- Atlassian report claims the number of developers saving significant time through AI is actually dropping
- Finding information and participating in meetings are major human bottlenecks
- More then four in five organizations still don't have a clear AI strategy
Nearly nine in 10 (87%) developers surveyed by Atlassian now sayAI is helping them to write and ship code faster, but despite offering significant speed improvements across certain parts of the workflow, three in four (72%) also report either the same or greater friction across the entire lifecycle.
Longer-term optimism around AI's ability to save time is also dropping, with just 25% of developers saving 10+ hours a week when using AI, compared with 68% in 2025.
Ultimately, Atlassian's findings point toward AI simply shifting the bottleneck further downstream rather than tackling time constraints altogether.
Is AI actually helping developers?
The data shows that faster code generation isn't translating into faster software delivery, with the reality of AI-assisted coding now highlighting other areas where AI could actually help. For example, half of developers struggling with finding information (51%) and meeting frequency (50%)
"Faster code is not the same as better software delivered sooner," Atlassian Customer CTO Andrew Boyagi explained. "If teams still struggle to find information, clarify requirements and move work through the delivery process, speeding up coding alone won’t translate into business outcomes."
However, many developers simply feel that AI is being implemented too quickly without consideration for its impacts. Three in five say merely using AI is seen as a positive metric – but its real-world impacts don't fall under that scope.
Companies with an effective strategy were found to be in the best position to reduce friction, where twice as many workers reported positive effects. Still, not even a fifth (17%) of developers say their employer has a clearly communicated and implemented AI strategy.
Now that we're beginning to emerge from the early, experimental years of artificial intelligence, and that use cases are starting to be identified, the report calls for greater consideration and clearer strategies from the top down in order to unlock the real benefits.
