According to recruitment platform Handshake, graduate job postings in the US slumped 15 percent between July 2024 and mid-April 2025 compared with a year earlier. Employers can therefore afford to be more demanding.

Claire McCartney, Senior Policy Adviser at the Chartered Institute of Personnel and Development (CIPD), the body for HR and people development, points to research showing that half of UK employers don’t think young people are ready for the world of work. Many say graduates still lack the workplace behaviours and social skills they need to succeed.

“A key challenge is that fewer young people are combining work and learning while in full-time education, reducing their ability to develop the broader workplace skills that employers value,” McCartney says.

Lucy Bisset, Senior Director at recruitment firm Robert Walters, says the graduate jobs market is caught in a paradox. Employers are cutting entry-level hiring because of rising costs, economic uncertainty and AI, while expecting graduates to arrive with more workplace skills and experience.

“Graduates can only gain those skills if organisations are able to provide the initial opportunities,” says Bisset. “As entry-level roles continue to decline, this skills gap will only widen.”

The squeeze is already apparent. According to the job site Indeed, graduate vacancies in the UK have fallen to their lowest level since 2018, with employers advertising a third fewer entry-level roles than a year ago. That makes the question for universities less about whether they should prepare students for work, and more about how they should do it.

The answer doesn’t lie in creating a bigger university careers service. It’s in embedding employability much more deeply into the degree itself.

AI makes that shift more urgent. As routine work becomes increasingly automated, the skills that make graduates valuable are changing at pace. Adzuna estimates UK job postings for graduate roles have fallen by almost two-thirds since ChatGPT launched in late 2022 — roughly twice the decline seen across entry-level jobs overall.

“Being able to use AI to automate many tasks that predominantly involve routine processes or administration means that ‘human skills’ will increasingly be seen as how high-performing professionals differentiate themselves,” says Gemma Kenyon, Director of Employability at City St George’s, University of London.

Teamwork, creativity, leadership and critical thinking, she argues, will become even more valuable.

That doesn’t mean graduates can ignore AI. Quite the opposite. “Graduates will need to be confident using AI to improve their efficiency at the routine tasks,” she says. That means understanding different AI tools, writing better prompts and, just as importantly, knowing when not to trust the answer.

“We will need to teach students critical thinking and a solid grounding in the fundamentals... to interrogate AI outputs and spot when AI has made errors,” Kenyon says.

And that means AI can no longer be taught as a specialist subject. “AI training will need to become a mainstream part of the content of all university courses, not just in STEM disciplines,” adds Kenyon.

Universities also need to blur the line between study and work. Assessments should reflect the problems graduates will face, while placements and employer projects become a much larger part of the degree. The shift has already begun: UK universities have pledged work experience for more than 1 million students amid growing concern that too many graduates still enter the workplace unprepared.

McCartney at the CIPD says universities could do more in this respect by “partnering with employers to provide work exposure and experience” while embedding it throughout degree programmes. It also means developing “soft skills such as communication, self-management and problem solving alongside core capabilities including digital skills”.

Placements are one of the best ways to bring those skills to life. Bisset at Robert Walters says they give students something employers increasingly value: real experience.

“A professional work placement not only provides students with workplace experience and practical application of their area of study, but also allows them to expand their professional networks, develop key soft skills and gain a better understanding of the career paths available,” she says.

The benefits don’t stop at graduation either. “It’s also very common for high-performing students to secure a position with their placement company after graduation,” Bisset adds. Employers gain, too.

“Providing student placements is a key opportunity to identify future talent and build a pipeline of talented young professionals before they enter the job market,” she says.

But work experience alone is not enough. If the workplace is changing faster than universities can update what they teach, employability also becomes a curriculum problem.

Keeping pace will require universities to adapt courses more quickly and strengthen links with employers so curricula evolve alongside the workplace rather than behind it.

The QS World Future Skills Index 2027 suggests many institutions are struggling to keep up, with employer demand for human, AI, digital and green skills changing faster than higher education can respond. Human skills, in particular, remain among the hardest capabilities for employers to find.

City St George’s Kenyon argues AI literacy and ethics should become part of every degree, with teaching built around the kinds of problems graduates will actually face at work. Some universities are already taking that approach.

At Canada’s University of Waterloo, AI is used to track changing employer demand, identify emerging skills gaps and help students understand which capabilities employers are looking for. Those same insights can then feed back into curriculum design, helping courses keep pace with a labour market that is changing very quickly indeed.

This all points to a bigger shift. In many ways, higher education is entering a new era of accountability.

Students are paying more and entering a tougher job market, and competing in workplaces being upended by AI. In return, graduates increasingly expect universities to help bridge the gap between education and employment.

That means making employability part of the degree rather than something students discover in the careers office during their final year; bringing employers into the curriculum; giving students meaningful time in the workplace; and ensuring every graduate knows how to use AI, question it and apply the human judgement it still can’t replace.

Universities won’t necessarily create more graduate jobs. But they can give graduates a better chance of competing for the ones that exist.

For as the cost of university continues to rise, students won’t just pay for a degree. They’ll expect a career.