AI has made writing code faster, but it has not made software arrive sooner. That is the central finding of a new Atlassian report, which shows the productivity gains celebrated by the industry are stalling before they reach the finish line.

What You Need to Know

Atlassian's report surveyed developers and found that 87% now say AI helps them write and ship code faster. However, 72% report the same or greater friction across the entire software lifecycle. Only 17% of developers say their employer has a clearly communicated and implemented AI strategy. The disconnect suggests AI is shifting bottlenecks rather than removing them.

The Speed Paradox in AI-Assisted Coding

Nearly nine in 10 developers surveyed by Atlassian say AI helps them write and ship code faster. The figure, 87%, paints a picture of a workforce racing ahead. That picture, however, is incomplete.

Three in four developers also report either the same or greater friction across the entire lifecycle. Faster code generation is not translating into faster software delivery. The bottleneck has simply moved downstream.

87%

Developers say AI speeds coding

72%

Report same or worse friction

17%

Have a clear AI strategy

Longer-term optimism around AI's ability to save time is dropping. Just 25% of developers are saving 10 or more hours a week when using AI, compared with 68% in 2025. That is a collapse in expected returns.

Where the Real Bottlenecks Live

The data points to two human bottlenecks that AI has not touched. Half of developers struggle with finding information, and 50% cite meeting frequency as a major drag. These are not coding problems.

Perceptions of AI deployment are also shifting. Three in five developers feel AI is being implemented too quickly without consideration for its impacts. Merely using AI is seen as a positive metric, but its real-world effects do not fall under that scope.

  • Finding information: 51% of developers say this is a major source of friction.
  • Meeting frequency: 50% cite too many meetings as a drag on delivery.
  • Lack of strategy: Only 17% say their employer has a clear AI strategy in place.
  • Perceived haste: 60% feel AI is being implemented too quickly without considering impacts.

Companies with an effective strategy are in the best position to reduce friction. The Atlassian report found twice as many workers reported positive effects in those organizations.

Why This Matters

The Atlassian findings challenge the assumption that faster code generation equals faster software delivery. For businesses that have invested heavily in AI coding tools, the return on investment may be far lower than expected if workflows remain clogged by information silos and meeting overload.

Developers are already adjusting their expectations. The drop from 68% to 25% saving 10 or more hours a week suggests the initial hype is giving way to a more sober assessment. Companies that fail to pair AI tools with clear strategies may find themselves with faster typists but slower teams.

The report calls for greater consideration and clearer strategies from the top down. Without that, AI risks becoming a costly solution to the wrong problem.

What Leaders Should Do Next

Atlassian Customer CTO Andrew Boyagi framed the challenge directly. "Faster code is not the same as better software delivered sooner," he said. "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."

The prescription is not more AI. It is better integration. Organizations need to treat AI as one part of a larger system, not a silver bullet. That means investing in knowledge management, reducing meeting load and setting clear expectations for how AI fits into the development process.

The report signals a shift from experimentation to accountability. As the early years of AI adoption fade, the pressure is on leadership to deliver real productivity gains, not just faster code.