Discover how Databricks Genie transforms data exploration by letting you ask questions  in plain English. Learn to navigate Genie spaces, refine your prompts for precise answers,  and understand how Genie translates natural language into SQL behind the scenes.

Your first questions

Your first conversation with Genie

Unlocking data without code

Options

Ready for Genie

Needs Refinement

Genie-ready questions

The art of asking better questions

Effective Prompt Formula

Building better Genie prompts

Refining your Genie queries

Reading the generated SQL

How Genie works

The Brain (LLM)

The Map (Unity Catalog)

The Muscle (SQL Warehouse)

Which component does what?

Genie's three components

Genie's Processing Pipeline

How Genie processes a question

Getting Started with Genie

Build your first Genie space from scratch. Configure data sources, craft example queries  that guide users toward valuable insights, and write clear instructions that help  business users get the most out of conversational analytics.

Creating a space

Prerequisites for Genie spaces

Creating your first Genie space

Space creation steps

Adding example queries

Evaluating example query quality

Adding your first example query

When to add an example query

Guiding users with instructions

Effective instructions

Observing instructions in action

Evaluating space instructions

Building Genie Spaces

Curate your Genie space to speak your organization's language. Define synonyms and  descriptions for technical columns, establish table relationships, and create SQL  examples that teach Genie how to answer domain-specific questions accurately.

Synonyms and descriptions

Querying the curated space

Matching business terms to columns

The synonym gap

Defining relationships

Cross-table queries

Queries without defined relationships

Configuring a table relationship

SQL examples and functions

Querying a pre-defined SQL example

Classifying KPI complexity

SQL example or auto-generated?

Teaching Genie Your Business Language

Keep your Genie space performing at its best. Implement feedback loops to learn from  user interactions, leverage trusted assets for consistent answers, and benchmark  accuracy to drive continuous improvement over time.

The feedback loop

Identifying failed queries

The feedback loop steps

Diagnosing a failed query

Trusted Assets

Creating a Trusted Asset

Understanding Trusted Assets

Trusted Asset or regular example?

Benchmarking and continuous improvement

Viewing benchmark results

Prioritizing curation tasks

Reading benchmark results

Wrap-up

Monitoring and Maintaining Genie Spaces

Ever wished you could ask your data a question in plain English and get an instant, accurate answer? With Databricks Genie, you can. In this course, you'll go from asking your first natural-language question to building and curating production-ready Genie spaces. Learn to refine prompts, add example queries and instructions, teach Genie your business terminology through synonyms and relationships, and lock down critical metrics with Trusted Assets. By the end, you'll confidently set up, maintain, and improve Genie spaces that empower your whole team to explore data without writing SQL.

Business users shouldn't wait hours for a data team to answer straightforward questions. Databricks Genie turns plain English into instant, accurate SQL — this course shows you how to make the most of it.
<br><br>
<h2>Ask Questions and Refine Your Prompts</h2>
Ask your first natural-language questions in a Genie space and interpret the results. Learn the prompt formula — what, when, how, plus constraints — and refine vague queries into precise insights. You'll see how Genie's compound AI system combines an LLM, Unity Catalog metadata, and a SQL warehouse to deliver answers in seconds.
<br><br>
<h2>Build and Configure Genie Spaces</h2>
Create a Genie space from scratch: select data sources, write parameterized example queries that route common questions to verified SQL, and add instructions and filters that guide users toward accurate answers.
<br><br>
<h2>Teach Genie Your Business Language</h2>
Bridge the gap between business terms and technical column names with synonyms, descriptions, and table relationships. Define SQL examples for complex KPIs that Genie can't auto-generate, so every answer reflects how your organization measures success.
<br><br>
<h2>Monitor, Trust, and Improve</h2>
Spot failed queries on the monitoring page, diagnose root causes, and feed fixes back into your space. Lock down critical metrics with Trusted Assets and benchmark accuracy over time so your space keeps getting smarter.

Learn Databricks Genie hands-on: ask questions in plain English, build Genie spaces, curate business language, and monitor quality.

Introduction to Databricks Genie

Ask data questions in plain English with Databricks Genie - build spaces, curate business language, and monitor quality.

How many orders were placed in January 2025?

What should we do about customer complaints?

Genie-ready questions

Exercice interactif pratique