Depending on which survey you read, between 12 and 44 per cent of Australian businesses have adopted artificial intelligence. That is not a rounding difference. It is a definitional chasm, and it tells you more about the state of the evidence than any single headline figure.
The Australian Bureau of Statistics, surveying nearly 7,000 businesses for its 2024 to 2025 Business Characteristics Survey released on 25 June 2026, found around 12 per cent of Australian businesses reported using AI in their workplace. The National AI Centre's monthly SME tracker put small and medium enterprise adoption at 44 per cent in February 2026. MYOB puts it at 40 per cent. Weel, reading actual card transactions across more than 4,000 Australian and New Zealand SMBs, found 30.8 per cent were paying for AI tools in June 2026.
When four credible sources cannot agree on how many businesses are using AI, the odds of finding clean evidence that it is driving revenue growth are slim. That is exactly what the data shows: adoption is accelerating, and proof of revenue impact remains thin.
This is not an argument against AI. It is an argument for honesty about what the evidence supports and what a business owner should reasonably expect. If you are working out what to spend, our guide to budgeting for AI tools is the practical companion to this piece.
Published: September 2026
Each survey measures something different, and the differences are the story.
The ABS is the most conservative and the most official. Its 12 per cent figure comes from a workplace-use definition applied to a sample of nearly 7,000 businesses surveyed between October 2025 and February 2026. The size gradient is steep: around 35 per cent of large businesses reported using AI, up from 9 per cent in 2021 to 2022, 22 per cent of medium businesses, up from 3 per cent, and around 11 per cent of small and micro businesses. Information, media and telecommunications led, followed by professional, scientific and technical services.
One ABS finding deserves more attention than it has had. The AI adoption rate for innovation-active small businesses was 19 per cent, almost five times the rate of small businesses that undertook no innovation activity. AI adoption is not distributed randomly across the small business population. It tracks businesses that were already changing something.
The National AI Centre tracks intent and experimentation. Its SME AI Pulse, run by Fifth Quadrant with at least 400 businesses responding each month, reported 43 per cent adoption across the December 2025 to February 2026 quarter, rebounding to 44 per cent in February, against 45 per cent the previous quarter. That figure counts any level of adoption, which includes a business owner using a chatbot occasionally.
Weel reads the bank statement. Its Q2 2026 index found 30.8 per cent of Australian SMBs on its cards paying for AI tools in June, up from 22.1 per cent in January. This is transaction data rather than survey response, which removes the self-report problem, though it cannot see AI bundled inside a Microsoft or Google subscription.
Treasury is the sceptic. Advice to the Treasurer reported in late August 2026 described Australian AI uptake as widespread but shallow, with fewer than one in ten businesses reporting significant adoption, and noted that the predicted employment shock has not materialised.
Those four numbers are not in conflict. Roughly 44 per cent of SMEs have touched AI, around 31 per cent are paying for it, around 12 per cent use it in a way an official statistician recognises as workplace adoption, and fewer than one in ten would call their own use significant. That is a funnel, not a contradiction.
MYOB's April 2026 data found 40 per cent of SMEs adopting AI products or features, with 54 per cent of users reporting time saved and 34 per cent reporting productivity gains. It also found 46 per cent not using AI and not intending to in the next twelve months.
The measurement problem is the important part. Nearly half of Australian businesses using AI do not measure its impact at all. Among those that do, the most common method is checking whether objectives were met. These are perceptions, not measured outcomes, which means a large share of the positive sentiment in every survey sits on no evidentiary base at all.
Two enablement figures explain a lot: MYOB found 72 per cent of SMEs have no plans for AI training, and around two-thirds are not hiring for AI skills. Hays, in its FY26/27 Salary Guide, found 60 per cent of employees use AI regularly at work while only 22 per cent have received any training. Adoption without enablement is the most likely mechanical explanation for why so many organisations report no measurable financial benefit.
MYOB's most quoted 2026 finding is that SMEs using AI are growing 2.8 times faster than those that are not, drawn from aggregated product data across hundreds of thousands of businesses.
Read it carefully. This is vendor analysis of its own customer base, not independent econometrics, and the selection effect is obvious. Businesses that adopt AI are also more likely to be digitally mature, younger-owned, better resourced and already growing. The ABS finding that innovation-active small businesses adopt AI at nearly five times the rate of non-innovators is the same effect from a different angle. The correlation is real. The causal direction is not established.
The same caution applies to the most cited Australian projection: the Tech Council of Australia's estimate that AI could add $142 billion annually to GDP by 2030, against $21 billion today. That research came from Australia's AI Opportunities Report, which was funded by OpenAI, a company with a direct commercial interest in widespread adoption. The finding is not invalidated by its funding, but the disclosure is routinely omitted from media coverage and readers should have it. Getting from $21 billion to $142 billion in four years requires roughly sevenfold growth, which is an aggressive assumption on any reading.
A more defensible domestic anchor exists. The Productivity Commission assesses that broader AI adoption could drive up to 4.3 per cent labour productivity growth over the next decade in the market sector, worth around $116 billion in GDP. That is a public-sector estimate with no vendor funding attached.
Anthropic's economic research on Australian usage found something that reframes the whole SME conversation. Australia over-indexes heavily on per-capita Claude usage, ranking at or near the top globally on a population-adjusted basis. But the composition is unusual: Australian usage sits below the global average on computing and mathematical tasks and above it on back-office work, including invoicing, inventory and administration. Anthropic's Australia and New Zealand general manager described the same pattern publicly in August 2026, noting Australian usage over-indexes on back office and under-indexes on coding.
More than half of Australian usage is classified as augmentation, meaning a human stays in the loop rather than handing the task over. The state distribution is also uneven, with New South Wales and Victoria accounting for the large majority of Australian conversations and mining-heavy states under-indexing relative to their economic output. That is a workforce composition effect, not an income effect.
The SME story in one sentence: Australian businesses are using AI on the administrative work that clogs a small business, not on building software. That matches the National AI Centre's use-case data, where content generation and data analytics each lead at 54 per cent of adopters, with cybersecurity at 48 per cent, while agentic AI, supply chain optimisation and AI-assisted human resources remain largely untouched.
The National AI Centre identified three barriers, and none of them is price.
Trust is the largest. Around 65 per cent of non-adopting businesses cited either distrust in AI decision-making or a strong preference to keep humans in control. That is a confidence problem, and it is consistent across business sizes and geographies.
Relevance is the most fixable. More than half of non-adopters, 54 per cent, said AI is not relevant to their business. This is most pronounced in construction and agriculture, where fewer than 30 per cent of businesses are adopting, against more than half in health, education and services. The difference is context, not capability. These businesses have not seen a relatable example of what AI looks like in a business like theirs.
Nineteen per cent do not know where to start, up two points on the previous quarter. They are not opposed. They are disoriented.
The most commercially relevant finding in the current Australian data has nothing to do with adoption rates.
Among businesses currently using AI, the most common safeguard is checking AI outputs before they affect customers, and roughly half have that in place. Practices around telling customers that AI is in use, and giving customers a formal way to raise a concern about it, lag well behind. The National AI Centre describes this as a gap between internal operational safeguards and outward-facing accountability, and flags it as a likely source of risk as regulatory expectations and public scrutiny grow.
For anyone in a regulated profession, that gap has already closed. The Tax Practitioners Board published TPB(GS) 55/2026 on 22 July 2026, setting out how existing obligations under the Code of Professional Conduct apply when a tax or BAS agent uses AI. Using AI does not reduce or transfer a practitioner's responsibilities, and Code item 6 prevents disclosing client information to a third party without permission, which means putting client data into an AI tool can itself be a disclosure requiring consent. We cover what that means for your books in who owns your ledger.
The direction of travel is clear. Internal safeguards are not enough. The businesses that will be comfortable in two years are the ones documenting what they use, on what data, with what review, now.
Adoption is rising. Every source agrees on the direction even where the levels differ by a factor of three.
The primary use case is efficiency, not revenue. MYOB frames it as controlling costs and managing administrative workload. Only around 7 per cent of Australian SMEs have built AI into their products or services, meaning the other 93 per cent are using it internally. This is a cost story.
Measurement is largely absent, and training is worse. There is no widespread, data-supported evidence of a direct AI-to-revenue link for Australian SMEs.
Benefits accrue to better-resourced businesses. Mid-market firms with more infrastructure, management capability and reporting frameworks report better outcomes, which is consistent with technology adoption research across every era.
Start with measurement. Before investing in any tool, establish a baseline for the process you want to improve. If you cannot measure the current state, you cannot measure the improvement, and you will end up in the roughly half of businesses reporting positive impact with no evidence.
Focus on efficiency before revenue. The data supports AI as a time and cost tool for SMEs. Content generation, document handling, first-draft analysis and administrative reduction have the most consistent positive feedback. Revenue-generating applications require more sophistication than most SMEs have available.
Build the operational foundation first. The performance gap between mid-market and smaller firms is not primarily about technology. It is about the management and financial infrastructure that makes any technology investment productive. Clean books, accurate reporting and defined processes make every subsequent investment more effective. That is the compounding value of an bookkeeping support for ai adoption, and it is the same reason our outsourced bookkeeping in Sydney work starts with the ledger rather than the tools.
Budget for training, not just seats. With 72 per cent of SMEs having no training plans and only 22 per cent of employees receiving any, enablement is the cheapest available lever on the return.
Check the tax treatment. AI subscriptions are generally deductible in the year incurred, and instant asset write-off provisions may apply to associated hardware. See our guide to the instant asset write-off for 2026 to 2027 and confirm your position with your adviser.
The revenue case may be unproven, but the risk of standing still is real. More than half of Australian SMEs are not meaningfully using AI. As adoption rises among competitors, the productivity gap between adopters and non-adopters widens even if no single business can prove causation.
The businesses most exposed sit in the middle: large enough that ignoring technology has a cost, small enough that they have no dedicated capability to implement it. Those are the businesses that benefit most from structured support rather than another tool subscription.
Treasury's assessment is the right closing note. Uptake is widespread but shallow, and fewer than one in ten businesses report significant adoption. On that reading, almost nobody in Australia is late. Most are just early in a way that has not shown up on a profit and loss yet.
What percentage of Australian businesses use AI in 2026?
It depends entirely on the definition. The ABS Business Characteristics Survey found around 12 per cent of all Australian businesses used AI in the workplace in 2024 to 2025, rising to 35 per cent of large businesses. The National AI Centre's SME tracker reported 44 per cent in February 2026 using a broader definition, and Weel's card data found 30.8 per cent of SMBs paying for AI tools in June 2026.
Does AI adoption increase revenue for small businesses?
There is no strong, independently verified evidence of a direct link for Australian SMEs. MYOB reports that AI-adopting SMEs grow 2.8 times faster than non-adopters, but that is vendor analysis of its own customer base and does not control for the fact that adopters tend to be more digitally mature and already growing.
What are Australian SMEs using AI for?
Internal efficiency, overwhelmingly. Content generation and data analytics each lead at 54 per cent of adopters. Only around 7 per cent have built AI into their products or services. Australian usage over-indexes on back-office work such as invoicing and administration, and under-indexes on coding compared to global averages.
Why do AI adoption surveys disagree so much?
Because they measure different things. Official statistics use a conservative workplace-use definition, monthly trackers count any level of adoption including occasional experimentation, and card data only sees direct subscriptions. A business owner using a chatbot once a week counts as an adopter in one survey and not in another.
What stops Australian SMEs adopting AI?
Not cost. Around 65 per cent of non-adopters cite distrust of AI decision-making or a preference for human control, 54 per cent say it is not relevant to their business, and 19 per cent do not know how to start. Fewer than 30 per cent of construction and agriculture businesses are adopting, against more than half in health, education and services.
How much could AI add to the Australian economy?
The Tech Council of Australia estimates $21 billion annually today, rising to $142 billion by 2030, from research funded by OpenAI. The Productivity Commission's independent estimate is up to 4.3 per cent labour productivity growth in the market sector over a decade, worth around $116 billion in GDP.
Do Australian businesses need to tell customers when they use AI?
Practice is behind expectation. Around half of AI-using businesses check outputs before they reach customers, but far fewer disclose AI use or offer a way to raise concerns. Registered tax and BAS agents now have explicit guidance in TPB(GS) 55/2026, published 22 July 2026, and client permission is required before disclosing client information to a third party, which can include an AI tool.
Should a small business wait?
Waiting is a decision with a cost, but so is buying tools you cannot measure. The most useful first step is establishing a baseline for one process you want to improve, then trialling one tool against it. Businesses with clean financial reporting extract more from any technology investment than businesses without it.
Scale Suite is a Sydney-based provider of outsourced finance teams and fractional CFO services for Australian SMEs. We deliver weekly bookkeeping, payroll, BAS/IAS lodgement, cashflow reporting, management accounts, and strategic fractional CFO oversight, all as a fully embedded team that works inside your business.
CA-qualified, Xero Certified, and registered BAS Agents, we replace fragmented bookkeepers and once-a-year accountants with one responsive finance function at a fraction of the cost of full-time hires. We serve growing businesses across Sydney, Melbourne, Brisbane, and Perth, with packages starting from $1,500 per month and no lock-in contracts.
See outsourced bookkeeping instead of a hire.
We review and check this guide periodically. At the time of writing (September 2026), all information was current. Scale Suite is a registered BAS Agent, not a licensed tax advisor or financial advisor. This content is general information only and does not constitute professional tax, financial, or legal advice. Some details may change over time.
Scale Suite is a Sydney-based provider of outsourced finance and HR services for Australian SMEs. We deliver bookkeeping, financial reporting, payroll processing, fractional CFO support, recruitment, employee onboarding, people and culture support, and fractional HR oversight, all as a fully embedded team that works inside your business.
Employment Hero Gold Partner, CA-qualified, and Xero Certified, we replace fragmented finance and HR processes with one responsive, senior-level function at a fraction of the cost of full-time hires. We serve growing businesses across Sydney, Melbourne, Brisbane, and Perth, with packages starting from $1,500 per month and no lock-in contracts.
30 minutes with our team.
We will review your current finance setup, compare the full cost of an internal hire against our embedded team, and show you exactly what your finance function should cost at your stage of growth.
You will leave with a clear view of what is working, what is missing and where you would save.
No lock-in contracts. 30-day money-back guarantee.
Prefer to book directly? Grab a time here.
