Atlassian’s 2026 State of Developer Experience report reveals that while AI helps 87% of developers ship code faster, 72% still face the same or increased friction, highlighting systemic constraints beyond individual coding tasks.
- 87% of developers say AI speeds up shipping code, but 72% face unchanged or more friction.
- Clear AI strategies are lacking, with only 17% of developers seeing defined organizational plans.
- Top friction points remain: information access, meetings, context switching, and unclear requirements.
What happened
Atlassian’s latest report on the developer experience shows a paradox where AI tools accelerate coding activities but do not reduce overall friction in the software delivery process. A majority of developers report faster output, yet many experience the same or higher levels of bottleneck issues downstream, such as code reviews and coordination complexities.
The data reveals a sharp decline in the percentage of developers gaining significant time savings from AI—from 68% saving 10+ hours weekly in 2025 to only 25% in 2026. Furthermore, 65% of developers now regularly spend time fixing errors from AI-generated code, indicating that faster generation does not translate into faster delivery.
Why it matters
These findings demonstrate that optimizing individual stages in development, like coding speed, without addressing systemic constraints leads to inefficiencies and potential overload for downstream teams. Just as improving a part of a Formula 1 car can create bottlenecks elsewhere, increasing developer output via AI shifts workloads rather than alleviating friction.
Persistent challenges such as difficulty finding information, excessive meetings, context switching, technical debt, and ambiguous requirements remain the top irritation points for developers. These are areas where AI could unlock real value by improving coordination, reducing repetitive tasks, and enhancing information flow, but organizations have yet to capitalize fully on this potential.
What to watch next
Organizations should focus on developing clear, communicated AI strategies that address the entire software delivery ecosystem rather than treating AI usage purely as a productivity metric. Only 17% of developers currently report that their organizations have such strategies in place, signaling a significant opportunity for leadership to align AI efforts with genuine developer needs.
Further progress will come from leveraging AI to reduce the major sources of friction rather than just accelerating code writing. Tracking how AI impacts bottlenecks like code reviews, cross-team dependencies, and information discovery will be key to turning increased developer output into faster delivery of high-quality software.