How it works public Slack question → collaborative investigation or code change.
Receive a request An employee mentions River in a public Slack channel.[Source 1
Investigate and act River reads files, runs tests and queries, and posts findings; it can open pull requests.[Source 1
Continue collaboratively Colleagues add constraints and redirect the work within the same thread.[Source 1
Where people stay involved People review, approve, or guide the agent’s work here.
Reported: The CEO reports review of River-authored pull requests. Oversight of other River tasks is not specified.
Supervision level
Not classified
public Slack question → collaborative investigation or code change[Source 1 Supervision level
Work-product review
A person reviews the draft or implementation.
3/5 Level 3 of 5
River-authored code change → reviewed pull request[Source 3 How supervision levels work
Implementation details Model Not reported Harness An agent profile on Aquifer.[Source 1 Sandbox Not reported Tool access Code, tests, data warehouse and production traces.[Source 1 Knowledge World monorepo, skills and shared Slack transcripts.[Source 1 Context management Teams can pre-load channel instructions and skills. River also uses memory that updates across conversations.[Source 3 Credentials Not reported Interfaces Mechanisms Work in public River declines direct messages and asks employees to start a public channel.[Source 3
Carry knowledge forward Teams pre-load channel context, and River updates its memory from shared work.[Source 3
Validation and failure handling Not reported: The available sources do not describe how the team checks the output.
Reported observations These results come from the sources. Notes below each result explain what it covers and any reported limits.
Key observation 59,918 sessions across 5,170 Slack channels in a reported 30-day period; 3,536 River-coauthored PRs merged.[Source 1
Reported by Shopify Scope River: 59,918 sessions across 5,170 Slack channels in a reported 30-day period; 3,536 River-coauthored PRs merged. Method river_sessions domain table written by River each session The source does not report the denominator of this figure.
About this result Category Adoption output Evidence type Reported measurement What it covers River: 59,918 sessions across 5,170 Slack channels in a reported 30-day period; 3,536 River-coauthored PRs merged. Key observation 5,938 employees worked with River across 4,450 Slack channels in the 30 days before 2026-05-09.[Source 3
Reported by Shopify Scope Employees who worked with River and distinct Slack channels with River activity Method Company usage data summarized by the CEO; 30 days ending 2026-05-09 The source does not report the denominator of this figure.
Observed on 9 May 2026
About this result Category Adoption output Evidence type Reported measurement What it covers River employees and Slack channels in a 30-day period ending 2026-05-09. Key observation 1,870 pull requests opened by River in the main monorepo in the week before 2026-05-09.[Source 3
Reported by Shopify Scope Pull requests opened by River in the main monorepo in one week Method Company usage data summarized by the CEO; week ending 2026-05-09 The source does not report the denominator of this figure.
Observed on 9 May 2026
About this result Category Adoption output Evidence type Reported measurement What it covers River pull requests opened in the main monorepo in the week ending 2026-05-09. Key observation River pull request merge rate rose from 36% to 77% over two months, according to Shopify's CEO.[Source 3
Reported by Shopify Scope Merge rate of River pull requests over two months The source does not report the denominator of this figure.
Observed on 9 May 2026
About this result Category Adoption output Evidence type Reported measurement What it covers River pull request merge rate over two months, as reported by Shopify's CEO. Lessons Shared River threads let Shopify employees learn from other teams' requests and reuse their skills.[Source 3
Shopify's CEO attributes River's rising pull request merge rate to public observation and written skills.[Source 3
Back to agents Sources and research details Citations link to the original publisher. Each source also keeps a preserved copy in the repository, so a changed or removed page stays checkable.
Under the River https://shopify.engineering/under-the-river Engineering blog · First party · Last source verification: 2026-08-31 Tobi Lütke on The Knowledge Project: AI Agents, Better Decisions, and the Future of Work https://fs.blog/knowledge-project-podcast/tobi-lutke-3/ Podcast · Direct participant · Last source verification: 2026-09-15 Tobi Lütke on X: Learning on the Shop floor https://x.com/tobi/status/2053121182044451016 Social post · Direct participant · Last source verification: 2026-09-15 Question coverage and scope purpose Reported workflow Reported human involvement Reported: The CEO reports review of River-authored pull requests. Oversight of other River tasks is not specified. implementation Reported validation Not reported observations Reported lessons Reported Research details for every claim on this page Summary Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Harness Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Tool access Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Knowledge Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Context management Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Interfaces Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Receive a request Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Investigate and act Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Continue collaboratively Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Supports Under the River Part II: River only works in the open; At scale; River is one profile Work in public Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Carry knowledge forward Statement type Fact Provenance Reported Confidence Not assessed Confidence reason No explicit confidence assessment is recorded for this claim. Key observation Statement type Metric Provenance Reported Confidence Medium Confidence reason Self-reported by the source; no independent verification. Reported by Shopify Scope River: 59,918 sessions across 5,170 Slack channels in a reported 30-day period; 3,536 River-coauthored PRs merged. Denominator Not reported Method river_sessions domain table written by River each session Observation date Not reported Supports Under the River At scale: recent 30-day river_sessions domain-table counts Key observation Statement type Metric Provenance Reported Confidence Medium Confidence reason The CEO reports exact counts without a measurement method. Reported by Shopify Scope Employees who worked with River and distinct Slack channels with River activity Denominator Not reported Method Company usage data summarized by the CEO; 30 days ending 2026-05-09 Observation date 2026-05-09 Key observation Statement type Metric Provenance Reported Confidence Medium Confidence reason The CEO reports the weekly count without a measurement method. Reported by Shopify Scope Pull requests opened by River in the main monorepo in one week Denominator Not reported Method Company usage data summarized by the CEO; week ending 2026-05-09 Observation date 2026-05-09 Key observation Statement type Metric Provenance Reported Confidence Medium Confidence reason The CEO reports the change without defining the merge rate or its measurement method. Reported by Shopify Scope Merge rate of River pull requests over two months Denominator Not reported Method Not reported Observation date 2026-05-09 Lesson Statement type Fact Provenance Reported Confidence Medium Confidence reason The CEO describes employees learning from public River threads and sharing skills across teams. Lesson Statement type Fact Provenance Reported Confidence Medium Confidence reason The CEO attributes the reported increase in merge rate to employees updating River's skills. Operating model assessment Statement type Inference Provenance Catalog judgment Confidence Unverified Confidence reason The source describes a colleague optionally redirecting a public thread, but does not require continuous steering or a universal PR approval checkpoint. Observation date 2026 Supports Under the River River only works in the open: well-going thread includes a colleague redirect and continued collaboration Operating model assessment Statement type Inference Provenance Catalog judgment Confidence Medium Confidence reason The CEO says employees reviewed River-authored pull requests. This supports review of that work product, but does not define oversight for other River tasks. Observation date 2026-05-09 