River

River is the name of an AI agent used inside Shopify. Its existence has been made public through the company's engineering blog and posts from its leadership, and it embodies a way of working in which employees delegate parts of their daily work to an AI agent.

It is not a publicly available product but an internal system at Shopify. Detailed internals and architecture go beyond what has been published, so there is only so much that can be stated definitively from the outside.

River stands out as a case where an ecommerce platform vendor itself has spoken publicly about weaving AI agents into its internal operations. It reflects a stance that treats AI not merely as a feature shipped inside a product, but as infrastructure that reshapes how an organization works.

That message carries weight for partners and agencies in the Shopify ecosystem. If the platform side moves toward development and operations that assume AI agents, the teams building and maintaining on top of it are likely to face the same shift.

We covered River in our three-part “AI-first” column series — how it relates to Shopify's company-wide monorepo “World,” how human feedback accumulates as codebase knowledge through “pebbles,” and the design of Aquifer, the agent infrastructure that runs it. Part 3, “From DX to AX (Agent Experience),” goes into the most depth.

Flagship, too, is building AI agents into its own development workflow — design, implementation, and review — and treats River, a case of the platform itself shifting to agent-first ways of working, as a reference point for how a partner team should be organized. River itself is an internal Shopify system; it is not something we use or deploy directly.