of digital card transactions flagged as fraudulent were actually false positives, causing significant merchant losses1.

Companies like PayPal and Citigroup initially used AI for customer service chatbots and payment process automation, focusing on improving existing human-driven workflows2.

Now, with 75% of financial institutions in Hong Kong implementing or piloting generative AI use cases as of May 20253, the industry has reached a point where AI agents can autonomously complete purchases.

This evolution aligns with Visa CFO Chris Suh’s prediction that generative AI could transform commerce similarly to how e-commerce revolutionized shopping4.

2️⃣ Traditional payment infrastructure requires fundamental redesign for AI agents

Conventional payment systems were built around human decision-making patterns, requiring multiple clicks during checkout processes that don’t align with how AI agents operate.

Antom’s solution addresses this mismatch by building payment flows specifically designed for dialogue-based interactions, where purchase intent is confirmed through natural language conversations rather than traditional form-based checkouts.