Measuring and orchestrating AI across the full engineering lifecycle

Atlassian has announced its agreement to acquire DX, a developer intelligence platform, to address a growing disconnect between AI tool adoption and measurable engineering outcomes. While AI coding assistant adoption now exceeds 95% among software engineers, high-level innovation output has remained largely unchanged. Engineering leaders are shifting focus from individual productivity metrics to system-wide capacity, flow, and business impact to bridge this gap.

DX’s data from over 500 organizations reveals that engineers using AI coding assistants save more than six hours per week, median pull request throughput has doubled, and onboarding time for a newly onboarded engineer to merge their 10th pull request has dropped by approximately 57-58% since early 2024. Documentation satisfaction has improved by roughly 20% as AI reduces the overhead of maintaining technical knowledge, though these gains primarily reflect localized efficiency rather than systemic delivery acceleration.

Research indicates that focusing solely on code volume can obscure broader bottlenecks in the delivery pipeline. Writing code accounts for only about 16% of an engineer’s typical working week, with the remaining 84% spent on requirements refinement, testing, security reviews, and operational tasks. AI-driven acceleration of code authoring compresses only a fraction of the total delivery lifecycle, often shifting bottlenecks rather than eliminating them.

Uma Namasivayam, Senior Director of Engineering Productivity at Dropbox, highlights that tripling code throughput without upgrading upstream and downstream stages risks overloading the software development lifecycle. Dropbox implemented a holistic value stream strategy, emphasizing multi-layered measurement frameworks that track cycle time, change failure rates, and PR review cycle times alongside code generation metrics to ensure AI investments translate into sustained business value.