AI is already helping the work of millions of Australians. But it is not yet transforming productivity for most of them.
New research from The AI Breakpoint, based on a survey of more than 1,000 Australian workers aged 18 to 64, finds nearly three in four workers (72%) say AI now assists with at least some of their work. Based on Australia’s workforce, this indicates more than 10 million workers nationally now have AI helping with some part of their job. But only 9% of workers – around 1.3 million – say AI has made them much more productive.
Closing the gap between these two figures, equivalent to roughly 9 million workers, is the real productivity challenge for Australian workplaces.
The gap shows that adoption alone is not the same as productivity. Overall, two in five workers (41%) say AI has made them more productive over the past 12 months, but most of these gains are still moderate, with 32% saying AI has made them somewhat more productive.
Workers who have embedded AI more deeply into their day-to-day workflows are more likely to report productivity gains. Among those who say AI assists with a great deal of their work, 90% say it has made them more productive, compared with 31% among those who use AI only a little for work. More frequent AI use, however, does not automatically translate into deeper productivity. What matters is whether AI is being applied to the kinds of tasks where it can create deeper, repeatable value.
Work tasks AI is being used for (BAR CHART)
At work AI is most widely used for general knowledge-work tasks such as research, summarising, emails and writing, but stronger productivity gains appear in more structured or role-specific workflows. While just 9% of workers say AI has made them much more productive, this rises to 27% among workers using AI for project planning or strategy, 25% for data analysis, 23% for coding or technical work, 23% for presentations and 22% for creative work or design.
“This is the difference between AI helping work and it changing work,” Breakpoint founder Ian Walker said.
“The productivity opportunity is not simply about getting more workers to use AI more often, it is helping workers move from occasional prompts and light task support into role-specific AI capability.”
The AI Productivity Ladder represents how this plays out.
At the base of the ladder, 78% of Australian workers use AI in some form for personal or professional reasons. The next step is workplace use, with 68% using AI for work.
From there, the ladder narrows. Only 43% say AI assists with a moderate amount or more of their work. Then 32% say AI has made them somewhat more productive. And just 9% say it has made them much more productive.
The pattern suggests the deepest gains are concentrated further up the curve by a smaller group of workers using it for strategy, data, coding and creative work.
“For many workplaces, the gains are being discovered by workers before they are being designed by employers,” Walker said.
“That means productivity is happening, but it is uneven. The next stage is turning scattered use into repeatable capability.”
The challenge for employers is not whether workers will use AI. Many already are. The question is whether workplaces can identify where AI creates the strongest workflow gains, then provide the training, rules, approved tools and quality checks needed to make those gains consistent.
The report reveals Australia is now past the first stage of AI adoption at work and the next test is productivity maturity. More workers have AI assisting their work than ever before, but the real value will come when more of them can move up the ladder from light work to meaningful productivity gains.
