What Happened
NAB Economics and Markets Research published new analysis in late May 2026 examining how artificial intelligence is beginning to register in Australia's economic data. The research covers investment flows, business adoption rates, and early signals in the labour market.
The headline finding is straightforward: AI is no longer just a boardroom conversation. Software investment now accounts for about 10% of total private sector business investment, and demand for digital infrastructure - data centres in particular - is contributing to broader business investment growth. Separately, NAB research on small and medium enterprises found 42% of SMEs are already using AI and a further 14% are planning to.
NAB Chief Economist Dr Sally Auld put the situation plainly: "AI is increasingly showing up in economic data, particularly through investment and early labour market trends." She was careful to add that "the scale and timing of the impact remain uncertain, and the effects will vary across industries, firms and workers."
Why It Matters
Australia's economy is growing slowly. NAB's own forecast puts GDP growth at 1.5% this year, with only a marginal improvement expected next year. Against that backdrop, AI investment is one of the few areas showing genuine momentum - and the distribution of that momentum tells you a lot about where the risks sit.
Two-thirds of businesses in real estate and accounting are using AI. In manufacturing, retail and transport - sectors that are less digitised and employ large numbers of workers - adoption rates sit between 35% and 21%. That gap matters. The businesses and workers least prepared for AI are concentrated in sectors where adoption is slowest, which means the productivity gains are likely to arrive unevenly.
The labour market picture is similarly mixed. About 15% of Australian jobs are highly or significantly exposed to AI. Since ChatGPT launched in late 2022, there is early evidence of slower employment growth in those more-exposed occupations. Dr Auld acknowledged the complexity: "While it is difficult to attribute these changes solely to AI, there are signs that labour market dynamics are starting to shift in areas with higher exposure."
For businesses trying to figure out where to focus their AI strategy, this research is a useful reality check. The gains are real but not yet broad-based. Getting ahead of the curve now - through genuine capability building rather than surface-level tool adoption - is what separates businesses that will benefit from those that will feel left behind.
Key Details
The research draws on NAB's own operations as a case study. The bank's analysis notes that its bankers can spend up to 50 per cent of their time on administrative tasks to set up or service accounts and loans for customers. AI is being deployed to reduce that burden, freeing bankers to spend more time on customer relationships and business development. That is a concrete, measurable use case - not a projection.
On the broader productivity question, Dr Auld's assessment was measured: "The evidence suggests this will be a gradual shift. Productivity gains are expected over time, but there is still little sign of a broad-based lift in aggregate productivity so far." The long-run view is more optimistic - "historically we've seen technology create more jobs than it displaces in the long run" - but that historical pattern offers cold comfort to workers in exposed occupations right now.
The research also flags that "over time, the net impact will depend on how productivity gains flow through to wages, demand and job creation across the broader economy." That is not a given. It depends on policy, on how businesses invest their efficiency gains, and on whether workers in displaced roles can access retraining.
AI training programs and genuine workforce upskilling are part of how that transition gets managed well rather than badly. The NAB research makes the point directly: "These findings reinforce the importance of broadening AI knowledge and capabilities across the economy, so more businesses and people are able to benefit from AI instead of feeling left behind."
Background and Context
The NAB analysis sits within a broader pattern visible in Australian economic data. Household spending has been under pressure, with consumers tightening budgets in response to cost-of-living pressures. Business investment in physical assets has been patchy. Software and digital infrastructure are the exception - they are growing, and AI is a significant driver of that growth.
The unemployment rate is expected to rise to 4.7%, which adds urgency to the labour market questions the research raises. A rising unemployment rate combined with AI-driven displacement in specific occupations is a combination that warrants attention from policymakers and businesses alike.
For Australian businesses - particularly SMEs - the data suggests a clear split is forming. Those that have started building AI automation capabilities are already seeing operational benefits. Those that have not are falling behind on adoption rates that are now measurable at the economy-wide level.
The financial services sector is further along than most, partly because the regulatory and compliance burden creates obvious use cases for automation, and partly because firms like NAB have the scale to invest in custom AI deployments. Smaller businesses in less-digitised sectors face a harder path, which is why accessible AI training matters more than the headline adoption numbers suggest.
What Comes Next
NAB's forecast of 1.5% GDP growth this year reflects an economy that is not yet seeing the productivity dividend from AI at scale. The research is explicit that broad-based productivity gains are still ahead rather than already arrived.
The near-term indicators to watch are software investment as a share of total business investment, employment trends in AI-exposed occupations, and whether SME adoption rates continue to climb past the current 42% mark. If the 14% of SMEs currently planning to adopt AI follow through, that would push the majority of Australian small businesses into active AI use within the next year or two.
For businesses and workers in sectors with lower current adoption - manufacturing, retail, transport - the window to prepare is open but not indefinitely. The AI strategy decisions made in the next 12 to 24 months will shape competitive positioning for years. Getting practical about capability building now, rather than waiting for the productivity data to become undeniable, is the more defensible position.
Mindiam works with Australian businesses on exactly this kind of transition - from initial AI strategy through to automation implementation and workforce capability building. You can read more about our approach on the about page and our editorial standards.
Frequently Asked Questions
How much of Australian business investment is going into software and AI?
According to NAB's AI in Australia report, software investment now accounts for about 10% of total private sector business investment. This figure reflects not just direct AI tool purchases but also the broader digital infrastructure - including data centres - that AI workloads require. It is a meaningful share of total investment and has been growing as businesses move from evaluating AI to actually deploying it.
What share of Australian SMEs are using AI?
NAB research on small and medium enterprises found that 42% of SMEs are already using AI, with a further 14% planning to adopt it. Adoption is not uniform across sectors. Two-thirds of businesses in real estate and accounting are using AI, while in manufacturing, retail and transport - sectors that are less digitised - adoption rates range between 35% and 21%. The gap between high-adoption and low-adoption sectors is one of the more important findings in the research.
How exposed is the Australian workforce to AI displacement?
The NAB research found that about 15% of jobs in Australia are highly or significantly exposed to AI. Since late 2022, when ChatGPT was introduced, there has been early evidence of slower employment growth in occupations with higher AI exposure. Dr Sally Auld noted that while it is difficult to attribute these changes solely to AI, "there are signs that labour market dynamics are starting to shift in areas with higher exposure." The long-run historical pattern suggests technology creates more jobs than it displaces, but the transition period carries real risk for workers in affected roles.
What is NAB's economic outlook for Australia in 2026?
NAB forecasts GDP growth of 1.5% in 2026, with a marginal improvement expected the following year. The unemployment rate is expected to rise to 4.7%. These are modest numbers that reflect an economy still working through cost-of-living pressures and not yet seeing a broad-based productivity lift from AI. The research is clear that the productivity dividend from AI is expected to build gradually rather than arrive all at once.
How is NAB itself using AI in its operations?
NAB's research uses the bank's own operations as a concrete example. NAB bankers can spend up to 50 per cent of their time on administrative tasks related to setting up or servicing accounts and loans. AI is being used to reduce that administrative burden, with the goal of freeing bankers to spend more time on customer relationships and identifying ways to help customers grow their businesses. It is a practical, operational use case rather than a speculative one.
