Atlassian 2025: AI saves developers time — organizational friction wipes much of it out

# What the report says

# The problem: equal losses elsewhere Those gains sit next to a countervailing trend. Half of developers report losing 10 or more hours per week to nontechnical inefficiencies, and 90% lose six or more hours. In aggregate, the time saved by AI is frequently being absorbed by organizational friction, producing a near-zero net improvement in weekly capacity for many teams.

# Where time is lost Atlassian identifies consistent sources of developer friction across the software development lifecycle (SDLC):

• Finding information (services, docs, APIs) — searching for authoritative, up-to-date references is a common blocker.
• Adopting new technology — onboarding and integration gaps cause churn and slow delivery.
• Context switching between tools — moving among code, issue trackers, docs, and CI causes lost flow time.
• Collaboration with other teams has moved into the top friction points in 2025, even as tech debt fell out of the top five.

# The empathy gap between leaders and developers A growing split exists in perceptions. In 2024, 44% of developers said leaders didn't understand their pain points. In 2025 that rose to 63%. The report links this to leaders celebrating AI-derived time savings without simultaneously fixing the structural and process issues that consume developer time.

# Practical actions recommended Atlassian's recommendations are straightforward and specific:

• Start by talking with developers. Gather direct, concrete descriptions of where time is lost and quantify impact in hours or blockers.
• Test solutions with developers. Pilots and feedback cycles show whether a proposed fix actually reduces friction.
• Use AI where it addresses real friction across the SDLC. AI investments that only speed individual coding tasks won't reduce the broader time sinks developers face.
• Build simple non-AI fixes when they work better. Self-serve enablement materials, clearer documentation, and centralized discovery can be cheaper and faster than a tooling overhaul.
• Encourage developers to frame problems in impact terms so leaders can prioritize fixes rather than dismiss complaints.

# Bottom line AI is delivering measurable time savings for developers, but organizational inefficiencies — information gaps, onboarding, tool switching, and cross-team collaboration — are consuming much of those gains. Closing the gap requires disciplined listening, cross-functional fixes, and choosing AI only when it targets the friction that actually costs developer hours.