What does the ABS AI adoption data really tell us about Australian businesses?

About 12% of Australian businesses used AI in their workplace in 2024-25, according to the Australian Bureau of Statistics (ABS) — up from just 1% three years earlier, before generative AI tools were widely available. That figure looks low next to industry surveys that put adoption closer to 44%, but the two count different things: the ABS measures AI embedded in day-to-day operations, not one-off experimentation.

The gap between 12% and 44% comes down to definitions. "Adoption" can mean anything from a single ChatGPT prompt to AI running in production across core business workflows. Knowing which definition a survey uses determines whether your business is genuinely behind or simply measured differently.

This post unpacks what the ABS data actually counts, how adoption varies by business size and industry, what is blocking the other 88%, and what the governance gap means for businesses using AI today.

What the ABS AI adoption data actually measures

The ABS AI adoption figure comes from the Business Characteristics Survey, which asked nearly 7,000 businesses about workplace AI use between October 2025 and February 2026. It counts AI running in production rather than the occasional chatbot prompt, which is why the headline sits at about 12%.

That definition explains why credible sources report very different numbers. The National AI Centre's SME tracker puts adoption at 44%, and Weel's card-transaction data shows 30.8% of small and medium businesses paying for AI tools. Each measures a different stage of the same funnel.

Read the numbers as a sequence, not a contradiction:

• Roughly 44% of businesses have touched AI in some form.
• Roughly 31% pay for at least one AI tool.
• About 12% use it in a way the ABS counts as workplace adoption.
• Fewer than one in ten call their own use significant.

Treasury advice to the Treasurer in August 2026 described Australian AI uptake as widespread but shallow, which matches that funnel. One distinction matters more than any single percentage: whether the number counts experimentation or production use. Headline figures tend to measure the first and imply the second.

How AI adoption differs between large and small Australian businesses

Company size is the strongest predictor of AI adoption in Australia. Large businesses reported 35% adoption in 2024-25, roughly three times the rate of small and micro businesses at about 11%, according to the ABS.

• Large businesses: 35% adoption, up from 9% in 2021-22.
• Medium businesses: 22% adoption, up from 3%, a more than sevenfold rise.
• Small and micro businesses: about 11% adoption.

The gap is not mainly about budget. Large enterprises have dedicated technology teams, established data infrastructure, and the capacity to run pilots, measure results, and adjust.

Resources alone do not guarantee depth. The Reserve Bank of Australia surveyed 105 medium-to-large firms and found most had not moved past digital assistants, with adoption shallow even among well-funded companies.

Momentum matters more than headcount. Innovation-active small businesses adopted AI at 19%, almost five times the 4% rate of small businesses that ran no innovation activity, according to the ABS.

The practical result is a widening divide. Firms that started early are already on their second or third implementations, building know-how that late movers need a year or more to match.

Which industries are leading AI adoption in Australia

Information, media and telecommunications leads AI adoption in Australia at 38%, more than triple the national average, according to the Australian Bureau of Statistics. Professional services and financial services follow, each at 24%.

Adoption clusters in industries that work with digital information rather than physical goods. The pattern holds across every credible dataset.

Highest adoption sectors

• Information, media and telecommunications: 38%, the highest of any industry.
• Professional, scientific and technical services: 24%.
• Financial and insurance services: 24%, up from 1% in 2021-22, roughly a 24 times rise and the fastest growth of any sector.

Sectors trailing the national average

Physical-work industries lag behind. In the National AI Centre's broader adoption tracker, construction and agriculture sit below 30%, held back by workflows that offer few relatable AI use cases. By contrast, the National AI Centre found the health, education, and services sectors are leading adoption, with more than half of businesses in those sectors using AI, so the gap is about context and relatable use cases rather than capability.

More than half of non-adopters (54%) said AI is not relevant to their business, most pronounced in physical-work industries, according to the National AI Centre.

What Australian businesses actually use AI for

Content generation and data analytics each lead at 54% of adopters, with cybersecurity at 48%, per the National AI Centre. Agentic AI, supply chain optimization, and AI-assisted human resources remain largely untouched.

Anthropic's economic research found Australian usage over-indexes on back-office tasks like invoicing, inventory, and administration, and under-indexes on coding relative to global averages.

Only about 7% of Australian SMEs have built AI into their products or services. The other 93% use it internally, which keeps current adoption focused on cost and efficiency rather than new revenue.

Why the productivity claims need more scrutiny

Treat the headline productivity claims about AI in Australian businesses with caution, because the most-cited figures come from vendors analyzing their own customers.

MYOB's most-quoted 2026 finding, that SMEs using AI grow 2.8 times faster than those that don't, is analysis of MYOB's own customer base, not independent econometrics. The selection effect is large: businesses that adopt AI tend to be more digitally mature, younger-owned, better resourced, and already growing.

Measurement is thin even among adopters. Nearly half of Australian businesses using AI do not track its impact at all. Among those that do, the most common method is checking whether objectives were met, which captures perceptions rather than measured outcomes.

The cost of reviewing and correcting AI output rarely enters return calculations, and that labor often goes untracked.

Macro projections deserve the same scrutiny. The Tech Council of Australia's estimate that AI could add $142 billion a year to GDP by 2030 came from an OpenAI-funded report, a disclosure often left out of media coverage.

A steadier benchmark comes from the public sector. The Productivity Commission estimates broader AI adoption could add up to 4.3% to labor productivity over the next decade in the market sector, with no vendor funding attached.

What is blocking the other 88% of Australian businesses

The main barriers for the businesses not using AI are distrust, a belief that AI is not relevant, and not knowing where to start. A wide skills gap makes all three worse.

Trust and control

About 65% of non-adopting businesses cited distrust in AI decision-making or a strong preference to keep humans in control, the single largest barrier, according to the National AI Centre. The Reserve Bank of Australia notes that across advanced economies, trust and adoption move together, and Australia ranks relatively low on both.

Enablement and skills

Around 72% of SMEs have no plans for AI training, and about two-thirds are not hiring for AI skills, per MYOB. The Hays FY26/27 Salary Guide found 60% of employees use AI regularly at work while only 22% have received any training. That distance between everyday use and formal training is the likeliest reason so many businesses report no measurable financial benefit.

Decision paralysis

About 19% of non-adopters say they don't know where to start, up two points on the prior quarter, according to the National AI Centre. These owners have not rejected AI. They lack a clear starting point, and the conversation tends to fixate on tools rather than the workflows a business actually runs.

The governance gap that matters more than adoption rates

Most Australian businesses using AI have thin oversight, and that governance gap carries more practical risk than the adoption rate itself.

Among businesses currently using AI, the most common safeguard is checking outputs before they reach customers, and only about half have even that in place, according to the National AI Centre. Telling customers when AI is in use, or giving them a way to raise concerns, lags well behind internal checks.

Professional obligations are already catching up. The Tax Practitioners Board published TPB(GS) 55/2026 on July 22, 2026, setting out how the existing Code of Professional Conduct applies when a tax or BAS agent uses AI. Using AI does not reduce or transfer a practitioner's responsibilities.

Regulation is uneven. Australia's National AI Plan relies on voluntary standards with no audit pathway, logging requirements, or model-governance obligations. The Security of Critical Infrastructure Act already mandates evidence-based risk programs, incident reporting, and auditability for systems that influence critical infrastructure.

Document what AI you use, on what data, and with what review process. Doing it now, before the rules harden, positions your business to operate comfortably within two years.

How to read the data and act on it

To act on the ABS AI adoption data, start with measurement, work on efficiency before revenue, and fund training rather than tools alone. That sequence separates businesses that report real gains from those that only report activity.

• Baseline first. Measure the process you want to improve before buying any AI capability, or you join the roughly half of adopters claiming impact with no evidence.
• Start with efficiency. Content generation, document handling, first-draft analysis, and administrative work show the most consistent gains.
• Pick one workflow. Choose a single revenue-chain step, such as the quote, the missed call, or the close, rather than buying every tool at once.
• Fund training. With around 72% of SMEs lacking training plans, enablement is the cheapest available lever on return.
• Build the foundation. Clean data, accurate reporting, and defined processes make every later technology investment more productive.

The competitive window is still open. On Treasury's reading, uptake across Australian businesses is widespread but shallow, and fewer than one in ten report significant adoption.

Frequently asked questions

What percentage of Australian businesses are currently using AI?

The ABS reports about 12% of Australian businesses used AI in the workplace in 2024-25, rising to 35% among large firms. Broader industry surveys report 43% to 50% because they count any level of use. The ABS figure is the most rigorous measure of genuine workplace adoption.

How does AI adoption differ between large and small businesses in Australia?

Large businesses adopt at roughly three times the rate of small and micro firms, 35% versus about 11%. The difference comes from data infrastructure, dedicated technology teams, and the capacity to pilot and iterate, not budget alone. Adoption also tracks a company's existing innovation activity.

What industries in Australia are leading in AI adoption?

Information, media and telecommunications leads at 38%, followed by professional services and financial services at 24% each. Financial and insurance services grew fastest, rising about 24 times from just 1% in 2021-22. Physical-work industries like construction and agriculture trail behind.

What challenges do Australian businesses face in adopting AI technologies?

The biggest barriers are distrust and a preference for human control (about 65% of non-adopters), a belief that AI is not relevant (54%), and not knowing where to start (19%). A skills gap compounds all three, since around 72% of SMEs have no training plans.

What are the implications of the ABS AI adoption data for future business strategies?

The data points to three moves: start with internal efficiency use cases, fund training alongside the tools you buy, and document your governance before regulation requires it. The advantage now sits with businesses that move from experimentation to measured deployment on one specific workflow, rather than adopting broadly and tracking nothing.

The gains go to businesses that measure one workflow and deploy AI where the numbers hold up. When you reach that point, we can help: we connect your company's knowledge and return cited, permission-aware answers your team can trust and act on. Request a demo to see how we can put your company's knowledge to work.