When Poets&Quants reported on the AI Business School Index this August, the headline number was hard to miss: 42% of the 60 schools assessed still require no AI coursework at all. It is a striking figure, and it deserves the attention it received. But I want to argue that it describes only the visible half of the problem. Even where AI is now on the syllabus, much of what we teach is AI strategy. What our graduates will be paid for is AI judgment.
The difference matters. Strategy is about where AI could create value: which processes to automate, which products to rebuild, which markets open up. Judgment is about whether to approve this specific bet, now, with this data, this vendor, these people and this level of risk, and about knowing when to stop. Strategy can be learned from frameworks and cases. Judgment is formed by making decisions under pressure, getting feedback and living with the consequences.
GRADUATES CAN DESCRIBE AI. THEY ARE NOT YET READY TO GOVERN IT.
In my Executive MBA classrooms at IPADE in Mexico City, and in the programs I teach in Europe and Latin America, I see the same pattern. Participants arrive fluent. They can explain large language models, list use cases and quote adoption figures. Many have already run a pilot. Yet when we put them in front of a realistic board decision, such as a mid-sized company asked to approve a seven-figure AI investment with uncertain returns, the conversation quickly reveals the gap.
They ask what the technology can do. Fewer ask who will be accountable when it gets something wrong, what data the model relies on and who owns it, how the decision will be reversed if it fails, or what the investment means for the people whose work will change. Those are not technical questions. They are governance questions, and they are exactly the ones directors, CEOs and owners of family businesses are now being asked to answer.
This is not a failure of our students. It reflects how we have added AI to the curriculum: as content. A new elective, a module on generative tools, a guest speaker from a tech company. All useful. None of it, on its own, builds the capacity to say yes, no or not yet to a specific AI decision, and to defend that answer in front of a board.
WHAT BOARD-LEVEL AI OVERSIGHT LOOKS LIKE INSIDE AN EXECUTIVE MBA
Executive MBA participants are the right place to start. Many already sit on boards or management committees, or soon will. They do not need to learn to code. They need to practice the decisions that will land on their desks.
In my own teaching I use a simple discipline I call FLASH-Board: a short set of questions any board should ask before approving an AI initiative. Is the problem worth solving, and is AI the right tool for it? What does the organization actually own, in data, processes and talent, that makes this bet defensible? Who is accountable for the outcomes, including the ones nobody intended? What will we measure in 90 days to decide whether to scale, fix or stop? And how does this decision affect the trust of employees, customers and regulators?
The questions themselves are not the point. The practice is. When participants role-play a board meeting, defend a proposal against a skeptical chair, or vote against a project they championed an hour earlier, something shifts. They stop treating AI as a topic and start treating it as a decision. That is the muscle executives need, and business schools are well placed to build it, if we design for it.
THREE CHANGES SCHOOLS CAN MAKE BEFORE THE 2027 INTAKE
None of this requires a new degree or a large budget. Three changes can be made within a single academic year.
First, move AI out of the elective and into core decisions. Every core course already contains decisions where AI now matters: pricing in marketing, forecasting in operations, credit risk in finance, hiring in organizational behavior. Ask each core faculty member to add one session in which participants must decide whether and how to use AI in that decision, and justify it. This reaches every student, not only the enthusiasts who pick an AI elective, and it starts closing the 42% gap without waiting for a new course to be approved.
Second, assess judgment, not knowledge. We still measure AI learning with essays and reports that, ironically, AI can now write. Replace some of them with decision simulations and oral defenses: a mock board in which participants must approve, reject or redesign an AI proposal in front of peers and faculty acting as directors. Judgment is much harder to fake in a room than on a page, and it is far closer to the work graduates will actually do.
Third, bring the board into the classroom. Invite directors, family business owners and regulators, not only technologists, to bring live AI decisions they are facing. Executives learn fastest from real dilemmas with real stakes. In markets like Latin America, where family-owned firms dominate the economy and decisions often rest with a small group of owners, this is especially valuable: the person approving the AI investment is often the same person who will live with its consequences for a generation.
THE WINDOW IS SHORT
Business schools have spent the last few years asking how much AI to teach. The Index shows that many have still not answered that question. But the schools that will matter most in the next decade will be the ones asking a different one: are our graduates ready to decide about AI, not just to describe it?
Companies are already making these decisions every week, with or without our graduates in the room. The 2027 intake is our next chance to make sure they are there, and ready.
Javier González Núñez is Professor of General Management and Executive MBA Deputy Director at IPADE Business School in Mexico City. He was named one of Poets&Quants’ 40 Best MBA Professors Under 40 in the world in 2024, is a visiting professor at multiple Business Schools in the world, and has taught in more than 47 countries. He advises organizations and governments on strategy and innovation, and writes columns for Forbes México and Alto Nivel.
© 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.
