The conversation has shifted. It’s no longer if AI will replace jobs, but how many, how fast, and whose job is first on the chopping block. And while everyone nods along to the promise of new, exciting roles emerging from the ashes, very few are talking about the brutal transition period we are just beginning to enter.
The Elephant in the Server Room
So, how big is this problem? According to a recent and rather eye-opening survey by the consultancy BearingPoint, it’s already here. They polled over 1,000 global executives, and the results are a splash of cold water. As reported by The Register, around half of these business leaders admitted they already have a workforce “overcapacity” of between 10-19%. Let that sink in. They are essentially saying that one or two people in every team of ten are already redundant because of AI-driven tools.
The report gets even more bracing. Looking ahead to 2028, a staggering 45% of companies expect to have a surplus workforce of between 30% and 50%. This isn’t a gentle reshaping of roles; it’s a seismic shift. Alfred Obereder, a partner at BearingPoint, puts it bluntly: “Roles centered on routine analysis, process execution… are becoming increasingly redundant”.
And yet, in the other corner, we have a study from Yale University which claims to have found no “discernible disruption” to US jobs from AI so far. So, what’s going on? Are the executives at BearingPoint’s surveyed companies simply sabre-rattling, using AI as a convenient excuse for future cuts? Or is the Yale study looking in the rear-view mirror while the BearingPoint survey is looking at the hairpin turn just ahead? The truth, as ever, is probably wedged somewhere in between. The Yale data is historical, reflecting a world before generative AI truly hit its stride. The executive sentiment is forward-looking, a statement of intent.
First on the Chopping Block: Legacy IT Roles
When a storm is brewing, you can tell which trees will fall first. In the corporate world, the creaking, older trees are the Legacy IT Roles. These are the jobs that involve maintaining old systems, running routine diagnostic scripts, managing databases with predictable queries, and executing manual software testing. For years, these roles have been the dependable, if unglamorous, backbone of many IT departments.
Now, the Automation Impact is turning that backbone into a liability. Think of it like this: for decades, companies employed skilled mechanics to keep their fleet of petrol cars running. Then, the electric vehicle arrived. Suddenly, skills in oil changes and spark plug replacements became less valuable than expertise in battery management systems and software diagnostics. The old mechanics aren’t useless overnight, but their career path now has a very clear expiration date unless they retrain.
This is precisely what’s happening in IT. AI-powered tools can now monitor systems, predict failures, write and execute tests, and even manage database queries far more efficiently than a human. This doesn’t just improve efficiency; it fundamentally erodes the business case for employing large teams dedicated to these tasks.
The Great Skills Chasm
This leads us to the inevitable and much-discussed Skills Gap Analysis. The problem isn’t a simple lack of jobs; it’s a colossal mismatch between the skills people have and the skills companies now need. The demand for prompt engineers, AI ethicists, and machine learning operations (MLOps) specialists is soaring, but these are not roles you can simply parachute a legacy database administrator into.
The challenge is twofold:
– For employees: The onus is on them to engage in relentless, lifelong learning. The skills that secured a job five years ago might be obsolete in five years’ time. Waiting for your employer to provide a neat little training course is a losing strategy.
– For businesses: Smart organisations will see this not as a chance to simply cut costs, but as an opportunity to invest in their people. Reskilling programmes aren’t just a “nice-to-have”; they’re a strategic necessity to avoid a brain drain and retain institutional knowledge. But let’s be realistic, many won’t. The lure of a leaner payroll is often too strong.
The Corporate Playbook in Action
We’re already seeing this play out. Magic Circle law firm Clifford Chance announced it was trimming its London business professional staff by around 10%, a move happening alongside its adoption of AI tools. They might not draw a direct line between the two in their press release, but it’s not hard to connect the dots.
Then you have giants like Amazon, where reports have surfaced of managers being encouraged to find AI replacements to justify new hires. This is the AI Workforce Reduction strategy happening at a granular, team-by-team level. It’s a classic corporate move: frame automation as a tool that “frees up” employees for “higher-value work,” while simultaneously reducing headcount and claiming it’s all in the name of efficiency.
So, Are We All Redundant?
It’s easy to slide into fatalism, but the future isn’t necessarily a binary choice between human or machine. The most successful organisations will be those that redesign work around human-AI collaboration. The AI handles the grunt work—the data crunching, the first-draft writing, the code testing—while the human provides the strategic oversight, the creative spark, the ethical judgment, and the final sign-off.
This requires a fundamental rethink of what a “job” even is. Instead of a fixed list of tasks, roles will become more fluid, centred on solving problems using a suite of intelligent tools. A marketing manager won’t spend their day pulling performance metrics; they’ll ask an AI to analyse the data and then use their human expertise to craft a truly innovative campaign strategy based on those insights.
The transition, however, will be messy. Companies that handle it well will invest heavily in change management and reskilling, creating a culture where AI is seen as a co-pilot, not a replacement. Those that handle it poorly will see a collapse in morale, a loss of valuable expertise, and a workforce divided between the “AI-haves” and the “AI-have-nots”.
The data from the BearingPoint survey, as cited in The Register’s article, is a warning shot. It’s a declaration from the C-suite that the age of workforce bloat is over. The question now is whether this leads to a smarter, more productive human-AI partnership, or simply a leaner, more brutal form of capitalism.
What do you think? Is your role safe, or are you already looking at ways to outsmart the algorithm?
