Developers save time with AI, yet many still lose 10+ hours a week to friction

What Atlassian asked and found

Atlassian, the software company, ran its 2025 State of Developer Experience research. It surveyed 3,500 developers and managers across six countries. The aim was to learn how daily work for people who build software is changing as AI spreads, according to a write-up on atlassian.com.

The headline result sounds like a clear win. Almost every developer, 99%, now says AI tools save them time, and 68% say they save more than 10 hours a week. A year earlier, most developers had not yet seen real productivity gains from these tools.

The survey also looked past coding. It tried to measure how generative AI (software that produces text, code and other content on request) affects the whole working week, not just the time spent typing code.

Where the saved hours go

The author notes that the jump since 2024 is large, especially for tasks that are not coding. Time saved only matters if it is used well, and here the report is encouraging. Developers say they spend the extra time improving existing code, building new features and writing documentation.

So the first half of the story is mostly positive. The tools work, people use them, and the hours go to useful work.

The hours that still leak away

The second half is less cheerful. Developers lose time to tasks outside coding. In the survey, 50% report losing 10 or more hours a week, and 90% lose 6 or more hours. The report blames mostly organizational inefficiencies, meaning how companies are set up and run rather than how fast anyone types.

The author sums it up as breaking even: about 10 hours saved with AI, about 10 lost to friction. Developers rank the biggest time-wasters as finding information such as services, documents and APIs (the connection points between software systems), adapting to new technology, and context switching, which means jumping between different tools.

Two shifts from last year stand out. Tech debt, the buildup of old shortcuts in code that cause trouble later, dropped out of the top five. Working with other teams moved up. Coding itself appears on neither year's list. Developers only spend 16% of their time coding, so tools aimed purely at coding were never aimed at the main problem.

A widening gap with leaders

The report also measures how well managers understand the people they manage. 63% of developers now say leaders don't understand their pain points. Last year that figure was 44%.

The author's explanation is that leaders may be counting the hours AI saves while leaving existing friction untouched. The result is a false economy: pressure to deliver faster, while the obstacles stay in place. This is the author's interpretation, described as likely rather than proven by the survey.

What the report tells teams to do

The advice starts with a simple step: talk to developers. Leaders should learn where the friction is and test possible fixes with the people affected. AI will be part of the answer in many cases. In others, the fix might be as plain as self-serve guides that people can read without waiting for a colleague.

The author also puts some work on developers. They should explain their problems in terms of impact, so the issue reads as something to act on rather than a complaint to brush aside. The report argues that regular two-way conversation helps teams catch problems early and build trust.

The measure of success is straightforward. In a well-run organization, more developers would say their leaders understand them, and fewer would lose hours to inefficiency. Note that this is a company's own research, and the source gives no detail on the survey method beyond sample size and country count.