Australian businesses race to adopt Ai so they can…
One in eight Australian businesses now uses AI, and the adoption rate has tripled in three years — but the productivity data tells a very different story.
Australian businesses race to adopt AI so they can…
One in eight Australian businesses now uses artificial intelligence. That number has tripled in three years. The question nobody in the press release bothered to ask is: to what end?
Adoption is a measure of uptake, not of value
The Australian Bureau of Statistics surveyed nearly 7,000 businesses and found adoption is highest among large firms (35 per cent), in information and media (38 per cent), and in professional services and financial services (both 24 per cent). Financial services recorded the most dramatic growth, up 24 times from where it sat three years ago, when just one per cent of firms in the sector reported using AI at all. These are striking numbers. They are also, on their own, almost entirely uninformative.
Adoption is a measure of uptake, not of value. It tells you that businesses are installing something. It does not tell you whether that something is working. A firm that has licensed a suite of AI tools and deployed them to summarise meeting notes has "adopted AI" in the same way a firm that has restructured its entire logistics operation around machine learning has. The ABS data does not distinguish between the two, and the distinction matters enormously.
A firm that has licensed a suite of AI tools and deployed them to summarise meeting notes has "adopted AI" in the same way a firm that has restructured its entire logistics operation around machine learning has.
The pattern that does emerge from the data is instructive in a different way. AI adoption is far higher among businesses that were already innovating: 20 per cent of innovation-active firms report using AI, compared with six per cent of those that undertook no innovation activity. Among large, innovation-active businesses, the rate hits 37 per cent. This makes intuitive sense. Firms that were already investing in process improvement have the internal capability to identify where AI fits. The technology slots into an existing habit of asking "how do we do this better?"
AI's benefits are concentrating, not democratising
For everyone else, the picture is murkier. Small and micro businesses sit at 11 per cent overall, and non-innovation-active small businesses at just four per cent. This is the bulk of the Australian economy by business count, and for most of them AI remains either inaccessible, irrelevant, or both. The technology's benefits, to the extent they exist, are not yet democratising. They are concentrating.
None of this means AI is a fraud. The honest version of the case for it is genuinely compelling: AI tools can compress the time it takes to do knowledge work, surface patterns in large datasets that humans would miss, and reduce the per-unit cost of tasks that used to require skilled labour. In professional services especially, where billable hours are the product, that is a meaningful structural shift. A junior lawyer who can draft a contract in an hour instead of four is not a curiosity. Scaled across an industry, it changes margin structures.
Australia's productivity data doesn't yet confirm the promise
But "can" is not "does," and the productivity data that would confirm this at the macro level is not yet there. Australia's productivity growth has been anaemic for years. If AI were delivering the transformative efficiency gains its proponents promise, you would expect to see the early signals in aggregate output per hour worked. You do not, at least not yet. This is not unique to Australia. Economists have been tracking the same gap globally, sometimes called the AI productivity paradox, where adoption curves run well ahead of measured economic benefit.
There is a plausible explanation for the lag. New general-purpose technologies tend to take time to embed. Electrification took decades to show up in productivity statistics because factories had to be redesigned around electricity, not just wired for it. The same logic may apply to AI. The firms racing to adopt it now may be building the internal capability they will need to extract value from it in five years. That is a reasonable bet. It is also an expensive one, made without proof of return.
This is where the AI story sits right now: measurably accelerating, genuinely promising, and substantively unproven. Australian businesses are lining up for the interview. Whether the job turns out to be real is a question the current data cannot answer.
That gap, between the pace of adoption and the evidence of benefit, is precisely what makes careful analysis more valuable than the adoption curve itself. The number of firms using AI is rising fast. What the technology is actually doing to the economy remains, for the moment, an open question.
Sources
Frequently Asked Questions
What percentage of Australian businesses use AI in 2024-25?
Around 12 per cent of Australian businesses reported using artificial intelligence in 2024–25, up from negligible levels in 2021–22 — roughly a tripling in three years. Adoption is highest among large firms, at 35 per cent.
Why hasn't AI adoption shown up in Australia's productivity figures yet?
Australia's productivity growth has been weak for years, and there is no clear macro-level signal yet that AI is changing that. The likely explanation is a lag effect: general-purpose technologies like electrification also took decades to register in productivity statistics, because organisations had to be fundamentally restructured to extract the benefit, not just equipped with the new tool.
Are small businesses in Australia using AI?
Small and micro businesses report AI adoption of around 11 per cent overall, but non-innovation-active small businesses sit at just four per cent. For the majority of Australia's businesses by count, AI remains either inaccessible or irrelevant.
Which Australian industries are adopting AI the fastest?
Information and media leads at 38 per cent adoption, followed by professional services and financial services at 24 per cent each. Financial services has seen the sharpest growth rate, rising 24-fold from one per cent three years ago.
Does higher AI adoption mean businesses are getting more productive?
Not necessarily. Adoption measures whether a business is using AI, not whether it is benefiting from doing so. A firm using AI to summarise meeting notes counts the same as one that has rebuilt its entire logistics operation around machine learning — the data does not distinguish between the two.