Australian organisations are moving artificial intelligence out of the sandbox and into production faster than the global average, yet they are falling well behind their international peers on turning that activity into genuine business transformation, according to Deloitte Australia’s State of AI in the Enterprise report.
The study, drawn from a global survey of 3,235 business and technology leaders across 24 countries conducted in August and September 2025, found that 28 per cent of Australian respondents have shifted at least 40 per cent of their AI pilots into production. That edges past the 25 per cent global figure. A further 57 per cent expect to hit that mark within three to six months.
But scale is not the same as impact. Just 12 per cent of Australian leaders say generative AI is already transforming their business or industry, against 25 per cent globally. It is a gap that reframes the local story: plenty of deployment, far less reinvention.
Moving pilots to production, but not to transformation
Deloitte’s numbers show a workforce busy with AI. Some 61 per cent of Australian respondents report improved efficiency and productivity, 69 per cent are using agentic AI, and 57 per cent have deployed physical AI such as robots and automated machinery.
The catch is depth. Only 30 per cent are using AI to deeply transform how work is done, compared with 34 per cent globally, and a third remain focused on automating existing processes rather than rethinking them.
Investment appetite tells the same story. Just 65 per cent of Australian organisations plan to lift AI spending next financial year, against 84 per cent globally, a near 20-point gap that suggests local boards are still weighing the return.
David Alonso, Deloitte Australia’s National AI Market Lead, framed the challenge as one of ambition. As reported by Accounting Times, Alonso urged organisations to “stop treating AI as isolated use cases” and instead make enterprise-wide decisions that “scale AI across products, services and operations”.
Stuart Scotis, Deloitte Australia’s National Leader for Technology, Innovation and GenAI, put the workforce dimension bluntly in the same coverage, saying leaders need to “rethink how work is done in a world of abundant digital labour”.
The agent question forces a governance reckoning
The rise of agentic AI, systems that can plan and act with limited human oversight, is where Australia’s caution shows most clearly. While 69 per cent of local organisations are using agents, only 22 per cent report an advanced governance model to manage them.
More than half cite talent and skills as the barrier to going further with agents, and over 40 per cent point to cost and data availability, per Deloitte’s figures.
Dr Elea Wurth, Deloitte’s Lead Partner for AI across Asia Pacific and Australia, argued the shift raises the stakes. “When systems can decide and act, organisations must embed real risk management,” she said in the Accounting Times report, pointing to clear accountability, guardrails and intervention points.
That governance-first instinct extends to where the technology comes from. Some 72 per cent of Australian respondents now factor a vendor’s country of origin into purchasing decisions, and more than 80 per cent treat sovereign AI as a strategic priority.
Sovereignty moves from slogan to procurement rule
The sovereignty theme is hardening into practice. A separate report covered by SecurityBrief Australia in June found 52 per cent of Australian respondents evaluating sovereign AI technologies, with a further 40 per cent running early proofs of concept.
Australia’s stance reads as more measured than the wider region. That coverage noted 73 per cent of local respondents support agentic AI only when paired with strong oversight, against near-universal enthusiasm elsewhere in Asia Pacific, with national security, defence and healthcare flagged as the priority use cases.
Nicole Jefferson, Vice President of Global Government Affairs at Dell Technologies, told SecurityBrief the debate had matured, saying “the question is no longer whether Sovereign AI matters, but how to operationalise it”.
Skills remain the choke point across both studies. The SecurityBrief coverage recorded 86 per cent of respondents reporting digital skills gaps affecting their initiatives, echoing Deloitte’s finding that insufficient worker capability is the single biggest brake on progress.
Why it matters
For Australia, the pilot-to-payoff gap is a productivity question with national stakes. The economy has wrestled with sluggish productivity growth for a decade, and AI has been pitched by government and industry alike as a route out. Deployment without transformation risks banking the cost of the technology without the dividend.
The skills shortage is the clearest domestic pressure point. Deloitte reports 53 per cent of organisations are educating staff on AI fluency and 48 per cent are running upskilling programs, but demand for AI safety, data governance and sovereign cloud specialists is outstripping the local talent pipeline.
Sovereignty raises the stakes higher still. With more than 80 per cent of Australian firms treating locally controlled AI as strategic, and defence and healthcare named as priority domains, the choices made now shape whether critical capability sits onshore or is rented from offshore providers.
The measured Australian posture, heavy on governance and oversight, may prove prudent as agentic systems mature. It may also mean local firms watch faster movers overseas capture the first wave of AI-driven advantage.
The forward look
The next two quarters will test the local narrative. With 57 per cent of Australian organisations expecting to push 40 per cent or more of their pilots into production within three to six months, the second half of 2026 should reveal whether scaled deployment finally converts into the transformation figures that today lag the world.
Deloitte’s message to boards is that the window is one of activation, not experimentation. Whether Australian leaders make the bold, enterprise-wide bets the consultancy is calling for, or keep hedging with incremental automation, will determine if the country closes the gap or settles into a permanent second tier of AI maturity.
Sources: Deloitte Australia — State of AI in the Enterprise; Deloitte Australia press release (12 February 2026); Accounting Times; SecurityBrief Australia.
