If you haven’t already, check out Arun Ulag’s hero blog “FabCon and SQLCon 2026: Unifying databases and Fabric on a single, complete platform” for a complete look at all of our FabCon and SQLCon announcements across both Fabric and our database offerings.
Co-authors: Jene Zhang, Jenny Jiang, Qixiao Wang
Context awareness built in and immediately accessible
Microsoft Fabric notebooks offer a Copilot feature that adapts to complex tasks, providing helpful context-aware support as soon as you start using them.Copilot no longer requires a session to be started or for you to re-explain what you’re working on. Instead, it automatically understands your workspace, attached Lakehouse, notebook structure, and execution environment, so assistance is available immediately and remains relevant as your work evolves.
Copilot provides always-on, notebook-aware assistance and operates across entire notebook workflows, helping you write larger blocks of code faster, understand complex notebooks more easily, and make changes with confidence while staying in control of what context is used.
What this enables
Copilot has built-in awareness of your workspace and notebook context. Copilot understands notebook flow, attached Lakehouse schemas and tables, files, and runtime state.Copilot in notebooks can help you move through multi-step notebook work end-to-end, planning, generating, refining, and validating code as your notebook evolves.
You can use Copilot to:
- Develop faster: Generate cohesive code across multiple steps, not just single cells.
- Build complete workflows: Create multi-step transformations and validation logic directly in your notebook.
- Refactor with confidence: Rewrite existing logic to reduce duplication and improve readability.
- Understand at scale: Summarize complex or unfamiliar notebooks quickly.
- Debug and troubleshoot sooner: Use Fix with Copilot to get an error summary and recommended fix, then iterate with full context.
Understand performance as you build
Beyond generating code, Copilot in Fabric notebooks helps you understand how a notebook behaves as you work, so potential performance issues can be identified earlier.When prompted, Copilot draws on your notebook’s structure, data, and execution environment to surface opportunities such as:
- More efficient join strategies.
- Avoid unnecessary data shuffles.
- Refactoring logic into reusable functions.
- Highlighting potential data quality issues.
Diagnose code failures
Debugging code just got a lot easier. Copilot in Fabric notebooks helps you diagnose and resolve failures directly within your development workflow—no context switching required.Whenever a cell execution fails or a failed Spark job is detected, a Fix with Copilot button automatically appears below the affected cell. Selecting this, opens a Copilot panel which instantly pulls in the cell code, running context, and Spark execution details—no manual copy-pasting needed.
Figure: Fix with Copilot provides error summary and suggested fixes.
Copilot then generates a rich diagnostic summary drawing from Microsoft Fabric documentation, Spark community knowledge, and other public resources. You'll get:
- A clear summary of the error message.
- A root cause analysis.
- Actionable recommendations for fixes.
Beyond cell-level diagnostics, you can also trigger analysis using the /Fix command directly in the Copilot chat. This command lets you diagnose a specific cell or the entire notebook—giving you flexible control over how and where you investigate issues.
Getting started
Copilot now works alongside you as you write, understand, refine, and improve your notebook, helping you move faster from code to insight in Microsoft Fabric.Try the new Copilot experience today. To learn more, visit the Copilot for Data Engineering and Data Science documentation.