AI seems to be actually breaking even on time saved, a report from Atlassian shows, with 68% of developers saving more than 10 hours a week thanks to AI, but half are also losing over 10 hours to organisational inefficiencies.

Atlassian’s report on the developer experience does show a marked improvement in AI efficiencies from last year, with more developers citing increases in time saved thanks to the technology.

Developers say that they are using this extra time to improve code, create new features, and document their work.

However, this is where the good news kind of ends as far as AI efficiency improvements, the research shows.

Many companies invested their massive AI budgets into AI coding assistants. While this can be beneficial, developers apparently only spend about 16% of their time coding, which is not typically a friction point.

This is where AI can, conversely “enhance developer experience without necessarily improving it,” Andrew Boyagi, head of DevOps Evangelism at Atlassian said.

While coding may be enhanced by AI assistants, developers are often losing time to non-coding tasks. While 50% are losing over ten hours, 90% lost over six hours due to organisational inefficiencies, the research shows.

Coding was actually not cited as a time water by developers in 2024 or 2025. Finding information, switching between tools, and adapting to new technology, however, were top time wasters, according to the survey.

Developers also cited communication and collaboration struggles, as well as poor instructions when it comes to projects.

Recommended reading

• Organisations Are Failing on DevOps Experience and Maturity
• Developers Are Embarrassed Over Outdated Tech Stacks
• Developer Disconnect Could Waste Billions on Cloud Infrastructure

Perhaps any AI is being treated as an all-around fix, as the empathy gap between leaders and developers appear to be widening.

More than three in five (63%) of developers say their leaders do not understand their struggles, a sharp rise from 44% in 2024.

Treating any AI as beneficial does not go nearly as far in improving worker well-being and efficiencies as simply talking to workers, Atlassian advices.

“We’ve seen evidence that AI can improve developer experience if it’s used to address developer pain points,” Boyagi said.

Speaking with developers to actually address friction points can make a bigger difference than simply plugging in an AI assistant.