Utilizing AI: Companies reconsider hiring entry-level, graduate, and young talent

At London-based advertising agency Catalyst, one junior employee can now handle five different client accounts.

Founder Tobias Greene said it used to require several. Now, AI tools handle the investigation, make the first pass of copy, and collect the data.

“Then someone has to come and fix it,” he said, likening the arrangement to a “reverse Mechanical Turk.” AI works behind the scenes and humans are the face we show to our clients.

The way we look at entry-level roles in advertising agencies is changing. “When AI takes away the drudgery, the drudgery becomes irrelevant,” Green told CNBC. “They don’t need to learn those skills… they need to learn how to manipulate and leverage AI.”

Artificial intelligence is becoming increasingly adept at research, drafting, coding, and administrative tasks. All of these tasks have traditionally been given to junior employees in each industry.

While this could boost productivity, experts have raised concerns about what this means for career advancement and how young workers will acquire the skills they need in the first place.

The Open University’s 2026 Business Barometer report, which surveyed 1,500 UK business leaders, found that 51% of respondents said AI is changing the way they hire. Approximately 19% say they have reduced entry-level hiring, with 42% citing AI adoption as the reason.

Lucy Beaumont, global SVP of products at talent assessment firm SHL in Surrey, UK, said some large graduate recruiters are reducing the number of people they hire as they assess whether the role is really needed.

“If organizations used to be pyramid-shaped, we’re now seeing what we call diamonds, with more people in the middle,” Beaumont told CNBC. He added that this is a short-term solution and could create some kind of crisis if employers look to the future.

“These entry-level roles are your training ground, where you get your training wheels, earn your badge, pick up your war scars, make mistakes, and learn how to operate in a corporate environment,” Beaumont added. “If you take that away, it has a significant impact on every rung of the ladder.”

However, the junior role has not completely disappeared. In a survey of 500 mid-market leaders in the UK and Ireland conducted by technology consultancy Klarus, 45% said AI is being used to help junior staff work more efficiently and faster. Meanwhile, almost a quarter said new roles and opportunities were being created.

This is especially true in software development. Sheila Flavell, COO of business and technology consulting at London-based FDM Group, said tools such as Claude Code and Codex are making developers increasingly responsible for overseeing projects completed by both AI agents and humans.

“The skill set that’s going to be very important is understanding how to work with other people,” she told CNBC. Therefore, AI could encourage younger employees to supervise their work much earlier in their careers.

However, recruiting for this new way of working can be difficult. SHL’s Beaumont said some employers were hiring graduates for a broader range of abilities, rather than hiring them for pre-defined roles, and deciding on their subsequent placement.

“We’re hiring for potential,” she says. “Do you have the potential to reskill? Can you reskill another skill? It doesn’t really matter where you come from.”

Beaumont gave the example of one large global retail bank she works with, which is looking beyond its traditional talent pool to graduates in fields such as psychology and law. Banks reason that while changing technical skills can be taught, competencies such as critical thinking are difficult to build from scratch.

For Cheney Hamilton, a director at research firm Bloor, the question is whether junior employees can develop the judgment to oversee work they never learned to do themselves.

She argues that “grunt work” was never just a job, but a mechanism by which people accumulated context and judgment.

“AI raises the bar for what juniors can produce, but it doesn’t give them the judgment to know when their output is wrong. That judgment has historically come from repetition and consequences,” Hamilton, who leads Bloor’s research on workplace transformation, added in an interview with CNBC.

However, that doesn’t mean saving menial tasks for training. Hamilton said companies could treat entry-level work as “exposure” rather than “output,” with junior employees checking and modifying AI-generated work and senior employees intentionally teaching decisions about it.

Consulting firm FDM Group is conducting a similar experiment. Flavell said that instead of just teaching people how AI tools work, they are now training people using real business problems and asking them to use agent engineering to find solutions. As a result, companies may face a shortage of experienced workers in the coming years.

“This is why I argue that the real problem is not the time given to ‘early careers,’ but ‘early knowledge,'” Bloor’s Hamilton added. “Are we intentionally building the contextual knowledge and judgment that was previously available through osmosis? If not, we risk creating a generation of AI-enabled breadth but no depth.”