69% of Australian Small Businesses Use AI. Most Aren't Saving Time. Here's What Oz Oils Did Differently.
Most Australian small businesses now say they use AI, but in September Treasury and COSBOA warned that many are stuck at experimenting, and the productivity gain isn't showing up. Here's the difference between the businesses where it is, and the ones where it isn't.
TL;DR
- Most Australian small businesses now say they use AI, but in September Treasury and COSBOA warned that many are stuck at experimenting, and the productivity gain isn't showing up.
- The surveys disagree wildly: one puts regular use at 69%, the ABS puts workplace use at around 12%. The gap is the difference between someone using ChatGPT and AI built into how the work gets done.
- A chat tool saves one person a few minutes. A process rebuilt around AI saves the business hours, every day.
- Oz Oils went from about 6 hours of daily pickup scheduling to about 30 minutes. Ramjet cut its order admin time by roughly 80%.
- Both started with one repetitive process, a measured baseline, and a time saving agreed before anything was built.
- If you can't say how many hours AI saved your business last month, you're experimenting, not automating.
Why Do the AI Adoption Numbers for Australian Small Businesses Disagree So Much?
They're measuring different things. Intuit QuickBooks research found regular AI use among Australian small and mid-sized businesses rose from 40% in July 2024 to 69% in January 2026. The ABS, surveying businesses for 2024–25, found around 12% reported using AI in their workplace. The National AI Centre's tracker sits in between, at around 43–44%.
"Using AI" can mean one staff member drafting emails in ChatGPT. It can also mean AI reading every incoming order and entering it into your system. Both count as adoption in a survey. Only one changes how many hours your business spends on admin.
My hunch: if surveys only counted businesses where AI runs as part of a process, the number would be a long way below 69%.
What Do Business Owners Mean When They Say They Use AI?
Usually, they mean ChatGPT. When I ask owners on discovery calls whether they use AI, the answer is almost always "yeah, all the time." Then comes the example: it sharpens up their emails, or they use it to research something. I do the same, and it's genuinely useful.
But nothing about how the business runs has changed. Orders still arrive by email and get retyped. Scheduling still happens by hand. The AI sits beside the work, not inside it.
What Are Treasury and COSBOA Actually Worried About?
They're worried that small businesses stop at experimenting. In early September, COSBOA warned that Australia risks leaving its 2.7 million small businesses behind in AI. That followed Treasury advice that the economy could miss out on AI's broader benefits if businesses don't move past basic experimentation. A parliamentary inquiry is now examining what's stopping SMEs adopting AI more deeply.
I agree with them. The barrier isn't access: every owner can open the same chat window. It's that most owners haven't seen what AI can do to a genuinely manual process, so the tool never turns into a changed process.
What's the Difference Between Using AI and Automating With AI?
Using AI helps one person do a task faster; automating with AI takes the task out of someone's day.
| Using an AI tool | Automating a process with AI | |
|---|---|---|
| Who benefits | The person typing the prompt | Everyone who touches the process |
| Where it runs | In a separate chat window | Inside the systems you already use |
| How often | When someone remembers | Every time the work arrives |
| Accuracy | Depends who's doing it that day | The same result every time |
| Measured? | Rarely | Hours before and after, in writing |
| Typical gain | Minutes per task | Hours per day |
Speed is only half of it. A process that runs the same way every time also takes out human error: no mistyped order numbers, no missed pickups, no "I thought you'd done it." Over a year, that consistency is worth as much as the hours.
What Does It Look Like When AI Actually Saves Time?
It looks like one specific, repetitive job disappearing from someone's day. Two examples from our clients:
Oz Oils: 6 hours of daily scheduling down to 30 minutes. Oz Oils was spending about 6 hours a day scheduling pickups. It now takes about 30 minutes, and the system has run live since May. That's roughly 5.5 hours back every working day, or more than 1,300 hours a year across a five-day week. "We've consolidated 3 systems into one and now have full visibility of our fleet, something we've been chasing for 5 years," says owner Scott M.
Ramjet: roughly 80% less time on order admin. Ramjet's team used to retype every order email by hand. Roughly 80% of that time is now gone, and the tool runs in production every day. "We can spend more time moulding and less time on admin," says general manager Grant T.
Neither started with "let's use AI." Both started with a process that cost real hours every week, a measured baseline, and a target agreed before anything was built.
I see the same pattern on almost every build we do. Owners are keen but a bit daunted at the start. Once they see the tool is simple for their team to use, they stop asking whether AI will work and start asking where else in the business it can go.
How Does a Small Business Move From Experimenting With AI to Measurable Results?
Pick one process and measure it before you change it.
- Pick the most repetitive job: the same steps, every day or week, done by someone whose time is worth more elsewhere.
- Measure it now: hours per week, who does it, how often it goes wrong.
- Agree the target in writing before you build: for example, "6 hours a day down to under 1."
- Build it into the systems you already use, not a separate tool people have to remember to open.
- Measure again. We check at Day 14 and Day 60 against the Day 0 baseline.
Hypajump's recommendation: Stop counting AI tools and start counting hours given back. Choose one repetitive process, record how long it takes today, and agree the time saving in writing before anything is built. Until you can show the before and after, you don't know whether AI is helping.
Common Questions
Is my small business falling behind on AI? Probably not on tools: most of your competitors use the same chat tools you do. The gap that matters is whether any of your processes now run differently because of AI. If none do, you're level with most of the market, which is exactly what Treasury is worried about.
Do I need to replace my systems to automate with AI? No. The automations that save the most time run inside what you already use: your email, your accounting software, your job or order system. Replacing systems is expensive and slow; building into them is usually neither.
What is the Digital Solutions program, and can it help my business with AI? Digital Solutions is a federally funded program that helps small businesses adopt digital tools, and more than $25 million has gone into its latest round with a focus on AI. It's a useful place to get advice on where to start. Check business.gov.au for current eligibility.
How do I know if AI is actually saving my business time? Measure before and after. Record the hours a process takes in a typical week before you change anything, then measure the same way after two weeks and again after two months. If you can't produce both numbers, you don't know.