Before joining Microsoft as a junior software engineer last year, Ume Habiba, imagined spending most of her time fixing bugs and juggling other mundane tasks.
Instead, the University of Maryland graduate was charged with building a new feature for Azure Networking, one of Microsoft’s flagship products.
Habiba’s miscalculation? Not realizing she would get to offload the kind of grunt work that early-career developers have long been saddled with to AI tools such as Microsoft’s GitHub Copilot.
“It was crazy,” said Habiba, who is 24 and lives in New York. “I totally was not expecting to do a feature right off the bat.”
The Gen Zer’s experience shows that what has long been a rite of passage for entry-level office workers — tedious but foundational assignments that teach them the ropes — is increasingly being handled by AI.
In turn, Microsoft and other employers are now entrusting junior recruits with more advanced projects than previous young cohorts were given. It’s a recalibration that could make starter jobs more appealing, but also create a steeper, less forgiving learning curve.
“AI is changing the entry-level experience for an entire generation of white-collar workers,” said Peter Cappelli, a management professor and director of the Center for Human Resources at the University of Pennsylvania’s Wharton School. “Companies really need to think through how to support these new hires.”
Limited opportunities
AI may be reshaping not only what entry-level jobs look like, but also how many companies need.
In 2025, job ads for junior positions on Indeed declined 7% from a year earlier, while posts for senior roles rose 4%. Meanwhile, the unemployment rate for recent college graduates stood at 5.7% in the first quarter, compared with 4.2% for all workers, according to the Federal Reserve Bank of New York.
The divergence reflects economic uncertainty, combined with AI’s automation capabilities, and the cost of AI itself, said Laura Ullrich, director of economic research at Indeed.
While some employers are bucking this trend, former Cisco CEO John Chambers told Business Insider he expects AI to dent overall entry-level demand in the near term, before eventually creating new categories of work. He compared the AI boom to the rise of the internet, but said AI’s transformation will occur far faster and affect more industries.
“I’m the optimist on how this turns out,” said Chambers, who now leads the venture-capital firm JC2 Ventures. “But there will be a lag.”
AI’s productivity boost
For recent college graduates who do land entry-level jobs, some employers say AI is already making them more productive and less dependent on others.
At Okta, junior auditors used to spend hours reviewing compliance documents for inconsistencies, said Rebecca Port, the identity-management company’s chief people officer. Now, an AI assistant compares those documents against “gold standard” examples and flags anomalies in real time.
Port said the change is freeing junior auditors up to focus on higher-level analysis, such as evaluating why something isn’t compliant and how it could be improved.
“We’re leveraging AI to really automate some of the simpler tasks and the more mundane elements of work,” she said.
Likewise, Microsoft’s junior sales reps can now use AI tools to practice sales pitches before meeting with prospective customers, said Katy George, corporate vice president of workforce transformation at the tech giant. Previously, they would need to ask more experienced colleagues to role-play with them.
‘A grand experiment’
Despite AI’s time-saving benefits, some observers warn that automating too much entry-level work could harm junior recruits.
For example, Chambers said repetitive tasks help people understand how things work, spot mistakes, and build intuition.
“AI cannot replace experience,” he said.
What’s more, it can take years for people to develop the savvy needed to make tough decisions under pressure, navigate office politics, and grasp other nuanced workplace dynamics, said Jennifer Tosti-Kharas, a professor of organizational behavior at Babson College.
“We have to remember it’s still their first job,” Tosti-Kharas said. “AI doesn’t make them more mature.”
Another concern she sees is generational friction. Today’s youngest workers will now have early-career experiences that their millennial and Gen X superiors can’t relate to, Tosti-Kharas said. Older workers may see them as having not paid their dues, “which could breed resentment,” she said.
Employers are essentially testing a new model for developing young talent that may or may not work, added Tosti-Kharas. “This feels like a grand experiment,” she said.
Imposter syndrome
To circumvent any potential pitfalls, some employers are rethinking how to train and develop early-stage talent.
Accounting giants EY and KPMG recently told Business Insider they are experimenting with AI-driven training for entry-level recruits and, in some cases, experienced staffers as well.
EY said it plans to use simulated audit scenarios and embedded learning tools as it rolls out AI agents across its assurance business, while KPMG is testing a simulation tool to help junior staffers, in particular, learn the ins and outs of tax prep as AI begins to take over tasks.
Still, a big jump in responsibility can be intimidating.
Habiba, the Microsoft engineer, said she was thrilled to let AI help her write code, generate unit tests, and automate other parts of the development process. Yet as she tackled more challenging aspects of the assignment, such as network infrastructure and product optimization, “feelings of impostor syndrome definitely were coming up,” she said.
Ultimately, Habiba got the job done, crediting not just AI but also several of her more experienced colleagues for helping her along the way. It’s the kind of mentorship that Microsoft is leaning into through “multi-generational teams” that pair junior and senior employees, said George.
Working closely with senior colleagues also reinforced for Habiba that technical ability alone is no longer enough. She said communication, collaboration, and interpersonal skills are becoming more important, not less.
“Anyone can code now,” said Habiba. “What else do you bring to the table?
Read the full article here
