Only 18 percent of US retail purchases happen online. That number has moved slowly for years. Shopify's president, Harley Finkelstein, thinks agentic AI is what finally shifts it, and he is putting the company's infrastructure behind that belief.

Speaking at the Upfront Summit in Los Angeles, Finkelstein laid out the case: AI shopping agents will become a new entry point for ecommerce, one that bypasses the traditional storefront entirely. Shopify's internal data backs him up. Orders sourced from AI agents are up 14x year over year. AI-driven traffic to Shopify stores has surged 7x since January 2025, with purchases attributed to AI-powered search up 11x.

The numbers are small in absolute terms. The trajectory is not.

Merit-Based Shopping

Finkelstein introduced a concept he called "merit-based shopping." The idea: when an AI agent selects products on behalf of a consumer, it does not respond to banner ads, promotional placement, or brand loyalty built through decades of marketing spend. It responds to product data. Clean specifications, competitive pricing, accurate availability, strong reviews.

This inverts the power structure that has defined ecommerce for two decades. The merchants who win in an agentic environment will not necessarily be the ones with the largest advertising budgets. They will be the ones with the cleanest, most structured product data.

As we covered when Shopify moved the storefront inside ChatGPT, OpenAI tried building its own checkout before handing that infrastructure to Shopify. The division of labour is becoming clear. AI platforms handle discovery and conversation. Commerce platforms handle fulfilment, checkout, and payments.

In merit-based shopping, the best product wins. Not the best-marketed product. The implications for how merchants allocate budget are significant.

The Infrastructure Play

Shopify is not just talking about agentic commerce. It is building the rails.

In December 2025, Shopify launched Agentic Storefronts, machine-readable interfaces that let AI agents browse, query, and purchase without scraping HTML. In January 2026, Google unveiled the Universal Commerce Protocol (UCP), co-developed with Shopify, Walmart, and Target. UCP provides a standardised way for AI agents to discover merchants, apply discount codes, use loyalty accounts, and complete purchases.

This is the same protocol landscape we mapped in our analysis of the agentic commerce standards race. Google's UCP handles discovery and intent. Visa's Trusted Agent Protocol and Mastercard's Agent Pay handle trust and settlement. Shopify is positioning itself as the fulfilment layer that works with all of them.

Finkelstein has been explicit that AI agents will not bypass Shopify's checkout systems. The strategy is to ensure that wherever the conversation happens, whether in ChatGPT, Gemini, Perplexity, or any other agent, the transaction closes through Shopify.

The Gap Between Trajectory and Scale

The 14x growth figure demands context. AI-sourced orders are still less than one percent of total Shopify volume. The growth rate is extraordinary. The base is tiny.

This mirrors a pattern we have seen across agentic commerce. As we reported when agentic commerce moved from demo to live transaction, the infrastructure is shipping faster than the demand. Visa reported a 4,700 percent surge in AI-driven traffic to retail sites, but absolute transaction volumes remain a fraction of traditional commerce.

Shopify's bet is that the infrastructure must exist before the volume arrives. Build the agentic storefronts, standardise the protocols, ensure the checkout works, and the agents will come.

The risk is timing. If agentic shopping remains a novelty for another two to three years, the investment in infrastructure sits idle. If it accelerates the way mobile commerce did after the iPhone, the companies that built early will own the category.

What This Means

Shopify is making a calculated bet that the storefront, as a destination a human visits, will decline in importance. The store becomes a service layer that agents query. The brand becomes a data profile that agents evaluate. The checkout becomes an API call that agents execute.

For merchants on Shopify, the message is clear: invest in product data quality, structured metadata, and machine-readable catalogues. The AI agents choosing between your product and a competitor's are not reading your homepage copy. They are parsing your data feed.

For Shopify's competitors, the message is sharper. The company that powers over four million merchants is building native agent infrastructure while most platforms are still debating whether to allow agent access at all.

Sources

If AI agents choose products based on data quality rather than brand awareness, which merchants are actually ready for that shift?

Reply

Avatar

or to participate

Keep Reading