The debate about whether AI will replace human workers has dominated boardrooms for years. A new report from Boston Consulting Group’s Henderson Institute cuts through the noise with hard data, and the answer is more nuanced than either side of the argument typically admits.
BCG’s analysis of approximately 165 million US jobs, distributed across roughly 1,500 distinct roles, found that 50 to 55 percent of positions will be substantially changed by AI over the next two to three years. But “changed” is not the same as “eliminated.” Only 10 to 15 percent of US jobs, somewhere between 16 and 25 million positions, could disappear within five years.
The distinction matters enormously for how business leaders plan, hire, and invest.
Reshaping, Not Replacing
The BCG report introduces a framework that goes beyond the blunt question of “will AI take this job?” Instead, it asks three things about every role: how much of the work AI can do, whether AI is more likely to substitute for workers or augment them, and whether productivity gains would expand demand or simply reduce headcount.
This framework produces a more honest picture. A customer service manager won’t be replaced by AI, but their job will look fundamentally different. They will spend less time on routine queries handled by voice AI and chat agents, and more time on complex escalations, relationship management, and strategic process design. The job survives, but someone who doesn’t adapt won’t.
Contrast that with a data entry clerk or a first-level report writer. There, AI genuinely substitutes for the human, and demand shrinks as a result.
Entry-Level Roles Carry the Most Risk
The BCG report’s most striking finding is its warning about junior workers. Entry-level and early-career roles make up a disproportionate share of the positions most vulnerable to substitution.
The reason is structural: repetitive, rule-based work is precisely what AI does best. Many “learn as you go” roles, the kind that used to give new graduates a foothold in an industry, are shrinking because companies are eliminating the repetitive tasks that created those roles in the first place.
This creates a genuine pipeline problem for organisations. You can cut junior headcount in the short term and look productive on a spreadsheet. But the institutional knowledge, judgment, and client relationships that senior employees carry were built through years of doing those entry-level tasks. Skip that apprenticeship layer, and you hollow out your own bench.
BCG’s recommendation is unambiguous: the companies that get this right will reassign roles, not eliminate them. Those that cut beyond AI’s actual ability to substitute will see productivity drop, institutional knowledge erode, and critical talent leave.
What This Means for Business Leaders
The Skills Gap Is the Real Crisis
For most organisations, the challenge isn’t AI adoption. It’s skills transformation. If 55 percent of jobs are being reshaped, then more than half your workforce needs to learn new ways of working. That doesn’t happen by distributing software licenses and hoping for the best.
Companies that treat upskilling as optional are building a fragile foundation. The employees who understand how to work alongside AI, who can prompt it, critique its outputs, apply data judgment, and know when to override it, will become disproportionately valuable. Everyone else becomes a liability.
This is not a new idea. It is, however, one that is now urgent in a way it wasn’t 18 months ago.
The “Reshape vs. Replace” Decision Is a Strategic Choice
The distinction between a role being reshaped versus replaced is not determined purely by technology. It’s also determined by leadership choices. Two companies with identical workforces and identical AI tools can end up in very different places depending on how they deploy those tools.
One might automate the junior analyst function, lay off five people, and save money in Q2. Another might retrain those five analysts to focus on higher-value data interpretation, strategic reporting, and AI output auditing, roles that AI genuinely augments rather than replaces.
The first approach optimises for the short term. The second builds a more capable organisation.
Senior leaders who think their job is simply to “implement AI” are missing the bigger question: what kind of workforce do you want to emerge on the other side of this transition?
Advisory and Training Need to Run Simultaneously
One pattern BCG highlights is the danger of organisations adopting AI faster than their people can adapt. This creates internal tension, a disconnect between what tools can do and what employees actually know how to do with them.
Getting this right requires two things happening in parallel: a clear AI strategy from leadership (what we’re automating, what we’re augmenting, what that means for roles) and genuine investment in skills development for the people affected.
Neither works without the other. An AI strategy without training produces tools that sit unused or get misused. Training without strategy produces capable individuals who can’t apply their skills within an incoherent organisational direction.
The Opportunity Inside the Challenge
The BCG report is ultimately a message of conditional optimism. AI eliminates far fewer jobs than it reshapes, and most of the reshaping creates opportunity rather than loss, for those who prepare.
The organisations that build data-literate teams, invest in AI upskilling, and develop clear frameworks for human-AI collaboration will find themselves with a genuine competitive advantage. Those that treat AI adoption as a cost-cutting exercise will find themselves managing the fallout when institutional knowledge disappears and productivity gains don’t materialise.
The window to make those choices well is not unlimited. With 50 to 55 percent of jobs in motion over the next two to three years, the organisations that delay are increasingly operating on borrowed time.
Enterprise DNA offers two paths for organisations navigating this transition. EDNA Learn provides structured AI and data skills training for entire teams, from data literacy fundamentals to advanced AI tooling. For senior leaders who need a strategic framework before they can act, Omni Advisory provides a fractional AI advisor who helps define the roadmap, evaluate tools, and manage the organisational change that comes with serious AI adoption.