Is This The AI Tutor That Business Schools Have Been Waiting For?

When ChatGPT burst into the B-school classroom in late 2022, many in graduate business education responded with skepticism — or concern.

At the Indian School of Business, Professor Sanjay Kallapur saw something different: a long-awaited opportunity.

“Once it was clear that students could use GPT to generate written assignments — and plagiarism detectors couldn’t catch it — I realized the traditional approach wasn’t going to work anymore,” Kallapur says. “But instead of fighting the technology, I thought we could use it to our advantage.”

A GPT THAT QUIZZES, RESPONDS & GRADES — INSTANTLY

Kallapur, who teaches Corporate Governance and Business Ethics at ISB, had struggled with the challenge of making students engage deeply with difficult material. In previous semesters, he required written submissions — but in 2022 and 2023, AI made that unreliable. He then shifted to oral group presentations to reduce copy-pasting, but the workload of grading those became unsustainable.

That’s when he began developing a custom GPT solution, built entirely by himself, that could simulate personal tutoring at scale.

Kallapur’s custom-built web application, which interfaces with OpenAI’s GPT models, provides students with dynamic, personalized feedback based on course readings. For example, in one assignment built around Milton Friedman’s seminal essay, The Social Responsibility of Business Is to Increase Its Profits, each student is prompted with a unique set of three questions generated on the spot: one factual, one conceptual, and one asking students to apply Friedman’s ideas to a modern context.

The system then asks students to submit a response to each question — no copy-pasting allowed. For each answer, the GPT gives targeted feedback: what’s working, what’s missing, and how to improve. After students revise their submissions, the model generates a grade and overall assessment on the spot.

OVER 1,000 ASSIGNMENTS, AND ALMOST NO COMPLAINTS

“It functions like a private tutor,” Kallapur explains. “It gives you hints, it nudges you forward, and it gives immediate feedback — which is key to learning any skill, especially soft skills.”

In the last two years, more than 1,000 assignments have been completed through Kallapur’s app across four annual sections of his MBA course. Just three students, he says, have raised concerns about the GPT-generated grading — and even those were minor.

That reliability is especially impressive given that Kallapur — by his own admission not a computer science expert — coded the entire platform himself. He used GPT models to help write the code, drawing on prior experience from building a much earlier web application during his faculty years at Purdue University in the early 2000s.

“This was very different technology from what I’d used back then,” he says with a laugh. “But the GPTs helped me every step of the way. In a sense, the system helped build itself.”

SCALING THE OXFORD MODEL

Kallapur points to a foundational idea in educational psychology: the “two sigma” effect, first documented by Benjamin Bloom in the 1980s. Students who receive personal tutoring often perform two standard deviations above their peers in traditional classrooms. But the problem has always been scalability — until now.

“In business schools, we’ve always talked about the importance of judgment, communication, and critical thinking,” Kallapur says. “But we haven’t had a way to truly teach those at scale. Now we do.”

Unlike traditional adaptive learning systems, which often rely on fixed multiple-choice assessments, Kallapur’s tool works with open-ended, qualitative inputs. It evaluates soft skills like argumentation and ethical reasoning — skills that have become more important as AI automates hard skills like financial modeling and present value calculations.

A NEW TOOL, NOT A TEACHER REPLACEMENT

Kallapur is careful to draw boundaries: AI tutors like his aren’t a replacement for human instruction. Rather, they free up instructors to focus on richer class discussions and ensure students arrive better prepared.

“This system lets me make sure students have not just read the material but also engaged with it — before class even begins,” he says. “That makes the whole classroom experience more valuable.”

His advice to fellow educators: don’t fear the technology. “The opportunity is here. The need is real. And the tools are finally ready.”

DON’T MISS AT THE INDIAN SCHOOL OF BUSINESS, ONE PROF IS USING AI & VR TO BUILD SMARTER, MORE STRATEGIC MBAs