Teaching Operations in the Age of AI: How I’ve Embraced New Tools to Teach for an AI World

A mother sits in the car in the hospital parking lot, moments after learning that her child has cancer. Overwhelmed by fear and uncertainty, her mind races to one question — how will she afford treatment? With nowhere else to turn, she opens GoFundMe on her phone, unsure where to begin. An intelligent agent immediately guides her step by step, helping her create a campaign in minutes. Here, AI has demonstrated its positive side despite the worries of many.

The above scenario is presented by Darden alum Arnie Katz (MBA ’09), GoFundMe’s Chief Product and Technology Officer, as part a case I cowrote last year, “GoFundMe Inc: AI Opportunities.”

Katz introduces the example to illustrate the human element of the company’s proposed “Create Funnel,” using disguised data to examine the time as well as the drop-off rate for each step. I use the case to teach process analysis, just one of the cases I present in my new Digital Operations Strategy elective, a deep dive into learning both with and about artificial intelligence.

GenAI is positioned to transform how the University of Virginia Darden School of Business teaches, not merely by offering students a new tool, but also by forcing us to rethink the very underpinning of the School’s teaching: the case method and Socratic dialogue.

The method has always taught students what to think about, not what to think. But when students can prompt AI to build a queuing model, parse a spending cube, or draft a strategic recommendation in minutes, the question becomes: what, exactly, should we be teaching — and how?

My class explores an answer to that question, built around cases set in digitally-native as well as traditional businesses that are themselves experimenting with GenAI. At the same time, I’ve employed the open use of GenAI tools and a “live case” philosophy that pushes the bounds of student learning. I believe it offers a practical model for evolving the case method without abandoning what makes it powerful.

Classic Concepts in New Contexts

In this new digital world, the foundational tools of operations remain the same — we still explore essential concepts such as process analysis, queuing theory, experience curves and scale economics. Digital Operations Strategy, however, embeds those concepts in businesses in the process of technological transformation.

Another new case, “Shiftsmart: Opportunity for Fractional Labor,” features a crowdsourcing platform with origins in helping charities schedule volunteers.

Ultimately, the platform migrated to the commercial world, while still retaining its laudable business mission to treat gig workers as more than mere cheap labor to be deployed to meet uncertain demand. Instead, Shiftsmart aspires to help frontline workers piece together multiple shifts in a flexible schedule that fits their lifestyle, at the same time rewarding them for dependably completing defined tasks without supervision.

The case pushes students to employ queuing theory to ensure full-time employees cover cash registers, while also leveraging Shiftsmart members to cover other tasks such as bathroom cleaning and cooler stocking. But rather than an Excel spreadsheet model and a desktop word processor, students use GenAI agents to conduct desired analysis and help write a stakeholder memo.

Authorize GenAI — Then Raise the Bar

I write and teach cases and technical notes about artificial intelligence and other new digital tools such as AR/VR and humanoid robots. And I encourage students to use GenAI as an essential part of case analysis. That may seem anathema to some instructors, who have done everything they can to ban ChatGPT and other GenAI platforms from the classroom out of fear it will stop students from doing their own writing and analysis. But I am not alone at UVA, where I’m in my second year as a Faculty AI Guide, an initiative launched by then–vice provost Brie Gertler to empower faculty to integrate GenAI in the classroom.

Far from shortcutting learning, I argue that proper use of GenAI raises the bar for students to perform even more sophisticated work. It reminds me of my own experience as a Darden student in the early 1980s.

When I started, students composed papers with typewriters, wrote exams by hand in “bluebooks,” and used the then-cutting-edge HP-12C calculator for mathematical analysis.

Then in 1983, a dozen Apple II personal computers showed up in the library with 64K of memory, floppy disk drives and the original “killer apps” of the PC era: the spreadsheet program VisiCalc and the word-processing software WordStar.

As capabilities for analysis and composition rose, so did the expectations for students. We were soon required to conduct more complex data analysis using embedded formulas and held to a higher standard for writing given software now allowed us to revise without the help of whiteout and erasers.

In the same way, using GenAI raises the bar for classwork and accustoms students to the tools they’ll use in the workplace. Expanding the role of the learning community, teams submit their write-ups — prepared with the aid of AI tools — for peer reviews evaluating and ranking the effort prior to the class discussion.

The shift is deliberate: the deliverable is no longer "can you build the model?" but "can you frame the right problem, interrogate the output and make a defensible recommendation?" Class discussion moves up the cognitive ladder — from computation to judgment.

At the same time, I integrate GenAI into the tools students use for analysis of traditional cases. For example, students take on the role of a consulting team six weeks into an eight-week purchasing diagnostic.

The case, “Dovrex Procurement Diagnostic,” published in December 2022, includes a 39 Mb spreadsheet “spending cube” covering every line item of every purchase order for more than 2,000 suppliers.

Using GenAI, the students recommend specific spend categories for the implementation phase. One student produced a proposal letter nearly indistinguishable from what I used to write as a partner at Booz Allen.

But that’s just the point: as GenAI makes sophisticated analysis easier, the baseline changes, and even a strong letter backed by rich insight is no longer enough to drive implementation.

What matters now is credibility — the ability to act as a trusted advisor by framing the right problems, tailoring insights to organizational realities and persuading stakeholders to act.

Shifting to Live Cases

In addition to using new cases to teach classic concepts, and analyzing classic cases with new tools, I am also experimenting with “live cases,” using GenAI to process data that is shifting in real time.

Two new cases I am using for this endeavor are “Wonder Foods” and “Walmart EV Stations: Coming to a Store Near You?” Neither case contains an exhibit with data; rather, they both simply provide a short “backstory” on the company and the case protagonist. Students use GenAI to collect and analyze data needed to address the case questions, which themselves may shift as new information becomes available.

For example, the Wonder Foods case challenges students to apply my Operations Strategy framework, which aligns operational investment in assets and capabilities with a firm’s competitive strategy.

Using GenAI to capture data on the current footprint of stores, they then advise founder Marc Lore on investments in a second centralized kitchen. Because that data is constantly changing, the case will never be taught the same twice, giving a level of immediacy to the work that mirrors the kind of work students will be performing in the field.

Similarly, the new Walmart case challenges students to advise Adam Happel (MBA ’08) on how to prioritize the roll-out of EV stations to stores.

In my first application at Darden, we challenged students to focus on Virginia, which has yet to have the new EV chargers deployed. Students develop their own framework, for example prioritizing EV registrations in a store catchment area versus building a network for long-distance travelers or focusing on markets with high EV penetration or emerging markets underserved by existing charging networks. Students then collected data using GenAI to identify their recommended top ten stores with supporting analysis to justify their choices.

Even the GoFundMe case exploring Katz’s “Create Funnel” has refused to stand still. Though the case is set less than two years ago, the company has since implemented many of the improvements it explored.

My students discovered this firsthand when, encouraged to apply the Toyota Production System principle of Genchi Genbutsu — “go and see for yourself” — they visited GoFundMe and launched their own fundraiser using the platform’s tools. I am now exploring how to convert the case to a shorter “live case” for the future.  

The Road Ahead

As these examples illustrate, the case method endures not because it resists change, but because it adapts to new technologies and approaches of analysis. Its deepest value is forcing future leaders to commit to a position under uncertainty, defend it against intelligent peers and learn to act based on that conviction. All of those skills are more relevant in an AI-augmented world, where information flows at an ever-faster pace, and operations managers must process an incredible amount of data in real time in order to make decisions.

The question is how our andragogy best rises to meet the moment, adapting to prepare students for the realities they will face upon entering the world beyond their MBA. I present my Digital Operations Strategy course as one answer — not the only one, but a deliberate one. The cases are live. The tools are authorized. And the cold call still demands that you know what you think and why.