Responsible AI

Australia lags on AI governance as adoption races ahead

Only 22% of Australian organisations are mostly or fully ready on AI governance, well below the 33% global average, as AI task support climbs to 29%.

Australia lags on AI governance as adoption races ahead

Key takeaways

  • Only 22% of Australian organisations reported being mostly or fully ready on AI governance, against a 33% global average, according to SAP research conducted by Oxford Economics across 2,600-plus business leaders in 13 countries.
  • AI now handles 29% of tasks in the average Australian business, up from 25% a year ago, with leaders expecting that share to reach 48% within two years.
  • 43% of Australian organisations lack human-in-the-loop processes for agentic workflows, and 49% reported AI agents had already taken incorrect actions during pilots or live deployment.
  • Australian AI spend is forecast at AUD $35.5 million this year, up from AUD $27.5 million, yet still below the global average of AUD $40.4 million.
  • Fewer than half of surveyed Australian companies have a dedicated AI leader, and only 41% offer training on AI use and risk.

What Happened

In-body image for: Australia lags on AI governance as adoption races ahead
Illustrative AI-generated image by Mindiam (Flux 1.1 Pro Ultra)

SAP, working with Oxford Economics, surveyed more than 2,600 business leaders across 13 countries and published findings showing Australia's AI governance readiness trails the global average by a significant margin. The research, reported by CFOtech Australia on 20 July 2026, found only 22 per cent of Australian organisations described themselves as mostly or fully ready on AI governance. The global figure was 33 per cent.

The report's framing was blunt. "Think of this as Australia's AI school report: improving, but still not working to its potential," the research noted. Adoption is clearly moving: AI now supports 29 per cent of tasks in the average Australian business, up from 25 per cent a year earlier, and leaders expect that share to climb to 48 per cent within two years.


Why It Matters

The gap between adoption pace and governance maturity creates concrete operational risk. Some 42 per cent of respondents said they were deploying AI agents faster than they could standardise and govern them. A further 54 per cent said employees were increasingly accepting AI outputs without enough scrutiny.

The agentic AI numbers are particularly sharp. Forty-three per cent of Australian organisations have no human-in-the-loop processes for agentic workflows, and 49 per cent reported that AI agents had already taken incorrect actions during pilots or deployment, typically resulting in rework and delays. These are not theoretical risks; they are already showing up in operations.

The research called out a structural problem in how organisations are set up to manage this. Fewer than half of surveyed Australian companies had a dedicated AI leader. Only 33 per cent had leadership key performance indicators tied to AI outcomes, and just 41 per cent offered any training on AI use and risk.


Key Details

Spending is rising. Australian organisations expect to put about AUD $35.5 million into AI this year, up from AUD $27.5 million last year. That is still below the global average of AUD $40.4 million, which mirrors the governance gap: Australian businesses are investing, but not yet at the scale of their international peers.

Expected returns are improving. Australian companies forecast an AI return on investment of 19 per cent this year, up from 15 per cent last year, with that figure projected to reach 37 per cent within two years. The trajectory is positive, but the governance shortfall puts those returns at risk if incorrect agent actions and unchecked outputs accumulate.

The research pointed to workforce readiness as a core issue. "These findings point to an urgent need for role-specific training, stronger safe-use guidance and change management that brings employees along as AI becomes embedded in daily work," the report stated.


Background and Context

Australia does not yet have binding AI-specific legislation equivalent to the EU AI Act. The Office of the Australian Information Commissioner (oaic.gov.au) has published guidance on how the Privacy Act 1988 applies to AI systems, and the Department of Industry, Science and Resources has released voluntary AI Ethics Principles, but mandatory governance frameworks remain under development. That regulatory gap gives organisations discretion over their own governance timelines, which the SAP data suggests many are using to prioritise deployment over oversight.


What Comes Next

The SAP research points to two years as the window in which AI task support is expected to nearly double, from 29 per cent to 48 per cent of work. If governance readiness does not keep pace, the proportion of organisations running agentic systems without adequate controls will grow. The report's call for role-specific training and change management suggests the near-term priority is internal capability rather than waiting for regulatory mandates.

Sources & citations

  1. Mark Tarre, "Australia lags on AI governance as adoption races ahead," CFOtech Australia, 20 July 2026
  2. Office of the Australian Information Commissioner, AI and privacy guidance
JUST THE WEEKLY ROUNDUP

One Friday email. The five things AU operators actually need to know.

Operator-tested, primary-source linked, citation-first. Written by an operator, not a marketing team. Or, for a personalised view first, take our 90-second quiz.

Unsubscribe anytime. No spam. See our privacy policy.