FlowPilot -- Intelligent Flow for Software Teams

Inspiration

Modern software teams don’t struggle because they lack tools — they struggle because their tools don’t talk to each other intelligently.

While using Jira and Bitbucket, we noticed a recurring pattern:

Jira issues stuck in In Progress for days
Pull requests waiting endlessly for reviews
Developers context-switching between Jira, Bitbucket, and dashboards
Managers relying on manual follow-ups instead of actionable signals

This inspired FlowPilot — a lightweight, developer-friendly app that focuses on workflow flow, not micromanagement.
The goal was simple: remove friction so teams can focus on shipping quality software.

About the Project

FlowPilot is an Atlassian Forge app that intelligently connects Jira Software and Bitbucket to help software teams:

Detect workflow bottlenecks early
Reduce PR review delays
Keep Jira issues and code changes perfectly in sync
Visualize sprint health in real time

Instead of static reports, FlowPilot actively nudges teams at the right moment, making collaboration smoother and faster.

What We Learned

Building FlowPilot helped us deeply understand:

Atlassian Forge’s security-first, serverless architecture
How to design event-driven productivity tools
Balancing automation without overwhelming developers
The importance of authoritative data when integrating multiple platforms

We also learned that developer experience (DX) matters as much as features — tools should help quietly, not interrupt constantly.

How We Built It

FlowPilot is built entirely on Atlassian Forge, ensuring:

No external servers
Secure, permission-scoped access
Native integration inside Jira and Bitbucket

Core Architecture

Custom UI (React) for dashboards and panels
Forge backend functions for workflow analysis
Jira REST APIs for issue data
Bitbucket REST APIs for pull request insights
Scheduled triggers to detect bottlenecks automatically

This allows FlowPilot to compute:

Average PR review time
Issue cycle time
Stuck or blocked issues
Reviewer overload patterns

Challenges We Faced

Understanding the separation between frontend Custom UI and backend Forge functions was initially challenging, especially resolver wiring.
Keeping Jira and Bitbucket data consistent required careful handling of identifiers, permissions, and API scopes.
We had to ensure FlowPilot supports developers instead of feeling like a monitoring tool. This led us to focus on team-level insights, not individual rankings.