October 8, 2026

Is AI Coming for Your Job? History Might Have The Answer

Every time a major new technology arrives, we seem to ask the same question: what happens to the people whose work it can suddenly do? AI is no different.

Dorian Trevisan

Depending on what you read, it is either about to make us dramatically more productive or suddenly make half the workforce redundant.

The reality is likely to be a little less dramatic, and much more interesting.

History gives us some useful clues. 

When Barclays introduced one of the world's first ATMs in north London in 1967, customers could withdraw cash directly, without speaking to a bank teller. A simple process: put into the machine a special voucher, receive your money, and off you went.

It wasn't difficult to imagine where this was heading. If a machine could perform one of the teller's core tasks, surely banks would eventually need far fewer tellers?

Except technology rarely changes businesses quite that neatly.

ATMs certainly changed banking, but they didn't simply make the people behind the counter disappear. They changed the work those people did and, importantly, changed what banks themselves were capable of doing.

We're now asking remarkably similar questions about AI. But for business leaders, I think there's a much more useful question than "which jobs can AI replace?".

How do our business systems evolve when technology becomes not just capable of doing more, but doing so with discernment and autonomy?

A job is a set of tasks, not a single activity

Think about what you or your employees actually do during an average working day.

There are probably dozens of individual tasks involved. Searching for information, updating a spreadsheet, writing emails, preparing reports, checking data, chasing approvals, moving information between systems and occasionally getting your hands dirty to build, create or deliver something. Some of those activities require experience, judgement and context. Others simply need to get done.

An interesting example comes from aviation. Modern aircraft have spent decades becoming increasingly automated. Autopilot, fly-by-wire systems and automatic landing technology can now handle significant portions of a flight. Yet we still have pilots.

The technology didn't necessarily remove the role. It changed which parts of that role required a person.

The same thinking is useful when looking at AI inside a business. Rather than starting with roles and asking which ones could disappear, start with the work happening inside those roles.

Where are people spending time on repetitive activities that require very little judgement? Where is information still being manually copied between systems? Where are people doing something simply because "that's how we've always done it"?

Those are all the typical places the experts are looking for AI and automation opportunitie, easy low-hanging fruits that make their efforts look worth the price tag.

Automation creates capacity, not just cut costs

One of the more interesting possibilities with AI receives much less attention than job replacement.

What if the real opportunity is being able to do things your business simply couldn't justify doing before?

Most organisations have a backlog of good ideas sitting somewhere. Improvements customers would value. Services they'd like to personalise. Internal processes they'd love to tidy up. Experiments they'd like to run.

They often don't happen because the time, cost or complexity involved makes them difficult to justify. AI can change that equation.

It might make it economical to offer something customers couldn't previously access. It might remove enough administration from an existing service to make it viable at scale. Or it might allow you to test an idea properly before committing significant time and money to it.

That's why I think one of the most useful questions around AI isn't simply "what can AI replace?" but "what does AI make possible?"

‍

The technology […] changed which parts of that role required a person.

Dorian Trevisan

Dorian Trevisan

What about your Frankenstein systems?

There is a catch, of course.

Adding AI to a messy business doesn't suddenly give you beautifully designed systems.

If responsibilities aren't clear, processes change depending on who is doing them and information is scattered across six platforms, three spreadsheets and Karen's mysterious folder called "FINAL FINAL USE THIS ONE", AI isn't going to solve the underlying problem.

This is where businesses need to be careful with the current enthusiasm around AI tools. According to McKinsey's 2025 global survey, 88% of respondents said their organisations were regularly using AI in at least one business function, yet only around one-third said their companies had begun scaling AI programmes across the organisation.

Experimenting with tools is relatively easy. Building AI into the way a business actually operates is much harder because it requires us to understand the processes, people and systems surrounding the technology.

Before reaching for another AI platform, it is worth asking what outcome you're actually trying to improve. Where is the friction in the current process? Which parts could technology reasonably handle? And where does human judgement still genuinely add value?

That is the difference between adopting AI because everyone else seems to be doing it and deliberately designing AI-enabled business systems.‍

The alternative

We've seen this pattern plenty of times before.

Spreadsheets changed accounting. Word processors changed writing. Digital editing transformed filmmaking. Nail guns changed carpentry. In each case, technology removed or simplified parts of the work while also changing what people could accomplish.

AI is likely to do the same, although probably on a much broader scale.

That doesn't mean every role will remain untouched. Some tasks will disappear. Roles will evolve. Expectations will change and, yes, some jobs may eventually become unnecessary as we currently see and experience them.

But there is an important difference between automating parts of someone's work and removing the person from the system entirely.

For growing businesses, that's where the opportunity becomes interesting.

If technology can take care of more of the repetitive grind, your people can spend more time on judgement, relationships, creativity, problem-solving and all the delightfully complicated things humans tend to be rather good at.

The goal shouldn't be to squeeze people out of every process simply because technology makes it possible. It should be to build customer-centric technology and business systems where people and technology are each doing the work they're best suited to.

A useful question to take back to your team

Looking back at what automation took away from our ancestors, we can all agree those are all tasks we are glad we no longer need to do ourselves.

Is your team regularly complaining about an annoying task no one loves doing?

Maybe it's repetitive. Maybe it involves too much administration. Maybe it requires information to bounce between three different systems before somebody can make a decision.

Then resist asking "can AI do this job?" and instead ask "which parts of this system should a person still be doing?".

That small change in perspective can lead you towards a very different conversation about AI, one focused less on replacement and more on designing better ways of working.

If you're trying to work out where AI genuinely belongs in your business, The AI Compass is a useful place to start. It's designed to help you assess AI opportunities deliberately, rather than collecting shiny new tools and hoping they somehow turn themselves into a strategy.

The technology will keep changing, but the goal is always the same: better systems, better work and, ultimately, people delivering on your business’ purpose in better ways.

About the Author

Dorian is an expert software advisor with a development background that provides a detailed and comprehensive understanding of systems and processes.

Dorian Trevisan

Dorian is an expert software advisor with a development background that provides a detailed and comprehensive understanding of systems and processes.

Email

Keen to continue the conversation?

Further Reading