SprintGPT

Sprint Intelligence Dashboard with Health Score, Metrics, Insights, and Pattern Detection
Confluence Retrospective Page - Auto-generated with Sprint Metrics and Action Items
One-click Confluence Page Creation from Jira Dashboard
Confluence Content Action - Insert Sprint Retro Report into any page

Inspiration

Sprint retrospectives are broken. Teams spend hours in subjective discussions with no data to back up observations. The same issues resurface sprint after sprint. Action items get forgotten. Scrum masters waste time manually gathering metrics.

I wanted to fix this by bringing data-driven insights to retrospectives - like how F1 race engineers analyze telemetry data to improve performance.

What it does

SprintGPT is an AI-powered Sprint Retrospective Engine that:

Calculates a Sprint Health Score (0-100) using velocity, completion rate, cycle time, and blockers
Provides a Rovo AI Agent for natural language sprint analysis
Detects patterns across multiple sprints automatically
Generates Confluence reports with one click
Auto-triggers analysis on sprint completion

How I built it

Platform: Atlassian Forge (Node.js 22)
Frontend: React 18 Custom UI with F1-themed styling
APIs: Jira Agile REST API for real sprint data
Storage: Forge Storage for persisting sprint history

Challenges I faced

Getting real data - Making sure all metrics came from actual Jira APIs, not mock data
Health Score algorithm - Balancing the weighted formula to give meaningful scores
Installation link expiry