Shopify says AI agents will not bypass its checkout systems

Shopify is betting that the rise of AI shopping agents will not push it out of online commerce, but strengthen its role as the infrastructure that powers transactions behind the scenes. According to the transcript of Shopify’s Q4 2025 earnings call, the company laid out what executives described as foundational work to prepare for “agentic commerce,” a shift in which AI interfaces such as ChatGPT, Google’s AI search results, and Microsoft Copilot can discover products, fill carts, and complete purchases on behalf of consumers.

The e-commerce platform used the call to highlight the importance of its Universal Commerce Protocol (UCP), a new standard created with Google and other partners. Shopify has previously outlined the technical foundation of UCP in its engineering blog, describing it as a shared protocol designed to let AI agents transact directly with merchant infrastructure. “UCP is infrastructure. It’s not a product. It’s the common rails agentic commerce runs on,” Shopify president Harley Finkelstein said.

This is an important point. As AI tools begin to assist people in finding products and even making purchases, commerce platforms may risk losing control, profit, or direct customer contact. Shopify’s position is that even if the shopping experience moves to AI, the system handling the actual transactions stays under its control.

Universal Commerce Protocol: what was said on the call

In the earnings call, executives explained that UCP will connect any AI agent with any merchant without requiring brands to rebuild checkout systems for each new interface. “It’s an open standard for any agent to connect with any brand on the internet,” Finkelstein explained.

Because UCP is an open standard rather than a standalone product, the experience depends on implementation by merchants and integration by platforms. A simplified example would work like this:

• A merchant publishes product data in a UCP-compatible format, including SKU, price, variant options, availability, and checkout endpoint.
• Platforms like Google, participating apps, or AI assistants ingest these standardized product feeds.
• When a user expresses interest in buying through search, an AI prompt, or a marketplace interface, the platform can invoke the checkout endpoint using the UCP specification.
• The user completes checkout through a consistent protocol that interacts directly with the merchant’s system, preserving payment and fulfillment flows.

Unlike simple product feeds or embedded checkout links, UCP will handle the full transaction journey, from search and cart creation to checkout and post-order management, preserving merchant logic such as subscriptions and shipping rules. Independent breakdowns of early implementations suggest the protocol enables AI agents to pass structured purchase intent directly into merchant systems rather than redirecting users through traditional storefront flows.

“UCP is the only protocol that covers the full commerce journey, end-to-end, from search to cart, then checkout to post-order,” Finkelstein said. “It keeps the merchant’s essential checkout logic intact without forcing them to rebuild their customizations over and over again.”

This distinction is central to Shopify’s pitch. AI may control the front-end experience, but Shopify intends to remain the system of record behind it. 

Shopify’s key monetization point: AI does not bypass checkout

A major concern among analysts was whether AI agents might disintermediate Shopify from transactions. If that happened, Shopify would lose control of the transaction layer and potentially lose revenue, data, and influence.

Finkelstein pushed back on that concern, arguing that while large language models may increasingly shape how shoppers discover products or build carts, they do not replace Shopify’s checkout. The interface may evolve, but the transaction layer remains intact. “LLMs do not bypass Shopify’s checkout,” he said.

Finkelstein framed it as a distinction between front-end and back-end infrastructure. “Under UCP, Shopify still powers the overall experience, but the merchant gets to keep their own checkout system on the back end.” 

He also used ChatGPT as an example to describe a split architecture. OpenAI controls the interface, while Shopify runs the transaction infrastructure. OpenAI may control the front end that shoppers interact with, but Shopify still runs the back end. Everything behind the scenes, including order processing and payment handling, is managed through Shopify’s systems, keeping it in control of the transactions. He added that for merchants, the economics remain unchanged. “The economics are the same as if the transaction happened in the online store.”

Shopify earns revenue through subscription fees and payment processing. If AI checkout redirected transactions outside Shopify’s payment rails, it could compress margins. Executives argue that UCP prevents that outcome by integrating Shopify deeper into the transaction flow rather than routing around it.

Orders from AI search are rising, from a small base

Finkelstein pointed to early evidence that AI surfaces are already influencing commerce. “Since January 2025, orders coming to Shopify stores from AI search are up 15x,” he said. “Now, that’s on a small base, but that’s still a really big jump in 12 months.”

The qualifier is important. AI-driven orders remain a small share of overall volume. But the growth trajectory is steep enough for Shopify to treat it as a structural shift rather than a novelty.

For 2025, Shopify reported gross merchandise volume of $378 billion, up 29% year-over-year, and revenue exceeding $11.5 billion, up 30%. Free cash flow surpassed $2 billion. The company entered 2026 with momentum and the balance sheet to invest in long-term infrastructure bets.

Agentic Storefronts expand distribution

Alongside UCP, Shopify launched Agentic Storefronts, allowing merchants to syndicate products to AI systems, including Google’s AI model, Gemini, ChatGPT, and Microsoft Copilot. 

“One click, and our merchants get instant access to millions of potential buyers who are actively looking for their products,” Finkelstein said. Brands such as Glossier, Spanx, Vuori, Stanley, and Steve Madden are already participating, according to the company. 

Shopify also introduced an Agentic plan aimed at merchants not currently on Shopify, enabling product distribution into AI interfaces without a full platform migration. The move suggests Shopify sees AI discovery as both a defensive play and a merchant acquisition channel.

The larger bet

Debates are already emerging around how AI-driven shopping systems may influence pricing transparency and consumer data use, particularly as large platforms expand AI-powered shopping integrations. The larger bet is that AI will reshape the interface, not the infrastructure, and that infrastructure is where the durable value sits.

Shopify is betting that shopping will increasingly happen through AI agents and conversational interfaces, and that product discovery, recommendations, and even cart-building may become automated. But the actual transaction layer, including checkout, payments, order routing, and merchant systems, will remain essential and defensible.

If AI agents become the new “browser,” Shopify wants to be the layer those agents plug into, not the platform they replace. That preserves its control over payments and checkout, merchant relationships, data, and platform revenue.

The expectation is not that AI will leave commerce unchanged. AI is expected to increase transaction volume, and Shopify will still own the pipes.