Atlassian’s State of Developer Experience 2026 suggests the promise of AI in software delivery is becoming more complicated: the share of developers saying they save ten or more hours a week with AI fell from 68% to 26% in one year, even as AI adoption among engineering teams climbed to near-universal levels.
AI use is up, but the time savings are shrinking
The report says 95% of developers and 96% of engineering leaders now say their organization actively uses AI in engineering. That makes the headline drop in reported weekly time savings more striking, especially because Atlassian frames the data as a shift in where work is spent, not simply whether AI is being used.
Taroon Mandhana, Atlassian’s CTO for AI and Teamwork, summarized the change by saying that “the bottleneck has moved to managing AI’s work.”
What Atlassian says the survey measured
The report is based on responses from 1,155 individual contributors and 405 engineering leaders surveyed during 2026. Atlassian also includes an unusually explicit caution: its findings “reflect self-reported experience and perception, not system-level telemetry such as deployment frequency, cycle time, or revenue impact,” and associations in the data “should not be interpreted as causal proof.”
That matters because the headline number is easy to overread. The survey captures what people say they feel, not a direct measurement of throughput or output.
Developers say AI helps, but it also adds new work
Even with the decline in reported time savings, the report says AI still provides clear day-to-day benefits. Eighty-seven percent of developers say AI helps them ship faster, and 80% say it helps them deliver more without longer hours.
At the same time, only 28% say AI has noticeably reduced friction, while 50% say friction is about the same as before. Atlassian describes that as a different kind of friction rather than a simple reduction in it.
Common friction points remain familiar
- Finding information: 51%
- Meetings: 50%
- Context switching: 45%
- Technical debt: 44%
Those pain points have barely shifted over three years. Finding information also topped the 2025 list and ranked second in 2024, while coding has never appeared on the list, according to the report.
AI itself is now part of the overhead
One of the report’s most notable findings is that the tools meant to speed work can create their own management burden. Forty-nine percent of developers say they spend four or more hours per week managing AI tools, and 65% say they spend time correcting AI-generated work that missed the original intent.
Both figures are higher for engineering leaders, at 57% and 72% respectively. Twenty-nine percent of respondents also identify reviewing or managing AI-generated code as a direct source of friction.
Measurement may be shaping behavior
The report suggests that how organizations measure AI use may be influencing how people behave. Sixty percent of developers say AI use is treated as a productivity signal in their organization, and 41% say they sometimes use AI more than necessary because they feel expected to.
Only 42% of leaders say they have the right metrics to separate real gains from activity. That gap has prompted criticism that teams may be rewarding visible AI usage more than actual business impact.
Shan Huang, who builds production AI agents for enterprise workflows, argued that the issue may be a “metric problem” rather than just a tooling problem. Taylor Johnson of Athena Intelligence put the point more directly: “AI activity is not impact; measure accepted work, not usage.”
Clear strategy appears to matter
Atlassian’s report finds a relatively small share of organizations have a clearly defined AI strategy across the software delivery lifecycle. Seventeen percent of developers and 23% of leaders say their organization has one that is actively implemented.
Among developers who do, 41% report noticeably less friction than before AI. That compares with 22% among those who describe their strategy as inconsistent or poorly communicated. Atlassian’s Andrew Boyagi, head of DevOps evangelism, points to that gap as the differentiator, while the report itself stops short of claiming causation.
The leadership gap keeps widening
One trend has continued in the same direction for three years: more developers feel leadership does not fully understand the challenges they face. That figure was 44% in 2024, 63% in 2025 and 66% in 2026.
That result helps explain why the report’s most important theme may be less about AI capability and more about organizational design. If teams are expected to use AI without changing review, measurement or workflow practices, the gains can be hard to hold onto.
Why the headline drop matters
The decline from 68% to 26% does not mean AI has stopped being useful in software development. Instead, it suggests that the first wave of enthusiasm may have given way to a more realistic picture of the work AI creates around the code, including review, correction and coordination.
Atlassian’s own caveats also make the result harder to overstate. The comparison is based on self-reported survey data, and the company notes that the numbers should not be treated as proof of cause and effect. Still, the shift is notable because it comes alongside near-universal adoption, not in spite of it.
In practical terms, the report points to a familiar enterprise lesson: adopting a new tool is not the same as redesigning the process around it. For many teams, the next productivity breakthrough may depend less on getting AI into more hands and more on deciding how to manage the work it creates.
Source: Original report
Was this helpful?
Explore more: AI Automation Services More Software Development Tech News
Last Modified: October 27, 2026 at 10:32 pm
1 views
