If you run a small business in Australia and you have not started using AI yet, here is the first thing you should know: most people in your position have not either, and the most common reason is simple. They do not know where to start.
That is not a guess. It is the finding of a Deloitte Access Economics report released in November 2025, commissioned by Amazon and based on a survey of more than 1,000 Australian small and medium businesses. One-third of the businesses not currently using AI told Deloitte the same thing: we do not know where to start.
You are not behind. You are in the middle of the curve. The people selling you AI tools want you to feel late. You are not.
What the Deloitte report actually found
Two-thirds of Australian small and medium businesses are using AI in some form. Only 5 per cent are using it well enough to capture its full potential. The gap between "using" and "using well" is where the money is.
Deloitte's modelling found that a small business moving from basic AI use to intermediate use can expect a 45 per cent increase in profitability. A business moving from intermediate to fully enabled can expect 111 per cent. If just one in ten Australian SMBs advanced one step on the AI maturity ladder, the country's GDP would grow by $44 billion a year.
That is the upside. The downside is what is keeping you from it. Deloitte identified five barriers, and the first one — the one one-third of non-users named — is not knowing where to start.
Why nobody told you where to start
Because nobody's job is to tell you.
Your accountant is paid to do your books. Your IT person, if you have one, is paid to keep the email working. The vendor on LinkedIn who says "I will build your AI in two weeks" is paid to sell you the most expensive thing you will say yes to. Nobody in that list has a reason to spend an hour helping you understand a tool you might never buy.
This is the gap PRODUCT3 was built for. Not because we have a magic answer. Because the gap exists.
Where to start, in plain words
You start with one task. Just one.
Pick the smallest thing you do every week that you resent. Not the most important thing. The most boring thing. Replying to the same five customer questions on WhatsApp. Drafting the same overdue-invoice email. Writing the social media post you keep meaning to write and never do.
Open ChatGPT (chatgpt.com) or Claude (claude.ai). Both have free versions that work for what you are about to try.
Type the task in plain English. Not "as a marketing assistant, please generate" — just "write me a polite WhatsApp reply to a customer asking if we are open Sunday." Read what comes back. Change a word. Use it.
Do that for two weeks. One task. Same tool. See what happened to your time.
That is the entire starting point. No subscription, no agency, no $5,000 setup fee, no decision about which AI is best. Just one task, one tool, two weeks.
What you are actually testing
You are testing three things at once, and you do not have to know it.
- Whether AI saves you time on this specific task. Some tasks it helps a lot. Some it helps a little. Some it does worse than you. You will find out which is which.
- Whether you can trust what it writes. The honest answer is sometimes yes, sometimes no. AI sometimes makes things up. After two weeks you will have a feel for when to read carefully and when to send.
- Whether the tool fits how you work. Some people love ChatGPT and find Claude cold. Some people are the opposite. There is no universal answer.
Two weeks of ten minutes a day teaches you more than ten articles.
What the Deloitte report says about the cost of staying still
Deloitte's barrier list is honest about why people do not start. Beyond not knowing where to begin, the report names four other reasons: limitations in business systems and data, workforce skills, funding constraints, and uncertainty about responsible AI use.
All of those are real. None of them is a reason to skip the two-week experiment with one task. The experiment costs nothing, requires no new systems, demands no team training, and risks no money.
The cost of staying still is harder to see. Deloitte put a number on it: a small business that does not advance from basic to intermediate AI use is leaving 45 per cent profitability growth on the table, on average. The competitor down the street who does run the experiment will get there first.
What we publish next
This post is the first in a series for Australian small business owners who are not in tech, do not want to be in tech, and want to understand what is happening without being sold to. Coming up:
- Your team is already using AI without telling you. What's at risk, and how to handle it without becoming the boss who bans things.
- The unsexy AI wins. Five boring tasks where AI gives you back five hours a week.
- The AI agency selling you a chatbot in two weeks. What to ask before you sign anything.
- Real WhatsApp results from real small businesses. Numbers, not hype.
If you want the next one in your inbox when it goes up, the email box is at the top of the page. We do not send anything else. We do not sell your email. We just publish.
You are not behind. You are early enough to learn before the wave of bad implementations breaks. The Deloitte numbers say it clearly: one in three of the businesses not using AI yet are exactly where you are. The other two-thirds are using AI, and most of them are using it badly. The room to do it well is wide open.
Pick a task. Open a tab. Two weeks.