75% of Australian SME Finance Leaders Want AI. Only 1 in 4 Are Using It. Here's Why.
A report dropped last week that managed to be both unsurprising and worth talking about. Budgetly's 2026 CFO Survey found that three in four Australian SME finance leaders want AI automation in their function — but only one in four are actually using AI tools today.
That gap — 75% intent, 25% action — is one of the more interesting numbers to come out of Australian finance this year. Not because the adoption lag is shocking (technology adoption curves rarely are), but because of what's sitting inside it. The gap isn't about access. AI tools are cheap, widely available, and require no procurement cycle. It isn't about awareness either — clearly, three quarters of finance leaders already know they want this. So what is it?
Having spent time with finance teams across sectors and having been deep in the middle of building automation into my own workflows, I think the gap comes down to three specific things — none of which are the ones most commonly cited.
|
75%
Of Australian SME finance leaders want AI automation in their function — Budgetly CFO Survey, May 2026
|
25%
Are actually using AI tools today — leaving a 50-percentage-point intention-action gap in SME finance
|
Barrier 1: They're Waiting to Feel Ready — Which Never Comes
The most common thing finance professionals say when asked why they haven't started with AI is some version of "I want to understand it better first." That instinct is completely understandable — finance teams are trained to understand things thoroughly before acting on them. Due diligence is the job.
The problem is that AI literacy doesn't really work that way. Unlike a new accounting standard or a regulatory change, where you can read the guidance and arrive at a definitive understanding, AI fluency is experiential. You don't understand what an AI tool can do for your month-end close by reading about it. You understand it by trying it — on a real task, with real data, and accepting that the first few attempts will be imperfect.
Finance teams that are waiting until they feel fully prepared are, in practice, waiting indefinitely. The teams making progress are the ones that started with a single low-stakes task and built from there.
Barrier 2: The "Where Do I Even Start" Problem
For finance teams that want to use AI but haven't started, the most paralyzing question is often the first one: which task? AI tools can theoretically be applied to almost everything in a modern finance function — variance analysis, management commentary, reconciliations, payroll checks, forecasting, grant reporting — and that breadth can itself become an obstacle. When everything is a candidate, nothing gets started.
The teams that have moved from intent to action tend to share one characteristic: they identified a single task that had three specific properties. It was repetitive — done frequently enough that any time saving would compound quickly. It was well-defined — clear inputs, clear expected outputs, easy to verify. And it was low-risk — if the AI output was wrong, the finance team would catch it before it caused a problem.
Management reporting commentary is a good example. Drafting the narrative section of a monthly management reporting pack from a set of variance numbers is repetitive, structured, and easily reviewed. A finance manager who uses AI to produce a first draft of that commentary — which they then edit and verify — has immediately reclaimed an hour or two per month, and built the kind of hands-on AI experience that accelerates everything else.
Budget variance analysis summaries, supplier payment run checklists, and cash flow forecast refresh notes are other tasks in the same category. The entry point doesn't need to be dramatic.
Barrier 3: The Governance Anxiety Is Real — But Often Misapplied
The third barrier is the one finance professionals talk about least openly, but which I suspect is the most significant. There's a genuine anxiety about governance, privacy, and what it means to introduce AI into a function that handles sensitive financial and payroll data.
That anxiety isn't wrong. It's actually the right instinct — just sometimes applied to the wrong part of the problem. ASIC's 2026 key issues outlook flagged variable maturity in how Australian businesses manage AI governance risks, and finance functions that handle sensitive data have a legitimate obligation to think carefully about which tools they use and what data goes into them.
The important clarification: not all AI use in finance involves feeding sensitive data into a general-purpose tool. Using an AI tool to help draft management commentary based on numbers you've already summarised is categorically different from feeding payroll records or client financial data into an AI system whose privacy terms you haven't read. The governance question matters most for the latter category.
For most SME finance teams, the practical answer is to start with tasks that don't involve raw sensitive data — commentary drafting, process documentation, analysis framing — and build a clear internal protocol for what data can and cannot go into AI tools. That protocol doesn't need to be a lengthy governance document on day one. It can be a single-page policy that the finance team agrees on and revisits quarterly.
A word on privacy, worth stating clearly: if you're exploring AI tools for your finance function, always review whether the tool uses your data to train its models, and whether that can be turned off. For any tool handling real financial, payroll, or client information, this isn't optional due diligence — it's basic data governance.
What the 25% Doing It Well Have in Common
Across the finance teams making genuine progress with AI automation, a few patterns stand out consistently.
They started narrow and specific, not broad and aspirational. A decision to "use AI in finance" is not actionable. A decision to "use AI to draft the variance commentary in the management reporting pack by next month-end" is. The specificity is what converts intent into action.
They built in human review from the start. The teams having the most success aren't treating AI outputs as final outputs. They're treating them as high-quality first drafts that a finance professional reviews, adjusts, and approves. That review step isn't a concession to AI's limitations — it's how the finance team maintains accountability while capturing the time saving.
And they normalised imperfect early results. The first time a finance team uses AI for a task, the output is often 70–80% of the way there and needs editing. That's still a significant time saving, and the quality improves quickly with iteration. The teams that stuck with it through the imperfect early attempts are now operating meaningfully faster than those that tried once, found it imperfect, and stopped.
Curious about where agentic AI — the kind that can take multi-step actions autonomously — fits into this picture? This post on agentic AI in finance covers the next wave of automation and what it means practically for finance teams.
The Cost of Staying in the Intention Zone
There's a temptation to treat AI adoption as a nice-to-have that can wait until the timing is better — until the team is less busy, the tools are more mature, the governance is clearer. That logic has been operating in SME finance for the better part of three years now, and the result is the gap this survey just quantified.
The cost of staying in the intention zone isn't dramatic in the short term. No single month-end is noticeably worse for not having AI assistance. But the compounding effect of reclaiming two or three hours per month across a finance team, and progressively directing that capacity toward higher-value analysis and strategic support, is significant over a year. The teams that started twelve months ago are already operating differently from those that haven't.
The 50-percentage-point gap in this survey will close. The question for finance leaders is whether they want to be in the group that shaped their AI capability intentionally, or the group that eventually caught up reactively.
PFL helps SME finance teams move from AI intent to practical, working automation — starting with the tasks that will have the fastest impact on your team's capacity and reporting quality. If you're not sure where to start, that's exactly the conversation we're built for.
Talk to PFL →
Comments
Post a Comment