Case studies / Shopify

River, a Slack-native agent on the Aquifer substrate

Shopify built River, an AI agent that only operates in public Slack channels. It reads code, runs tests, opens pull requests, queries the warehouse, and inspects production traces. Underneath, Aquifer is the internal platform (durable session, harness, sandbox, gateway, credentials proxy, observability). The May 2026 engineering post is explicit that the 2024 monorepo-and-Nix work was done so agents would have a substrate.

Dated 28 May 2026

Scale claims

• Verified One in eight merged pull requests across Shopify is coauthored by River.

  shopify.engineering/under-the-river — Shopify Engineering, 28 May 2026, “Under the River.” First-party. No independent audit.

• Verified Median River session is 19 minutes with a median of 50 tool calls.

  shopify.engineering/under-the-river — Same post.

• Verified In a recent 30-day period, 59,918 River sessions ran in 5,170 Slack channels, touching 7,000+ people.

  shopify.engineering/under-the-river — Same post. Shopify says the numbers come from a river_sessions table River writes itself.

• Verified 3,536 River-coauthored pull requests merged in that 30-day window.

  shopify.engineering/under-the-river — Same post and same table. The authors note the usage numbers were already moving up as they published.

Sources

• https://shopify.engineering/under-the-river
• https://shopify.engineering/building-an-agentic-harness-that-outlasts-the-model
• https://shopify.engineering/introducing-roast

Named tools

Only products the primary source names, and only if we list them. claude-code