3 lessons from rebuilding Atlassian’s product development lifecycle for people and agents

AI tools now handle much of the routine coding and testing, but developers report spending more time managing AI workflows than saving time. The company’s 2026 State of Developer Experience research shows a drop from 68% in 2025 to 26% of developers saving 10+ hours weekly with AI, while 49% now spend four or more hours weekly on AI management. As agents accelerate output, the challenge shifts to defining what to build and ensuring quality, requiring deeper collaboration between product, design, and engineering teams.

Atlassian is redesigning its product development lifecycle to integrate AI without adding layers of supervision. In security workflows, automated pipelines now identify vulnerabilities, propose fixes, and run tests before engineers review them, handling 52% of vulnerabilities and saving roughly 50 engineering days weekly. Engineers focus on validating solutions rather than investigating routine issues, while leaders redefine roles to emphasize human judgment where it creates the most value.

The company is adopting spec-driven development to maintain shared context across teams and agents. Requirements and decisions are captured in evolving specifications that feed directly into Jira tasks executed by AI, reducing the risk of misaligned work. In a Confluence pilot, engineers delivered two production features in 2 and 4 developer weeks, compared to historical estimates of 30 and 40 weeks, demonstrating faster alignment and execution.

Atlassian is expanding who can contribute to product development while maintaining quality standards. An internal initiative enabled over 50 product managers and designers with no prior coding experience to ship more than 100 engineering-approved production changes in three months, with engineers providing oversight. The goal is to embed expert judgment across more stages of work, creating a shared language that accelerates better outcomes without compromising standards.