Atlassian’s Forward Deployed Engineering Playbook

By the numbers

How it works

Atlassian FDEs are senior engineers, bringing AI engineering and applied AI together in a single role and building alongside your teams in your environment.

Atlassian FDEs start with a business outcome, then trace the work required to achieve it: who is involved, which systems hold the necessary knowledge, where handoffs break down, what permissions apply, and what an agent needs to know and do. Then they build. FDEs write production code, create agents, connect systems, and redesign workflows around AI.

Lessons from the field

• Pick the right problem.
  Focus on issues that are frequent, painful, and measurable, then work backward to figure out what AI needs to solve them.
  
• Treat governance as part of the architecture.
  Build permissions, security, and auditability in from day one, increasing autonomy as trust grows.
  
• Design AI to become part of how work gets done.
  Curiosity drives initial adoption. Value drives sustained usage. The most effective, widely used AI is embedded in the tools and workflows where people already collaborate.
  
• Optimize the system, not just the model.
  At enterprise scale, economics matter. Reuse context, match compute to the task, and measure AI against the work it replaces.

Results

• Technology company: $20M+ saved
  FDEs deployed AI that autonomously resolves tier-1 and tier-2 issues, cutting mean time to resolution from 18 days to 10 minutes while maintaining customer satisfaction.
  
• Semiconductor manufacturer: ~50% less triage effort
  FDEs built a custom triage agent while navigating complex enterprise governance requirements, unblocking production AI rollout across 40,000 employees.
  
• Travel company: ~6,700 estimated hours saved annually
  FDEs co-built two production Rovo agents, one that archived ~600,000 pages and one for work-item triage, estimated to save 6,700+ hours a year across 20+ teams.