Blog Post

Power BI Community Blog
2 MIN READ

Data Days | SQL + AI Promptathon

slindsay's avatar
slindsay
Icon for Community Admin rankCommunity Admin
2 months ago

Return to Data Days Homepage


SQL + AI Promptathon - now closed

 

Congratulations to everyone who participated. Read on to learn more about the winners and the judges favorites!

 

👀 Watch the kickoff livestream, either live or on-demand!

 

 

Build with Copilot. Powered by SQL. Are you ready to go beyond a single prompt?


The SQL + AI Promptathon is all about building something real with Copilot, SQL, and agentic workflows. This is a multi-turn experience where you collaborate with an AI agent, explore live data, and create something meaningful along the way.

 

👉 Get started here: https://aka.ms/dd/promptathon

 

What to Expect

 

Everything you need lives in the repo. Just open it and go.

  • A ready-to-use Codespaces environment
  • A preconfigured SQL database
  • Missions to guide your experience
  • Submission templates and resources

No setup friction. Just open Copilot Chat, switch to agent mode, and start building.

 

Quick Start

  1. Fork the GitHub repo
  2. Open the repo in Codespaces
  3. Let the environment finish loading (it takes ~5 minutes)
  4. Open Copilot Chat and switch to agent mode
  5. Work through a mission and capture your journey

Choose Your Mission

Step into the role of a principal on Zava’s data team:

  • Data Engineer – Build a reusable support intelligence data mart
  • Data Scientist – Create and evaluate a voice of customer system
  • Data Analyst – Investigate a hidden product quality issue
  • Bring Your Own – Define and solve your own problem

Explore all missions here: View Missions 


What You’ll Do

This challenge is about the process, not just the outcome.

  • Set a goal and pursue it across multiple turns
  • Ask Copilot to work against live SQL data
  • Build something real (notebook, model, report, tool)
  • Show your full journey, including pivots and learnings

What to Submit

Submit via GitHub Issues in the repo.

Your submission should include:

  • Your goal and approach
  • Evidence from SQL MCP tool calls
  • A working artifact
  • A simple architecture or data flow view
  • Reflections on what worked and what didn’t

Start here: https://aka.ms/dd/promptathon

 

Resources

You’ll find everything in the repo, including:

  • SQL database guide
  • Copilot documentation
  • AI + SQL learning resources

 

Ask Questions. Share Your Journey.

If you get stuck or want feedback, ask in the comments on this post.

Chances are someone else is working through the same thing, and that’s what makes this community strong.

 

Final Thought

This isn’t about getting the perfect answer. It’s about learning how to work with the agent. Experiment, iterate, and share what you learn!

We can’t wait to see what you build 💙

Updated 1 month ago
Version 6.0

19 Comments