The automation wave is not theoretical anymore. Research for ServiceNow and Pearson estimates that by 2027, about 1.3 million existing full‑time Australian roles will be automated to a significant degree, roughly 10 per cent of current jobs. Behind that headline are some very specific targets: repetitive, rules‑based work in offices, call centres, back‑office finance and retail.[6][7]
The data break it down clearly. Up to 45.8 per cent of bank‑worker roles, 38.2 per cent of accounts‑clerk roles, 36.9 per cent of checkout‑operator roles and 36.8 per cent of bookkeeper roles are flagged as highly exposed to automation. These are not sci‑fi jobs; they are everyday positions built around consistent processes: entering data, checking forms, following scripts, reconciling transactions. Large language models, workflow engines and robotic process automation can now do a lot of that faster, cheaper and without breaks.[6]
At the same time, around 6.6 per cent of jobs are expected to be heavily “augmented” by AI rather than replaced, meaning a big chunk of tasks inside those roles will be handed to machines but the job title stays. Think migration agents, network and systems administrators, hardware technicians and telecom engineers: people who will still exist, but whose day‑to‑day work includes overseeing, checking or steering automated systems instead of doing every step manually.[7][6]
For Australian workers, the distinction between “automated” and “augmented” matters less than it sounds. If nearly half of bank‑worker tasks can be done by code, banks do not need the same number of people in branches or back offices. If a supermarket can run more self‑checkouts and AI‑assisted inventory, it can cut shifts. Employers interviewed in the same research admit they are worried about job loss perceptions, but they also see clear cost‑saving and productivity gains from cutting routine labour.[6]
The story is not only white‑collar. Blue‑collar roles that mix physical work with standardised processes—warehouse picking, basic assembly, some logistics tasks—are increasingly exposed as robotics and machine vision get cheaper. The difference is that in those areas, capital cost and site‑specific complexity still slow adoption. It is easier to roll out AI in an office than to retrofit a messy warehouse or mine site, at least for now.[6]
The standard reassurance is that new roles will appear: AI specialists, data scientists, prompt engineers, people who maintain and audit the systems. That is true, but the volume is uncertain and the match with displaced workers is not automatic. A 50‑year‑old accounts clerk whose role is redesigned may not find it trivial to move into machine‑learning operations, even with “upskilling” programs on offer.[7][6]
Calling this inevitable is not alarmist, it is just reading the numbers. Almost 10 per cent of current full‑time jobs in Australia will change materially within a few years, and many of the most exposed roles are exactly the ones people have treated as safe office work. The question is not whether AI will replace tasks—that is already happening—but how bluntly businesses will convert those efficiency gains into headcount cuts, and how quickly governments and training systems can adjust.[7][6]
Sources (links)
https://www.technologydecisions.com.au/content/it-management/article/ai-to-automate-1-3m-aussie-jobs-by-2027-servicenow-89435762[6]
https://www.servicenow.com/content/dam/servicenow-assets/public/en-us/doc-type/resource-center/analyst-report/ar-building-a-successful-team-in-the-age-of-ai.pdf[7]
https://www.abc.net.au/news/2025-09-24/australia-in-dangerous-place-as-ai-adoption-ramps-up/105807430[8]