AI Adoption Barriers: What Holds Australian Businesses Back

Most Australian businesses are not blocked by a lack of AI tools. They are blocked by distrust, unclear strategy, thin skills, security worries and, most of all, a lack of visibility into what their own people already do with AI. This guide sets out the barriers Australian data actually shows, how to tell which ones apply to your business, and what to do about each in the next 90 days.

Calibreforumcalibreforum.auSee Where AI Adoption StandsCalibre Forum’s AI readiness assessment combines employee insights into a scored report and a clear 90-day action plan for your business.Explore the AI readiness assessment

TL;DR

Around 65% of non-adopting Australian SMEs named distrust in AI decision-making or a preference for human control as their main reason, according to the National AI Centre’s latest quarterly insights.
Other barriers recur across Australian surveys: security and compliance worries, a digital skills shortage, no formal strategy, difficulty integrating with existing systems and no way to measure return.
Staff exclusion is a barrier you can fix cheaply: involve people in how AI is deployed and trained.
You cannot remove a barrier you cannot see. Start by mapping what tools your staff use today, then pick two or three barriers to attack.

What the numbers say about AI adoption in Australia
The main barriers, ranked by what Australian research shows
Barriers inside the business versus barriers in the market
Why visibility comes first
A 90-day approach to removing your barriers
Our perspective
See your own barriers in a scored report
FAQ
Sources

What the numbers say about AI adoption in Australia

Before the barriers, the baseline. The Australian Bureau of Statistics reports that about 12 per cent of Australian businesses used AI in the 2024-25 financial year. Use varies sharply by industry: the same release puts Information, media and telecommunications at 38 per cent, with Professional, scientific and technical services and Financial and insurance services both at 24 per cent.

AI adoption rates by Australian business industry

You will also see much bigger figures quoted. Treasury analysis reported by the ABC noted that about two-thirds of Australian businesses report some level of AI adoption, but fewer than 10 per cent report significant adoption. The two figures do not conflict. They measure different things: formal business use in a survey year versus any level of use, including a single person trying a chatbot. Treasury’s own description of uptake as “widespread but shallow” is the most useful summary: lots of light use, little deep integration, and not enough investment in organizational change and workforce skills.

That gap between light use and embedded use is where the barriers sit. The ABS release is also the best public benchmark you have, because it breaks adoption down by industry and employment size. Compare your business to firms like yours, not to a headline average.

The main barriers, ranked by what Australian research shows

No single survey ranks every barrier, and the studies use different samples. Read the list below as a map of recurring themes, not a league table.

1. Distrust and a preference for human control

The National AI Centre’s quarterly insights found that around 65% of non-adopting SMEs cited distrust in AI decision-making or a preference for human control as their primary reason. Difficulty seeing how AI applies to their specific business and not knowing where to start were the other common answers.

This is a leadership barrier, and it is often reasonable. A manager who has watched a chatbot invent a fact has good cause for caution. The fix is not persuasion but small, bounded trials where a person still checks the output.

2. Security and compliance concerns

In The State of AI in Australian Business 2026, a vendor-published report, security and compliance is reported as the top barrier for Australian SMEs, ahead of cost, data quality and integration. Treat it as one survey’s finding rather than settled fact, but it matches what owners tell us: the worry is that staff paste client or financial data into public tools.

That worry is justified by behavior. CPA Australia’s public sector work found that 40 per cent of public sector employees use publicly available, often unsanctioned AI tools, while only 24 per cent use employer-sanctioned ones. CPA Australia’s experts describe a governance gap: when policies are not widely understood, people reach for unverified tools, which exposes organizations to security and privacy risk. That is the public sector, but private businesses without a clear policy face the same dynamic. Our guides on AI data leakage, AI data classification and AI access control cover the practical controls.

2. Security and compliance concerns — overview diagram

3. A digital skills shortage

The 2026 Unlocking Australia’s Potential report puts a shortage of digital skills as the single biggest barrier for 41% of businesses. Skills here means more than prompt writing. It means knowing which tasks suit AI, how to check output and what must never be pasted into a tool.

Training sticks when it is tied to a real task that people do every week. The same principle shows up in other professions, for instance in this look at AI for language teachers, where the move from experimenting to useful practice depends on fitting AI into the actual work of the day.

4. No formal strategy

The same report found only 22% of businesses have a formal and comprehensive AI strategy, unchanged from the year before. Without one, AI spend becomes a scatter of individual subscriptions, and nobody owns the outcome. A strategy for a 40-person business does not need to be long. It needs a named owner, a short list of approved tools, two or three target workflows and a date to review.

5. Integration and unmeasurable returns

A Governance Institute of Australia survey reported that 88% of respondents had difficulty integrating AI with existing systems and 93% said they could not effectively measure return on investment. Those respondents skew toward organizations already deploying AI, so the figures describe the struggles of people who have started, not the whole market. The lesson still holds: decide before you start what you will measure, such as hours saved on a named task, turnaround time or error rate. Otherwise the return question cannot be answered later.

6. Staff left out of the process

This barrier is the least discussed and the cheapest to fix. A 2024 UTS study of nursing, retail and federal public service workers found that failing to include employees in AI deployment and training creates significant barriers to adoption. The researchers’ point is that employees are not inherently opposed to AI. Low trust and lack of involvement are the problem, and both are addressable.

Barriers inside the business versus barriers in the market

Some barriers are outside your control: the cost of enterprise tools, the pace of regulation, the quality of vendor products. Most of those in the list above are not. Distrust, skills, strategy, ownership and staff involvement are internal, and internal barriers move fastest when you can see them.

Barrier	Where it comes from	Quickest first move
Distrust of AI decisions	Leadership experience, accuracy worries	Pilot one low-risk task with human review
Security and compliance worries	No clear policy, unsanctioned tools	Approved tool list and a one-page usage rule
Skills gap	No task-based training	Short training on one real workflow per team
No strategy	No owner, scattered subscriptions	Name an owner, set two target workflows
Integration and ROI	Legacy systems, no baseline	Record a baseline before the pilot starts
Staff exclusion	Top-down rollout	Ask staff what they already use and what slows them

Why visibility comes first

Most advice on overcoming AI barriers starts with “develop a strategy”. That skips a step. In most small and medium businesses, leadership does not know which AI tools staff already use, for what, or with what data. A strategy written without that picture rests on guesses, and the guesses tend to be too optimistic about who uses AI and too relaxed about where company data goes.

The National AI Centre’s Guidance for AI Adoption, published in October 2025, sets out six essential practices: decide who is accountable, understand impacts and plan accordingly, measure and manage risks, share essential information, test and monitor, and maintain human control. Every one of them assumes you know what AI is in use. You cannot assign accountability for a tool you have not heard of, or measure the risk of a workflow you have not seen.

A 90-day approach to removing your barriers

Ninety days is long enough to change something real and short enough to keep attention. A workable sequence:

Days 1 to 14: find out what is actually happening. Ask every staff member, anonymously, which AI tools they use, for what tasks and with what kind of information. Anonymity matters: people who fear blame will under-report, and you will get a tidier picture than the real one.
Days 15 to 30: name an owner and write the one-page rule. One person is accountable. The rule covers approved tools, data that must never go into public tools and who to ask when unsure.
Days 31 to 60: fix the biggest security gap and train for one workflow. Usually this means moving staff from personal accounts to approved ones and teaching one team how to use AI on one task, including how to check the output. Our AI security checklist and AI security awareness guide give you a starting list.
Days 61 to 90: measure and decide. Compare the pilot against the baseline you recorded. Keep, change or stop. Then repeat the staff survey and see what moved.

If you are weighing risk more broadly, AI data governance and AI model risk management are the next topics to read.

Our perspective

We built Calibre Forum around one observation: the barrier most businesses name is not the one that is stopping them. Owners say “we’re not ready” or “we don’t know where to start”, which the National AI Centre data confirms is common. But when we look at how an organization really works, the first problem is nearly always that nobody has a shared picture of current use. Some staff use AI daily, some never, and management sees neither.

That is why we survey every staff member, anonymously, and combine the answers into one scored report. We do not install software, we do not collect individual data, and we make no compliance or legal claims. We do this because honest answers are the whole point. If staff think a survey is monitoring, they tell you what they think you want to hear, and you are back to guessing.

Our advice to any manager is plain. Do not start with the tool or the strategy document. Start by asking your people what they already do, treat what they say as useful information rather than a confession and pick two barriers to remove this quarter. Small, visible progress builds the trust that the survey data says is missing.

— Anthony

See your own barriers in a scored report

If you want the picture without building the survey yourself, our AI readiness assessment is built for Australian and New Zealand businesses. Staff complete an anonymous survey, and you receive a scored report that maps current tool use, flags risks and sets out a tailored 90-day action plan. You can see what is in the readiness report and open a sample report before you decide.

Calibreforum

The one-off report costs A$1,990. If you want to repeat the assessment every quarter and track progress against your 90-day plan, the annual membership is A$2,388 per year. We also publish industry versions for aged care and NDIS providers and RTOs and training providers, among others.

FAQ

What is the biggest barrier to AI adoption for Australian businesses?

It depends on the survey. The National AI Centre found distrust of AI decisions or a preference for human control was the leading reason among non-adopting SMEs, at around 65%. Vendor-run surveys point to security and compliance, and others to skills. All agree the barriers are mostly about people and process, not technology.

How many Australian businesses use AI?

The ABS reports about 12 per cent of businesses used AI in 2024-25. Treasury analysis suggests about two-thirds report some level of adoption, but fewer than 10 per cent report significant adoption. The difference is the definition of use.

How can I benchmark my business against others?

The ABS Business Characteristics Survey reports AI adoption by industry and employment size, so it is the best public starting point. For an internal benchmark, survey your own staff now and repeat in a quarter to see your own movement.

Where should a manager start?

Find out which AI tools your staff already use and what data goes into them. Then name one accountable owner, write a one-page usage rule and pilot one workflow with a baseline you can measure against.