
It’s one of the first questions owners ask us, and the office manager usually asks it straight after: is AI safe for business, or are we about to hand our customer list to someone we’ve never met?
In a survey of more than 1,000 Australian small and medium business owners, reported by Electrical Connection in September, data privacy was the most common reason given for not using AI, at 39%.
Our honest answer: it depends less on the AI and more on how it’s set up. Here’s the plain-English check we run before recommending anything.
Start With What Your Office Actually Holds
Most trade businesses hold more personal information than they realise. Not just names and numbers, but gate codes, notes about the dog in the backyard, photos from inside someone’s home, payment details and staff records.
Before any AI tool goes near that, we list what lives where: the job software, the inbox, the shared drive, the phones in the crew’s pockets. You can’t decide what’s safe to share until you know what you’ve got.
Four Questions to Ask Before Any Tool Goes In
The Australian Cyber Security Centre (ACSC) has a short AI checklist for small business. Here’s the same ground in trade terms.
Where Does the Information Go?
When the office connects an authorised inbox or pastes a customer email into an AI tool, where does that information go? Is it stored in Australia or overseas, and for how long? Check that before anything is connected. The answer is usually buried in the terms. We read them.
Is It Used to Train the Tool?
The ACSC notes that some AI providers may use what customers type in to train or refine their models, depending on the settings or the type of subscription. Often it can be switched off, if someone knows where to look.
Who and What Can It Reach?
An AI assistant that can read every inbox in the business is a very different risk to one that only sees the shared enquiries inbox. The ACSC and its overseas counterparts recommend never giving AI agents broad or unrestricted access, especially to sensitive data, and widening access only as trust builds. We start narrow.
Who Checks What Goes Out?
AI drafts can sound right and still be wrong. A person should check anything going to a customer, especially quotes and prices, unless you’ve agreed clear sending rules for that job.
What the Privacy Rules Say, in Plain English
We’ll keep this short. We’re not lawyers, so run it past yours.
Under the Privacy Act, a small business is one with annual turnover of $3 million or less, and most of those aren’t covered. But plenty of trade businesses with 20 or 50 staff sit above that line, and some smaller ones are covered anyway, such as those working under a Commonwealth contract. The Office of the Australian Information Commissioner (OAIC) has a free checklist to help work out where you stand.
If your business is covered, two things matter for AI:
· Personal information still counts when it goes into an AI tool. The OAIC says privacy obligations apply to personal information entered into AI systems. As best practice, it recommends not putting personal information, particularly sensitive information, into publicly available AI tools.
· A new transparency rule starts on 10 December 2026. Covered businesses that use a computer program to make, or substantially help make, decisions that could significantly affect someone, using that person’s information, will need to explain it in their privacy policy. The OAIC’s fact sheet says this can include AI, and even off-the-shelf software you’ve configured.
For most everyday office jobs, like drafting a reply or summarising a supplier agreement, that rule may not come into it. But if AI starts deciding who gets a call back or who’s offered payment terms, it’s worth checking. We’ll flag those points so you can take them to your lawyer.
How We Handle It
Is AI safe for business, then? It can be, when someone has done the checking. That’s a big part of how we work at AI4Tradies:
1. Find where AI fits. We look at where the office loses time, and what information each job would touch.
2. Set it up right. We pick ready-made tools, check where your data goes before anything is switched on, and give each tool only the access it needs. If nothing ready-made is safe enough, we build it properly.
3. Show your team how. Including which accounts to connect, what not to paste in, and how to strip out names when they aren’t needed.
4. Keep an eye on it. Tools update and settings change. We keep watching and tell you straight when something needs a look.
The National AI Centre suggests naming a person who is accountable for each AI tool a business uses. Often that’s the office manager. We make sure they’re not doing it alone.
A Simple First Step
Not sure where AI would fit in your office, or what it would touch? Our capacity calculator shows where your office hours are going and the first thing we’d look at.