Looker
Google's semantic-layer BI platform with Gemini-powered conversational analytics.
About
Looker is Google Cloud's enterprise BI platform, built around LookML, a semantic modeling layer that defines metrics once so every dashboard, report, and AI answer stays consistent. At Google Cloud Next 2025, Google made Gemini-powered AI available to all Looker-hosted customers.
Conversational Analytics lets users ask questions in plain language and get charts, tables, and narrative explanations grounded in the LookML model, with existing row-level permissions enforced. The LookML Assistant generates LookML code from natural-language descriptions, the Visualization Assistant customizes charts conversationally, and the Formula Assistant writes calculated field syntax. A Code Interpreter (experimental) generates Python for forecasting and anomaly detection.
The major caveat: AI features require Looker-hosted instances on Google Cloud; customer-hosted deployments can't use Gemini features. Looker also carries a steep learning curve (SQL/LookML knowledge expected) and opaque pricing, with independent analyses putting entry around $35,000-60,000/year. It's built for data teams and enterprises, not casual business users.
What it can do
- Conversational Analytics (multi-turn natural language querying grounded in LookML)
- LookML Assistant for generating semantic model code
- Visualization Assistant for conversational chart customization
- Formula Assistant for calculated fields
- LookML semantic layer with row-level security
- Embeddable conversational analytics via API
Setup
Deploy Looker on Google Cloud (hosted instance required for AI features). Model data in LookML with Explores. Enable Conversational Analytics; users ask questions in plain language and get governed answers. Optionally embed via API into internal apps or chat interfaces.
Prerequisites
Looker hosted on Google Cloud (customer-hosted instances excluded from AI features); LookML-modeled Explores; Vertex AI access; SQL/LookML familiarity for modeling.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Google Cloud enterprise BI with LookML semantic layer; Conversational Analytics (Gemini-powered, Next '25) documented on official Google Cloud blog; AI features require hosted instances - honestly disclosed limitation; Opaque pricing ($35-60K/yr entry per independent analyses) disclosed.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.