What Happened
The June 2026 edition of Deloitte Access Economics' quarterly Employment Forecasts report landed with a clear message: AI's structural impact on Australia's workforce is no longer theoretical. It is showing up in the data.
Deloitte Access Economics assessed occupations by how far their core tasks can be automated by AI versus how much they depend on human judgement, empathy, or interpersonal skill. The result was a list of 82 "AI-disrupted" occupations - roles where AI can replace enough of the work to put employment levels at genuine risk over the coming years.
The headline jobs number adds context. Annual employment growth in the year to April 2026 was 0.9 per cent, a significant slowdown from the 1.9 per cent average recorded over the previous three years, when hiring across the public sector provided stronger support. That deceleration is not solely an AI story, but Deloitte's analysis suggests AI is a contributing factor in specific occupational categories.
David Rumbens, partner at Deloitte Access Economics, noted the pattern so far has been more cautious than some forecasters predicted. "Limited evidence of widespread job losses could suggest that AI is currently playing more of an augmentative role in the Australian labour market, with Australians less likely to use AI primarily for automation," Rumbens said.
Why It Matters
The distinction between augmentation and automation is not a minor semantic point - it has real consequences for how businesses plan and how workers prepare.
Right now, most Australian workers using AI are doing so to get through their own workload faster. They are drafting documents, summarising meetings, running analysis. The job still exists; the person just does more of it. Rumbens put it plainly: "These augmentation-focused usage patterns suggest AI could deliver much-needed productivity gains while job gains are still seen."
That is the good news. The harder truth is that augmentation can be a precursor to automation. Once workflows are redesigned around AI-assisted output, the number of people needed to produce that output can shrink. As Rumbens noted, "This may also simply reflect the current stage of adoption, with AI tools still used mainly to boost individual productivity before a later phase of more significant change as work processes are reorganised."
For Australian businesses, the implication is direct. Waiting to see how this plays out is itself a strategic choice - and not a safe one. Rumbens was pointed on this: "Executives should not be complacent in their existing business models, work design and workforce strategies, and workers should seek to upskill and retrain in AI."
Our AI strategy work and AI training programmes are built around exactly this transition - helping organisations move from ad hoc AI use to deliberate work redesign before the pressure becomes acute.
Key Details
Deloitte's methodology separated occupations by the degree to which AI can handle their core tasks without human involvement. The 82 occupations flagged as most at risk are those where the work leans heavily on repeatable, rule-based, or information-processing tasks - the kind of work AI handles well.
Job vacancies in some of these disrupted occupations have already started to fall. Employers are hiring fewer people into roles that AI can increasingly cover. Employment levels in those same occupations have continued to rise to date - existing workers are not being let go in large numbers yet. The disruption is arriving through the front door, not the back: fewer new hires, not mass redundancies.
The investment case for getting this right is substantial. Research cited in the Deloitte report showed that "organisations are twice as likely to exceed their return on investment expectations for AI when they prioritise work redesign and thoughtfully redesign human and machine interactions." That is not a marginal improvement - it is the difference between AI being a cost centre and a genuine competitive advantage.
Our AI automations service focuses on this exact redesign challenge, working through which tasks to hand to AI, which to keep human, and how to restructure the handoffs between them.
Background and Context
Australia's labour market has been running hot for several years, supported by strong public sector hiring and a tight post-pandemic jobs market. The slowdown to 0.9 per cent annual employment growth in the year to April 2026 reflects both the fading of that public sector tailwind and the early signs of AI beginning to reduce demand for certain types of work.
Deloitte Access Economics has been tracking AI's labour market effects as part of its quarterly Employment Forecasts series. The June 2026 edition is notable for moving beyond scenario modelling into observed data - vacancies falling in specific occupations, adoption patterns showing up in survey responses, and the augmentation-versus-automation split becoming measurable.
The Australian Government's National AI Plan identifies spreading the benefits of AI across the economy as a core objective, including making sure productivity gains translate into broader workforce opportunity rather than concentrated gains for a small number of firms or workers. That policy framing matters for how businesses in professional services and other AI-exposed sectors approach their workforce planning.
Rumbens flagged the pace risk clearly: "The volume of investment in AI over the past years suggests that AI will likely move along the adoption curve much faster than previous technologies, creating winners and losers in the process."
What Comes Next
Deloitte Access Economics expects the structural impacts to become more noticeable over the next few years. The current augmentation phase - where AI makes workers more productive without reducing headcount - is likely a transitional period rather than a stable end state.
The firms and workers that come out ahead will be those that treat this window as preparation time. For organisations, that means auditing which roles sit in the 82 disrupted occupations, redesigning workflows before external pressure forces it, and investing in the capability to manage human-AI collaboration deliberately. For workers in exposed roles, it means building skills that sit outside AI's current reach - judgement, client relationships, complex problem-solving, and the ability to direct and check AI output.
Mindiam's AI strategy and AI training teams work with Australian businesses on both sides of this challenge. The organisations that act during the augmentation phase will be far better placed when the automation phase accelerates.
Frequently Asked Questions
What does Deloitte Access Economics mean by "AI-disrupted" occupations?
Deloitte Access Economics assessed each occupation based on how far its core tasks can be handled by AI without requiring human judgement, empathy, or interpersonal skill. The 82 occupations classified as "AI-disrupted" are those where AI can replace a significant share of the work, putting them at the highest risk of declining employment as adoption deepens. These are not necessarily roles that will disappear overnight, but they are the ones where hiring demand is already starting to soften and where the long-term employment outlook is most exposed to AI-driven change.
Why are employment levels in at-risk occupations still rising if AI is disrupting them?
The disruption is currently showing up in recruitment rather than redundancies. Employers are posting fewer vacancies in AI-exposed roles - they are hiring less - but they are not yet letting existing workers go in large numbers. Deloitte Access Economics describes this as consistent with an augmentation phase, where AI is making current workers more productive rather than replacing them outright. The expectation is that this changes as work processes are reorganised around AI capabilities, which is why the report urges businesses and workers to prepare now rather than wait for the redundancy phase to arrive.
What can Australian businesses do to get ahead of this shift?
Deloitte's research points to work redesign as the critical lever. Organisations that thoughtfully redesign how humans and machines interact - rather than simply layering AI tools onto existing workflows - are twice as likely to exceed their return on investment expectations for AI. That means auditing which tasks in each role are genuinely better handled by AI, restructuring the work around those boundaries, and investing in the human skills that complement rather than compete with AI. Mindiam's AI strategy and AI automations services are designed to support exactly this kind of structured transition.
How fast is AI adoption expected to move in Australia?
Deloitte Access Economics flagged that the volume of investment in AI over recent years suggests it will move along the adoption curve much faster than previous technologies. That speed creates both opportunity and risk - businesses that adapt quickly can capture productivity gains and competitive advantage, while those that delay face being caught in a rapid transition without the workforce capability or operational structures to respond. The current augmentation phase may be shorter than many organisations are assuming.
