Generative AI has set a new tempo in business. Managers no longer need to wait for a meeting to generate a market analysis. Founders don’t need to spend countless hours and resources developing a prototype.
Hard work still remains though: Leaders must decide whether an answer is right, the prototype can scale, the economics make sense, and the organization is ready to change.
At Columbia Business School, that reality shapes an expanding AI curriculum that includes more than 25 courses, with more to come. The aim is to prepare students to lead in a business environment where AI shrinks the distance between idea and implementation.
Existing courses, like “Leadership Intelligence in the AI Era,” are being adapted for the AI era. At the same time, CBS is introducing new courses — including three this fall — built around the challenges now emerging inside companies: “Vibe Coding,” “AI Solutions: From Diagnostics to Deployment,” and “AI in the Private Firm.”
The courses are at the forefront of the broader shift in business education. Tomorrow’s leaders increasingly need to learn how to question AI outputs and test whether solutions can scale.
What comes after the prompt?
Leadership Intelligence in the AI Era, co-taught by Oded Netzer, the Arthur J. Samberg Professor of Business, Christopher Frank, and Paul Magnone, has sharpened its focus since generative AI made it easier to produce analysis and recommendations.
The course posits that today’s leaders are awash in information and need to hone the judgement required to use it well. In the era of AI, the instructors argue, human judgment and decision-making are more, not less, important.
“The course is focused on the hardest thing that a leader needs to do — the most important thing, actually, that the leader needs to do — which is making decisions,” Netzer says.
That means teaching students to slow down before acting on a fast answer. Through a framework the instructors call Quantitative Intuition, students learn to ask better questions, pressure-test assumptions, recognize algorithmic and cognitive biases, and move from what the data says to what it means and what to do about it. The course includes exercises in precision framing, AI interrogation, and executive storytelling — skills meant to help leaders stay accountable for decisions even when AI is part of the team.
From idea to working product
Vibe Coding flips the script on business students who often think of themselves as nontechnical.
Taught by Professor Mattan Griffel, the course equips students to design, build, and deploy lightweight software products using AI coding tools like Lovable, Claude Code, and Cursor. It assumes no technical background. Instead, the course runs like a lab, with structured tool explorations, weekly build sprints, and a final demo day where students present working applications.
“At the end of the day, it’s really about turning people into builders,” Griffel says.
For Griffel, vibe coding is part of a longer evolution in the technical skills business students need. Excel gave managers a way to model business problems. Python helped them work with larger datasets and more complex analyses. AI-assisted coding now allows students to give computers instructions in natural language and see those instructions become functioning software.
That speed opens creative possibilities, but Griffel emphasizes that it also raises the value of technical judgment. Students need to understand what goes on behind the scenes: How to debug, test, evaluate tools, make architectural choices, deploy safely, and understand the risks attached to what they ship. A bad technical decision can compound swiftly when code is generated faster.
The course is designed to give future founders and product managers enough fluency to build quickly and collaborate more effectively with technical teams. Students are graded not only on final products, but on how they engage with AI tools to get there.
From prototype to payoff
Once students can prototype with AI, the next question is whether that prototype can work inside a real company with real data, costs, privacy constraints, and performance expectations.
That is the focus of AI Solutions: From Diagnostics to Deployment, co-taught by Omar Besbes, Vikram S. Pandit Professor of Business and Reynolds Family Professor of Digital Economy, and Amine Allouah (PhD ’19), co-founder of MyCustomAI. The course is aimed at students who expect to lead, evaluate, manage, or advise on AI initiatives. It reflects the rise of new roles at the intersection of business and AI, including forward deployed engineers and AI product or transformation leads.
Besbes describes the course as focused on “the last mile of AI.”
“We have an idea, we have a prototype,” he says. “Now what does it mean to bring it to life within the company at the scale of the company?”
A prototype can use the most powerful model available because the stakes and costs are limited. At enterprise scale, the same choice may raise concerns around token costs, inference, maintenance, data privacy, reliability, and whether a cheaper model can handle simpler tasks.
In the course, students examine how firms can use proprietary data without leaking it, how to evaluate deployment options, how to define accuracy for a specific business use case, and how to determine when additional optimization stops being worth the investment. By learning to evaluate those tradeoffs — rather than simply track the latest model release — students develop the kind of AI acumen Besbes believes will last beyond the current model cycle.
AI adoption poses a leadership test
In private and family firms, AI adoption can mean navigating legacy, identity, trust, and a family’s values.
In AI in the Private Firm, Professor Gaia Marchisio teaches students involved in the ownership, governance, or operation of family firms. The course, co-taught with Marc De Kuyper, helps students lead responsible AI adoption without compromising what makes their businesses distinctive.
“AI should be treated like a full-blown strategic element of the company,” says De Kuyper, founding partner at Praexo Management and a member of the 11th generation of De Kuyper Royal Distillers, one of the oldest family distilleries in the world.
For De Kuyper, the urgency is clear. “It’s how quickly you can adopt AI, not a question of if you adopt,” he says.
Within the course, students learn to identify and prioritize AI use cases, assess organizational and cultural readiness, design low-risk, “small t” transformations, build intergenerational buy-in, and scale responsibly through governance structures. Their capstone deliverable is a board-ready AI adoption roadmap tailored to each student’s enterprise.
Speed alone is not enough, Marchisio emphasizes. For next-generation leaders, AI may be a chance to bring something new to the business. In private and family firms, however, innovations have to be introduced with an understanding of who holds authority, what they value, and how deeply their identity may be tied to the enterprise.
“[AI adoption] could be a great opportunity,” says Marchisio, Faculty Director of the Global Family Enterprise Program. “But if [users] don’t understand the context, and if they don’t manage the process … there can be a huge, quick pushback and rejection.”
That is why the course treats adoption as a leadership challenge and not just a technical problem. Students are asked to think about how to gain credibility, earn mandate, and frame AI as a way to strengthen the enterprise while not dismissing the judgment of the previous generation.
The work AI leaves to leaders
The courses are part of a broader vision for AI education at CBS. Students need to know how to question the outputs, build with the tools, evaluate whether solutions can scale, and lead adoption across organizations with different cultures, constraints, and incentives.
Those skills mattered before generative AI too. What has changed is how quickly leaders now have to use them. A business leader may need to interrogate an AI-generated analysis, prototype a product, assess deployment costs, and persuade people to adopt a new workflow in the same strategic cycle.
That is what the School’s AI curriculum prepares students to do: decide what to trust, what to build, what can scale, and how to bring an organization into the future.
