Replace Or Reinvent? B-Schools Confront Uncertainty Around AI, Jobs, And The MBA Pipeline

The MBA students and alumni came to London to talk about the future of work, though no one can predict what that will look like anymore. They filled the theater at IET London: Savoy Place, the historic headquarters of the Institution of Engineering and Technology, located on the River Thames. Somewhat ironic that the theater was named for Alan Turing, the father of artificial intelligence.

Michael Nowlis, director of MBA programs at Imperial Business School in London, started the discussion by reading a New York Times’ headline from that very morning, April 23: “Job Cuts Driven by AI Are Rising on Wall Street.”

The article outlined how JPMorgan Chase, Citi, Goldman Sachs and other bulge bracket banks – traditional entry points for many MBAs at the start of their careers – posted a combined $47 billion in first-quarter profits while cutting roughly 15,000 jobs.

“So that’s where we’re going; That’s the reality,” Nowlis said. The audience chuckled uncomfortably.

“It caught me for a moment, knowing I was coming here this evening to talk about this whole issue.”

Nowlis was moderating “Replacement or Reinvention? How AI is redefining skills, work and careers,” a panel discussion co-hosted by the EDHEC Alumni and the EDHEC Artificial Intelligence Centre. Panelists included industry leaders at the forefront of AI transformation from Microsoft, QS, Artefact and Busuu/Chegg.

While the World Economic Forum expects AI to create 78 million net new jobs globally by 2030, no one can yet say what those jobs will look like, who will be transplanted, or how companies will create value with entirely new business models. If AI makes companies magnitudes more productive, what happens to the people now doing the work?

“Replace or reinvent, I think, is a subject all of us are having conversations around,” said Michelle Sisto, Associate Dean & Director of the EDHEC AI Centre. She was also the event’s keynote speaker.

“Do we replace certain tasks, do we reinvent them? What about work flows, teams, the actual essence of our businesses?”

THE DEATH OF THE CORPORATE PYRAMID

Panelist Stephen Bennett says the structural shift is already underway. The traditional corporate pyramid is stretching into something closer to a diamond. Fewer entry-level and junior roles at the base, with a wider middle layer of expertise and people who can interpret, challenge, and act on AI-driven outputs.

“Most people think of AI as an opportunity to reduce cost and take people out. That is not actually 100% true. It’s not even 50% true,” said Bennett, a partner and leaders of agentic AI transformation at Artefact. He formerly worked at McKinsey and spent 20 years delivering digital services in the British Army.

“We as humans are going to remain important, because we’re the decision makers … So, McKinsey’s view is, ‘we will reduce the number of business analysts directly out of school, but we are not going to reduce the number of associates, we’re not going to reduce the number of managers, because we will generate more work (with AI). And therefore, as a partner at McKinsey, I will have more teams working for me, but those teams will be smaller.’”

The problem with the diamond structure is the sharp drop in junior roles where business graduates often start their careers. This is where mistakes, friction, and on-the-job learning teach the judgement and perseverance required for leaders 10 years or so down the line. If those roles disappear, the pipeline breaks.

“We might be creating a generation of future workers that never had actually to go through something difficult,” said panelist Karine Allouche Salanon, General Manager of Chegg, Inc., leading Busuu’s global language learning business, and founder of Human x AI.

That’s where it gets messy, and where companies are just starting to experiment with solutions.

At Microsoft, for example, teams have slowed down parts of the development process on purpose. They separate learning from execution while creating parallel tasks to build mastery.

“We created a second AI system that is actually providing tutoring and small tasks that you do on the side, not as part of your core job, that are still useful to create the mastery,” said panelist Luca Cassani, Head of AI, Compete & Transformation at Microsoft UK&I.

It’s an example of deliberately slowing down what AI allows you to do because it’s more beneficial long term than capturing all the value immediately.

SOFT SKILLS MAKE A COMEBACK

For all the talk about technical disruption, the skills employers want have not changed that much. If anything, they have become more important.

“In the UK, employers feel that the real skills gap between what they need and what they’re seeing in their graduates are in areas like critical thinking, communication, and resilience,” said panelist Jessica Turner, CEO of QS Quacquarelli Symonds where she navigates AI’s impact on global education. She is a former leader at Clarivate Analytics, Thomson Reuters, and McKinsey & Company.

Even in technology roles, employers want problem solving, emotional intelligence, and creativity, the kind of soft skills business schools have traditionally emphasized. AI can generate answers, but it cannot take responsibility for them.

Allouche Salanon agreed. Access to AI is no longer the differentiator. The advantage comes from how companies use the tools available to them. AI makes it easier to produce content, analyze data, and scale operations, but it does not automatically create value.

That requires judgment, learning loops, and system design. In practice, it means redesigning workflows, deciding where humans stay in the loop and defining how decisions get made. That work still belongs to people.

AI IS REWIRING HOW PEOPLE THINK

In her keynote, Sisto set the context for the discussion on how AI is already impacting how humans think. For decades, researchers have described human cognition according to Daniel Kahneman’s two-systems framework: System 1 thinking – fast, automatic, intuitive, emotional – is used for our everyday decisions. System 2 – slow, deliberate, analytical, logical — handles complex problem-solving.

Now there is a third system. In February, Wharton scholars Steven D Shaw and Gideon Nave proposed what they call “artificial cognition,” where generative AI functions as an external system people consult when making decisions.

The risk is what researchers call “cognitive surrender.” That happens when people accept AI-generated responses with little scrutiny, bypassing both instinct and deeper analysis.

In the study, participants scored higher on reasoning tests when AI gave correct answers, with accuracy jumping by 25%. But when AI produced faulty answers, performance dropped by 14% because people followed the output anyway.

“It showed very clearly that there was a default toward what the AI assistant was providing,” Sisto said. “Kahneman taught us that fast thinking feels right even when it’s wrong, and AI makes that even more irresistible.”

For students, the effects are already visible.

In a survey across six business schools, nearly one in four undergraduates said they use AI because they fear falling behind. One in five said it affects their confidence in completing work without it.

That creates a kind of feedback loop: Less confidence leads to more reliance, more reliance limits the development of real skills. “So for educators, part of the challenge is really thinking about what are the skills that we need to teach to help students get beyond this,” she said.

EDHEC is responding by embedding AI across its curriculum. Students build their own AI assistants, test and challenge them, and then reflect on how they use them.

It launched the EDHEC Artificial Intelligence Centre to drive AI strategy across its education, research, and global engagement. It is also a founding member of the Responsible AI Consortium alongside Imperial College, LUISS Business School in Rome and QS. Nine other global universities have since joined.

The challenge in AI education is pace. By the time a course launches, parts of it may already be outdated. The value is no longer just knowledge, but in context, community, and creating a culture of lifelong learners committed to the constant upskilling the digital transformation will require. Events like this one, bringing students, alumni, and industry into the same room, are a prime example.

“If you think about why do people continue to need to go to business school, it’s because you’re in an environment where you can be thinking about these things,” Turner said.

THE MILLION-DOLLAR QUESTION

Of course, the speed of the AI transition is its own disruptor.

In response, the UK has partnered with industry to launch one of the most ambitious workforce training efforts in decades. The AI Skills Boost program aims to equip 10 million workers with practical AI skills by 2030 through short-form courses delivered online. The program is free for every adult with a goal of making Britain the fastest AI adopting country in the G7. The effort reflects a broader shift away from traditional education models toward what Cassani called “a system wide workforce transformation.”

“The UK Government believes AI to be one of the growth levers for the UK economy,” said Cassani who is leading Microsoft’s commitment to train 1 million workers as a program partner. Microsoft will provide tools, training, and credentials at no cost.

“Every quarter, there is what feels to be a seismic shift in this technology,” Cassani said. “Skilling [is] a force multiplier of social mobility.”

But what about the people sitting in the audience of the Alan Turing theater? The MBAs who will soon be looking for junior roles and the alumni already in them?

One EDHEC MBA stood up to ask the million-dollar question: “How do you get to the middle of the diamond when there are no junior jobs anymore? How do you actually build experience within a company?”

Unfortunately, no one yet has a million-dollar answer.

“I don’t think I have the answer,” said Allouche Salanon. “I think we all have a piece of the answer.”

Inside Busuu/Chegg, teams are mapping current roles against future needs to identify which work stays, which disappears, and which evolves. They have introduced new layers like “architect” roles, focused on building systems, and “oversight” roles, focused on checking outputs. They are also looking at workflows to ensure employees still gain the experience they will need later.

The shift is not clean, she said. Not everyone is ready to step into higher-level work that requires judgment and design. But that is where demand is growing.

There is also a cost equation. Hiring AI-native talent can come at a premium that is 40% to 50% higher in some cases, creating pressure to train from within.

“For our company, we’re becoming, whether or not we want it, a learning institution,” she said.

Bennett agreed. While AI may reduce some entry-level tasks, it does not eliminate the need for the decision makers.

“If we’re going to generate more and more and more, we’ll still need more and more humans.”

DON’T MISS: STANFORD GSB LEADS MBA PROGRAMS IN AI COURSE OFFERINGS, WITH DARDEN & WHARTON CLOSE BEHIND and THE AI GENERATION IS NOT ATHEIST. IT IS SOMETHING MORE CONCERNING 

© Copyright 2026 Poets & Quants. All rights reserved. This article may not be republished, rewritten or otherwise distributed without written permission. To reprint or license this article or any content from Poets & Quants, please submit your request HERE.