A US study of 21,559 firms found the heaviest AI spenders grew headcount 10.2% in two years. Here's what that means for smaller Australian businesses.
For two years the loud prediction has been that AI would thin out payrolls. Buy the software, cut the staff. A new US study points the other way.
Corporate card firm Ramp and workforce analytics firm Revelio Labs matched spending records to headcount data for 21,559 US firms, covering January 2021 through February 2026. The firms that spent the most on AI didn’t shrink. Over two years their headcount grew 10.2%, and the growth reached right down to entry-level roles.
If you run a smaller business here in Australia, the study wasn’t about you. Most of the heavy adopters were large US firms. But the pattern underneath it is worth your time. Picture a small insurance brokerage in Brisbane, the kind of firm with a handful of brokers and a couple of support staff buried under renewals and paperwork. The study is really a story about what a business like that could gain by putting AI to work properly, rather than treating it as a way to cut heads.
What the study found
The researchers split firms by how hard they leaned into AI. High-intensity adopters, which Ramp defines as roughly the top third of per-employee AI spending, were anchored around thirty US dollars per employee each month in the first three months after they started. That group is the one that grew.
The growth wasn’t only at the top. Entry-level headcount rose 12%, while managers and officers rose 6.7%. That runs against the common worry that AI would eat the junior jobs first. If anything, these firms hired more of the people everyone assumed would go.
The timing is worth noting too. The hiring gains didn’t show up straight away. They emerged six to twelve months after firms adopted AI, which fits the idea that a business needs time to work the tools into how it operates before that effort turns into more work, and more people to do it.
Where this happened matters as much as the headline. The heaviest adoption, and the hiring that followed, sat in information, finance, and insurance. Sectors like construction, healthcare, arts and entertainment, and food services lagged behind.
Why you should read it carefully
Here’s the part the excited headlines skip. This is a correlation, not proof. The researchers were clear that the firms adopting AI were already larger, more engineering-heavy, more often venture-backed, and growing faster before they ever bought a single model subscription. So some of that hiring was always going to happen. Buying more AI seats didn’t magically create jobs at a steel fabricator in Ohio.
There’s more the study can’t yet tell us. We don’t know whether those hires last beyond the two-year window. We don’t know how many of the new roles are AI-related work rather than ordinary backfill. And we don’t know how the picture looks once you take out the venture-funded growth. The honest read is that AI spending and team growth move together in a thriving slice of American firms, and it’s too early to stretch that from finance and information across the whole economy.
Our Brisbane brokerage isn’t venture-backed, and it isn’t sitting in a fast-growing US tech sector. So the caution cuts both ways for a smaller Australian firm. You can’t assume that spending on AI will grow your team. But you also don’t have to fear that doing it well will cost you people.
What this means for you
So what does a smaller firm do with this? Not rush out, spend thirty dollars a head on subscriptions, and wait for staff to appear. The lesson isn’t the spending. It’s the sequence. The firms that grew put AI onto real work, gave it six to twelve months to bed in, and let the freed-up time turn into more business.
Go back to that brokerage. Say they point AI at the routine first pass on renewals, the sorting and the chasing that eats a broker’s week. That might free up, for the sake of example, [most of a day a week] for each broker. That figure is only an illustration to show the shape of the gain, not a result from the study. But the freed time is the real point. Used well, it lets the same team quote more, look after more clients, and take on growth that would otherwise need another hire just to stand still. A person still signs off every policy and every client call. The AI clears the routine so the team can do the work that actually wins business.
That’s what getting the strategy right looks like, and it’s where our work starts. If you want to see where AI could take routine work off your team’s plate, and where a tool like Claude fits, have a read through our AI strategy service. The study is a US headline. The chance to plan this properly is yours.
