Looker Agentic Workflows Automate Data Monitoring & Root-Cause Analysis
Quick answer
Looker Agentic Workflows automates data monitoring and root-cause analysis, delivering insights to Slack or email. Preview now available.
Traditional BI alerts are like a capybara splashing in the swamp—they tell you something moved, but not why. Today, Google Cloud introduces Looker Agentic Workflows in preview, turning those noisy alerts into intelligent, automated investigations.
From Ad-Hoc Questions to Continuous Monitoring
With Conversational Analytics, you could already ask questions in plain English. Now, you can turn those questions into persistent monitors—just say “Monitor return rates weekly” or “Notify me if average order value exceeds $1,000.” The agent interprets your intent, confirms thresholds, and generates a workflow plan for your approval.
Root-Cause Analysis Delivered to Slack or Email
When a metric crosses a threshold, the agent doesn’t just ping you. It runs a Key Driver Analysis (KDA) across your data model to pinpoint the culprit—like a specific product category or customer cohort. The full diagnostic lands in Slack or email, no manual digging required.
Dive Deeper with Central Oversight
Each notification includes a link back to Looker, where you can ask follow-up questions and test hypotheses. Admins get a central dashboard to review, edit, or disable workflows across the instance—balancing user flexibility with enterprise governance.
Get Started
Looker Agentic Workflows is available in preview for Looker 26.08+. Admins enable it in Gemini settings, then users with chat_with_agent and create_alerts permissions can start building. Check the documentation to configure your first workflow.
Original announcement published on Google Cloud.