AI Agents Revolutionize E-Commerce: A New Era for Product Leaders
E-commerce is experiencing a monumental shift driven by the rise of artificial intelligence (AI) agents. This transformation goes beyond simple machine-to-machine (M2M) payments, evolving into autonomous “AI-to-AI” transactions, according to Sherman Jiang, a seasoned product executive in consumer finance, fintech, and banking.
This evolution necessitates a fundamental rethinking of payment systems and presents product leaders with novel challenges.
In his recent analysis, Jiang explores the future of e-commerce and the profound impact of agentic AI, highlighting the obstacles that lie ahead.
Unlike traditional, rule-based M2M payments, agentic payments involve sophisticated reasoning, continuous learning, and purposeful action. These systems surpass basic reordering when a sensor reaches a threshold. Instead, they autonomously detect when a user is running low on a product, identify top-rated options from local suppliers, negotiate discounts based on flash sales, place the order, and execute the payment.
For product managers, this represents an unprecedented paradigm shift where the “customer” is no longer an individual but an intelligent, dynamic AI agent demanding new tools, infrastructure, and safeguards. Product executives must develop advanced wallets capable of securely delegating payments, enabling truly autonomous commerce.
Navigating Fraud and Liability
While the transition to agentic commerce offers significant opportunities to enhance customer support, boost efficiency, and personalize customer experiences, it also introduces critical challenges, as Jiang points out.
AI agents operate at machine speed and scale, which can trigger false positives under current fraud detection models. This necessitates a complete overhaul of fraud prevention strategies, moving from know-your-customer (KYC) to “know-your-agent” frameworks.
Building a robust trust infrastructure is crucial for distinguishing between “good” and “bad” AI, likely leveraging cryptographic tokens to verify identity and authority.
For instance, a User ID Token could authenticate a user, confirming their identity. An Agent ID Token would serve as a digital passport for the AI, establishing its identity within the system. A Delegation Token would then link the user and the AI, clearly defining the AI’s authorized actions. Together, these tokens form a trust chain between the human user, the AI agent, and their defined relationship.
The rise of agentic AI also brings up questions concerning liability and dispute resolution, especially when an AI agent makes unintended purchases.
Existing legal frameworks rely on intentional or negligent actions. However, AI agents can make technically authorized purchases that contradict the user’s actual intent, creating a legal grey area.
Adding to this is the “agentic loyalty problem,” where AI agents might prioritize the interests of their deploying platform over the user’s best interests. For example, an agent might select a more expensive item from a partner company despite cheaper, more suitable options being available elsewhere.
Addressing Incompatibility and Infrastructure Gaps
The growth of agentic commerce highlights limitations in legacy financial systems. Systems built for human-centric, high-value, low-frequency transactions struggle with the high-frequency, low-value transactions that AI agents require. This includes paying for individual API calls, data, or computing resources on a “pay-per-use” basis, often involving fractions of a cent.
To overcome this, crypto-native technology firms are introducing blockchain-based micropayment systems. Coinbase, for example, launched x402 in May 2025, a payment protocol that facilitates instant stablecoin payments over HTTP, enabling seamless transactions for APIs, apps, and AI agents.
Blockchain infrastructure offers secure ledgers, verifiable identities, and smart contracts for autonomous value transfer.
The Rapid Adoption of AI Agents
AI agents are rapidly entering digital commerce, fueled by increasing customer demand. A July 2025 US consumer survey by Boston Consulting Group (BCG) revealed that 81% of consumers anticipate shopping using agentic AI.
BCG projects that agent-assisted spending could exceed US$1 trillion in the coming years, representing approximately 50% of current e-commerce expenditure. Early adoption will likely focus on routine purchases, like household supplies, restaurant orders, personal care, and supplements, rather than luxury goods or medical devices.
Major players like Google, PayPal, Visa, and Mastercard are actively competing to capitalize on the agentic AI shift.
Payment Webinar
Fintech News Singapore will host a webinar on September 24, 2025, at 3:00 PM to discuss the evolving APAC payment landscape. This webinar will examine the potential of Click to Pay and advancements in passwordless authentication, featuring insights from industry leaders at Worldpay, Visa, and Thales.
Featured image by tete_escape on Freepik
