Agentic Commerce: When Your Storefront Becomes an API

In the Agentic era, the user interface dissolves. Merchants must stop optimizing for eyeballs and start optimizing for machines.

1. The Agentic Infrastructure Arms Race

Over the past year, I have been in the trenches launching pilot programs for Agentic Commerce. Looking at the broader industry, it feels like an infrastructure arms race. There is immense pressure across the ecosystem—from startups to tech giants—to invest and scale the technology before "mass adoption" has truly been achieved, and it seems primarily driven by a classic Game Theory scenario.

We are seeing this play out in real-time: Claude launched the Model Context Protocol (MCP); Google introduced the Agent-to-Agent (A2A) framework; and OpenAI is actively rolling out agentic commerce protocols with major partners. This space will continue to evolve and move at a fast pace.

My personal view is cautiously optimistic. While the pace is frantic, it is necessary. We are trying to solve a "Cold Start" problem: You cannot have adoption without the underlying technology working well, but building technology without adoption is daunting.

I would characterize the current state of Agentic AI as an "A+ for Engineering Effort," but we are still in the early innings of solving core user problems and anticipating changes in user behavior. We have built the engine; now we need to build the road.

2. The Current Friction: Why 70% of Carts Fail

To understand why Agents are inevitable, we have to look at the "Unit Economics" of a modern user session.

Users generally fall into two categories:

1. The High-Intent Shopper: They know exactly what they want ("Nike Pegasus 40, Size 10").
2. The Window Shopper: They are browsing broadly ("I need a gift for my dad").

The current internet fails both. According to the Baymard Institute, the average cart abandonment rate sits at nearly 70%. That represents trillions of dollars in intent that evaporates simply because the friction is too high.

• The High-Intent Shopper is subjected to a time consuming process. They are the manual link connecting the search engine to the merchant, forced to re-enter the same shipping and billing data they have already typed into a hundred other sites
• The Window Shopper is served products to cross-sell, upsell, or just more Ads—not solutions.

These journeys have been optimized over the years, and people have gotten used to it. Just like before the current state of the internet, people using dial-up were still "ok" with it because they didn't know better. We have moved from filling out forms to autofill and using e-wallets in the checkout flow, but that is just a faster horse. If we can build a 1-click checkout, why are we still asking users to act as the router?

3. The Shift: From Aggregating Supply to Aggregating Intent

I predict that users will eventually trust AI Agents the same way they learned to trust saving their address or payment information with their browsers and merchants.

The shift we are seeing is the move from Traffic to Action.

• Old World: The goal was to aggregate demand and send traffic to a distributed network of websites.
• New World: The goal is to aggregate Intent and execute a transaction within the interface.

If I can say, "Buy me the best running shoes under $150," and the Agent executes that task, the value proposition changes. The user is no longer paying for "Information" (Search); they are paying for Autonomy.

4. The Invisible Mall: The Storefront Moves Inside “AI”

This brings us to the most disruptive shift. The storefront isn't dying; it is migrating. Merchants have spent 20 years optimizing their "Storefronts" for human eyeballs (beautiful images, newsletters, reviews). But an AI Agent doesn't care about your newsletter. It cares about Structured Data, Real-Time Inventory, and Price.

This creates a new paradigm: The API-fication of Commerce.

• The "Front End" becomes the AI interface (ChatGPT, Gemini, etc.).
• The "Back End" becomes the Merchant infrastructure.

For merchants, this is a double-edged sword.

• The Risk: They lose the direct "Brand Experience." The user may never see their homepage.
• The Opportunity: The Customer Acquisition Cost (CAC) could plummet. Industry data suggests CAC has risen by roughly 60% over the last five years as ad inventory saturates. If a merchant exposes a clean API, the Agent will choose them over a competitor with a messy site. You stop competing on "Ad Spend" and start competing on "API Latency" and the agent's ability to check out accurately.

5. Who Wins? Whoever Can Build Trust

The winner of this arms race won't necessarily be the one with the smartest LLM. It will be the one that solves Trust.

In a world of infinite AI content, the ability to verify user intent and complete a transaction becomes the scarcest commodity. The platforms that can combine Identity (verifying the user is real), Payments (verifying the funds are real), and Reliability (agents successfully executing on the user’s behalf) will become the gatekeepers of the Agentic Economy.

We are moving from a world of "Browsing" to a world of "Delegating." And once a user experiences the magic of delegating a purchase, they may never go back.