AI for Australian SMEs: from tools to outcomes

Most of what gets sold to a small business owner under the banner of AI is, when you look closely, just software with a smarter interface — a tool that does a task faster if you feed it the right inputs, sit next to it, and check its work. That's genuinely useful, and I don't want to be sniffy about it. But it isn't the change that matters, and I'd argue treating it as the change is exactly how owners end up busier rather than freer.

The real shift — and I do think the world is genuinely changing here, not in a hype way but in a boring, structural way — is that the operating model of a small business can now be different from what was ever possible before. For most of the last few decades, a business of ten or twenty people could only look one way: a founder or two doing the thinking, a handful of people doing the doing, and a back office (bookkeeping, payroll, compliance, admin) that was either an overworked internal person or a patchwork of outsourced bits that never quite talked to each other. The size of what you could run was capped by how much coordination one or two people could hold in their heads.

What's different now is that a lot of that coordinating, chasing, reconciling and routine judgement can be carried by systems rather than by people — but only when those systems share the same context. And that last clause is the whole game. So the question I'd actually be asking, if I were an owner today, isn't 'which AI tool should I buy?' It's 'what could this business be if the repeatable machinery ran itself, and I spent my time on the handful of things that genuinely need me?' That's a different question, and it leads somewhere different. It's the thinking behind our whole approach to a connected back office (/connected-back-office) and, more broadly, how we think about technology (/technology).

The uncomfortable truth is that AI is only as useful as the context it can see. An AI tool that knows about your invoices but not your contracts, or your payroll but not your leave policy, or your customer emails but not your actual order history, is working half-blind — and it will confidently produce answers that look right and aren't. The intelligence isn't really in the model; it's in the model plus the shared, connected picture of how your business actually runs. Take that context away and you've got a very articulate assistant guessing.

So here's the trap. Most small businesses are already fragmented — the accounting lives in one system, payroll in another, the CRM somewhere else, half the real knowledge in someone's inbox and the other half in their head. When you bolt a separate AI tool onto each of those silos, you don't reduce the fragmentation; you add a layer of confident automation on top of it. Now you've got five AI tools, each seeing a slice, each making local decisions that don't add up, and a new coordination job — reconciling what all these tools did — that lands right back on you. I've watched owners buy their way into more work this way, with the best of intentions.

This is why I'd argue the sequence matters enormously. Connect the business first, then apply the intelligence — not the other way around. When the underlying picture is joined up, AI becomes genuinely transformative, because it's finally reasoning over the whole thing rather than a fragment. When it isn't, more AI is just more surface area for things to go quietly wrong. It's the difference between an AI-enabled back office (/ai-enabled-back-office) built on a connected foundation, and a pile of clever tools sitting on a shaky one. If you take one thing from this page, I'd like it to be that ordering: context before cleverness.