Google Cloud adds agentic workflows to Looker analytics
Wed, 29th Jul 2026 (Today)
Google Cloud has introduced Looker Agentic Workflows in preview for its Looker analytics platform, adding automated metric monitoring and root-cause analysis.
The update extends Looker's existing Conversational Analytics tool, which lets users query business data in natural language. With the new workflow feature, users can turn one-off requests into ongoing monitoring routines from the chat interface.
Instead of setting up conventional alerts that only flag when a number changes, users can ask the system to watch a metric on a recurring basis or notify them when a threshold is crossed. For example, they can ask it to monitor return rates each week or send a notification if average order value rises above a set level.
Google Cloud said the system interprets the request, confirms the threshold conditions, and produces a workflow configuration plan for review before activating the monitor. Users must approve the setup rather than have it run automatically from the first prompt.
Alert analysis
When a monitored metric breaches a defined threshold, the background agent can go beyond a basic alert. It can run a Key Driver Analysis across the underlying data model to identify factors linked to the change, including product categories or customer cohorts.
The system then sends a diagnostic summary through Slack or email, reducing the need for analysts or business users to search dashboards manually after an alert is triggered.
Each notification also includes a link back into Conversational Analytics in Looker. Opening it starts an interactive session with the diagnostic findings, allowing users to ask follow-up questions and investigate further inside the platform.
Oversight controls
The feature includes a central workflow management interface designed to give administrators visibility into active monitors. Business users can view and edit their own workflows, while administrators can review, adjust, or disable workflows across a Looker instance.
This structure addresses a common tension in business intelligence software between self-service access for staff and central control over how data tools are used. In practice, Google Cloud is packaging automated analysis in a way that still leaves oversight with platform administrators.
The feature arrives as software vendors continue adding AI-based assistants to data and analytics products, with a particular focus on natural language interfaces and automating routine reporting tasks. In many organisations, business intelligence alerts have remained limited to threshold warnings, leaving analysts to investigate the causes of changes through separate dashboard work.
Looker Agentic Workflows is available in preview for Looker version 26.08 and later. Administrators can enable it through the Gemini in Looker settings page, and users with chat_with_agent and create_alerts permissions can begin creating workflows once the setting is turned on.
The feature is designed to turn conversational prompts into persistent monitoring tasks and combine them with automated diagnostic work when exceptions occur. It is initially being introduced as a preview feature for customers running recent versions of Looker.