Selling to machines: How AI will change e-commerce

How did you get to this article? Maybe you opened a link in an email, or you navigated from the Fast Company home page. Perhaps you Googled “agentic AI” and this figured in the results. The point is, you almost certainly clicked, scrolled, tapped, or typed your way here, because that’s the digital grammar that shapes nearly every online experience.

But that 30-year-old paradigm is about to change. Agentic AI is ripping up the rulebook, by creating a new layer of intelligent, autonomous mediation between us and the digital world. Personal shopping agents will handle routine purchases, while in the workplace, agents will automate workflows and streamline procurement.

Investors are excited, of course. Earlier this year, more than half of the Y Combinator startups chosen for its accelerator program were working in agentic, while the three biggest Q1 acquisitions in the AI sector involved companies building agentic enterprise technologies. This is because agents are about to transform how work gets done, how businesses operate, and how systems interact. And nowhere are the consequences more profound than in e-commerce.

CLICKLESS COMMERCE

Today, e-commerce success depends on online visibility, with search driving half of website traffic and fuelling the $75 billion search engine optimization (SEO) industry. But agentic AI upends the traditional business model. For agents, displays and brand content are less important. Merchants must focus on making product data machine-legible—accurate, structured, and accessible—because machines won’t browse sites. They’ll connect directly to sales platforms and databases via APIs, creating a parallel e-commerce track that serves agents rather than human shoppers.

Building out that machine track will be a critical job for merchants, fintechs, software companies, and financial institutions. For merchants, an early task is compiling and structuring data in a way that’s visible and relevant to agents. Another is creating seamless connections between merchant systems and agent systems. The good news is that the connective tissue is already here. New protocols—model context, agent-to-agent, and agentic payments—now enable agents to connect, communicate, and transact autonomously.

MARKETING TO MACHINES: A NEW INDUSTRY

For retailers, one question looms: How do I market to machines? As SEO evolves for human audiences, new disciplines are emerging to optimize digital environments for AI agents. Agent engine optimization helps make digital spaces easier for agents to understand and use. Agent interaction design focuses on how they communicate with platforms, APIs, and other agents to get things done for users. These areas connect with generative engine optimization, which improves content so AI systems can better generate recommendations and make decisions. Taken together, they signal a new ecosystem—one that will create fresh roles, unlock new value streams, and redefine how businesses compete in an AI-driven economy.

Another way of attracting machine buyers is to reward them. This is the agentic version of e-commerce’s affiliate system, where publishers, partners, and influencers are paid a commission for driving sales. There’s feverish speculation as to how an attribution mechanism may work, and who will get paid. There are plenty of challenges too—not least verifying agents’ identities, knowing with confidence which agent influenced the sale, and preventing fake agents from gaming the system. But the opportunities are apparent. The affiliate market is predicted to double in size by 2030, even without agents.

FACE THE HEADWINDS

Certain issues will require regulatory clarity. Agentic referral will have to be transparent, for example, to ensure fair competition. Then there’s the matter of trust. We’ll need confidence in the agents and the ecosystems they operate in. Giving an autonomous bot spending authority is a big commitment, but getting it right is key to scaling agentic commerce. Ensuring agents can operate across different payment systems, platforms, and legal jurisdictions will also pose technical and regulatory challenges.

That said, we’re about to experience the biggest transformation in e-commerce since it began. This promises to be an era of great innovation—a time to build new tools, new ecosystems, and new ways of creating value. Agentic commerce isn’t just a tech shift; it’s a reimagining of how we buy, sell, and connect in a digital world.

Ken Moore is chief innovation officer at Mastercard.

View the full article