AI Agents Are Buying for You: The Agentic Commerce Shift

Alejandro Martínez · 2026-04-28

Agentic commerce is a fundamental shift. AI agents can now orchestrate the entire purchase journey, from discovery to payment, all inside a single conversational interface. Driven by new infrastructure and partnerships between platforms, payment systems, and marketplaces, this model is already reshaping distribution channels and redefining customer relationships. In this fast-evolving ecosystem, big questions remain: How do we build trust? Who governs agent behavior? Who controls the customer experience? How is value shared? Check out the summary and video of the fireside chat "Agentic Commerce" with Laurent Curny, Country Lead France at Stripe, by Chloe Rossignol, Founder of Paiement.fr, at FinTech R:Evolution | #FFT26 | Flight to Quality. ### What Is Agentic Commerce? Agentic commerce means AI agents can independently manage the entire buying process — from finding a product to completing the payment — without a human stepping in at every step. It's a complete shift in how we, the system, and the transaction relate. In traditional online shopping, you interact directly with a screen: you search, compare, pick, and confirm. With agentic commerce, that interaction layer is handed off to an AI agent. The agent figures out what you want, does the research, weighs options based on your rules, and sometimes even buys stuff for you. You set the limits and rules; the agent works within them. Right now, we're in a transition phase. The tech exists and works in the U.S. for specific use cases. But full automation is still held back by regulations — like needing a human to approve certain payments — and by trust and legal responsibility issues that aren't fully sorted out yet. Widespread deployment is still a few years away, but getting the infrastructure ready needs to start now. ### Why Agents Need a Common Language to Scale Stripe is tackling the fragmentation of agentic systems by building a universal protocol layer — basically, a standard way for AI agents and merchant systems to talk to each other. Different AI models (like GPT, Gemini, Claude) use different architectures and languages. Without a common protocol, every merchant would have to build custom integrations for each agent, which just isn't practical or affordable. This push for standardization is a lot like what card networks did for physical payments back in the day: create an interoperable infrastructure that the whole ecosystem can use. The difference is that this will spread way faster, and the governance questions — who controls the protocol, who sets the rules — will be even more critical. ### From Interfaces to Data: How Visibility Changes Agentic commerce changes not just how we buy, but how merchants get noticed. - In the current model, a merchant's visibility depends on optimizing its user interface: good design, easy navigation, strong SEO. - In the agentic model, the user interface fades away. What matters is the data layer. Agents search for products based on structured data, not fancy layouts. This means merchants need to focus on making their product data clear, complete, and machine-readable. Things like product descriptions, pricing, availability, and shipping details need to be structured so agents can find and compare them easily. It's a big shift. Instead of competing on design, you're competing on data quality and how well your systems talk to agents. ### Trust, Control, and the New Rules With agents making decisions and handling payments, trust becomes everything. Users need to know the agent will act in their best interest, not just push the highest-paying merchant. This raises questions about transparency — can you see why an agent chose one product over another? There's also the question of control. Who's responsible if an agent buys the wrong thing or makes a bad deal? Is it the user, the agent developer, the merchant, or the platform? These legal and ethical issues are still being worked out. And finally, value distribution. In the old model, merchants paid for ads and SEO to get seen. In the agentic model, who gets paid? Will agents charge merchants for listing their products? Will platforms take a cut of each transaction? The economics are still being figured out. ### What This Means for Professionals For anyone in payments, e-commerce, or digital strategy, agentic commerce isn't just a trend — it's a structural change. You need to start thinking about how your systems will interact with AI agents. That means investing in clean data, understanding protocol standards, and preparing for a world where the customer's first interaction might be with an agent, not your website. The shift is coming. The smartest move is to start preparing now.