Tuesday 27 October 2026 • By Timothy, CPA — Managing Director, Professional Financelink (PFL)

A stack of translucent layers where one middle layer glows brighter than the rest, adding its own pattern to the light passing through

Big Companies Are Cutting Jobs and Naming AI. In a Small Finance Team, the First Role at Risk Belongs to Whoever Still Answers "I Don't Know"

The AI gap between large and small organisations is wide, and the layoff headlines come from the large end. The question for everyone else isn't whether AI can do your job. It's whether you're adding anything to what it already does.

Late last week, two large financial companies announced job cuts within days of each other. Barclays disclosed plans to cut about 5,000 roles, roughly 5% of its workforce, citing "organizational simplification and technology change efficiency driven by AI". American Express announced cuts of about 5% of its global workforce. Its CFO said the restructure would partly help the company "prepare for AI" by reducing management layers before automating. Barclays said about 60% of its cuts would fall on support functions.

Headlines like these now arrive most weeks. If you work in finance, the natural question is: how long until it's my job?

I'll give you my honest answer. Then I'll give you something practical to do about it this week.

35% vs 11%

Australian large businesses using AI in 2024-25, against small and micro businesses combined (ABS, released June 2026)

86% vs 15%

US small businesses in a Goldman Sachs program survey that use AI, against those that have fully integrated it into core operations

The gap is real, and it cuts both ways

The Australian Bureau of Statistics found about 35% of large businesses used AI in 2024-25, up from 9% three years earlier. Among small and micro businesses combined it was about 11%. Other surveys, such as the National AI Centre's monthly SME tracker, measure differently and report much higher use. But every measure points the same way. Large organisations are further ahead, and the job-cut announcements come from that end.

The Goldman Sachs survey of 2,473 alumni of its small business program in the US, run in August, shows the other half of the picture. Most of those businesses use AI in some form. Few have built it into how the business actually runs.

For a small or mid-size organisation, that gap is both a warning and a breathing space. The warning is that the large-company pattern of flattening first, then automating, will reach smaller teams as the tools get easier. The breathing space is that it hasn't arrived yet for most of you. What you do in the meantime is the part you control. In a team of four or five, nobody announces an AI restructure. Roles get reshaped quietly, one vacancy at a time, around the people who've shown they can do more with the tools.

What I've seen decide who gets replaced

Here's the view I've formed from years of managing finance teams, and from watching how that's changed in the last eighteen months.

When a manager asks someone for something they haven't done before, there are now two kinds of answer.

The first is: "I'm not sure, I've never done that." Being unsure isn't the problem — in finance, saying so is often the prudent answer. Stopping there is. Two years ago that was a reasonable reply, and it often led to a useful conversation. Today, the manager can type the same question into an AI tool and get something like 70–80% of an answer in a few minutes. Not perfect, but a solid start. So "I don't know" now means the person adds less than the assistant on the manager's own laptop. If that keeps happening, the manager stops asking the person and goes straight to the assistant. Over time, that's how a role shrinks, and eventually how a layer disappears. Nobody needs to announce a restructure.

The second answer is: "I hadn't done that before, so I used AI to learn the basics and work through it. I checked it against the source, and here's what I think it means for us." That person can't be replaced by the tool, because they're the one making the tool's output usable. They used AI to do something they couldn't have done alone, and then they added the part the tool can't supply: knowledge of our business, judgement, and accountability for the answer.

The difference between those two people isn't intelligence or qualifications. It's whether they treat AI as a smart assistant or as an excuse.

The mistake that's worse than "I don't know"

There is one answer worse than "I don't know". It's forwarding an AI's output with no review. Sometimes it comes with "the AI says…" in front, as though that makes it someone else's responsibility.

The biggest bottleneck in most AI-assisted work today isn't generating the output. It's a person reviewing it. If you pass AI output up the line without checking it, you haven't saved time. You've moved your review work onto your manager, who is usually the most expensive and least available person in the chain. You saved your time at your boss's expense, and your boss will notice.

Outsource the typing, the first draft and the research legwork. Never outsource the thinking. The moment you can't explain why a number is right, it isn't your work any more. It's just something you passed along.

A practical habit: the four-part reply

Here is a simple structure I'd encourage every finance team member to use when they hand back AI-assisted work. It works for a variance explanation, a policy question or a new reporting request:

1. What I found. The answer, in your own words, in two or three sentences.

2. How I checked it. Which figures tie to the ledger, which statements you read at the source (the legislation, the ATO page, the contract), and which parts the AI drafted.

3. What I think. Your view on what it means for this organisation. The AI can't write this part well, because it doesn't know your funders, your board or last year's history.

4. What I'm still unsure about, and how I'd close it. This is the honest replacement for "I don't know". "I couldn't confirm whether the new rate applies from the first pay period or the first full one. I've asked our payroll provider and will have it by Thursday" is a strong answer. It tells your manager exactly where the risk is.

If you manage a team, ask for this structure explicitly. It makes review faster, it shows you who is thinking, and it gives people a clear standard to grow into.

The part AI genuinely can't replace

One last thing, and it isn't technical. The most durable skill in any team is how you treat the people around you. Explaining something patiently to a junior colleague and backing a peer under pressure count for more than people think. So does being the person others want to work with when month-end goes wrong. No model does that. In my experience, when organisations decide whose role to keep, that counts for a great deal more than it shows in a job description.

So use AI as hard as you can. Learn with it, draft with it, test your thinking against it. Then add your judgement, own the answer, and be kind to the people around you. That combination is very hard to replace.

Note: The scenarios in this post are based on real experiences — mine and those shared by colleagues across the sector. Details might have been changed and modified slightly to protect confidentiality, and mostly used 1st person perspective for convenience.

A note on AI tools and your data: learning with AI often means pasting in real work, such as a payroll query, a client file or a set of figures. Before you do, check whether your organisation has approved the tool and whether the vendor trains its models on customer inputs. The safest posture is a tool where your data isn't retained for training. Remove personal and client details when you're only trying to learn a concept.

Is your finance team using AI to think better, or to think less?

The difference shows up in review time, error rates and how much your senior people can trust what lands on their desk. PFL provides senior-level outsourced finance, management reporting, and AI automation for Australian NFP, NDIS, and SME organisations, including helping small teams build AI habits that make their work more reliable rather than just faster.

Talk to PFL →

Timothy, CPA is Managing Director of Professional Financelink (PFL), providing senior-level outsourced finance, management reporting, and AI automation for Australian NFP, NDIS, and SME organisations. 20+ years in finance leadership across NFP, NDIS and SME.

Sources

Tomorrow: childcare centres keep closing. We dig into why, and where AI can genuinely lift a centre's efficiency.

Search This Blog

Finance Intelligence