ATLANTA — (Fast Company) At McKinsey & Company, generative AI is now part of the daily workflow for analysts, a role long seen as a proving ground for next generation business leaders. Research and the creation of slide decks that once filled months of early careers can now be assembled in days.
That shift matters not because it can be done faster, but because of what it replaces. The system that once produced seasoned professionals is changing.
McKinsey isn’t alone. Recent reporting from The Wall Street Journal and The Economist shows entry-level hiring declining in white-collar fields exposed to generative AI, even as firms maintain output. Companies aren’t shrinking their ambitions. They’re shrinking the base of the staffing pyramid.
What’s breaking down isn’t just tasks. It’s the apprenticeship model. For decades, professionals built their judgment through repetition and small mistakes. What replaces it could broaden access to careers that the old grind kept out of reach. Or quietly thin out expertise when organizations need it most.
Across industries, the pattern is clear: The repetitive work that taught people how organizations function is fading. Investment banks close deals with fewer analysts. Law firms rely less on first-year associates for document review. Architecture firms no longer center junior roles around documentation work that once defined early careers.
THE COMPETENCY INVERSION
As the traditional entry-level workload shrinks, a different pattern is emerging inside firms.
In a Harvard and BCG study of nearly 760 consultants, those using ChatGPT-4 completed more tasks, finished faster, and produced higher quality work on problems within the AI’s range. But when assignments fell outside that boundary, performance declined. Junior consultants generated polished analyses yet struggled to explain where recommendations might break down.
The same concern is surfacing in design. A widely cited Adobe survey found many user experience professionals worry that newer designers in AI-integrated environments are missing core skills. Ben Anthony, design director at Dentsu Creative, said it best: “I don’t want professional prompters. I want designers. AI might help you get somewhere quicker, but if you skip the fundamentals, you’ll only ever be average.”
Senior leaders are taking notice. As the Irish Times recently reported, several major consultancies have frozen starting salaries and are reassessing graduate hiring as AI pressures the traditional pyramid model. Former Deloitte partner Mark Bunker argued in that article that as routine work fades, experienced judgment becomes more valuable, not less.
That’s the competency inversion: AI can accelerate performance early in a career but it cannot accelerate the accumulation of judgment. Organizations built on judgment cannot afford to confuse speed with depth.
HOW TO BUILD JUDGMENT
What’s changing isn’t just the pace of work. It’s how people learn in the first place.
In the old model, early-career professionals built their judgment slowly. By the time they were trusted with bigger decisions, they had already seen how things break.
That foundation is eroding. New hires are exposed to real decisions earlier, often before they understand the systems behind them. Instead of building intuition over time, they’re expected to manage outcomes almost immediately.
Managers can no longer assume an employee’s time in their role builds experience. Judgment has to be developed on purpose. Feedback matters. So does creating room for reflection in a culture that rewards speed.
Some firms are experimenting with how to do this. Spotify’s Technology Fellowship brings in self-taught programmers and bootcamp graduates and moves them quickly into hands-on work, while still emphasizing fundamentals. PwC has shifted audit training toward overseeing AI-generated work, teaching professionals how to spot bias, missing context, and answers that sound convincing but aren’t correct.
Others are adding friction back deliberately. Some teams still require manual modeling or hand drawing before digital tools take over. When you haven’t done the work yourself, it’s difficult to know when the machine is wrong.
The real question isn’t whether people can produce output. It’s whether they’re given the time to build judgment alongside it. If they aren’t, the consequences won’t be obvious at first. They will show up years later, when leadership demands instincts that were never fully formed.
ROLES ARE CHANGING
As learning changes, careers start to look different too.
The old model was simple. Learn the basics. Move up. Eventually manage the people doing the work you once did. That ladder only worked because the bottom was wide. It depended on layers of repetition before responsibility.
When that repetition disappears, the ladder changes with it.
Harvard Business Review has described organizations moving away from the classic pyramid toward something flatter and more fluid. Some professionals build and maintain AI systems. Others translate outputs into decisions that make sense in context. Senior leaders spend more time on relationships, risk, and judgment.
These roles don’t stack neatly anymore. They sit alongside each other.
Hiring patterns point in the same direction. LinkedIn’s 2026 report, Jobs on the Rise, highlights roles like AI engineers and AI consultants among the fastest growing in the U.S. Many of these positions barely existed a few years ago.
Which raises a tougher question: If careers are no longer built on mastering yesterday’s repetitive work, what qualifies someone to lead?
THE TRADITIONAL CAREER LADDER IS REDESIGNED
It’s easy to frame this shift as decline. Fewer entry-level roles. Less time spent learning the basics.
But the old system wasn’t perfect.
Those early years of repetitive work came at a cost: long hours, low pay, limited control. Many talented people never made it through, not because they lacked ability, but because they couldn’t afford the grind. That model rewarded endurance as much as talent.
AI doesn’t remove the need for judgment. It changes how people get there.
Early-career professionals now gain access to meaningful work sooner. That brings real risk but it also opens doors that were once closed.
The tradeoff is real. Depth does not build itself. Organizations will have to be deliberate about how judgment is developed. Individuals will have to be intentional about what they choose to master.
The traditional ladder isn’t vanishing. It’s being redesigned in real time as the apprenticeship model fades. What replaces it will determine whether AI strengthens professional expertise—or slowly erodes it.