If you've opened ten tabs trying to work out where to start with AI, you're not behind — you're in the majority. Among Australian small businesses not yet using AI, about one in three say the same thing: they don't know where to start. The honest first step isn't a tool. It's one task worth handing over. This guide shows you how to pick it — and, just as importantly, where to stop.
Ready to go a level deeper once you've picked your task? See the companion guides: 7 jobs to hand to AI and 8 workflows that run themselves.
No jargon, no tool list to wade through. Five short sections to take you from stuck to a clear first move.
If you've been saving AI posts you never read, signing up for free trials you've forgotten about, and watching competitors talk about "leveraging AI" while you quietly wonder what you're missing — that's the group you're in, and it includes a lot of very capable business owners.
It's not you. It's that there are thousands of AI tools out there, most of them shouting for your attention, and almost none of them will tell you the honest thing: a tool on its own won't fix anything. The proof is in the numbers. Deloitte's 2025 study of more than 1,000 Australian SMBs found that while about two-thirds are already using AI, only 5% have it genuinely embedded in how they work. Most businesses have dabbled. Very few have handed a real job over.
Source: Deloitte Access Economics, The AI Edge for Small Business (commissioned by Amazon), November 2025 — see the notes at the foot of this page.
The question to start with isn't which tool. It's which task. Starting with a tool is like buying a power drill before you know whether you need a shelf. You'll end up with the drill, a few expensive subscriptions, and the same problems you started with.
Here's what usually happens when businesses get this the wrong way round: they see a slick demo, sign up, open the tool once, don't quite know what to do with it, and cancel two months later. Nothing changed because nothing was pointed at a real job.
So flip it. Start with the task that's eating your time, not the tool. Once you know what you want to fix, the tool choice gets obvious fast — and often cheap.
For most Australian small businesses, these four are the best place to start. They're high-volume, low on creativity, and structured enough to hand over safely — the classic "same shape every time, different details" jobs.
You probably lose the first 30 to 90 minutes of the day to a wall of unread mail before you've done anything that actually moves the business forward.
Same structure every time, different inputs — exactly what AI is built for. And the bit most owners never get to: chasing the quote that went quiet, where the work quietly evaporates.
"Are you open?" "What's it cost?" "Do you do X?" The same handful of questions, over and over, across email, web chat and DMs.
Copying customer details from your inbox to your CRM to your accounting software. Dull, error-prone, and it drains an hour or two a week from someone who should be doing something more valuable.
None of these are glamorous. That's the point. Start with the dull, repeatable one where the leak is biggest for you — not the impressive one from the demo.
Pick one task in your business right now and answer four questions about it. This is the whole test — no maturity model, no jargon.
Four yeses, and that task is a strong candidate. No to most of them, and you're not ready to automate that one yet — and no tool will change that. If nothing in your business is repeatable, that's a useful answer too: AI isn't your next investment.
Take the hours the task eats each week, multiply by a modest cost per hour, then by the weeks you actually work in a year. Then halve it — because AI saves you a good chunk of the time, not all of it. Here's a worked example, with the assumptions shown so you can argue with them.
| Example: answering the same questions | Figure |
|---|---|
| Time on it each week | ~3.75 hrs |
| Your cost per hour (modest) | ~A$60 |
| Cost per year (× 48 weeks) | ~A$10,800 |
| Recovered, halved to stay conservative | ~A$5,400 / yr |
That's one task, on deliberately cautious assumptions. The figure is an illustration, not a promise — your real number depends on your actual hours, your actual rate and your actual tools. A generic guide can show you the pattern; it can't tell you your number.
Run the maths on your own task with the free AI ROI calculator — or just time the task by hand for a week and count what it costs. That number tells you whether to bother.
Handing off the repetitive parts of a job is smart. Handing off the judgement is how it goes wrong. Whatever you start with, it should always stop short of anything that:
The rule that keeps you safe is simple, and it's the same one all the way up: AI does the doing; you keep the judgement.
Deloitte's modelling puts a number on the upside: an SMB that moves from basic, dabbling use to AI genuinely embedded in the work could see profitability rise materially — they model the lift from basic to intermediate use at around 45%. Treat that as a direction of travel, not a guarantee; your number depends on your business.
The warning sits right beside it. An MIT report widely cited through 2025 (The GenAI Divide) found that roughly 95% of enterprise generative-AI pilots delivered no measurable impact on the bottom line. The reason wasn't the technology — it was that most stayed at "someone opens a chatbot now and then" and never got wired into a real workflow. That's a big-company study, but the lesson is exactly the same for a small business: dabbling doesn't pay; a task properly handed over does. Which is the whole argument for starting with one clear job and doing it well.
Spend five to seven minutes telling us how your business runs, and we'll send back a short, branded AI Opportunity Map built only on your answers: your top opportunities ranked, the hours and dollars each could save, the data watch-outs for your situation, and the one task worth starting with. No charge, no obligation, no sales pitch.
Book a free AI Opportunity Audit