Why Building AI Agents Is No Longer the Hardest Part of Agentic Commerce
Watch more: The Agentic Experience With Mastercard’s Sabrina Therani Agentic commerce is becoming an infrastructure race. The biggest agentic constraint today is not whether AI can shop but whether the systems surrounding that AI can determine what an agent is allowed to do, who is responsible when something goes wrong and how a machine-initiated transaction […] The post Why Building AI Agents Is…
Agentic commerce is transitioning from a challenge of AI capabilities to an infrastructure issue. The primary obstacle is determining what an AI agent can do, who is accountable for errors, and how transactions initiated by machines can safely flow through merchants, banks, and payment networks. Sabrina Tharani, Mastercard's senior vice president of Global Fintech Programs, stated that agentic commerce is more of a new interface rather than a new shopping channel.
The first mainstream adoption of agentic transactions is likely to be low-value, high-frequency purchases where consumers prioritize speed and convenience. Mastercard has launched a cohort of 22 companies through its Start Path program to foster agentic commerce and services. The initial agentic transactions are expected to resemble mundane tasks such as recurring payments, essential household items, and seasonal shopping moments.
These categories have relatively low costs associated with mistakes and high value in saving time. Consumers are already partially automating these behaviors through subscriptions, saved carts, and recurring orders. The delegation process can be gradual, starting from the transaction funnel. However, each layer of autonomy requires stronger infrastructure to support the agent.
The final piece of the puzzle is ensuring consumers have recourse if an autonomous decision leads to an incorrect outcome. This requires merchants to recognize agents, financial institutions to understand permissions, payment systems to authenticate credentials and enforce limits, and loyalty programs to verify if benefits still apply.
These are primarily commerce infrastructure challenges, not AI challenges. Agentic commerce will be built upon the existing networks, merchants, bank accounts, and credentials, with the layer interacting with them evolving. This framing places agentic commerce within a sequence of interface changes. The shift from traditional shopping to AI agents could redefine who or what carries the consumer's intent into the transaction.
For Mastercard, this means focusing on whether startups can operate at a network scale rather than just showcasing compelling technology.
Written by urgent.news from PYMNTS's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
