Commentary: The MBA Is Splitting In Two. Here’s Which Side Your School Is On

The MBA is splitting in two. Not by ranking. Not by brand. By whether schools are structurally redesigning how they teach – or simply adding AI to an unchanged core.

During my years running the MIT Startup Exchange, connecting hundreds of corporations with over 1,000 startups, I heard the same complaint from company after company: new hires were excellent analysts but struggled to reason across functions under time pressure. They could build a model. They couldn’t anticipate what happened when that model met the supply chain, the regulatory environment, and the customer simultaneously.

The issue wasn’t intelligence. It was training.

We had prepared them for a slower, more compartmentalized world than the one they entered. For that reason alone, companies increasingly seek out schools looking for startup founders not for MBA-students as their source of top talent.

THE INNOVATORS ARE MOVING FROM STATIC ANALYSIS TO DYNAMIC SIMULATION

Some schools have responded structurally. At Columbia, Professor Dan Wang’s CAiSEY platform – a voice-to-voice AI discussion partner – now reaches 3,000 students across eight business schools. Students use it to stress-test arguments before class, and Wang found something unexpected: students with dyslexia and those who speak English as a second language reported it was the first educational tool that truly accommodated their learning style. This isn’t adding AI. It’s rethinking how preparation happens.

At Yale, the School of Management’s “raw cases” replace the traditional 20-page written packet with multimedia environments – video interviews, contradictory documents, financial spreadsheets, and competing stakeholder perspectives. Students must prioritize what to examine. Cases are co-taught across disciplines. The result: an approximation of how information actually arrives in the real world, not sanitized and neatly organized.

At UVA Darden, CAiSEY is now required in the core strategy course. Using AI is not only encouraged, it’s part of the requirement. The school’s LaCross Institute for Ethical AI is embedding responsible AI thinking across the curriculum – not in one elective, but systemically.

MIT Sloan’s System Dynamics group – founded by Jay Forrester in 1956–has taught cross-functional reasoning longer than anyone. The Beer Game, a supply-chain simulation that demonstrates how local decisions create system-wide chaos, remains a rite of passage for incoming MBAs. Students participating in MIT’s Action Learning labs have completed over 2,500 real-world projects with partner organizations ranging from startups to global NGOs. The motto, mens et manus – mind and hand – reflects what most MBA curricula have forgotten: analysis without application isn’t education.

THE LAGGARDS

Then there are schools where the vast majority of instruction still follows a model developed in the 1920s. Students read a 20-page written case, discuss it in a Socratic seminar, and move on. The method is refined. It is also structurally unchanged.

These schools aren’t failing. They continue to attract talented applicants and place graduates into prestigious firms. But they are training students for a decision environment that no longer exists – one where information arrives neatly packaged, where decisions unfold over weeks, and where the person with the best-argued memo wins.

Andrew Hoffman recently argued in Poets&Quants that business schools must blow up the curriculum around character and purpose. I’m making the companion argument: they must also blow it up around tempo and consequence literacy. Character without speed is noble but unemployable. Speed without character is dangerous.

THE REAL GAP

The divide isn’t whether a school has an “AI lab” or an “innovation center.” It’s whether AI and systems thinking are embedded in the core – in strategy, operations, finance, and organizational behavior – or relegated to electives that self-selected students take.

Most schools have chosen the additive path. A new workshop here. A certificate program there. An AI-focused elective that 12 percent of students elect to take. This is understandable: structural change requires faculty consensus, curriculum committee approvals, and the willingness to retire courses that tenured professors built their careers on.

But additive change isn’t keeping pace with workplace reality. In conversations with executives across private equity, strategy consulting, manufacturing, and tech, three capability gaps keep surfacing in early-career hires:

• Systems awareness: Seeing how decisions reverberate across functions.
• Consequence literacy: Anticipating second- and third-order effects.
• Human-AI judgment: Knowing when to trust or override algorithmic output.

These aren’t boutique skills. They show up in deal evaluation, pricing, budgeting, and strategic planning – the daily work MBAs actually do. One tech CEO put it bluntly: “Our MBAs come in excellent at analysis and weak at connecting consequences.” Well, depending on where those MBAs are from.

STUDENTS SEE IT FIRST

But the gap isn’t just visible in the boardroom. It’s visible in the classroom.

My daughter recently called me three weeks into her MBA program. She had spent hours on a pricing strategy case – mapping assumptions, building scenarios, drafting a memo. Then she asked an AI the same question.

“It gave me twelve scenarios in four seconds,” she said. “So why am I doing all this manually?”

She wasn’t criticizing rigor. She was diagnosing a mismatch between classroom tempo and workplace reality. When we ask students to produce three-page analyses over multiple days while AI generates 90 percent of the same structure instantly, they draw the obvious conclusion: the assignment is teaching format, not judgment.

THE REFORMS THAT MATTER

Structural innovation doesn’t require new buildings or another AI elective. It requires three shifts:

1. Un-silo the core. Strategy, operations, finance, and OB should all contain interaction-mapping components. Every case discussion should include a simple systems diagram showing how choices propagate. Columbia and Darden are doing this. Others should follow.

2. Teach AI as a co-decision maker. Students need practice verifying outputs, checking boundary conditions, and choosing when not to use AI. Wharton’s AI major – which launched Fall 2025 with required ethics courses – is one model. But this should be in the core, not a specialized track.

3. Shift assessments toward consequence literacy. Evaluate who can anticipate second- and third-order effects, not just who recommends the most elegant solution. Require a “downstream impact” section in written analyses. When employers say they want “strategic thinkers,” this is what they mean.

THE WINDOW IS CLOSING

Programs that redesign their core now can credibly tell applicants: “We prepare you for real-world decision cycles.” Programs that don’t will continue signaling – implicitly or explicitly – that they train students for a slower, more predictable economy than the one graduates will actually enter.

The Columbia story is not just about one school. It’s about what happens when institutions treat AI as a catalyst for pedagogical redesign rather than a threat to be managed.

The schools that are redesigning structurally will produce graduates who can operate at the tempo employers now require. The schools still adding electives will produce graduates who can talk about AI.

That’s the gap rankings don’t yet capture – but employers already see.

---

Trond Undheim is the author of The Platinum Workforce: How to Train and Hire for the 21st Century’s Industrial Transitions (Anthem Press, November 2025). He is a former research scholar at Stanford and MIT Sloan lecturer, and the former director of the MIT Startup Exchange. He is an advisor to Answerr, an edtech startup that enables trusted AI-based learning.

© 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.