Business professors at Drexel University surveyed 647 employers this year and found something that should worry every business school in the world. Companies still rank communication, professionalism, teamwork, reliability, and critical thinking as the top qualities they look for in new graduates. AI skills made the list too, but they were not what separated the candidates employers actually wanted to hire.
That finding lands at an odd moment. Entry-level jobs have quietly become the first casualty of workplace AI. A Stanford working paper tracking the job market found that employment for workers between the ages of 22 and 25 in the most AI-exposed fields fell by 6 percent since 2022, while older workers in those same fields saw their employment rise over the same period. Job boards across the UK, the US, and several EU countries show similar drops in postings for junior tech and marketing roles.
The reason is simple. AI is very good at the exact tasks that used to be handed to the newest person in the room: drafting a first version of a report, building a rough financial model, sketching out marketing ideas before a meeting. Those tasks were never really about the output. They were how a 22 year old learned to think like a professional, under the eye of someone more experienced who corrected their mistakes.
Take that work away and you take away the training ground, not just the job. A junior analyst who reviews an AI-generated forecast instead of building one from scratch never learns why certain assumptions are wrong. A junior marketer who rates AI-generated campaign ideas instead of writing the first version never develops a feel for what actually works. The skill still gets used somewhere, by someone. It just is not being built anymore.
This creates a real problem down the line, and it is not one AI can solve on its own. Senior employees became senior by making mistakes early and having someone senior correct them. If that cycle breaks now, in five or ten years there may be a shortage of experienced judgment, not because people got worse at their jobs, but because fewer people ever got the chance to practice.
Business schools are responding by teaching students to use AI tools well. That is necessary but not sufficient. The employers in this survey are not asking for AI fluency as the deciding factor. They are asking for the same human qualities they always wanted: people who communicate clearly, work well with others, show up reliably, and think critically about what they are told, including what AI tells them.
For any company that hires young people, the practical takeaway is straightforward. If AI is handling the first drafts, someone still needs to design a way for junior staff to practice judgment on real stakes, even smaller ones. Companies that figure this out first will have a deeper bench of experienced managers a decade from now. Companies that do not will eventually run short of people who know how the business actually works.