UK workers report faster work from AI; firms see limited gains
Atlassian survey: 84% of UK knowledge workers use AI and 71% say it speeds tasks, but organisations report little enterprise productivity uplift as processes stay the same.
Recent research by Atlassian surveyed UK knowledge workers and found 84% now use AI, with 71% reporting faster task completion.
Despite widespread individual use, many organisations reported little or no enterprise-wide productivity improvement because existing processes and workflows were not updated to match AI-driven outputs.
Respondents said faster individual outputs created more downstream work, increased pressure on teams and workers, and produced duplicated effort when departments did not share context. About 70% rated their organisation’s processes and workflows as “okay” or “poor.”
Only around 14% of organisations reported they have fundamentally changed collaboration processes to account for AI. Those organisations reported broader productivity gains and highlighted practices such as shared team context, clearly defined roles for humans and AI, and a culture of experimentation and learning.
Molly Sands, head of Atlassian’s Teamwork Lab, said “AI has delivered on its promise of speed, but speed alone doesn’t create business value.” She added companies should move beyond using AI as an individual productivity tool and integrate it into team collaboration so people and AI systems work together more effectively.
An April analysis by Accenture reached similar conclusions, finding that AI benefits often remain at the individual or small-team level and do not extend across the enterprise. Accenture attributed the gap to organisational change lagging behind technology adoption and a focus on individual productivity metrics and cost reduction rather than redesigning infrastructure and processes.
Atlassian recommended firms stop treating AI as only an efficiency tool for individuals and instead empower teams to redesign end-to-end workflows. The company cited steps such as establishing common sources of context, clarifying handoffs between teams and AI systems, and creating incentives for experimentation and knowledge sharing to prevent faster outputs from generating duplicated work.








